ED 428 761

AUTHORTITLE

INSTITUTIONISSNPUB DATENOTEAVAILABLE FROM

PUB TYPEJOURNAL CITEDRS PRICEDESCRIPTORS

IDENTIFIERS

ABSTRACT

DOCUMENT RESUME

IR 057 302

Barrett, G. Jaia, Ed.ARL: A Bimonthly Newsletter of Research Library Issues andActions, 1998.Association of Research Libraries, Washington, DC.ISSN-1050-60981998-00-00106p.; For the 1997 issues, see ED 416 902.Association of Research Libraries, 21 Dupont Circle,Washington, DC 20036 (members $25/year for additionalsubscription; nonmembers $40/year).Collected Works - Serials (022)ARL; n196-201 Feb 1998-Dec 1998MF01/PC05 Plus Postage.Academic Libraries; Competition; Copyrights; DocumentDelivery; Electronic Journals; Electronic Text; HigherEducation; Information Industry; Information Policy;*Information Services; Interlibrary Loans; MeasurementTechniques; Newsletters; *Research Libraries; ScholarlyJournals*Association of Research Libraries; Digitizing; Library ofCongress; License Agreements; Performance Levels

This document consists of six issues of the ARL (Associationof Research Libraries) Newsletter, covering the year 1998. Each issue of thenewsletter includes some or all of the following sections: "Current Issues,"reports from the Coalition for Networked Information and the Office ofScholarly Communication, Office of Leadership and Management Services(formerly the Office of Management Services), and Coalition for NetworkedInformation, "Preservation," "Federal Relations," "Statistics andMeasurement," "Diversity," "ARL Activities," and calendar of events. Onespecial issue on measures (April 1998) focuses on the issues and activitiesin the area of performance measurement in research libraries. The secondspecial issue on journals (October 1998) discusses views of the currentmarketplace for scholarly journals, including what publisher profits revealabout competition in scholarly publishing, value and estimated revenue ofscientific/technical journals, and non-commercial alternatives to scholarlycommunication. Topics of other issues include the Scholarly Publishing &

Academic Resources Coalition (SPARC); the Dublin Core Metadata Program;E-information policies; an update on Uniform Commercial Code Article 2B;high-performing interlibrary loan/document delivery operations; librarydirector evaluation; salaries in research libraries; university funding forresearch libraries; collaboration on reduced-cost journals; the Library ofCongress mass deacidification program; the Digital Millennium Copyright Act;access management for networked information; and electronic theses anddissertations. (AEF)

********************************************************************************

Reproductions supplied by EDRS are the best that can be madefrom the original document.

********************************************************************************

ARL: A Bimonthly Newsletter ofResearch Library Issues and Actions

Nos. 196-201February-December 1998

"PERMISSION TO REPRODUCE THISMATERIAL HAS BEEN GRANTED BY

C.A. Mandel

TO THE EDUCATIONAL RESOURCESINFORMATION CENTER (ERIC)."

G. Jaia BarrettEditor

2

U.S. DEPARTMENT OF EDUCATIONOffice of Educational Research and Improvement

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERIC)

0 This document has been reproduced asreceived from the person or organizationoriginating it.

0 Minor changes have been made toimprove reproduction quality.

Points of view or opinions stated in thisdocument do not necessarily representofficial OERI position or policy.

FL

r-4

mEn

A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONSCurrent Issues

ARL PROMOTES COMPETITION THROUGH SPARC:THE SCHOLARLY PUBLISHING & ACADEMIC RESOURCES COALITIONby Mary M. Case, Director, ARL Office of Scholarly Communication

The system of scholarly publishing is acomplex process that is centered on theintellectual property of the faculty author.

Until recently, it was routine for faculty in theprocess of publishing to transfer the copyright ofthis property to a publisher. The intellectualproperty and the attendant rights form the eco-nomic basis of a publisher's investment. Formany generations, this system, dominated pri-marily by scholarly societies and university press-es, appeared to serve authors, publishers, and theeducation and research community well. Duringthe 1960s, however, these traditional outlets couldnot expand fast enough to provide the increasedcapacity for publishing generated by a system ofhigher education that was rapidly expanding.Commercial publishers stepped in to fill the need.Some of these publishers discovered that journalpublishing, especially in science, technology, andmedicine, could be very lucrative. A few haveexploited the market and are doing everythingthey can to ensure future profits through aggres-sive pricing strategies, publisher consolidation,and influencing legislation to ensure greaterintellectual property protection.

One key strategy in counteracting thesetrends toward higher prices is to provide addi-tional prestigious and cost-based outlets for thebest faculty work. To this end, ARL recentlyapproved the development of a project calledSPARCScholarly Publishing & AcademicResources Coalitiona project that seeks toencourage the development of competition inthe scholarly publishing marketplace.

9EST CuPY AVAILABLL-3

SPARC is conceived as a partnership projectof ARL and other educational and research orga-nizations. Its mission is to be a catalyst:

To create a more competitive marketplace forresearch information by providingopportunities for new publishing ventures,endorsing new publications and informationproducts, and recruiting authors, editors, andadvisory board members.To promote academic values of access toinformation for research and teaching, thecontinuation of fair use and other library andeducational uses in an electronic informationenvironment, and the ethical use of scholarlyinformation.To encourage innovative uses of technology toimprove scholarly communication bycollaborating in the design and testing of newproducts; advancing new publishing modelsas appropriate applications of electronicnetworks, such as Internet2; and developingsystems and standards for the archiving andmanagement of research findings.

Marketplace TrendsMost faculty and librarians are now very familiarwith the annual cycle of serials cancellations inacademic and research libraries. What has alsobecome clear is that the cancellation projects arenot one-time adjustments to local circumstances,but manifestations of a marketplace that is pricingsome resources beyond the reach of the educa-tional community. Over the past decade, ARLstatistics show that research library expendituresfor serials have increased almost 10% a year and

2

CURRENT ISSUESContinued

--------

that the unit cost of a serial title has increased by 147%.Since 1986, while ARL libraries have canceled hundredsof thousands of dollars worth of journals, they havespent 124% more on serials to purchase 7% fewer titles.During the same period, expenditures on monographshave increased 29% and ARL libraries have purchased21% fewer titles.

While some price increases can be justified by theincrease in numbers of pages and articles included with-in a volume to accommodate the increased output offaculty, there is evidence to suggest that some commer-cial publishers charge prices that significantly exceedthe costs of production. In 1989, Economic ConsultingServices, Inc. (ECS) concluded an analysis for ARL ofthe trends in average subscription prices and publica-tion costs over time. The analysis focused on over 100titles published by four major publishers in the U.S. andEurope. The result of the ECS report found that from1973-1987, publisher profits increased between 40% and137%. The results also indicate that the cost increasesfaced by these publishers did not fully justify the priceincreases paid by research libraries.

One of the most profitable scholarly publishers isReed Elsevier, one of the world's largest publishingand information companies, located in the UnitedKingdom. Its 1996 Combined Profit and Loss Statement(which is included on its website <http://www.reed-elsevier.com/combstat.html>) reports a gross profit of£2,082 million (about $3.5 billion) on sales of almost

£3,400 million ($5.7 billion). The net operating expensesare £1,248 million (or $2.1 billion). This appears to be afairly substantial margin (£834 million or $1.4 billion).After taxes, £604 million (or approximately $1 billion) isleft to pay out dividends to shareholders (£348 million)and to transfer funds to reserve (£256 million). Thestatement at the bottom of the report notes: "Thehistorical cost profits and losses are not materiallydifferent from the results disclosed above."

Elsevier Science, an international publisher of scien-tific information headquartered in the Netherlands, isone of the several publishing companies owned by ReedElsevier. Preliminary analysis of figures recently gath-ered by ARL shows that, on average, an ARL libraryspends almost $628,000 dollars a year with ElsevierScience to obtain 378 titles. That is 3.5% of the serialstitles subscribed to by a library but almost 21% of theannual serials expenditures. Extrapolating from theseaverages, the 121 ARL libraries spend over $75 million ayear on Elsevier Science titles. This is a significant por-tion of U.S. and Canadian research library materialsbudgets going to one company.

To illustrate, Elsevier Science publishes some of themost expensive serials subscribed to by researchlibraries. The accompanying table provides a sample ofthese journal titles and the increases in their subscrip-tion prices between 1995 and 1998. The increases rangefrom a total of 43.5% (or roughly 14.5% a year) to 65.8%(or 22% a year).

SAMPLES OF SUBSCRIPTION PRICE INCREASES

1995 1996 Change

95 to 96

1997 Change

96 to 97

1998 Change

97 to 98

Change

95 to 98

Brain Research $10,181 $12,234 20.2% $14,919 21.9% $15,428 3.4% 51.5%

Biochim. Biophys. Acta $7,555 $8,837 17.0% $10,528 19.1% $10,839 3.0% 43.5%

Chem. Phys. Letters $5,279 $6,569 24.4% $7,818 19.0% $8,060 3.1% 52.7%

Eur. Jrnl. of Pharmacology $4,576 $5,680 24.1% $6,431 13.2% $6,702 4.2% 46.5%

Gene $3,924 $5,069 29.2% $6,144 21.2% $6,433 4.7% 63.9%

Inorganica Chim. Acta $3,611 $4,476 24.0% $5,283 18.0% $5,540 4.9% 53.4%

Intl. Jrnl. of Pharmaceutics $3,006 $3,915 30.2% $4,691 19.8% $4,983 6.2% 65.8%

Neuroscience $3,487 $4,001 14.7% $4,543 13.5% $5,073 11.7% 45.5%

Theoretical Computer Science $2,774 $3,425 23.5% $3,835 12.0% $4,059 5.8% 46.3%

Jrnl. of Exp. Marine Bio. & Eco. $1,947 $2,445 25.6% $2,811 15.0% $2,931 4.3% 50.5%

Solid State Communications $1,945 $2,327 19.6% $2,602 11.8% $2,871 10.3% 47.6%

Note: The actual subscription price paid by an individual library may vary. The prices above for 1995-1997 reflectactual experience in two ARL member libraries; 1998 prices are taken from the Elsevier Science website.

A R L 1 9 6 F E B R U A R Y 1 9 9 8 4

Maximizing the MarketThrough ConsolidationThere is also anecdotal evidence that another trend inthe publishing industry may be affecting the price oflibrary materials: the consolidation of publishers of coreproducts. Reed and Elsevier merged in 1993 and havesince pursued an aggressive corporate strategy of acqui-sitions and mergers. Reed Elsevier now owns publish-ing companies throughout Europe and North America,including Bowker/Saur, Butterworth's, Shepard's,Lexis-Nexis, and the Congressional Information Service,to name a few. Its corporate global strategy, as stated onits website, is to concentrate on "must have" informa-tion and to continue development through "organicgrowth and acquisition." This includes not only buyingother companies, but also buying or contracting topublish society and association journals.

In October 1997, Reed Elsevier announced a plan tomerge with Wolters Kluwer, which itself owns Lippin-cott/Raven and a number of other European companies.A new company, Elsevier Wolters Kluwer, would,according to The New York Times (29 December, 1997),"create the world's largest publisher of professional andtrade journals, with estimated combined 1997 revenuesof $8 billion...." Regulatory agencies in both Europe andthe United States are concerned about and are investigat-ing this proposed merger. But that has not stopped addi-tional acquisitions. In December 1997, Elsevier Sciencereleased a statement that it was in negotiation with theBeilstein Institute, a non-profit German organization, toobtain an exclusive commercial license to the BeilsteinDatabase and Handbook, the standard reference work fororganic chemistry. In late January 1998, EngineeringInformation, Inc. notified its customers that it was sellingvirtually all of its assets, including its name, to the newlycreated Elsevier Engineering Information, Inc., a wholly-owned subsidiary of Elsevier Science.

Publisher consolidation is also underway in thefields of law and business, where the Thomson Corpora-tion of Canada has moved aggressively to acquire otherpublishers. Thomson owns the Institute for ScientificInformation, Gale Research, West, and Warren GorhamLamont. When Thomson purchased West Publishing in1996, it was required by regulatory agencies to sell off anumber of its law journals. These titles were purchasedby Lexis-Nexis, owned by Reed Elsevier. In anotherswitch among the giants, Thomson announced inDecember that it had agreed to sell its ThomsonScience subsidiary to Wolters Kluwer. (Data from theThomson website shows that the specialized informa-tion/publishing companies within Thomson made$747 million, or 19% operating profit, before amortiza-tion in 1996; see <http:/ /www.thomcorp.com/press/858133513.html>.)

As might be anticipated with the recent takeover ofLexis-Nexis by Reed Elsevier, pricing and access optionsare changing. As of September, one ARL memberlibrary estimated that it would have to pay 3.5 times thecurrent cost for the new Lexis-Nexis educational service,which includes less access than its previous subscrip-tion. The library would now pay a total of $87,000 ayear versus the current price of $25,200. An example ofthe impact on the pricing of an individual title takenover by Elsevier in 1997 is the subscription price for theJournal of Supercritical Fluids. The journal was publishedby Polymer Research Associates and cost $275 for ayearly subscription. A subscription for 1998 costs $657.

Faced with rising prices that show no sign of mod-erating, libraries have undertaken a number of strate-gies to try to manage their budgets and increase theirbuying power. These include journal cancellations,with reliance on document delivery and cooperativecollection development for low use materials, andconsortial licensing of electronic resources. These gen-erally local strategies, however, have had limited effec-tiveness in moderating prices and have had no impacton the growing concentration of the publishing indus-try. Cancellations only result in increased prices for therest of the subscribing community, exacerbating thealready high prices.

Expanding Product ControlThrough LegislationBut higher prices are not the only concern arising fromthe increased concentration of publishing in the handsof a few large commercial companies. There is also theissue of how much exclusive control the copyright hold-er may assert over access to and use of copyrightedinformation. The last year has seen accelerated effort onthe part of many stakeholders in the scholarly commu-nication process, including libraries, to ensure continua-tion of balanced copyright law in a digital and net-worked environment. However, the same commercialpublishers that have successfully pursued a profit-maxi-mization strategy in the print environment are aggres-sively advocating legislation that would strengthen thecopyright owner's monopoly and control over intellec-tual property. Some of the legislative proposals beingpromoted would transform current copyright law into adramatically more restrictive instrument, narrowing oreliminating fair use and other provisions that providefor educational use of copyrighted materials.

As an example, access to information for researchand education could be significantly restricted underH.R. 2652, The Collections of Information AntipiracyAct, a bill introduced into the U.S. House ofRepresentatives in October 1997. Reed Elsevier andThomson are major proponents of the bill, which wouldamend the Copyright Act to "protect databases from

A R L 1 9 6 F E B R U A R Y 1 9 9 8

Continued

misappropriation." While it is not unreasonable forpublishers to want to protect their products, there is noconvincing evidence, especially in light of the publish-ers' profitable track records, that data is being "misap-propriated" on a scale that justifies this legislativeaction. The legislation as written is extremely broad,covering everything from facts to journal articles, andthe exemptions for educational, scientific, or researchuses are too narrow to be useful. C. Judson King,Provost and Senior Vice President of Academic Affairs,University of California system, believes that "H.R. 2652would impede the free exchange of information andideas that is the engine of theacademic enterprise, therebyslowing the dissemination andcreation of new knowledge" (let-ter to The Honorable HowardBerman, 29 October, 1997). Notonly are Reed Elsevier andThomson intent on maximizingtheir profits through pricing andconsolidation, they are activelypursing legislative agendas to ensure long-term protec-tion of their revenue streams, but at the expense of edu-cation and the public good.

The conditions of the scientific and scholarly pub-lishing marketplace make it impossible for higher edu-cation to sustain library budgets on a scale sufficient tokeep up with increasing prices and the level of output.At the same time, access to and use of digital informa-tion resources could be restricted by legislative attemptsto expand the copyright holders' control over access tointellectual property. The need for alternative publish-ing outlets"new entrants," "greater competition," and"alternative models"particularly in the serials market-place, has become increasingly obvious and increasinglyurgent.

Why SPARC and Why Now?Increased competition was identified as a key strategyfor addressing rising prices for serials ten years ago, butchange has been slow in coming. A number of criticalfactors, however, have now converged to make thetransformation of the scholarly communication systempossible:

Alternative models of scholarly communication arenow economically and technologically feasible. Therise of the Internet and World Wide Web have madeit possible for anyone to publish.Capable partners are ready to join with researchinstitutions to create new publishing alternatives.Libraries and universities are prepared to redirectbudget resources to support new forms of scholarlypublishing. This was confirmed at a recent meetingof academic administrators who indicated their

willingness to devote resources to address the prob-lems in scholarly communication. Academic admin-istrators have come to realize that the strategies thatlibraries can pursue on their own cannot adequatelyaddress the fundamental problems in the currentsystem of scholarly communication. Collective effortby the community is required.Faculty and academic administrators will supportinitiatives that offer realistic alternatives for dissemi-nating research findings and scholarship.

On June 30, 1997, representatives from 45 ARL insti-tutions met to discuss a proposal to create an electronic

publishing fund and to seekpublishing partners interestedin entering the serials marketin areas in which prices arehighest and there is the greatestneed for alternative models ofresearch communication. Themeeting resulted in the forma-tion of a small working groupof library directors who

volunteered to work with ARL staff in developingan action plan.

The Working Group identified potential partnersorganizations that may be interested in forming part-nership relationships with libraries and educationalinstitutions. These include:

Professional societies and university pressesinterested in launching new publishing initiatives.Start-up electronic publishers that have alreadycreated publications in subject fields dominated bycommercial publishers.For-profit enterprises that offer new strategies forcontrolling costs and improving access to researchinformation."Visionary" enterprises, including both disciplineand institution-based server models, seeking tocreate entirely new economic models for scholarlycommunication.The partnerships forged by SPARC will depend on

the readiness of potential partners to achieve the goalsof the coalition in terms of content and infrastructure.In addition, potential partners will be evaluated onwhether: they share the values of the scholarly and aca-demic community, they have credibility in the scholarlymarketplace, their leadership is committed to the goalsof SPARC, they have publishing experience, and theyare willing to commit resources.

The key criterion that is critical to the success ofthese partnerships is shared values. The values pro-posed for SPARC partners include support for:

Fostering a competitive market for scholarly publish-ing by encouraging new participants in the publishing

The key criterion that iscritical to the success of

these partnerships isshared values.

A R L 1 9 6 F E B R U A R Y 1 9 9 8

field that are committed to principles of cost-basedpricing.Implementing policies for intellectual property man-agement emphasizing broad and easy distributionand reuse of material and the ethical use of scholarlyresources.Encouraging innovative applications of availableinformation technology to enrich and expandresearch and scholarship and the available means fordistribution.Assuring that new channels of scholarly communica-tion sustain quality requirements and contribute topromotion and tenure processes.Enabling the permanent archiving of research publi-cations and scholarly communication, includingthose published in digital formats.

Last October, the ARL Board gave the go-ahead tothe Office of Scholarly Communication to develop aplan to make the SPARC concept a reality. An ad hocWorking Group guiding the next phase of SPARC'sdevelopment will propose a membership strategy thatinvites the participation of other library, educational,and research organizations in SPARC. The WorkingGroup is also charged with the responsibility to developa business plan and action agenda to support at leastfive alternative publishing ventures by the end of 1998.The first major action of the Working Group this yearwas to invite ARL libraries to become FoundingMembers of SPARC. At press time, almost 50 ARLlibraries had responded. Next on the Working Group'sagenda will be hiring a SPARC project manager.

ARL recognizes that SPARC is only one of a numberof strategies that must be undertaken simultaneously toensure long-term access to scholarly research. Otherstrategies include working aggressively in the legisla-tive arena to ensure fair use and other educational andlibrary uses of copyrighted works in the digital environ-ment; investigating options for faculty and the universi-ty to retain and better manage intellectual propertyrights; and the decoupling of the academic credentialingprocess from formal publication. None of these strate-gies will work without the support and involvement offaculty, academic administrators, and the research andscholarly community. With these constituencies, ARLwill pursue SPARC as one effort to build the partner-ships necessary to create a future for scholarly commu-nication that is robust, innovative, and affordable.

Thanks to Carson Holloway for gathering information oncorporate ownership included in this article. Moreinformation about SPARC can be found at<http:11www.arLorg/sparcl>.

,

THE DUBLIN CORE DESCRIPTIVEMETADATA PROGRAM: STRATEGICIMPLICATIONS FOR LIBRARIES ANDNETWORKED INFORMATION ACCESSby Clifford Lynch, Executive Director, Coalition forNetworked Information

This article explores the potential applications ofthe Dublin Core descriptive metadata program forlibraries, museums, and other networked informa-

tion providers. It includes a brief summary of the currentthinking that has emerged from the Dublin Core initia-tive, including the broader metadata framework capturedby the Warwick Framework, and also some considerationof the relationship with the Resource DescriptionFramework under development by the World Wide WebConsortium, but the central focus is on applications sce-narios for Dublin Core metadata. The approach here isstrategic rather than technical; readers interested in theprecise definitions of the individual Dublin Core data ele-ments will need to consult the actual Dublin Core docu-ments or other articles. My hope is that it will be helpfulto library administrators and technology managers whoare trying to understand and evaluate the implications ofthe Dublin Core both for access to existing resources andfor practices of description that will be used to create,maintain and provide access to new resources.

MetadataMetadata is literally "data about data," information thatqualifies other information. Bibliographic description isa form of metadata, so also is information about intellec-tual property rights and terms of use, formats of elec-tronic information, reviews, errata, abstracts and sum-maries, provenance information, and a host of otherdata. Some metadata can be derived mechanically fromobjects; other metadata has independent standing asintellectual creation in its own right. It should be clearthat the set of metadata associated with an informationobject is unbounded. The division between data andmetadata is somewhat arbitrary and highly situational;information will be used as data in one setting andmetadata in another.-

At least in my view, discussions of metadata inde-pendent of context and purpose are of little interest; it ismost productive to speak of various kinds of metadata inconjunction with the processes that they are intended tosupport or facilitate. There are certainly types of meta-data that have been developed for various specific pur-poses and which it is now proving possible to repurpose,particularly in the digital environmentindeed, creativerepurposing and reuse of metadata is emerging as a keyidea in the development of sophisticated informationorganization, retrieval, and management systems. But

A R L 1 9 6 F E B R U A R Y 1 9 9 8

Continued

the point is that metadata is created and takes on impor-tance through its ability to support activities; for exam-ple, the point is not to describe but to support discoveryand other processes.

For a more extended discussion of the nature ofmetadata and its interactions with the processes of thediscovery and retrieval of networked information, read-ers might consult the draft CNI white paper on the topic<http: / / www.cni.org /projects / nidr / >.

The Warwick FrameworkWhile the Warwick Framework (named after the meetingin Warwick, England where it was developed) actuallypostdates the beginning of work on the Dublin Core(DC) described later, it is useful to discuss it first becauseit provides a broad framework in which to define sets ofmetadata.

The basic motivation for the work on the DublinCore was to develop a set of simple data elements thatcould be used to describe document-like networkedinformation objects in support of discovery (searching)activities. It rapidly became clear that there were anynumber of legitimate, important requirements for typesof metadata that went beyond the scope of the DublinCore; the problem was that because the Dublin Core wasan active effort, and also because it was not clear how touse the DC in conjunction with other sets of metadata,there was considerable pressure to extend the scope ofthe actual DC effort almost without boundaries. Thisthreatened the effectiveness of the Dublin Core program.To address this problem, an architecture called theWarwick Framework was developed that described howvarious sets of metadata for different purposes might bedefined and maintained by appropriate communities ofexpertise. Collections of data elements from thesediverse sets of metadata would be assembled into "pack-ages" (one package per metadata set). The frameworkdescribes container structures whereby a digital objectand a collection of such packages can be linked together.Each package is independent of all of the others, andsoftware systems that understand specific metadata setscan extract packages that are based on those sets andexamine them, bypassing other packages based on unfa-miliar sets. Individual packages can even be encryptedindependently. Containers can also refer to remotepackages stored independently on the network, and arerecursive: a container can include other containers,allowing for the construction of complex compositeobjects.

In designing the Warwick Framework, there was arecognition that division of the universe of metadata intopackages would be imperfect; there would be some over-lap between packages, and the content of one packagemight, in some cases, be derived computationally fromanother. There are also a number of research questions

about how relationships among packages are expressed.The importance of the Warwick Framework is

twofold. First, it provides a broad architectural frame-work for defining and using,metadata of various types.Second, it allows developers of metadata sets that havespecific purposes to limit and focus their work byappealing to the Warwick Framework as an overarchingcontext within which other groups interested in metada-ta can independently make progress on their own needs.

More information on the Warwick Framework canbe found in the article by Carl Lagoze in the July 1996issue of D-Lib Magazine <http://www.dlib.org/dlib-july96/lagoze/O7lagoze.html> and the references there.

The Dublin Core.The Dublin Core is a set of fifteen data elementseach ofwhich is both optional and repeatablethat wasdesigned to be used as metadata to describe a broadclass of information objects. The description applied toobjects through the Dublin Core data elements is notintended to be comprehensive or exhaustive; it does notseek to capture everything that can be asserted about anobject. In particular, the DC is designed to support dis-covery of information objects of interest using searchtools and systems; it is not designed to provide compre-hensive support for access, management, use, or assess-ment of networked information (though some of themetadata to support discovery is also important in thesecontexts). To give just one example of this distinction,the Dublin Core includes a data element for terms andconditions. This is provided primarily because some lim-ited information on terms and conditions of useforexample, that an object is not copyrighted, or that it canbe used without restriction in educational settingsisactually important in finding objects of interest. There iswork underway to develop very complex codings toexpress terms and conditions that might be used in con-junction with electronic rights management systems; thisdata would be essential for use and management appli-cations, but is probably too detailed and specialized tobe of much use in the discovery process, particularlygiven the current immature state of both standards andconceptual understanding in rights management specifi-cations. The Dublin Core is not intended to carry thiskind of very specific functional rights managementterms and conditions metadata.

The DC was developed to describe what have beencalled, for want of a better term, document-like objects.These have the characteristics of being relatively fixed,although they need not be textual (images or soundrecordings are easily within scope). They may haveinternal sub-structurefor example, an object withcomponent objectsbut the main focus of the DC to datehas been to describe objects as opposed to collections ofobjects. The primary concern has been to ensure that the

A R L 1 9 6 F E B R U A R Y 1 9 9 8

DC is serviceable for a rather broad range of commoninformation objectsfor example, a workshop was heldwith the specific focus and outcome of extending andvalidating the DC as a means of describing a large classof visual resourcesrather than drawing precise bound-aries for what is out of scope.

The DC clearly can be usefully applied to collectionsor to very complex dynamic objects or information ser-vices, but it probably does not do a completely adequatejob of describing such objects and services to supportdiscovery.

The development of the DC has had a very strongtheme of codifying practice rather than research: meth-ods for satisfactorily describing these complex new class-es of dynamic digital objects and network services is still,at least in my view, a research problem. Further, whilethere is a great deal of consistency across the class ofdocument-like objects that were the objective of the DC,there seems to be tremendous variation in the kinds ofdescription needed for the still evolving menagerie ofnew digital resources. And, at least today, there are a lotmore document-like objects than anything else on thenetwork; they are the rule rather than the exception.

Data ElementsThe fifteen Dublin Core data elements are: title; creator(author); subject and keywords; description; publisher;other contributer; date; resource type; format; resourceidentifier; source; language; relation; coverage; andrights management.

It should be clear from an inspection of this list thatthe DC is designed to serve as a sort of lowest commondenominator form of description. It does not, at leastdirectly, accommodate discipline-specialized descrip-tion; indeed, some of the data elements, such as "date"are so vague that they are of limited utility without somefurther scoping. The DC was designed so that data ele-ment values for an object could reasonably be defined byits author, or by a site manager, rather than by a trainedspecialist cataloger or indexer. DC relies very heavily onnatural language, and retrieval systems for the DC willhave to rely heavily on heuristics and language parsing;not only does the basic DC framework ignore specializedthesauri and subject classifica Lion, for example, but itdoes not even make assumptions about the format thatwould be used to list dates or personal names.

QualifiersOne of the ongoing tensions and controversies in thedevelopment of the DC has been its lack of precision.The most basic version of the DCcalled the UnqualifiedDublin Coredoesn't carry any information about theformat of the data element values, their source or con-text, or the specifics of their semantics beyond the verybroad definitions of the basic data elements. To address

this need various qualifiers have been proposed to servesuch functions as indicating the language or syntax inwhich the data element values are expressed, or to con-strain the semantics of the data elements (for example,indicating that a date is a date of creation, or that a cre-ator is a corporate author, or that a topic value is takenfrom a specific thesaurus). Obviously, use of qualifierswill tend to reduce interoperability, because participat-ing systems will need to understand much more thanjust the fifteen basic data elements in order to interpretthe semantics properly. To address this problem, a basicrule has been established for all types of qualifiers: ifone ignores the qualifier, the data element value must beconsistent with the basic definition of the data element'ssemantics in the DC. Thus, qualifiers can only constrainor refine the semantics of the DC data elements; theycannot be used to alter their meanings so that they areinconsistent with the original definitions. Definition ofthe data elements which should under normal practicebe qualified, and what the appropriate values of thesequalifiers should be, is a subject of ongoing work withinthe DC community; in a sense, this can be viewed as adiscussion about how to extend the DC beyond the origi-nal fifteen elements in practice without destabilizing theoriginal definitions, although some qualification (forexample, to indicate the language or format of the dataelement value rather than the meaning of the data ele-ment itself) really has a different and less semanticallysignificant character. It remains to be seen how the useof qualifiers will evolve within the various communitiesof DC users.

Relationship to SurrogatesAnother controversial issue in the definition of theDublin Core has been how use of the DC elementsshould interact with the surrogates that are so common-place in the digital environment. For a document thatwas created as a digital object, matters are simple: theDC data elements describe the document. The creator ofthe document is the person who authored it. But consid-er this common case: there is a painting hanging in amuseum that was created by artist X; fifteen years ago,photographer Y took a picture of this painting; last week,curator Z digitized Y's picture of X's painting. What arethe semantics of the DC metadata associated with thedigitized image? The answer is that there should bethree groups of DC metadata: one for the painting, onefor the photograph, and one for the digitized image. Thefirst would list X as creator; the second Y, and the thirdZ. The three groups of DC metadata would be connect-ed through the "relation" data element. This has theadvantage of being conceptually simple, albeit a bit ver-bose, for those creating metadata (though this can clearlybe mitigated by a well-designed data entry system forDC metadata). It also places a considerable burden on

9 ARL 196 FEBRUARY 1998

Continued

the design of retrieval systems to behave intelligently: tomany users, the conceptual distinction between painting,photo, and digitized image is at best murky, and an end-user query will often ask for the painting when what theuser really wants is a digitized image of the painting.Retrieval systems will need to be able to retrieve clustersof groups of DC data elements and present them to theuser in a comprehensible fashion.

The Evolution and Documentationof the Dublin CoreTo date, the DC has been developed informally by aloose international consortium of interested partiesthrough a series of five workshops: Dublin, Ohio (fromwhich the core takes its name); Warwick, England;Dublin, Ohio again (a meeting focused specifically on therole of the DC in describing visual resources); Canberra,Australia; and Helsinki, Finland. The sponsors of thesemeetings have included OCLC, the National Center forSupercomputing Applications, the National ScienceFoundation and the Coalition for NetworkedInformation in the U.S., UKOLN in the U.K., theAustralian National Library, the National Library ofFinland, and many others. Stuart Weibel of OCLC hasbeen the leader of the effort since the beginning.

At the conclusion of the Helsinki meeting in late1997, a series of working groups were chartered to con-tinue efforts to extend and refine various aspects of theDC. It is likely that work in 1998 will proceed through aseries of smaller meetings focused on specific issues,concluding with a sixth plenary meeting late in 1998.

At present, the Dublin Core is documented in aseries of meeting reports and articles in D-Lib Magazine,and in working documents on the DC website<http:/ /purl.org/metadata/dublin_core/>; this siteincludes extensive information on the meetings, bibli-ographies, and other useful links. A series of informa-tional (not standards-track) IETF RFCs are in preparationand should be released within the next few months.There are ongoing discussions about progressing the DCthrough the U.S. National Information StandardsOrganization as a formal standard, and also aboutwhat should be done to provide an ongoing "home"and maintenance agency for the standard if and whenit is finalized.

Note should also be made of the work of the WorldWide Web Consortium, which is working on a programthey call the Resource Description Forrnat (RDF). Whilethis work is not directly driven by the DC, and in fact hassome of its roots in extending earlier Consortium effortsto develop PICS (the Platform for Internet ContentSelection) for rating and content filtering applications, thegroup working on RDF includes heavy representationfrom the DC community. The goals of the RDF effortinclude the definition of general mechanisms for attaching

metadata of all kinds to web pages composed using thenew Extended Markup Language (XML) defined by theConsortium, including DC metadata, the development ofschema definitions for metadata sets, and query facilitiesfor metadata. This work will likely be central to facilitat-ing the large scale use of DC within the Web.

More information on PICS, RDF and XML can befound at the Consortium's website <http:/ /www.w3c.org/>.

Machinery Needed to Supportthe Use of the DCAt one level, the Dublin Core is a conceptual construct;it captures the idea that there are pieces of text that can beassociated with an information object with agreed-uponsemantics such as those of "creator" or "relation." In orderto make this conceptual construct concrete and to apply itin the networked information environmentwhich ischaracterized by the cooperation of large numbers ofautonomous machines and agencies, and the sharingof information among themthere is need for a varietyof supporting machinery. This machinery is codified insupporting standards and practices. It's important torecognize that, in a sense, the Dublin Core transcends spe-cific machinery, and that many different mechanisms canlegitimately be developed within different communities ofpractice and implementation to meet these requirements.It is likely that over time new mechanisms will continueto develop as a result of the overall evolution of architec-tures and standards for the networked informationenvironment.

The three major classes of mechanisms are: encodingand transfer syntaxes; methods of associating or attachinggroups of Dublin Core data elements with the informationobjects that they describe; and, more generally, methodsof retrieving or querying Dublin Core data associatedwith objects or groups of objects.

Several methods have been proposed for encodinggroups of DC data elements for storage and inter-systemexchange: these include the use of HTML META tags intoday's HTML-based web pages; the use of XML struc-tures as specified by the World Wide Web Consortium'sResource Description Framework (RDF) as part of futureXML-based web pages; and the incorporation in SGML.Several of these proposals also address the problem ofassociating DC elements with objects in a very direct way:in the Web setting, they are simply incorporated as partof web pages. There is also a way of requesting DC infor-mation for an object via HTTP (thus leaving it up to theweb server to maintain the linkage between DC elementsand the base object internally); this mechanism can alsobe used to query third-party metadata servers for DCmetadata.

One of the key deployment scenarios envisionedfor the Dublin Core is that web pages will increasingly

A R L 1 9 6 F E B R U A R Y 1 9 9 8 1 0

incorporate DC data elements as part of the pagesusingeither direct coding in META tags for current HTMLpages or the new RDF structures for pages in the newlydefined XML formatand that the familiar web indexingprograms (or their successors) will be upgraded to cap-ture this metadata and incorporate it into their web index-es, so that one could query a system like Lycos or AltaVista for pages that have a specific creator, for example.This metadata might be created by the authors of thepages, by website managers, or by third partyindexers/catalogers. Complementing this, we are likelyto see third party databases of DC metadata developwhich simply refer to and describe web content and otherinformation objects.

It's essential to recognizethat while the Weband in par-ticular the static, visible web ofHTML pagesis a key applica-tions environment for DC, it isnot the only one. It is perfectlyreasonable to think in terms ofdatabases containing objectsdescribed by DC data elements;here the DC data elementswould be encoded and linkedthrough some local data struc-tures. The retrieval of an objectfrom such a databaseaccom-plished through an interactive forms-oriented queryinterface or an inter-system query protocol like Z39.50might cause the retrieved object to be encoded as a well-known, common format, such as a page that includedXML tags for the relevant DC elements. Similarly, onemight want to associate DC elements with an entirewebsite or database; here one would need a mechanism(perhaps akin to some of those used in the Harvestsystem) that could be used by network resource index-ing systems that build site or database directories. (Formore on these issues, see Clifford A. Lynch's "Searchingthe Internet," Scientific American 276.3 [March1.997]: pp.52-56, available at <http://www.sciam.com/0397issue/0397lynch.html>.)

At present, query facilities for Dublin Core dataelements are very diverse. There are a number of inter-active query systems that offer DC data elements asaccess points to specific databases or other informationcollections. Several Z39.50 attribute setsnotably GILSand BIB-1are incorporating the DC elements as accesspoints that can be used in query construction, and, aspart of the migration to the new Z39.50 attribute archi-tecture, it is likely that a separate Z39.50 attribute setwill be defined. Part of the RDF work program includesthe definition of query facilities for metadata; however,work on this is only at the earliest stages.

Applications Scenarios for LibrariesThe Dublin Core has two different basic applications forlibraries. The first is in permitting library databases tobecome part of broader network search services, or toallow libraries to provide their patrons with consistentviews of both library and non-library databases. Thesecond is in describing new resources that cannot becost-effectively supported through traditional catalogingapproaches.

Use of DC in Federating Existing ResourcesOne of the key notions in networked information discov-ery and retrieval is that of federating disparate, indepen-

dently maintained databasesscattered about the network.Users should be able to search ,

such constellations of databasesas if they were a single, consis-tent, unified informationresource. In order to do that, itis necessary to provide a com-mon semantic view of the vari-ous databases involved, eventhough they may have radicallydifferent access points and datastructures, and may be accessedthrough different search proto-cols or other query mechanisms.

Because the Dublin Core is designed as a lowestcommon denominator descriptive approach, it offers avery flexible and general context to support federation.Traditional library catalogs or abstracting and indexingdatabases can clearly support queries constructed usingDublin Core data elements (albeit with some reductionin the precision that queries can express as compared toqueries formulated using the database's native searchlanguage, unless qualifiers are used extensively); thus itis possible to build a software layer that permits suchdatabases to participate in federations that use theDublin Core data elements. These interfaces will usemappings or crosswalks to translate from Dublin Coredata elements to the actual access points in the database.Mappings have already been developed from DC toMARC fields.

I think it is likely that libraries will use this capabili-ty to make their databases visible in database federationsthat operate outside of the traditional library systemsand servicesthese databases could be searched as partof a distributed search system that also encompassedweb-based resources outside of the library. To illustrate,suppose that a group of art history scholars developed adatabase of digitized images that were described usingDC, and build a search system for that database. Bymaking a DC-based "view" of a resource like a library

I think it is likely thatlibraries will use this

capability to make theirdatabases visible in

database federations thatoperate outside of the

traditional librarysystems and services...

A R L 1 9 6 F E B R U A R Y 1 9 9 8

-CURRENT ISSUESContinued

catalogue or an art history abstracting and indexingdatabase available on the Net, it would be possible toeasily extend their system to also consistently searchacross the database and catalogue as supplementalresources. Or a system designed to search digitalinstructional media might be extended to also searchlibrary holdings through the same interface.

Conversely, because the Dublin Core is applicable toso many information resources, a library might developa search interface and distributed search service thatoffered patrons a federated view of a very diverse set ofdatabases, including not only traditional library databases,but also databases from other sources, such as govern-ment databases, databases produced by the next genera-tion of web indexing services, or special purpose scholar-ly databases. While such a search service would noteliminate the need for much more precise and capabledomain-specific and database-specific search facilities, itwould be very useful to some users both in identifyingdatabases of interest which they might then search direct-ly for more comprehensive and precise results, or indoing very broad (not but necessarily precise or exhaus-tive) searches across a wide range of resources. In thisconnection, it is interesting to note that the InstructionalManagement System (IMS) being developed byEducom's National Learning Infrastructure Initiative(NLII) is using a descriptive scheme based in part on theDublin Core for instructional media; this is a good exam-ple of a possible new resource that libraries may want tobring under the umbrella of a search system that also cov-ers their catalogue and abstracting and indexing databases.(For more details on the NLII and its IMS project, see<http://www.imsproject.org>.)

Use of the Dublin Core in Describing New ContentThe Dublin Core--perhaps supplemented by additionalmetadata packages defined within the WarwickFrameworkwill be used to describe content wheretraditional cataloging approaches are too costly, or wherethere is a need to create metadata for content that is notwell served by current cataloging practices. The NLIIIMS is a good example: many of the key things that usersneed to know in searching for instructional media canonly be captured by traditional cataloging in unstruc-tured textual notes. The IMS supplements the DC ele-ments with an additional descriptive package designedspecifically for instructional media. For digitized imagesor other materials, whether created directly in digitalform or digitized from other media (e.g., special collec-tions), full bibliographic cataloging is particularly expen-sive because most of these items are unique, and librariescannot use the system of shared copy cataloging to con-trol and distribute costs. It's important to note that, whilethe Dublin Core was designed to be simple and thusmuch less expensive to apply than traditional AACR-2

A R L 1 9 6 F E B R U A R Y 1 9 9 8

based original cataloging, there is relatively little experi-ence with it, particularly when the DC is supplementedwith additional metadata packages. One effort that needsto take place over the next few years as part of the experi-ence in using Dublin Core is some measurement of thecost savings over traditional cataloging for varioustypes of material. We also will need to understand howretrieval quality varies with the different descriptiveapproaches.

The use of the DC and the Warwick Framework giveslibraries the ability to design supplementary metadata setsdescriptive and otherwiseto characterize materialsthat either require more depth or precision of descriptionthan the Dublin Core alone can offer, or need not onlydescriptive but also other types of metadata associatedwith them in order to support processes that go beyonddiscovery (e.g., management, use and reuse, or rightsclearance). Instructional media objects are a good exam-ple of such a category of materials; statistical datasets areanother. Museums will likely make substantial use of theDC plus additional metadata packages. To a great extent,I suspect that library use of the DC for description will bedetermined by the policy choices that libraries make abouttheir role in creating descriptions for materials that havenot historically been part of the mainstream of librarycollections, as opposed to simply making use of descrip-tions for these materials created by other (non-library)organizations.

ConclusionsThe Dublin Core is clearly, in my view, going to be impor-tant for libraries both as an engineering tool for federatinglibrary and non-library databases, and also as a lower-costalternative for describing materials. The creation ofDublin Core descriptions is going to be of particular inter-est for libraries expanding their collections with largeamounts of digital content: images, sound recordings,video recordings, and new genres of networked informa-tion. In the longer run, I think it will also be important forlibraries to track the work on the implementation of theWarwick Framework and to monitor the definition ofadditional metadata sets within that framework, whichwill be needed to address issues such as provenance,integrity, and management of digital content.

Copyright @ by Clifford Lynch. The author grants blanket per-mission to reprint this article for educational use as long as theauthor and source are acknowledged. For commercial use, areprint request should be sent to the author <clifford@cni.org>.

Editor's Note: NINCH, the National Initiative for aNetworked Cultural Heritage, is the U.S. distributor of a1997 U.K. report Discovering Online Resources Across theHumanities: A Practical Implementation of the Dublin Core.Contact ARL Publications for order information<pubs@arl.org>.

t-'PRESERVATION-----

RECON PROJECT FOR PRESERVATIONMICROFILM MASTERS COMPLETEDby Jutta Reed-Scott, Consultant for Preservationand Access Services

Over the past decade, ARL has managed a seriesof cooperative projects for creating more than579,000 online records for monographic and ser-

ial preservation microform masters that were listed inthe National Register of Microform Masters (NRMM).These projects have been funded by grants from theNational Endowment for the Humanities, Division ofPreservation and Access, as well as by an initial grantfrom The Andrew W. Mellon Foundation and majorcontributions by project participants. In December 199ZARL completed the final steps in the retrospective con-version of the NRMM. The NRM/VI reports are nowavailable online in the OCLC and RLIN databases andare accessible to the broadest possible community ofnational and international scholars.

Project HighlightsThe NRMM machine-readable records were createdaccording to national standards for national and inter-national distribution. With some minor exceptions, allreports included in the NRMM have been processed,representing:

549,147 monographs (over 12,000 in non-Romanlanguages);22,729 serials (Harvard, New York Public Library,and Library of Congress holdings); and7,083 musical scores.

A striking aspect of this recon project is that the newonline records describe titles held by more than 230 dif-ferent libraries and other institutions that contributedreports to the NRIVIM. The Library of Congress (LC),the largest noncommercial producer of microfilm, theNew York Public Library, as well as Harvard UniversityLibraries and several other ARL libraries played domi-nant roles in contributing records to the NRIVIM.Among vendors, University Microfilms International,General Microfilm Company, and Research Publicationswere major contributors. While large research librariesand vendors reported the majority of NRMM reports,an array of university and college libraries, historicalsocieties, archives, and state and public libraries alsocontributed a significant number of titles to the NRMM.

Equally diverse are the languages of the titlesdescribed by NRMM reports, with more than 200 differ-ent languages represented. As the accompanying chartillustrates, over 40% of the reports are in languagesother than English, concentrated primarily in WesternEuropean languages. All other foreign languages makeup about 10%. CJK (Chinese, Japanese, and Korean)reports were not converted by the project.

13

NRMM LANGUAGE DISTRIBUTION

English 57.5%

French 13.1%

Spanish 7.6%

Italian1.9%

Latin3.1%

Russian3.9%

Other 5.7%

German 7.2%

Not surprisingly, the NRMM reports reflect theclustering of publication dates in the late 19th and early20th centuries. More unexpected is that more than 16%of the reports show publication dates before 1850, withsome published as early as the 13th century.

Though these statistics cannot capture the vastscope of the materials described by records convertedthrough the NRMIvI Recon Project, they do documentthe extraordinary range of early preservationmicrofilming efforts.

Background and Project HistoryThe National Register of Microform Masters is itself alandmark in library cooperation. Keyes Metcalf firstproposed establishing a national register of microformmasters in 1936. The NRMM was finally begun in 1965as the direct result of a study conducted for ARL byWesley Simonton. I lis report summarized the well-known problems associated with rapid growth inpreservation microfilming without concomitant biblio-graphic control. The Library of Congress responded tothe call for action and took the lead in establishing theNRMM, the most significant effort by the library com-munity to collect, organize, and disseminate informa-tion about preservation microfilm masters. The NRMMMaster File, the largest single file of records for micro-form masters in the United States, consisted of reportsfor microform masters sent to the Library of Congressby hundreds of different libraries, historical societies,

IA R L 1 9 6 F E B R U A R Y 1 9 9 8

and publishers between 1965 and 1983. NRMM waspublished annually by LC until 1984, with one cumu-lation covering the years 1965 to 1975.

During its 18 years of existence, NRMM was thebest tool that could be produced for that time.NRMM's value to the library and scholarly communityis enormous, and the automation of NRMM makespossible a quantum improvement in the ease of useand ease of access to the records.

In 1986, with funding from NEH and matchingfunds from The Andrew W. Mellon Foundation, ARLand LC joined in the complex undertaking of creatingmachine-readable records of the NRMM reports.

The early road of the project was bumpy. Themajor problem was the default of the first project con-tractor in 1990. A second problem was the discoverythat the NRMNI file contained a significantly greaternumber of reports than had been estimated in the 1985sampling study. ARL overcame both problems.ARL first transferred the project to a new contractor,OCLC's Retrospective Conversion Department. Since1990, OCLC's staff have processed more than 518,000NRIVEM reports under an ARL contract. ARL alsosuccessfully obtained funding from NEH for process-ing the unexpected additional reports. While convert-ing the NRMM records took far more time and wasmore costly than initially projected, the goal of creatingmachine-readable records for all the NRMM reportsthat could be converted has been accomplished.

An accompanying chart (p.13) highlights progressduring the NRMM Recon Project.

Financing the ConversionMajor funding for the multi-year projects was provid-ed by the National Endowment for the Humanities.Over the life of the NRMM Recon Project, NEHawarded ARL a series of grants totalling $2,182,653.NEH's investment in the Project once more under-scores the vital importance of federal support fornational preservation efforts. The Library of Congressmade substantial contributions of technical assistanceand staff for the quality review throughout the pro-ject. A critical component of the project's financialsupport has been OCLC's in-kind contributions,which have substantially reduced the unit cost forcreating machine-readable records for the NRMM.OCLC's subsidy on behalf of research libraries hasexceeded $480,000.

Working TogetherThe Library of Congress was a vital partner in this pro-ject and played a unique role. While LC's primaryresponsibility was the establishment of the qualityassurance program, LC was also responsible for hold-ing the microform security copy of the reports and for

the distribution of project tapes. Most importantly,LC staff throughout the project provided advice on thetechnical specifications and conducted the extensivequality control.

In partnership with the Library of Congress, bothHarvard University and New York Public Library pro-vided significant staff support during the serials phaseof this project. NYPL also assisted in the review of itsnon-Roman records during the final project phase.

Many other libraries have also contributed to thesuccessful automation of the NRMM. At the end,every large-scale retrospective conversion projectconfronts a significant number of records that cannotbe processed within the standard workflow. For theNRMM project, these include reports with incompleteinformation, selected serials, and some non-Romanreports. Cards for records not converted werereturned to the institutions that created them for in-house review and processing, if required. Specifically,all East Asian reports have been turned over to institu-tions that created them for in-house processing.

ARL plans to set up a special web page for theNRMM Recon Project to provide detailed informationabout the scope and content of the automatedNRMM records and about the small percentageof NRMM reports that could not be processed aspart of the project.

Bibliographic Guidelines andQuality ReviewSpecial attention to quality control was a hallmark ofthis project. The project's initial bibliographical speci-fications were an important impetus for developingARL's Guidelines for Bibliographic Records forPreservation Microform Masters in 1990. Since then, theGuidelines have served as the project's bibliographicstandards.

Following agreed-upon criteria, LC staff membersreviewed a random sample of records in all batchesprepared by OCLC. The number of records in a batchand the interval between batches was determined byLC, ARL, and OCLC. Overall accuracy was set at aminimum of 94 percent. Each project phase includedan initial development period during which LC staffreviewed all records in each batch. For the productionperiods, the contract with OCLC specified that theentire batch of records would have to be redone if thebatch did not meet quality review accuracy require-ments. ARL never had to invoke this clause.

BenefitsThe automation of the NRMM files improves access topreservation microfilm masters. The linchpin of thenational and now international preservation effort isthe ability to showcase records of preserved titles.

A R L 1 9 6 F E B R U A R Y 1 9 9 8 14

Ongoing preservation programs that aim to reformatembrittled materials benefit significantly from theNRMM recon project because it makes it practical andeconomical to identify titles already filmed. For users,the online NRMM records expand access to specializedresearch materials that have been preserved.

The greatest benefit of the online NRMM lies inenhanced productivity. For all libraries, the machine-readable NRMM improves the efficiency of biblio-graphic searching and ensures against duplication ofmicrofilming. Not only was searching in the printed,multi-volume NRMM a time-consuming task (at leastone search in the 1965-75 volumes, and potentiallyeight additional searches in the annual volumes), butit was also easy to miss a title, since it is listed onlyunder a single access point. In contrast, the onlineNRMM makes possible a one-step search, which dra-matically reduces the amount of staff time needed forpre-filming searches. The danger of missing a record isalso minimized since the machine-readable records areaccessible by several access points.

Of equal benefit is the addition of thousands oforiginal bibliographic records for specialized researchmaterials to the national databases. The automation ofthe NRMM provided a large body of new cataloguerecords since many titles were not previously includedin OCLC or RLIN. That was especially true for the7,083 musical scores, of which 58% required originalrecords. Over time, the use of these converted NRMM

records for local cataloging will translate into reducedcataloging costs in libraries.

Future DirectionsOne challenge looms large: Providing rapid access tocopies of the microfilms. The user expects to get readyaccess, but several studies have documented the lengthydelays in obtaining filmed materials. While severalorganizations have explored centralized delivery ser-vices, the economic uncertainties of such services haveso far posed insurmountable bathers. However, theneed remains to improve interlibrary loan, documentdelivery, or other access services for materials preservedon microfilm.

In addition, digitization offers intriguing opportuni-ties to improve access to such resources. The technolo-gy of digitizing from microfilm has matured, but thefinancial resources needed to digitize more than half amillion preservation microfilm masters would be formi-dable. One intermediate strategy may lie in digitizingthe table of contents pages from selected volumes andmaking them available on the World Wide Web forbrowsing by researchers.

The NRMM Recon Project has contributed substan-tially to building the infrastructure for bibliographiccontrol of preservation microform masters. The nextstep is to deliver on a broader vision: To allow users toaccess electronically the information now preserved onmicrofilm.

NRMIVI RECON PROJECT MILESTONES

RECORDS PROCESSED CUMULATIVE #

OF RECORDSPROJECT PHASES

16,239 monographs (12,851 Non-Roman) 579 42 Phase VIJuly 1996-December 1997

22,729 serials and 7,083 musical scores 63 1 Phase VJanuary 1994-October 1996

54,000 monographs -533,374 Phase IVJune 1993-February 1994

160,000 monographs 47 374 Phase HIFebruary 1992-May 1993

258,000 monographs 319374 Phase IIJune 1990-March 1992

61,374 monographs 61 374 Phase IJuly 1987-June 1990

15 A R L 1 9 6 F E B R U A R Y 1 9 9 8 3

Prudence S. Adler, Assistant Executive Director-Federal Relations and Information Policy

CHECKLIST FOR DRAFTING ELECTRONICINFORMATION POLICIES

In today's emerging electronic information environ-ment, characterized by a panoply of networks, soft-ware, and information content, it is important that

every research university and other institutions prepareand disseminate policies concerning the use, creation,and exchange of electronic information. Electronic infor-mation policies, though requiring unique elements, oughtto be an extension of existing information policies. Theseelectronic policies should describe roles and responsibili-ties of users and providers, and should address appropri-ate behaviors, not only on campus systems, but also onthe World Wide Web and the Internet generally.

The following is a checklist of issues to consider whendrafting electronic information policies. The checklist,prepared by the Information Policies Committee of theAssociation of Research Libraries, is designed as a guidefor universities and other institutions that are developing,reviewing, or revising electronic information policies.

1. An introductionAn introduction should state the need for and purpose ofthe policy, and contains an explanatory statement aboutthe underlying values and principles guiding the policy.

ValuesRespect for freedom of inquiry and expression, and avigorous opposition to censorshipRespect for the rights of others, especially rights ofprivacy and confidentialityRespect for intellectual propertyRespect for all members of the community of users,including minorsRespect for appropriate conduct in a public forum,including civility, respect for others, and respect fordiversityRespect for law, for due process, and for thepresumption of innocenceOther values of local importance

PrinciplesEveryone using university information resources isexpected to honor the values of her or his academic com-munity, whose existence makes the use of those resourcespossible. Every user is also expected to be considerate ofthe needs of other users by making every reasonableeffort not to impede the ability of others to use thoseresources. Examples of infringements of these principlesmay include, but are not limited to, activities that:

obstruct usage or deny access to otherscompromise privacy or confidentialityinvolve unauthorized use of computer accountsattempt to modify system facilities withoutauthorization

allow unauthorized access to private accountsinvolve unauthorized use of resources for commercialenterprisesviolate intellectual property rightsviolate license agreementsviolate university or institutional rules or policiesviolate local, state, or federal laws

2. Access issuesDelineation of who is assured access to what resourcesand within what time framesDifferent levels of access that are provided to differentcategories of usersUse of public facilities (e.g., computing labs)Circumstances under which private equipment can beattached to publicly-owned or grant-funded equipmentWhen access can be denied or revoked

3. Electronic mail issuesDelineation of when and to whom the institution candeny access to its electronic mail servicesThe circumstances under which the institution caninspect, monitor, or disclose electronic mail, and theconditions under which this can be done without theholder's consentRestrictions on personal useRestrictions on commercial, political, or religious useUse for union activitiesUse of anonymous mail, chain letters, mail reflectors,and listservsForwarded mail after someone has left the institutionArchivingPublic recordsLibel and obscenityAdditional issues raised because of the presenceof the institutional identifier within the domain portionof an electronic mail address

4. Websites and Web publishing issuesPolicies governing content on a websitePersonal homepages, if permitted (questions of contentneed to be managed carefully)Restrictions on pointers and links to other websitesRestrictions on the use of copyrighted material onwebsitesUse of university or institutional logos and trademarkson websitesDefinition of an "official" website of the institution, anda description of how such a designation is grantedLevels of use and trafficRestrictions on anonymous websitesUse of commercial servicesInternet gamblingLibel and obscenityAdlirrce to all local, state, and federal laws

ARL 196 FEBRUARY 1998

5. Network use issuesInappropriate use of the network (e.g., IP addressspoofing, sniffing, and other inappropriate actions)Limits on amount of useExpectations of the user concerning network security

6. Coordination on campusCampus units that may be involved in the process ofwriting the policies include:

Campus LibrariesInformation TechnologyAcademic ComputingAcademic AffairsLegal CounselPublic RelationsFacultyStudent Affairs

7. ArchivingThe policy should include strategies for archiving or pre-serving electronic resources. All relevant campus or insti-tutional units, such as Computing, Libraries, and Archives,should be consulted in order to determine the attendantresponsibilities and procedures for electronic archiving.

8. Due process issuesAn information policy should identify the forms ofredress available if policy violations occur and theprocesses that will be followed. It should delineatebetween those actions that transgress university rules andregulations concerning appropriate behavior and thoseactions that may involve violations of the law. In someinstances, such as child pornography, an institution'sdiscretionary latitude may be severely restricted byfederal statutes.

9. DefinitionsIt will be important for each campus or institution todetermine relevant definitions (e.g., user, authorizeduser) and to include those definitions in the policies.

10. Implementation and reviewInstitutions ought to set forth a process for implementa-tion and review, and determine how often the policy getsreviewed or revised. It is important that the policy beclear and unambiguous, and able to be understood by allusers of electronic resources. Moreover, the policy shouldbe made readily available to all usersfor example, post-ed at public terminals and on institutional websites.

Prepared by Paula T. Kaufman (University of Tennessee,Knoxville) and Gerald R. Lowell (University of California,San Diego) with the ARL Information Policies Committee.

IT

SAMPLE ELECTRONICINFORMATION POLICIES

AUBURN UNIVERSITY

http://www.auburn.edu/its/guide/policies.html

UNIVERSITY OF CALIFORNIA, DAVIShttp:/ /www.ucdavis.edu/AUP.html

UNIVERSITY OF CALIFORNIA, SAN DIEGO

http://www-acs.ucsd.edu/main/instsupp.html

UNIVERSITY OF CALIFORNIA, SANTA BARBARA

http://www.ucsb.edu/policy.shtml

UNIVERSITY OF GEORGIA

http://www.uga.edu/-ucns/sites/use.html

GEORGIA INSTITUTE OF TECHNOLOGY

http: / / www.gatech.edu /ifis /policy/ usage /contents.html

UNIVERSITY OF HOUSTONhttp: / /www.uh.edu/info_serv/users_guide/

guidelines.html

INDIANA UNIVERSITY

http://infotech.indiana.edu/policy/policy.html

JOHNS HOPKINS UNIVERSITY

http://www.jhu.edu/www/jhuniv/guidelin.html

UNIVERSITE LAVAL

http:/ /www.ulaval.ca/sg/reg/Politiques/index.html

LOUSIANA STATE UNIVERSITY

http://www.lsu.edu/OCS/homedocs/usage_policy.html

UNIVERSITY OF NEW MEXICO

http://www.unm.edu/cirt/info/general/ethics.html

NORTHWESTERN UNIVERSITY

http://www.nwu.eduM/policies

UNIVERSITY OF TENNESSEE KNOXVILLE

http://www.cas.utk.edu/CAS/casccp.html

ARL 196 FEBRUARY 1998 15

1

OMS RENAMED OFFICEOF LEADERSHIP ANDMANAGEMENT SERVICES

Effective January 1998, the ARL Office ofManagement Services (OMS) adopted the newname of Office of Leadership and Management

Services (OLMS). The new name corresponds to thechange in name for the ARL Committee on Leadershipand Management. The new title is intended to conveythe focus of the Office on advancing both the discourseand practice of shaping research library and informationservice organizations of the future. Use of the new titlehas begun and will be phased in during the year on pre-printed materials.

Also in January, the search for a Director of ARL'sOffice of Library and Management Services was sus-pended for at least six months. Changing circumstancesat ARL have prompted a broad review of the staffingrequirements at the Association over the short- andlong-term. OLMS will continue to operate under thevery able leadership of Kathryn Deiss, Senior ProgramOfficer for Training and Leadership Development.

EVALUATING THE PERFORMANCE OFTHE ACADEMIC LIBRARY DIRECTOR

Att he request of the ARL Committee on ResearchLibrary Leadership and Management, ARLundertook, in the fall of 1997, a study of the

performance evaluation processes for ARL librarydirectors. The study was conducted by George Soete,ARL/OLMS Organizational Development Consultant,with the advice of Paul Kobulnicky, Director ofLibraries, University of Connecticut, and SarahMichalak, Director of Libraries, University of Utah.The study entailed a survey of ARL Directors who wereasked to describe and assess current practices, to pro-vide suggestions for how the processes may be mademore effective, and to share relevant documentationfrom their institutions.

Snapshots of the survey findings follow.Formal reviews of Directors are now a permanentand fast-growing feature of the human resourcessystems of the parent institutions of ARL librariesof the 74 directors participating in the study, 84%have a formal performance review.Most reviews (74%) were initiated and conducted bythe person to whom the director reported, typicallythe provost or a vice-president, though this adminis-trator often involved other participants in theprocess.Director reviews are tending to have broader partici-pation, with the most frequently named participants

being librarians and library support staff, non-librarianfaculty, and other academic administrators.A majority of directors indicated there were no formalguidelines or criteria for the review processoften theseare developed ad hoc during the review.Only 21% of review processes had a salary decision asan outcome; more than half (52%) reported that contractrenewal or reappointment was an outcome related tothe process.

As part of the study, Mr. Soete has also prepared ashort document to guide those who are responsible for orwho will participate in the performance evaluations of aca-demic library directors. It is written as a checklist and isintended to foster both effectiveness and fairness in direc-tor evaluations and to strengthen library leadership as partof the overall process of strengthening our libraries.

The full study findings will be reported in two publica-tions forthcoming from the ARL: SPEC Kit #230 will pre-sent the results of the survey describing current processesand documentation from member libraries, and OLMSOccasional Paper #21 will address analyses and evaluationsof current practices and provide guidance for improvingthese processes.

OLMS CALENDAR OFTRAINING EVENTS1998March 10-13LIBRARY MANAGEMENT SKILLS INSTITUTE I:THE MANAGERSeattle, WA

April 28-30HUMAN RESOURCES INSTITUTEBaltimore, MD

May 4-7LIBRARY MANAGEMENT SKILLS INSTITUTE II:THE MANAGEMENT PROCESSLos Angeles, CA

October 7-9FACWITATION SKILLS LNSTITUTEWashington, DC

October 27-28LEADING CHANGEChicago, IL

November 16-19LIBRARY MANAGEMENT SKILLS INSTITUTE I:THE MANAGERWashington, DC

For registration information contact Christine Seebold via email<cseebold@arl.org> or phone (202) 296-8656.

A R L 1 9 6 F E B R U A R Y 1 9 9 8 1 8

De Etta Jones, Diversihj Program Officer

LEADERSI-BP AND CAREERDEVELOPMENT PROGRAM MATCHESPARTICIPANTS, RESEARCH TOPICS,WITH MENTORS

Work is underway on the research projects pro-posed by each of the 21 participants in ARL'sLeadership and Career Development Program.

Listed below are the topics and designated mentor of eachLCD Program participant. A program is planned for June1998, scheduled in conjunction with the ALA AnnualConference, where participants will present their projectsto their peers, mentors, and the larger library community.

Participant Maria de JesusAyala-SchuenemanAssociate ProfessorHead of Reference ServicesTexas A&M University-Kingsville

Research Project The Examina-tion of Reference Services asThey Relate to Faculty

Mentor: Emily Mobley, PurdueUniversity Library

Participant: Jon CawthorneReference/Coordinator forOutreach ServicesUniversity of Oregon

Research Project Spoke for theLibrary: Academic LibraryOutreach

Mentor: Paula Kaufman,University of Tennesseelibraries

Participant Vicki ColemanInstructor/Head, Electronic

Reference ServicesTexas A&M University

Research Project Public Servicesfor the Virtual Library Patron

Mentor: Dana Rooks, Universityof Houston

Participant: Patrick Jose DawsonLibrarian Ill ReferenceCoordinator,Head, Collección TloqueNahuaqueUniversity of California -Santa Barbara

Research Project Cataloging andDescription of DigitalResources

Mentor: William Gosling,University of MichiganLibrary

Participant: Tracey Joel HunterReference/ CollectionDevelopment LibrarianTemple University

Research Project A Study of theInformation Seeking Behaviorof African-American Students

at an Urban ARL Library -Temple University

Mentor: Shirley Leung,University of California-IrvineLibraries

Participant Glendora Johnson-CooperAssociate LibrarianUniversity of Buffalo

Research Project ReferenceServices and Digital Demands

Mentor: Nancy Eaton,Pennsylvania State UniversityLibraries

Participant: Kuang-Hwei (Janet)Lee-SmeltzerCopy Cataloging ManagerUniversity of HoustonLibraries

Research Project: A Survey ofSome Enhancements to theDublin Core and Their Effecton Access to InternetResources.

Mentor: Jennifer Younger,University Libraries of NotreDame

Participant Barbara (Barbi) LehnDirector of Library ServicesSinte Gleska University

Research Project To develop adocument for publication that isboth a collection developmentpolity to guide acquisitions andan evaluative tool to measureprogress toward the fulfillmentof the library's mission.

Mentor: Sherrie Schmidt, ArizonaState University Library

Participant Poping LinAssistant Engineering &Science Librarian for CoreInformation CompetenciesMassachusetts Institute ofTechnology

Research Project The Integrationof Information Competencieswith the Training of InnovativeEngineers

Mentor: Miriam Drake, GeorgiaInstitute of Technology Library

Participant Nerea Llamas'Assistant LibrarianUniversity of California-SantaBarbara

Research Project: Evaluating theInstruction need of Facultyand Students in theHumanities

Mentor: Joseph Branin, SUNY -Stony Brook Libraries

Participant Johnnieque(Johnnie) Blackmon LoveReference/Cultural DiversityLibrarianUniversity of Kansas

Research Project Designing aProgram Evaluation Model forAssessing Effectiveness ofDiversity Programs in anAcademic Library

Mentor: Joan Giesecke,University of Nebraska-Lincoln Libraries

Participant: Thura MackAssociate ProfessorUniversity of Tennessee

Research Project ProvidingResearch and DevelopmentServices for Selected ScientificFields Using Intelligent Agents

Mentor: Pamela Andre, NationalAgricultural Library

Participants: Teresa NeelyAssistant ProfessorCSU/Teaching FellowColorado State University

Research Project: Promotion andTenure and ElectronicPublications

Mentor: Scott Bennett, YaleUniversity Library

Participant: Neville DurrantPrendergastAssociateLibrarian/CoordinatorInformation ManagementEducationUniversity of Buffalo

Research Project: TheDevelopment of A"TeachingLibrary" Within the HealthSciences Library

Mentor: Stella Bentley, AuburnUniversity Libraries

Participant: Gloria RhodesMulticultural OutreachLibrarianCalifornia State University -San Marcos

Research Project InformationCompetencies: ABenchmarking Project

Mentor: Ann Prentice, Universityof Maryland-College Park, SIS

Participant: Carlos RodriguezScience Reference LibrarianUniversity of California -Riverside

Research Project: HowTechnology Shapes theQuality of UndergraduateResearch

Mentor: Camila Alire, ColoradoState University Library

Participant Janice Simmons-WelburnCoordinator Personnel andDiversity ProgramsUniversity of Iowa

Research Project A Study on theRelationship BetweenAcademic Libraries and theGoals on Diversity and theGoals of Their ParentInstitutions

Mentor: Carla Stoffle, TheUniversity of Arizona Library

Participant Denise StephensGeographic InformationcoordinatorUniversity of Virginia Library

Research Project IntroducingEmerging Technology asService: Implications forLibrary

Mentor: Susan Nutter, NorthCarolina State UniversityLibraries

Participant Elayne WalstedterReference/OutreachSpecialistFort Lewis College

Research Project: Recruitment ofNative Americans and Latinosinto the Library Profession,Primarily in the AmericanSouthwest

Mentor: Nancy Baker,Washington State UniversityLibrary

Participant Valerie WheatBranch LibrarianSmithsonian Institution

Research ProjectMuseum/LibraryCollaborations: A NaturalCultural Partnership

Mentor: Barbara von Wahlde,State University of New Yorkat Buffalo

Participant Mark WinstonAssistant University librarianValdosta State University

Research Project: A ResearchStudy of Science andEngineering Librarians:Recruitment, Demographics,and Professional Activities

Mentor: James Williams Ill,University of ColoradoLibraries

ARL 196 FEBRUARY 1998

I, ACTIVITIESG. Jaia Barrett, Deputy Executive Director

ARL DIRECTORY TRACKSGROWTH IN E-PUBLISHINGby Dru Mogge, ARL Electronic Services Coordinator

Now in its seventh edition, the ARL Directory ofElectronic Journals, Newsletters and AcademicDiscussion Lists continues to document the

growth in electronic publishing. This year's Directoryincludes over 3,400 titles, a 100% increase over lastyear's count of 1,689. The number of e-journals (whichincludes titles classified as "e-zines," or magazines)make up 72% of the total, with 2,459 listings, whilee-newsletters account for 955 entries. The lists section(electronic conferences, including discussion lists, newsgroups, interactive web chat groups, etc.) also grew,from 3,118 entries last year to 3,808 this year, a 22%increase.

One of the striking increases is in peer-reviewedtitles. In the first edition of the Directory, published in1991, 6.4% of the titles listed were identified as peer-reviewed; in 1993, 13.2%; in 1995, 20.6%; and in 1997,30.7% of the titles in the Directory are so identified.Electronic scholarly publishing is growing as more andmore traditional publishers find ways to make theirpublications available online. While counts are notavailable from the Directory for "electronic only" publi-cations, the trend is to continue to make both print andonline versions available.

A sign that commercial publishers are joining thetrend is the increase in fee-based titles. In 1991, onlytwo serial titles in the Directory charged for access. In1995 and 1996, just 10% of the Directory serial titlescharged a fee for online access; this year, almost 27%(912 titles) charge a fee.

Both the e-journals and the lists that make upthe Directory were analyzed for subject content andkeywords were assigned to entries based on a thesauruscreated especially for the Directory project. The piecharts show the distribution of subjects acrosse-journals, e-zines, e-newsletters, and lists.

A feature added this year is a complete web versionof the Directory. The online version offers users the abili-ty to browse through individual entries or to search forspecific items. Search options include searching by title,description, publisher, peer review basis, or subject.Also included online is the thesaurus, thereby allowing

TRENDS IN E-SERIALSE-Journals andE-Newsletters 1991 1992 1993 1994 1995 1996 1997Total 110 133 240 443 675 1689 3414

Peer-Review 7 15 29 73 139 417 1049Fee 2 2 6 29 72 168 912

Based on entries for e-journals and e-newsletters included in annualeditions of the ARL Directony

18 A R L 1 9 6 F E B R U A R Y 1 9 9 8

Technology21%

Social Sciences40%

NEWSLE11ERSArts and Humanities 6%

Life Sciences16%

PhysicalSciences6%

Recreation &General Interest11%

JOURNALSTechnology 1011mints and Humanities 13%

SocialSciences24%

Recreation &General Interest4%

Physical Sciences 13%

ZINES/OTHERArts and Humanities

Technology 179 21%

)Life Sciences36%

SocialSciences23%

Technology 19%

SocialSciences42%

LISTS

Life Sciences7%

PhysicalSciences 2%

Recreation &General Interest 30%

Arts and Humanities21%

20

Life Sciences10%

PhysicalSciences 5%

Recreation &General Interest 3%

users to search by specific keywords. All web-accessiblee-serials in the Directory have a link from the entry to thetitle's actual site. The electronic version of the Directoryis available as a stand-alone product, while purchasers ofprint copies automatically receive access to the e-version.The e-version is updated on an ongoing basis.

Each year, ARL chooses a particularly relevant ornoteworthy article on electronic scholarly publishing forinclusion in the Directory. The 7th Edition article isJudy Luther's "Full Text Journal Subscriptions: AnEvolutionary Process." Previously published in theJune 1997 issue of Against the Grain, Luther's articlereviews options offered to libraries by commercialpublishers, subscription agents, and not-for-profitpublishers. Luther addresses issues such as licensingpackages and pricing structures, as well as varioustypes of format.

As in previous editions, the Directory is divided intotwo major sections. Section One, Electronic Journalsand Newsletters, was compiled by Dru Mogge, ARLElectronic Services Coordinator, and the ARL DirectoryStaff. Section Two, Academic and ProfessionalDiscussion Lists, was compiled by Diane K. Kovacs,Kovacs Consulting, and the Directory Team. TheDirectory is available for purchase as either a print andelectronic package (ARL member libraries, $65; non-members, $95) or electronic access only (ARL memberlibraries, $50; nonmembers, $70). Consult either thewebsite <http:/ /www.arl.org/scomm/edir/> or email<pubs@arl.org> for order information.

TRANSITIONSUC-Berkeley: Peter Lyman announced his resignationas University Librarian, effective July 1, 1998. He willbecome Professor in the Berkeley School of InformationManagement and Systems full-time.Case Western: Joanne D. Eustis was named Director ofLibraries, effective August 1, 1998. She is currentlyDirector of Planning and Program Review, Departmentof Information Services at Virginia Polytechnic Instituteand State University.ARL Staff Transitions: In July 1998, ARL DeputyExecutive Director Jaia Barrett will leave Washington,DC to accompany her husband to Kiev, Ukraine, wherehe has a two-year assignment at the U.S. Embassy. Shewill use the network to continue to work for ARL dur-ing this period. Ms. Barrett came to ARL from DukeUniversity in 1984 as Federal Relations Program Officer,a position she held until 1989, when she took a leave ofabsence. On her return in 1991, she took on manage-ment of ARL's Access Program, the ARL newsletter,and the Office of Research and Development. In 1993she was named Deputy Executive Director.

NAILDD SALUTES DIGMEMBERS' ACHIEVEMENTS

The 1998 ALA Midwinter Meeting in New Orleansmarked the beginning of the five-year anniversaryof the ARL North American Interlibrary Loan and

Document Delivery (NAILDD) Project. The goal of theProject is to promote ILL/DD system developments thatwill improve the delivery of library materials to users atcosts that are sustainable to libraries. To seek theinvolvement of the private sector, NAILDD formeda Developers/Implementors Group (DIG); over 70for-profit and not-for-profit organizations and projectsnow participate.

A five-year status report for NAILDD highlightsand salutes the responsiveness of some of the moreactive DIG members to advance the Project's technicalpriorities. The organizations highlighted in the reportare: OCLC; RLG; The Library Corporation; Ameritech;A-G Canada; Innovative Interfaces, Inc.; NetworkSupport, Inc./Relais International; the National Libraryof Canada; and the British Library Document SupplyCentre. The sustained commitment of these organiza-tions, and the resources they invest to achieve theProject's goals, are acknowledged as key contributionsto the future success of library programs and services.The report was distributed during the ALA MidwinterConference and is published on ARL's website <http:/ /www.arl.org/access/naildd/overview/statrep/statrep-9801.shtmb.

While progress was made during the last five years,work still remains. The NAILDD Project will be fullysuccessful when a library user is able to:

have transparent access to a variety of local andremote library catalogs, citation databases, andelectronic resources,transfer bibliographic citations or details about non-bibliographic items into electronic requests or orders,pass requests or orders through the library onlinesystem to determine local availability,direct the request or order to the ILL/DD departmentof a library or to a document supplier,communicate electronically with the chosen supplier,andreceive print materials, multimedia, data, or fulltext/image copy of non-returnables directly at theirdesk or workstation.

For information about the NAILDD Project contactMary E. Jackson, ARL Access & Delivery ServicesConsultant <mary@arl.org> and see ARL's website<http:/ /www.arl.org/access/naildd/naildd.shtml>.

21 A R L 1 9 6 F E B R U A R Y 1 9 9 8

- ,

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-296-2296 FAX 202-872-0884<http://www.arl.org/>Copyright: © 1998 by the Association of Research Libraries

ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

ARL CALENDAR

1998March 19-20

April 14-15

April 15-17

May 12-15

June 25- July 2

July 27-28

October 13-16

October 13-16

December 8-11

ARL Licensing WorkshopChapel Hill, NC

Coalition for NetworkedInformation Spring Task ForceMeetingArlington, VA

Net '98Washington, DC

ARL Board andMembership MeetingEugene, OR

American Library AssociationWashington, DC

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingWashington, DC

Educom '98Orlando, FL

CAUSE '98Seattle, WA

Executive Director Duane E. WebsterEditor: G. Jaia Barrett, Deputy Executive DirectorCopy Manager: Karen A. WetzelDesigner: Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscription; Nonmembers$50 per year.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arLorg>.

1999February 11-12

May 11-14

July 26-27

October 12-15

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingKansas City, MO

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingWashington, DC

OLMS 1998CALENDAR OF TRAINING EVENTS

SEE PAGE 16.

22

1

A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONS

Special Issue:

ISSUES IN RESEARCH LIBRARY MEASUREMENTby Julia C. Blixrud, Senior Program Officer

performance measures, quality assessment,public accountability, benchmarkingthesehave become common words and phrases in

higher education and government literature in the1990s. The environment in which ARL membersand many other libraries operate has changedfrom one of natural acceptance of value by virtueof function to one in which all units must substan-tiate their worth. This is no easy task. For muchof research library history, the functions of build-ing, housing, and making collections availablewere what libraries were expected to do.Quantitative and extensiveness measures werethe means by which libraries were measured. Aslong-time readers of this newsletter know, ARLhas a history of providing descriptive (i.e., quanti-tative) data about Association members. Thesedata are, in fact, the oldest and most comprehen-sive continuing library statistical series in NorthAmerica and are widely used for tracking trendsin scholarly communication. Today, those samemeasures are also often seen as negative pressureson libraries to acquire printed materials in an agewhen resource sharing and access to electronicinformation are so prevalent.

Most recently, a Pew Higher EducationRoundtable, co-sponsored by ARL and theAssociation of American Universities, encourageduniversities and their research libraries to gradu-ate "from a mindset that accords status and pres-tige by 'the tonnage model'the sheer number ofvolumes and subscriptions a single collection con-tains." While the Roundtable stopped short ofoffering an alternative for quantitative measures,

23

it is a vivid example of how ARL's descriptivedata, absent any other measures, becomes adouble-edged sword.

For the last five years, the ARL Statistics andMeasurement Program has therefore beenengaged in efforts to investigate new measures.In a 1992 article, "New Directions for ARLStatistics," Sarah Pritchard, then ARL AssociateExecutive Director, reminded us that, thoughthere is a need to find new ways to assess libraryperformance,"some things are not measurable, areirrelevant or too difficult to measure, or are onlymeaningful in a local context" (ARL 161).Nonetheless, we have not been deterred from con-tinuing our investigation, with the hope of siftingthrough the possible measures in search of thosethat are both relevant and meaningful in either alocal context or in comparison with other libraries.

This issue provides a snapshot of issues andactivities in the area of performance measure-ment. However, as space and resources limit usfrom being able to provide a census of all knownmeasurement activities here, the ARL Statisticsand Measurement Program has built a website onperformance measures, available at <http:/ /www.arl.org/stats/perfmeas/index.html>, inorder to more fully account for activities in thisarea. The site includes a bibliography and linksto other information resources on this topic, suchas an annotated bibliography originally preparedfor those attending a session on performancemeasures at the April 1997 ACRL NationalConference. Readers are invited to suggestadditions for the site, especially examples of use

Continued

of new performance measures. Plans are also beingmade to hold an ARL conference on performancemeasures in the future.

This issue begins with an article by MarthaKyrillidou, Senior Program Officer for Statistics andMeasurement, providing a context for the increasedinterest in performance measures within the higher edu-cation community. This article, adapated from ARL'sannual publication, Developing Indicators for AcademicLibrary Performance: Ratios from the ARL Statistics, alsomakes recommendations regarding new performanceindicators that research libraries might consider.

The article, "In Search of New Measures," empha-sizes the need to balance continuing and emerging reali-ties in the assessment of libraries. Written by WilliamCrowe, University of Kansas, and Martha Kyrillidou,this article documents how ARL's Statistics andMeasurement Program is responding to the need fornew measures while retaining the important descriptivedata collection activities that have long served to high-light research library trends.

One area of particular interest is in the measure-ment of electronic resources. Tim Jewell, University of

CONTENTSSPECIAL ISSUE ON MEASURES

An Overview of Performance Measures inHigher Education and Libraries 3

In Search of New Measures 8

The ARL Membership Criteria Index 9

Interactive Peer Group ComparisonsThrough the Web 11

Understanding Electronic Resourcesand Library Materials Expenditures:An Incomplete Picture 12

Assessing the AcademicNetworked Environment 14

The State of Performance Measurementin Libraries: A Report from the2nd Northumbria InternationalConference 16

Washington and ARL Visiting Program Officer, andJulia Blixrud, ARL, describe an ARL project to trackinstitutional and library investments in electronicresources. This project, supported by the Council onLibrary and Information Resources, has been usingARL's supplementary statistics questionnaire as atestbed to gather data in this emerging statistical area.

Kendon Stubbs, University of Virginia, describeschanges made to the interactive website for ARL statis-tics, hosted by the University of Virginia. He furthersuggests ways in which these data can be used for localcomparative purposes.

Libraries are not the only part of the research insti-tution looking at how best to assess performance; othercampus officials are interested, as well. Joan Lippincott,Associate Executive Director, Coalition for NetworkedInformation, describes individual institution efforts toassess networked information services for the CNImeasurement project, an outgrowth of the publicationAssessing the Academic Networked Environment:Strategies and Options.

Much of the activity for performance measuresdevelopment is taking place outside the North Americanborders. Amos Lakos, University of Waterloo, attendedthe second Northumbria International Conferenceon Performance Measurement in Libraries andInformation Services on behalf of ARL this past fall.His report summarizes some of the conference's papersand discussions that included areas to which authorssuggested libraries should focus their attention.

ARL continues its exploration of performancemeasurement for research libraries. The articlespresented here are intended to stimulate interest andprovide information about these difficult issues. Theassessment movement in higher education has beensteadily gaining momentum, and several states havealready begun linking funding to performance (see"Assessment Policies in Higher Education," ChangeEMar./Apr. 19981). We in libraries have been successfulin using quantitative and extensiveness measures todifferentiate collections and services, justify fundingincreases, and describe libraries for many years, butcomparative effectiveness measures have remainedelusive. We must take up the challenge to developour own measures or they will be determined for usby others.

24

A R L 1 9 7 A P R I L 1 9 9 8

AN OVERVIEW OF PERFORMANCEMEASURES IN HIGHER EDUCATIONAND LIBRARIESby Martha Kyrillidou, Senior Program Officerfor Statistics and Measurement

A.n April 4, 1997 article in The Chronicle of Higher

Education reported that the South CarolinaGeneral Assembly approved a law instituting a

system in which state appropriations to a public collegewould be based on how well the institution performs.'That action is one of many pieces of evidence that highereducation in North America is being pressed for greateraccountability and improved attention to quality.Legislators in many states are moving toward perfor-mance incentives based, at least in part, on whether uni-versities and colleges are accomplishing stated goals.A public concerned with the balance between costs andbenefits of higher education demands more informationon institutional operations and outcomes. In particular,there is a great need to demonstrate the extent to whichinstitutions are meeting their goals and objectives, andwhether these goals and objectives are aligned with soci-ety's needs. A plethora of "useful" measures and otherefforts has flooded the literature of higher education.Ultimately, it is the responsibility of each institution todefine and describe its own goals, to place them in thecontext of peer group comparisons, and to demonstrateto the public the position it holds in higher education.

The concepts of accountability and quality assess-ment in higher education constitute an internationalphenomenon. National education systems call uponuniversities to establish performance indicators tomeasure progress towards the establishment of nationalgoals. Universities increasingly are asked to describe inspecific terms their contribution towards the nationalwelfare and the relation between the welfare of a coun-try and university teaching and research. In Europe andAustralia, central governments are involved directly inestablishing "indicators." In the United Kingdom, forexample, quality control, quality audit, and qualityassessment are being carried out by the HigherEducation Quality Council and the three HigherEducation Funding Councils. A new central agency togather and analyze data, the Higher Education StatisticsAgency (HESA), has also been established. More specifi-cally, library performance indicators have flourished inthe United Kingdom as the restructuring of the Britishhigher education system proceeds.' The EuropeanCommission has been supporting an effort to create areliable statistical base for libraries in Europe. InDecember 1997, the Commission hosted a workshop tofocus attention on statistics that address service quality.3

In the U.S., there have been discussions about a

25

greater federal role in institutional accreditation or if sucha system might be based on "results" and "performance."Whether it is the federal government or some other entitythat will undertake the responsibility to define "quality"for higher education in the U.S., critics of higher educationhave warned that, if "the academy does not respond, thepublic appetite for results will expand and crystallizearound the use of external performance indicators to mea-sure results. And the jury is still out on the resultsdesired."' To some extent, this is already happeningthrough the crude but widespread ranking systems thatpopular magazines like U.S. News and World Report are pro-moting. In the 1997 issue dedicated to ranking colleges, theeditors point out that "the nation cannot afford to let high-er education become less and less affordable for more andmore students. The high cost of college is no longer just anacademic affair; it is a national concern as well."'

A recent report that presents the results of a two-yearstudy by the Commission on National Investment inHigher Education highlights the fact that the "presentcourse of higher educationin which costs and demandare rising much faster than fundingis unsustainable."The authors call upon the "nation to address the fiscalcrisis now, before millions of Americans are deniedaccess to a college education" and they recommend"increased public-funding of higher education andwide-ranging institutional reforms." In particular,they articulate the following five recommendations:

America's political leadersthe President,Congress, governors, mayors, and other state andlocal officialsshould reallocate public resourcesto reflect the growing importance of education tothe economic prosperity and social stability of theUnited States.Institutions of higher education should makemajor structural changes in their governancesystem so that decision makers can assess therelative value of departments, programs, andsystems in order to reallocate scarce resources.As part of the overall restructuring, colleges anduniversities should pursue greater missiondifferentiation to streamline their services andbetter respond to the changing needs of theirconstituencies.Colleges and universities should develop sharingarrangements to improve productivity.It is time to redefine the appropriate level ofeducation for all American workers in the 21stcentury. All citizens planning to enter theworkforce should be encouraged to pursueas aminimumsome form of postsecondary educationor training.'

To some extent, these recommendations are theresult of a fundamental societal transformation from the

A R L 1 9 7 A P R I L 1 9 9 8

Continued

Industrial Age to the Information Age and the correspond-ing challenges and opportunities it presents for highereducation. Performance measures are becoming themethod of choice to track for the transformation of highereducation. Critics are calling for the development of acompelling vision for learning in the 21st century, a visionthat would transform higher education by realigningit with three conditions: "1) the changing nature of infor-mation, knowledge, and scholarship; 2) the needs ofindividual learners; and 3) the changing nature of workand learning."'

In the discussion regarding performance indicators inthe U.S., the primary focus has been on cost efficiency andaccess to undergraduate education as well as on the long-term transformation of higher education and its effect ongraduate education and research. There is a real push forhigher education institutions in the U.S. to be judged by adirect, observable connection to the countly's econbmicwelfare. Contemporary indicators that point toward thistrend are: a stronger emphasis on scientific and technicaleducation; efforts towards better management of the intel-lectual property produced at universities; investments inthe Next Generation Internet (NGI);8 private initiatives,like Internet2;9 and the privatization of the NationalInformation Infrastructure (NH).

Performance Indicatorsin Academic LibrariesAcademic research libraries also feel the pressures thathave resulted from the shift from a management systemaccustomed to increased revenue and growth to systemsthat demand more evidence of efficiency and effective-ness, accompanied by fundamental transformations.A 1992 study conducted by The Andrew W. MellonFoundation analyzed the economic trends of researchlibraries in the context of the larger academic and pub-lishing trends and identified historical and technologicalchallenges that affect and transform academic libraries!'The Mellon study found that the explosion in the quanti-ty of desirable published material and a rapid escalationof unit prices for those items jeopardizes the traditionalresearch library mission of creating and maintaininglarge, self-sufficient collections for their users. The studyalso recognized the potential of information technologiesto transform the ways libraries organized collections andservices. Updates of these trends are charted and pre-sented through the annual publication of ARL Statistics."

As in higher education, libraries have also recognizedthe need for "output and performance measures." ARLresponded to these calls by including data on circulation,instructional sessions, and reference transactions, togeth-er with interlibraiy loan and document delivery statisticsin the supplementary portions of ARL Statistics. Despitesome concerns about the validity and reliability of suchmeasures, these measures were added to the main ARL

Statistics in 1995. In 1994, ARL also began distributing anannual report on selected ratios!' Efforts here are devel-oping, in both senses of the word, i.e., they are stillprimitive and under development.

Institutional data collected through the ARL Statisticshave also been packaged into an electronic publicationthat offers interactive statistical analysis through whichone can compute any conceivable ratio or performanceindicator based on the data of the collected variables!'The interactive electronic edition of the ARL Statistics,prepared by the Geospatial and Statistical Data Center atthe University of Virginia, can best be described as a basicdecision support system (DSS) that can answer questionsmanagers may have at the cross-institutional level, forinstance, by comparing the performance of one institutionto another or to a peer group through a variety of simple(ratio analysis) or complex (multivariate analysis) statisti-cal techniques (see Kendon Stubbs' related article).

In addition to ARL's efforts, there have been a numberof projects by other organizations that have tried to devel-op indicators or "benchmarks" for academic library opera-tions, oftentimes within a larger institutional framework.It is important that, as such library indicators are devel-oped, they address the strengths and weaknesses of thedifferent measures.

Two important activities are taking place at theinternational level. One that does not limit itself toacademic libraries is the work done through ISO 11620,a recently approved international standard on LibraryPerformance Indicators!' It specifies a set of 29 indicatorsgrouped in the following areas: (a) user satisfaction; (b)public services, which includes general indicators as wellas specific indicators on providing documents, retrievingdocuments, lending documents, document delivery fromexternal sources, inquiry and reference services, informa-tion searching, and facilities; and (c) technical services,including indicators in the area of acquiring, processing,and cataloging documents. Notable points in this pro-posed standard are its initial emphasis on user satisfac-tion; its inclusion of cost-effectiveness indicators; its clearand distinct way of describing each indicator, accompa-nied by suggestions regarding the methodology to beused in collecting the data; and a description indicatinghow to most accurately interpret each indicator.

Related to the ISO 11620 effort, but with a specialemphasis on academic libraries, is the InternationalFederation of Library Associations and Institutions's(IFLA) development of international guidelines forperformance measurement in academic libraries!'Seventeen select indicators are identified, with anemphasis on indicators that could be applicable interna-tionally to all types of academic libraries, concentratingon measuring effectiveness (but not cost-effectiveness).'6

Both ISO 11620 and the IFLA guidelines are importantworks that bring attention to the issue of library performance

A R L 1 9 7 A P R I L 1 9 9 8 26

measurement at the international level with an aim topromote acceptance of performance measurement.However, both efforts tend to emphasize indicators thatrequire special effort to be collected, and, although theyare useful in making historical comparisons within a libraryas long as the individual library's policies do not change,their usefulness is limited at the cross-institutional levelsince local policies (such as loan periods, number of booksauthorized for borrowing simultaneously, differing policiesfor different constituenciesstudents, graduate students,faculty, etc.) invalidate such comparisons.

An ambitious effort undertaken by the NationalAssociation of College and University Business Officers(NACUBO) aims to develop benchmarks for 39 function-al areas in universities." The "library" is just one of the39 "functional areas" for which data were collected and issandwiched between "legal affairs" and "mail room."ARL advised NACUBO on the development of thelibrary portion of the survey and, as a result, theNACUBO library survey is almost a duplicate of theARL survey. Unfortunately, some have taken the datacollected by NACUBO as "indicators of efficiency" and"best practices," even as indicators of "quality," despiteARL's long-standing caution against such interpretations.Ratio analysis, which is the way most of the results werereported by NACUBO, is not benchmarking and does notanswer questions; ratios of this sort provide a basis uponwhich to ask questions.

Another organization conducting performancemeasurement initiatives is John Minter Associates. Theirefforts to develop indicators in colleges and universitiesare built upon the Integrated Postsecondary EducationData System (WEDS) and thus are published with thesame delay that afflicts WEDS surveys. Academic LibraryStatistical Norms 1994 is the latest of a series of publica-tions issued by Minter since 1988 using the biennial IPEDSAcademic Libraries datafile to report 101 "measures" onacademic libraries. The publication reports ratios fordifferent types of libraries in groups that are based on theCarnegie Classification System. The authors, however,understand the limitations of ratio analyses and clearlypoint out in the 1992 introduction that "each comparisontakes on meaning only in light of management goals.Does the measure exceed, meet, or fall short of the desiredgoal? Why? In the absence of a stated goal the questionthen becomes, 'Is the position of this measure where wewish it to be? Why?' Operating measures are not of equalimportance nor of the same importance to different insti-tutions. It is unlikely that an institution will give equalconsideration to all 101 measures. Institutional contextand administrative vision are two reliable guides to theimportance of particular measures. Over time, the focuson particular measures will shift as goals are achieved andinstitutional context changes."'

To protect the confidentiality of individual institu-tions, both Minter and NACUBO report ratios for groupsof institutions. Non-disclosure of institutional dataworks against the understanding of data anomalies andthe subsequent correction of reported errors. Althoughratios may be misinterpreted by those who are not famil-iar with an individual institution's goals and circum-stances, there is a value in disclosure. The challenge ofa disclosure strategy involving individual institutionaldata entails investment of effort in educating the public,legislators, and university administrators about how tointerpret numbers related to libraries and other highereducation functions.

Factors Affecting the Reliabilityand Validity of DataThere are at least three major issues that need to be takeninto account in assessing the reliability and validity ofdata generally and of academic library data in particular:consistency across institutions and through time; ease vs.utility in gathering data; and values, meaning, andmeasurements.

ConsistencyLack of consistency in the way data are collected frominstitution to institution and in the way data are collectedover time within the same institution creates problemsfor describing cross-sectional comparisons and time-series trends. With no processes in place to guaranteecompliance with standard definitions, comparability ofdata across institutions may legitimately be questioned.The existence of the extensive "Footnotes" section of theARL Statistics publication testifies to the importance ofrecognizing the limitations of reported data.

One possible way to overcome inconsistencies frominstitution to institution is to develop standards for report-ing data across common automated systems, such as thosethat have been developed in higher education for transfer-ring student records. In order to develop parallel applica-tions for libraries, at least to the level of sophistication thatexists for student records, concerns such as the confiden-tiality and privacy issues related to patron records andInternet transaction logs will have to be addressed.

Ease vs. UtilityPerformance indicators are being developed from datathat can be easily gathered. Of course, what is easy tomeasure is not necessarily what is desirable to measure.It is always tempting to set goals based on the data thatare gathered, rather than developing a data-gatheringsystem linked to assessing progress towards meetingestablished goals. For example, ARL's ratios report liststhirty ratios that are derived from the existing data thatARL collects on an annual basis. Because the ARL datareflect the historical and traditional roles of academiclibraries, the ratios calculated and printed in this report

27 A R L 1 9 7 A P R I L 1 9 9 8

116

1. I

A

Continued

are primarily input indicatorsrelated to levels ofstaffing, collections, and expenditures. The differencebetween these ratios and the raw data published in theARL Statistics is that certain ratios can reflect advance-ment towards specific, local objectives. The ratios can beviewed as supporting tools to assess progress towardsachieving a certain objective, but the final judgmentabout the importance of a specific indicator must alsotake into account environmental factors that are part ofthe local institutional culture.

Values and MeaningThere is a danger of blurring the distinction between thevalue system that is reflected in certain indicators andthe indicators themselves. For example, in developinga system of measures to track library performanceregarding the cost of serial subscriptions or of mono-graphs, there are certain values behind the numbersthat can be fundamentally different from library tolibrary. These values and the interpretation of the mea-sures therefore may have meaning only in the contextof local circumstances. For example, a low unit cost forserial subscriptions may be extremely important for oneinstitution, while another may assert that high-qualityservice can be guaranteed only by acquiring the mostcostly scientific and technical journals, thus yielding ahigher unit cost per serial subscription.

Another ratio that is often calculated is libraryexpenditures per student or faculty: Does the librarythat spends more per student or per faculty offer betterservice? Or is this a sign of inefficiency? What is therelationship between library spending levels, usage, andeducational achievement or user satisfaction? The dataARL collects cannot answer the latter questions; the mean-ing and value assigned to these ratios must be developedlocally. Thus, one of the limitations of this approach isthe absence of an interpretation for each indicator.

The movement calling for performance indicatorswhich appears to be a near-universal phenomenonderives in part from the need to define a value system forhigher education in an era of unprecedented change andtechnological innovation. As ARL further explores insti-tutional value systems and establishes measures thatreflect these values, the Statistics and MeasurementProgram hopes to be better able to define and measurequality in higher education and in academic and researchlibraries. As a first step, the ratios that ARL publishes canserve a dual purpose, although a limited one:

(a) to identify whether a relative position in therankings for a ratio is that expected and desiredfor an institution, and

(b) to compare an institution against its peers,especially over time.

ARL's previously mentioned electronic edition of theannual statistics allows a reader to move beyond those

thirty ratios published by ARL and calculate interactivelyany conceivable ratio among the ARL data elements.

Recommendations for Research LibrariesIn addition to the data currently collected in the ARLStatistics, it would probably be useful for ARL libraries tostart adopting some cross-institutional performance indi-cators from the recently approved ISO 11620 standardand the IRA guidelines. The major advantage of theindicators proposed through these sources is that there isa standard interpretation for them regarding the value ofservices. In particular, at the cross-institutional level ARLlibraries can identify those indicators from the abovesources that are impervious to variations in local librarypolicies and explore their usefulness.

From the list of twenty-nine indicators in the ISO11620 standard and the seventeen indicators listed in theIFLA guidelines, the following performance indicatorscould be easily adopted by ARL institutions:

(a) The IFLA guidelines define market penetration orpercentage of target population reached as theproportion of the library's potential users whoactually use the library!' Although it would bemore difficult to get an overall use measure of thevarious services (e.g., reference, circulation, in-house use), it should be relatively easy for a libraryto calculate with their online circulation system theextent of their circulation services' market penetra-tion for each primary user groupfaculty, gradu-ate students, and undergraduates. Most ARLlibraries, then, should be able to easily adopt sucha "market penetration of circulation" measure.

(b) Although less important than market penetrationand recognizing that ARL libraries have a strongarchival function, collection turnover or collectionuse would be a useful indicator. The IFLA guide-lines suggest combining the number of loans withina year and the number of in-house uses (which canbe problematic for those libraries that do not keepin-house use statistics). However, the ISO 11620standard restricts the definition of this indicator tothe number of registered loans in a specified collec-tion divided by the total number of documents inthe specified collection, ignoring in-house use.Also, to control variations in the loan period thatwould affect renewal numbers, it might be advis-able to restrict the total number of loaned items tothe number of initial loans, excluding renewals.

(c) Extremely important, although not as easily applic-able, is the measurement of user satisfaction as aperformance indicator. Its applicability acrossinstitutions needs to be further explored given thevariations in the services provided by each library,but it is nonetheless a critical indicator of whetherusers' expectations are satisfied or not. Both the

A R L 1 9 7 A P R I L 1 9 9 8

IFLA guidelines and the ISO 11620 standardrecommend a five-point scale and suggestmeasuring both general user satisfaction as wellas satisfaction with specific service areas. TheIRA guidelines describe the process of measuringuser satisfaction in more detail and recommend acombination of satisfaction and importance thatcan help decision-making and action-taking; fur-thermore, the measurement of user satisfaction isnot only recommended with local services, butalso with services offered for remote use.

Librarians may not feel entirely comfortable under-taking such initiatives on their own, but there is a verystrong influence towards this direction, partly comingfrom user-centered management practices. The ARLStatistics and Measurement Program has been providingworkshops that familiarize librarians with the variousaspects of the user survey research process, aiming toeither help them initiate such activities on their own orto work effectively with consultants. A number of ARLlibraries have been systematically applying resultsobtained from user satisfaction surveys when imple-menting changes and charting new directions for theirorganizations. Grace Anne DeCandido describes theresults of such efforts in ARL libraries in an ARL SPECpublication entitled, After the User Survey, What Then?'

Lastly, it should also be pointed out that work isunderway in the area of performance indicators for theelectronic library. ARL efforts to date have concentratedprimarily on measuring the monetary investmentslibraries make in electronic resources. Current work byTimothy Jewell, University of Washington, who ana-lyzed data collected through the ARL SupplementaryStatistics (an experimental testbed for new measures),has documented a clear trend of increasing investmentsin electronic resources that indicates ARL librariesinvested about 7% of their materials budget in electronicresources in 1995-96.2'

Other efforts that have emphasized a more generalevaluation of the academic network environment andinformation services of universities include Assessing theAcademic Network Environment, by Charles McClure andCynthia Lopata,' and Management Information Systemsand Performance Measurement for the Electronic Library:eLib Supporting Studies, by Peter Brothy and Peter W.Wynne.' Overall, there is general agreement that allthese efforts attempting to define indicators for electron-ic resources and services are at the early stages of devel-opment and much more work needs to be done beforemeaningful cross-institutional comparisons can bemade.

I Peter Schmidt, "Rancor and Confusion Greet a Change in SouthCarolina's Budgeting System," The Chronicle of Higher Education(4 Apr. 1997): A26.

2

3

4

5

6

7

9

10

11

12

13

14

15

16

17

18

19

20

21

22

See, for example, Ian Winkworth's, "Performance Indicators," inLibrarianship and Information Work Worldwide (Graham McKenzie andRay Prychirch, eds., London: Bower Saur, 1993: 171-191).An executive summary of the workshop can be found at:<http:/ /www2.echolu/libraries/en/statwks.html>.Gerald Gaither, Brian P. Nedwek, and John E. Neal, Measuring Up:the Promises and Pitfalls of Performance Indicators in Higher Education,ASHE-ERIC Higher Education Report No. 5 (Washington, DC:George Washington University, Graduate School of Education andHuman Development, 1994), v.America's Best Colleges (Washington, DC: U.S. News & World Report,1996), 11

Council for Aid to Education, Breaking the Social Contract: The FiscalCrisis in Higher Education. Online. Rand Corporation. Available:<http://www.rand.org/publications/CAE/CAE100/index.html>.3 Apri11998.Michael G. Dolence and Donald M. Norris, Transforming HigherEducation: A Vision for Learning in the 21st Century (Ann Arbor, MI:Society for College and University Planning, 1995): 22.FARNET's Washington Update, November 7, 1997 issue, informs usthat NGI's recent success in garnering $95 million will be allocatedon Internet issues relevant to each agency's "particular expertise andagency missionDARPA's focus will be on advanced networkresearch, NASA's on specialized network testbeds, NIST will concen-trate on standards development, NSF will continue to cultivate itsrelationship with the academic community, and the NIH will focuson health care applications." Copies of this newsletter are distributedthrough <cni-announce@cni.org> and this issue can be retrievedthrough the cni-announce archives.<http://internet2.edu/>Anthony M. Cummings, et. al, University Libraries and ScholarlyCommunication (Washington, DC: Association of Research Libraries,1992).

Association of Research Libraries, ARL Statistics (Washington, DC:Association of Research Libraries). Annual.Association of Research Libraries, Developing Indicators for AcademicLibrary Performance: Ratios from the ARL Statistics (Washington, DC:Association of Research Libraries). Annual.<http://www.lib.virginia.edu/socsci/newarl/>ISO 11620, Information and DocumentationLibrary PerformanceIndicators (Geneva: International Organization for Standardization).To be published June 1998.Roswintha Poll and Peter te Boekhorst, Measuring Quality:International Guidelines for Performance Measurement in AcademicLibraries (London: K.G. Saur, 1996).1) Market penetration, 2) opening hours compared to demand,3) expert checklists, 4) collection use, 5) subject collection use,6) documents not used, 7) known-item search, 8) subject search,9) acquisition speed, 10) book processing speed, 11) availability,12) document delivery time, 13) interlibrary loan speed, 14) correctanswer fill rate, 15) remote uses per capita, 16) user satisfaction, and17) user satisfaction with services offered for remote use.NACUBO, "Benchmarking for Process Improvement in HigherEducation: A Prospectus" Coopers and Lybrand with the assistanceof Barbara S. Shafer and Associates, FY 1994. Online. Available:<http://www.nacubo.org/website/benchmarking/index.html>Academic Library Statistical Norms 1992 (Boulder, CO: John MinterAssociates): 2.Measuring Quality, 45.

GraceAnne A. DeCandido, After the User Survey, What Then? ARL SPECKit 226 (Washington, DC: Association of Research Libraries, 1997).Timothy D. Jewell, "Recent Trends in ARL Electronic and AccessServices Data," a report submitted to the Association of ResearchLibraries, 1997. Available at:<http://www.arl.org/stats/specproj/etrends.htm>.<http://www.cni.org/projects/assessing/><http:/ /www.ukoln.ac.uk/dlis/models/studies/>

A R L 1 9 7 A P R I L 1 9 9 8

8, A R L 1 9 7 A P R I L 1 9 9 8

Continued

IN SEARCH OF NEW MEASURESThis article was originally prepared by Martha Kyrillidou,Senior Program Officer for Statistics and Measurement, andWilliam Crowe, Vice Chancellor for Information Services andDean of Libraries, University of Kansas, for discussion by theARL Board at their February 1997 meeting.

In 1994, a new strategic objective was adopted by theARL membership to describe and measure "the per-formance of research libraries and their contributions

to teaching, research, scholarship and community ser-vice." This action ratified new directions for the ARLStatistics and Measurement Program to expand beyondmeasures of "input" (such as collection size, number ofstaff, expenditures, etc.) and to search for new kinds ofmeasures of library performance and impact. This articledocuments the progress in meeting this objective.

Old Wine and New BottlesThe seeds for this initiative were rooted in an article bySarah Pritchard in the March 1992 issue of the ARLnewsletter, in which she concluded that, "ARL's activeprogram of statistical analysis, research and managementdevelopment" must center on "maintaining the usefulapproaches of the past and exploring responses to thechallenges of the present and the future."

In October of 1994, there was an ARL membershipprogram on the topic of performance measures asincentives for redesigning library services. WilliamCrowe, Committee Chair from 1992-96, set the stage forthe program by quoting from the then recently releasedAssociation of American Universities Research LibrariesProject task force reports, which state that "there is nolikely substitution of new measures for the old measures,but rather an additive function, a balancing function, aswe move in this transition period."

Striking the right balance between measuring thecontinuing and the emerging realities of the modernresearch library is at the cornerstone of the ARL Statisticsand Measurement Program operations. Researchlibraries' traditional realities drive the ARL measures ofprinted collections, budgets, and staffing. The emergingrealities drive ARL's agenda to seek credible indicators ofthe steady growth and high demand for the complex mixof new services, consortial arrangements, electronic infor-

' mation, the influence of the Internet, and the ways inwhich students and faculty interact with each other andthese newer channels of information.

Emerging Realities and New TrendsWhen exploring emerging realities, the first challenge isto pose the questions that most need to be answered inorder to describe the transformations underway. Howmuch do libraries spend on electronic resources? On con-sortia? Electronic serials? Computer hardware and soft-ware? On digitization for preservation? Interlibrary loan

and document delivery? Are these expenditures madewith funds diverted from traditional budget lines or arethey newly appropriated funds? How many libraries areoffering "innovative" services? Does the availability ofany new services have implications on library use orperformance? For example, in libraries that provideelectronic reserves or user-initiated interlibrary loan, hasuse of the services increased faster than in other librariesand/or are the materials available to users faster and/orat less cost? The Program has taken steps to address theseissues. A planning document was presented and ratifiedat the October 1995 membership meeting.' The Programhas also been successful in attracting external funding andtalented Visiting Program Officers to help refine thesequestions.'

Although there is a better understanding of the ques-tions that need to be asked in this changing environment,the answers continue to be elusive and/or unstable. Forexample, Timothy Jewell's analysis confirms that databeing collected about electronic resources expendituresshow that most ARL libraries are spending a relativelysmall portion of their budget on electronic resources.'Although this portion is increasing rapidly, the change isnot consistent from year to year nor-from member libraryto member library.

The challenge, then, is not in describing any singlechange, but rather to develop quantifiable trend analysisin multiple institutions that can be executed from year toyear in a consistent way. There is an often unarticulatedassumption that change in libraries is moving in onedirection. However, data from ARL libraries indicate thusfar that change seems to be happening in rather haphaz-ard and chaotic ways, both within individual libraries andacross institutions. Some leaders suggest that it is the rateof change creating confusion rather than change itself.

To illustrate the complexity of tracking the emergingrealities of research libraries and this rate of change, oneneed only look back to the early 1990s, when manylibraries provided access to bibliographic databases bytapeload to campus mainframes or by stand-alone ornetworked CD-ROMs. More recently, libraries havebegun to provide either gateway or direct access to ven-dor and publisher full-text databases of journal articlesand monographs. Libraries are also integrating access tovarious electronic information resources through a WWWinterface, making it possible for the traditional libraryOPACs to link to various full-text resources. It has beenimpossible to provide consistent, quantitative indicatorsof such trends because of the rapidity of the changes andbecause they are, by their nature, not comparable to previ-ous measures. We are faced with a series of qualitativerevolutions, basic "paradigm shifts" that are changingwhat research libraries do and in many respects arechanging how research libraries fulfill their mission.

3 0

Certainly, this environment is not conducive to identi-fying consistent measures among 121 research libraries.However, in response to those who seek quick answers tonew measures, there have been some ideas, although not awidespread acceptance of them (for example, a screen cap-ture could be considered a potential unit of measurement,similar to a library gate count). Various software "coun-ters" tell us, for instance, how many computers accessed aweb page, how many bytes were transferred, and howmany sessions were established. Are these useful mea-sures?' Some authors argue that they are and have pro-posed that "a centralized voluntary reporting structure forWeb server usage statistics, coordinated by the Associationof Research Libraries' (ARL's) Office of Statistics, wouldprovide a significant service to academic librarians."'

The User's Point of ViewMost libraries are aware of the need to measure not onlythe use of their resources, but also the effectiveness oftheir library services. In particular, how are library usersbenefiting from their interaction with the library? To helpaddress this question, the ARL Statistics andMeasurement Program initiated a series of training eventsto help library staff collect information from their users tobetter inform their management decisions. In the longrun, if higher education wants to measure library"impact" we will need to initiate longitudinal studies, forexample, by questioning and tracking individuals fromgrades K-12 through their undergraduate/graduate studyand as alumni in order to assess how their lives and workmay have been affected by their library experience.

In the meantime, many libraries have found value inconducting user surveys and using the results to assesscurrent and to devise new library programs and services.Some ARL members have used the same survey instru-ment, offering a possible opportunity for cross-institution-al comparisons. ARL showcased how libraries are mak-ing use of user surveys as new measures in a 1997 publi-cation and continues to promote this strategy by offering aworkshop on the methodology of user surveys.'

The HeritageA core question that has been posed by some ARLmem-bers also helps frame the issue: Should we stop collectingthe established annual data series and instead invest allstaff efforts in exploring new areas? Is collecting andpublishing data on collections, expenditures, staffing, andservices holding us back? What is the current value ofour investments in the ARL Statistics, ARL Annual SalarySurvey, "Library Expenditures as a Percent of E&G," andARL Preservation Statistics?

Recently, one of the Program's fundamental strategieshas been to maximize the usefulness of the annual pro-jects. Thus, from data collected in the ARL Annual SalarySurvey, Stanley Wilder was able to publish a report on

THE ARL MEMBERSHIPCRITERIA INDEX

The criteria for academic library membershipin the Association of Research Libraries arebased partly on quantitative data that provide

a view of the range of resources deployed among theexisting members of the Association. Statisticalanalysis shows a high degree of homogeneity inrespect to five data categories:

volumes heldvolumes added, grosscurrent serialstotal library expenditurestotal professional plus support staff

Each year ARL uses the statistical method ofprincipal component analysis to identify the corn-monalities in the membership. The analysis is con-ducted on the 35 charter members of ARL and pro-duces coefficients, or weights, for each of the fivedata categories. When the data for a given libraryare multiplied by the weights and summed, theresult is a "score" for that library. This process ofmultiplying by weights and summing is carried outfor each ARL academic library. The resulting scorescomprise what is known as the ARL MembershipCriteria Index. The term "score" in this context isnot a judgment about the library's quality or perfor-mance. "Score" is a term from principal componentand factor analysis that refers to the summation ofdata. The ARL index score in effect aggregates thefive measures of size and resources. Each year thecurrent year's data in the five categories are pub-lished in The Chronicle of Higher Education, arrangedin descending rank order by the ARL Index scores.

The weights and data categories can also beapplied to the data of non-ARL libraries. This tech-nique is one of the tests used to determine potentialmembers of the Association. Candidates for mem-bership are required to have a score on the ARLindex scale of at least -1.65 for the most recent fouryears in order to be considered. This criterion wasestablished to ensure that new members share theessential characteristics of the existing members inregard to the five measures of size. The membershipcriteria also include other requirements to ensure thehomogeneity of the membership. See <http:/ /www.arLorg/stats/qualcov.html> for the completedescription of membership criteria and <http:/ /www.arl.org/stats/factor.html> for more informa-tion on the membership index.

1 A R L 1 9 7 A P R I L 1 9 9 8

I.

ASURES---Continued

aging in the library profession and calculated retirementprojections.' Martha Kyrillidou received permission tostudy earnings differentials and hierarchical segregationin ARL libraries using the same dataset. Using the ARLStatistics, a secondary annual report is published contain-ing thirty selected ratios that ARL directors have identi-fied as useful indicators for examining progress towardslocal goals and objectives.' Maximizing the investmentsin ARL's annual statistical data has proven a very success-ful strategy that helps managers develop a better under-standing of current trends in their institutions.

ARL Membership IndexWhen reviewing the ARL Statistics and MeasurementProgram activities, one cannot escape a discussion of theARL Membership Criteria Index, one of ARL's most pub-licized and controversial products (see accompanyingsidebar). The Index serves as a measure of the common-ality of new members with the founding ARL memberlibraries. It is a composite measure of volumes held, vol-umes added gross, current serials, total expenditures, andprofessional plus support staff. The Index is not a mea-sure of a library's services, the quality of its collections, orits success in meeting the needs of users.

Three out of the five variables that comprise theMembership Index encompass investments in the emerg-ing library realities. For example, "current serials" is ameasure that includes not only print journals, but elec-tronic ones as well; "total expenditures" incorporatesexpenditures for electronic investments; and "staffing"is an essential composition of both the old and the new.

The ARL Membership Index is published by TheChronicle of Higher Education every year. A number ofvoices have called for a stop to the Index, declaring that itfosters a competitive posture in an era of increasing coop-eration. The Index is also seen as a threat to resource shar-ing because it appears to emphasize investments in localcollections. Further, it is said to de-emphasize the distinctinstitutional character of each library. It is perceived ascalling those universities that invest most heavily inlibraries "winners," implying top, or best ranked, schools.

If ARL membership were not based on the Index,what would take its place? A variety of membershipcommittees have attempted to answer this question, themost recent body submitting a thoughtful report in 1994proposing to supplement the quantitative membershipcriteria with qualitative factors that take into account,among other things, investments in electronic resources.In 1995, the membership adopted this proposal to amendthe membership criteria. This latest revision does noteliminate the pre-existing quantitative requirements, butallows for consideration of qualitative assessments aboutthe contributions a research library makes to NorthAmerican learning and scholarship when the quantitia-tive requirements are marginal.

Striking a BalanceActivities undertaken by the Statistics and MeasurementProgram include collecting, refining, and making use ofthe traditional data while simultaneously searching fornew measures.

The analyses to date indicate that access measuresare best developed locally. Ratios, user survey data,and service transaction data complement the traditionalquantitative data in providing an overall picture oflibrary input and output. As the library and its con-stituent community reach consensus on how to bestmeasure the expenditures, collections, and use of elec-tronic resources, these additional measures can beadded to a library's collective dataset, as well.

Our challenge each year is to learn from our experi-ences how to improve ARL measures. As our members'environment changes, ARL is adjusting program goalsand adapting measures to suit the emerging realities ofresearch libraries. Douglas Bennett, President, EarlhamCollege, and former Vice-President of the AmericanCouncil of Learned Societies, was invited to commenton ARL's search for new research library measures. Henoted that "we need goals in order to measure progresstoward them, but at present we do not have adequategoals, or ultimate goals, with regard to what universities[and libraries] should do." He further cautioned thelibrary community to "avoid premature closure. Keepexperimenting because we are unlikely to settle intocomfortable grooves anytime

' Proceedings of the 125th Membership Meeting, Association ofResearch Libraries, 1996. p. 35.

2 see <http://www.arl.org/stats/program/planning.html>3 For example, the Council on Library and Information Resources

and the University of Washington are supporting Timothy Jewell inan examination of the investment made by research libraries inelectronic resources. The Andrew W. Mellon Foundation fundedMary Jackson's two-year study of interlibrary loan performance.Jan Merrill-Oldham, Harvard University, led a project to revise theARL Preservation Statistics Survey to begin to collect data ondigitization for preservation.

4 Recent Trends in ARL Electronic Access Services Data. Available at<http://www.arl.org/stats/specproj/etrends.htm>.

5 Chen, Paul. "Hit-or-Miss Measurement." Marketing Tools.(Mar. 1997): 22-25.

6 Hightower, Christy, Julie Sih, and Adam Tilghman."Recommendations for Bcnchrnarking Website Usage AmongAcademic Libraries." College and Research Libraries (Jan. 1998):61-79.

7 GraceAnne A. DeCandido. Julia C. Blixrud, Ed. Advisor.Transforming Libraries 4: After the User Survey, What Then? (SPEC226). September 1997. This publication is also available online at:<http://www.arl.org/transform/us/index.html>.

8 Wilder, Stanley. The Age Demographics of Academic librarians:A Profession Apart. Washington, DC: Association of ResearchLibraries, 1995.

9 Developing Indicators for Academic Library Performance: Ratios for theARL Statistics. ARL annual report.

I° Proceedings, 33.

32A R L 1 9 7 A P R I L 1 9 9 8

INTERACTIVE PEER GROUPCOMPARISONS THROUGH THE WEBby Kendon Stubbs, Deputy University Librarian,University of Virginia

Anewly revised ARL Statistics site on the WorldWide Web offers interactive comparisons ofacademic ARL libraries, expressed both in

numbers and in graphs. The site makes available theARL data from 1962-63 through the latest year, 1996-97.The median, average, high, low, and other statistics canbe displayed for any variable of any academic library orgroup of libraries for the past 35 years. Users can alsoselect data from Canada or by census regions of the U.S.The URL for the site is: <http://www.lib.virginia.edu/socsci/newarl>. The site is also available through theARL homepage at <http:/ /www.arl.org/stats/arlstat/index.html>.

In addition to displaying raw data from the annualARL Statistics publication, the web pages also allowusers to create ratios between any two data categories orvariables. Thirty selected ratios are already publishedeach year in ARL's annual printed publicationDeveloping Indicators for Academic Library Performance:Ratios, but with the website it is possible to compute anddisplay individual library data for close to 2,000 ratios,many of which may be of considerable interest in bench-marking or measuring library performance.

Ihowed,,kina (0*.s

Association of Research Libraries StatisticsInteractive Graphics for Academic Institutions

90 to 1997

= 111153,.

Figure 1: Interactive Graphics for Academic Institutions

For example, one could click on "InteractiveGraphics" and then on graphs for multiple libraries tocreate comparative charts. For Duke University, theUniversity of North Carolina, the University of Virginia,and the median ARL library, one could graph the ratio ofpurchased monographs to Ph.D.s awarded for the last 10years. Figure 1 shows the resulting graph. At a glance,one can see that at Duke and Virginia monographic

BEST COPY AVNLABLE

acquisitions per Ph.D. are lower than they were 10 yearsago, while at North Carolina they are higher, but that allthree libraries provide more monographs per Ph.D.degree than the median ARL library.

Time series data for individual libraries can also bedisplayed on the Web. For example, Figure 2 shows theratio of materials expenditures per graduate studentfor the past five years for three libraries plus the ARLmedian library.

_

I-71 96044 *adAssociation of Research Libraries Statistics

Ranked Lists for Academic InstitutionsI Riot. ....1,4 9.0, .1999

pate List for Total Materials Expendituresover

Total Fulltiree Onaduate Student§. . . .

Pi-$/77 15-IIHE 4OR711 CAROI.R4A rfrFrCZAA''PE.DIAN1993 1460 13 1365 61 1lir-1833 )--"-S491

994 1558 83 1510 14 rrlr'4 16 riT3Y.67.1995 f 1607.34 1535 77 r-m-vir (-RitzF9it.;1-2T---- 1694 40 I- 1358 76 1M9 0111997 1 1775 89 1 1831 76 1-1T---005 rssiir--073

Figure 2: Ranked Lists for Academic Institutions

Also available on this site are ranked lists for all ofthe ARL data variables and for the 2,000 possible ratios.In addition to the ranked data published in the printedARL Statistics, website users can create rank-order tablesfor variables such as computer files, circulations, studentassistants, etc., or for ratios such as total staff to volumesheld or to circulations, interlibrary borrowing to pur-chased serials, operating expenses to total staff, etc.

Further, this site offers 12 years of the ARLMembership Criteria Index scores and ranks, as well asthe data on which the Index is based. A click of the mousein this section lists or graphs the Index data for any ARLmember library.

Finally, the last five years of ARL statistics canbe downloaded from the website in ASCII or .wklformats (for use with Microsoft® Excel or Quattro® Proapplications).

The ARL statistics represent the oldest continuinglibrary statistical series in North America, and serve asa comprehensive resource on trends in academic andresearch libraries. The website statistics provide userswith an opportunity to conduct analyses and create datasubsets of academic research library statistics to addresslocal needs.The ARL Statistics site is maintained by the Geospatial andStatistical Data Center of the University of Virginia Library(formerly called the Social Sciences Data Center).

3 3 A R L 1 9 7 A P R I L 1 9 9 8 11

t2

Continued

UNDERSTANDING ELECTRONICRESOURCES AND LIBRARYMATERIALS EXPENDITURES:AN INCOMPLETE PICTUREby Julia C. Blixrud, Senior Program Officer, andTimothy D. Jewell, Electronic Resources Coordinator,University of Washington Libraries

Questions asked both of ARL generally and ofARL member libraries individually are: Howmuch are research libraries spending for elec-

tronic resources collectively and how much on average?How does an individual institution's expenditures forelectronic resources compare with other researchlibraries? Such questions are to be expected as librariesmove into a digital environment, but simple and credi-ble answers remain elusive. Since fall of 1996, and withthe support of the Council on Library and InformationResources and the University of Washington Libraries,the ARL Statistics and Measurement Program has beenexploring the character and nature of library investmentin electronic resources in order to develop standard

. .

definitions to more systematically collect informationabout the transformation of research library collections.Timothy D. Jewell has been serving as a VisitingProgram Officer for the project by analyzing data fromARL's main and supplementary surveys and by talkingwith member libraries about how to measure the invest-ment being made in electronic resources.

ARL has had a supplementary statistics question-naire since 1984. This survey serves as a testbed fordeveloping questions and for gathering new forms ofdata. Questions asked on this survey can be subsequent-ly added to the main statistics survey, dropped from con-sideration due to an inability to collect reliable data, ormodified over several years while determining the bestmeans by which to ask them. The reports from the sup-plementary statistics survey are not generally publishedbecause, due to their experimental nature, the results arenot considered sufficiently reliable. They are, however,made available to member libraries for their information.Data on expenditures for electronic resources were firstrequested as part of this testbed survey in 1992-93, withquestions on expenditures for (a) computer files and

ELECTRONIC RESOURCES AND ARL LIBRARY MATERIALS EXPENDITURES/1992/93-1995/96

1992-1993

Computer File ExpendituresTotalAverageMedian

Electronic Serial ExpendituresTotalAverageMedian

$14,147,625$172,532$148,158

Electronic Resources Expenditures(total of above)

TotalAverageMedian

Library Materials Expenditures(for respondents)TotalAverageMedian

$14,147,625$172,532$148,158

$393,271,073$4,795,989$4,242,887

E-Resource $ as Percent of 3.60%Library Materials Expenditures

N. (of 108 Academic ARLs 82through 94-95, 109 in 95-96)

1993-1994 1994-1995 1995-1996

$20,132,553 $22,030,727 $24,609,821$236,854 $249,286 $253,709$212,936 $226,318 $210,890

$11,847,577 $15,170,972$188,057 $174,379$156,754 $148,166

$20,132,553 $33,878,304 $39,780,793$236,854 $349,261 $364,961$212,936 $278,404 $301,992

$425,287,651 $489,664,539 $571,145,986$5,003,384 $5,380,929 $5,654,911$4,527,122 $4,714,384 $4,975,353

4.73% 6.92% 6.97%

85 97 101

34 BEST CUPY AVAILABLEA R L 1 9 7 A P R I L 1 9 9 8

search services; (b) document delivery/interlibrary loan;(c) computer hardware and software; and (d) bibliograph-ic utilities, networks, and consortia. A fifth question onexpenditures for electronic serials was asked beginningin 1994-95.

The 1992-93 question on expenditures for "ComputerFiles and Search Services" was defined to collect data on"expenditures for software and machine-readable materi-als considered part of the collections, whether purchasedor leased," and expenditures for online database search-es. It excluded expenses for library system and staff soft-ware, and for "bibliographic utilities, networks and con-sortia," and, according to the instructions, only thoseexpenses that would have been counted in the mainARL Statistics survey as part of expenditures for "OtherLibrary Materials or Miscellaneous" were to be included.In 1994-95, the question on "Expenditures for ElectronicSerials" was added, covering subscriptions whose prima-ry format is electronic. The instructions stated that onlythose expenses that would have been reported in themain ARL Statistics survey as part of "Current Serials"were to be included.

Overall, there was a reasonably good (and improv-ing) response rate for these questions, with 82 of the 108academic ARL libraries (Academic ARLs) providing anon-zero figure for the "Computer Files and SearchServices" in 1992-1993, and 101 of 109 providing a non-zero response for either or both of the questions in 1995-1996. The responses themselves ranged widely, withsome very large figures reported. Although no respons-es were excluded on this basis, both average and medianfigures are provided in the accompanying table for betterunderstanding of the data. As can be seen, the reportedamount spent on computer files increased more than $10million overall between 1992-1993 and 1995-1996. Whenexpenditures for electronic serials are added for the lasttwo years, the total of what can be called "electronicresources" nearly tripled from more than $14 million in1992-1993 to nearly $40 million in 1995-96. The averageand median figures have both doubled in that period.The percentage of "Library Materials Expenditures"devoted to electronic resources, although still modest,has almost doubled, from 3.6% in 1992-1993 to nearly7.0% in 1995-1996.

Mr. Jewell's project has confirmed that thesefigures underestimate the actual expenditures by someunknown amount. The reasons stem from the difficultyin finding adequate definitions to collect data in a stan-dard and comprehensive way when libraries areemploying a wide variety of funding and budgetingstrategies to acquire electronic resources. For instance,the largest jump in reported expenditures occurred in1994-1995, when the question on electronic serialsexpenditures was introduced. This suggests that actual

library expenditures for the prior two years were higherthan reported. In addition, after having thoroughlyreviewed the survey questions'intent, and by reviewingone institution's responses and consulting with otherlibrarians, Mr. Jewell concludes that local record-keepingpractices made it difficult to respond to the questionaccurately and that, as a result, many investments inelectronic resources may have been understated. It alsoseems likely that some expenditures in this area areactually being paid out of a library's operations funds,which would theoretically not be reported in either ques-tion (this suggestion was endorsed by several peoplewho attended an ARL survey coordinator meeting), andpossibly being paid out of "Consortial and NetworkExpenditures." Beyond this, several ARL library surveycoordinators confirmed that their libraries provide accessto resources that are purchased on behalf of their institu-tions by a state-funded consortium, and that these expen-ditures also go unreported in the ARL SupplementaryStatistics survey. Based on these findings, Mr. Jewellrevised the 1996-97 ARL supplementary questionnaireto try to capture those figures that had been elusive inprevious surveys. Preliminary reports indicate that thisyear's results will show a significant increase in totalelectronic resource expenditures.

In Mr. Jewell's analysis of the 1992-93 to 1995-96survey data, he identified several other trends regardingthe data that member libraries were reporting. A fullreport was sent to each ARL member library as part ofthe Report on the 1995-96 ARL Supplementary Statisticsand is also available at <http:/ /www.arl.org/stats/specproj/etrends.htm>. After consulting with the ARLStatistics and Measurement Committee and ARL surveycoordinators, as well as with other individuals andgroups through meetings and presentations, severalquestions were revised for the 1996-97 SupplementaryStatistics Survey. Data from that survey is currentlybeing collected and analyzed. A copy of the survey isavailable at <http:/ /www.arl.org/stats/sup97/survey.htm>.

Mr. Jewell's investigation of data on research libraryexpenditures for electronic resources confirmed thatexisting data is neither comprehensive nor comparablefrom one library to another. The revised questions onthe 1996-97 ARL Supplementary Statistics survey areexpected to yield a more comprehensive picture oflibrary expenditures for electronic resources. However,the picture will remain incomplete until definitions areresolved and practical techniques are developed toensure accurate and consistent reporting. In the mean-time, the ARL Statistics and Measurement Program willcontinue to monitor this and other measures to trackhow research libraries are responding to the digitalenvironment.

3 5 A R L 1 9 7 A P R I L 1 9 9 8

SPEemtissuE-ON-MEAsuREfs---,

Continued

ASSESSING THE ACADEMICNETWORKED ENVIRONMENTby Ioan Lippincott, Associate Executive Director,Coalition for Networked Information

How are libraries thinking about assessment inthe networked environment? Is progress beingmade in measuring the impact of the availability

of networked information resources and services? Is anyattempt being made to demonstrate the impact that theinvestment of many hundreds of thousands of dollarshas made on the improvement of access to informationby those in the university community?

Seven institutions are participating in a Coalition forNetworked Information (CM) project on assessment,which was developed as an outgrowth of the publicationAssessing the Academic Networked Environment: Strategiesand Options, by Charles McClure and Cynthia Lopata(CNI, 1996). The manual describes the challenges ofassessing networks and networked services and offersguidance on approaches to developing measures. Theauthors describe sample measures in a variety of areas.The institutions participating in the CM project choseareas of assessment for their particular campus anddeveloped measures using the McClure/Lopatamanual as a starting point.

Several of the participating institutions tested mea-sures related to library and information resources andservices as part of the CNI project. Their initiatives aretailored to the needs of their own institutions, havedistinct flavors, and employ a range of assessment tech-niques. While the first round of implementation ofmeasures is not complete on all campuses, the followingsummaries of several of the efforts provide informationon the kinds of topics the libraries are measuring andreport some of the initial findings.

Reports from each of the institutions' initiatives andsupporting materials, including in many cases the sur-veys and other instruments used, are available on CNI'swebsite at: <http://www.cni.org/projects/assessing/>.

University of WashingtonThe University of Washington has an ambitious programof assessment initiatives including redesigning theirtriennial library use survey to include a focus on net-worked information, continued development of evalua-tion methods for the UWired teaching and learningprogram, and an examination of faculty and graduatestudent information use.

The UWired assessment plan is a collaborative effortof the Undergraduate Education, University Libraries,and Computing and Communications departments.Evaluation efforts include the use of a variety of tech-niques, including printed surveys, web-based surveys,e-mail questionnaires, and focus groups.

A R L 1 9 7 A P R I L 1 9 9 8

DATA ON USE QUALITY AND COSTSOF NETWORK SERVICES

Christopher Peebles, CM Visiting Fellow andAssociate Vice-President and Dean of InformationTechnology at Indiana University, has developedan impressive set of survey data that describes useand user satisfaction with an array of services,including IT user support, hardware and software,and e-mail. The materials he uses in his presenta-tions are available at: <http://www.indiana.edu/ucsdcas/jm/>.

To view nine years of Indiana University IT qualitysurveys, visit: <http://www.indiana.edu/uitssur/>.

To view the Activity Based Cost data for thecentral IT organization at IU visit: <http:/ /www.indiana.edu/ucs/business/scindex.htm>

To determine how faculty and graduate students inthe biological sciences are using information for theirresearch and teaching activities, the University ofWashington team is conducting focus group sessionsstructured around three areas:

How users identify, obtain, and use informationfor research and teaching activities.The ways users would ideally like to getinformation they need and why.The use of and perceived tradeoffs associatedwith electronic journals.

The team identified enablers and obstacles to their work.Enablers included strong support from the administrationfor some of the assessment work, previous experience withsurveys and data analysis, and the high priority placed bystaff on this work. Obstacles included the difficulty of devel-oping effective performance measures in a very dynamicand complex environment, the time-consuming nature ofmost assessment activities, and the difficulty in gettingusage statistics from vendors of information products.

The University of Washington is also taking advantageof the services offered by the Flashlight Project (see box onpage 15).

Virginia TechThe "moving target" of networked information measure-ment is also an issue on the Virginia Tech campus. In herreport on the project, Dean of University Libraries EileenHitchingham writes, "We look at the changing realities ofa few months ago to make best guesses about why thingsare happening today, or to better understand what mighthappen in the near future. The perspective is speculative,not conclusive. Still, making guesses from some informa-tion seems better than working with no information."

The Virginia Tech assessment measures included a

3 6

student survey that was developed by a number ofcampus units, including the library. Staff asked studentsto describe their use of links on the library's main webpage. Students reported use of the electronic reservesystem, institutional library catalog, regional and speciallibrary catalogs, and list of database resources. For thisgroup of students, use of the library's more than 100electronic journals was disappointingly low.

Two of Virginia Tech's measures addressed thephysical location of students and others using libraryand information resources. The survey found that manyof the students visited the library in person as well asused the resources through remote network connectionsfrom dorm rooms, off-campus housing, etc.

In a study that included a review of web log data,the team examined where users were located when con-necting to the library web pages and determined thatless than half were inside the actual library. An inten-sive analysis of the use of library web pages was difficultgiven the many changes taking place with the pagesduring the short period of time the log data examined.Also, the fact that different library network servicesreside on different servers made collecting data difficult.However, understanding where users are physicallylocated when using networked information resourceshas implications for providing services and instructionfor library users and is useful for future planning.

Gettysburg CollegeAt Gettysburg College, the assessment project focused onthe use and cost-effectiveness of the Electronic ReservesSystem. The Electronic Reserves System is part of abroader Curriculum Navigation Project (CNAV) that pro-vides a central source of information for the campus. ViaCNAV, students can access information about their cours-es, including class rosters, course homepages, course syl-labi, and electronic reserves. Access to electronic reservesmaterials is restricted to those enrolled in the course.

Through both telephone and electronic surveys,Gettysburg assessed why students used or did not usethe electronic reserves system, determined usage pat-terns, and examined both faculty and student satisfac-tion with the electronic reserves system.

As was evident at Virginia Tech, many students accesslibrary resources from their dorm room if that access isavailable; well over half of the Gettysburg students usingthe electronic reserves system did so. Most studentsfound the system to be convenient and easy to use.

Faculty were enthusiastic about the electronicreserves system because of the added value it brought totheir courses. In particular, they valued the ability toeasily make available current materials to studentsenrolled in their classes, and they liked the capability ofallowing many simultaneous users to access coursereserves, since that is the frequent pattern of use of such

THE FLASHLIGHT PROJECTThe Flashlight Project, headed by Steve Ehrmann ofthe TLT Group, provides a suite of evaluative tools,training, consulting, and other services. The work isbased on the Flashlight Current Student InventoryTM(CSI), which can be used to collect facts and opinionsfrom currently enrolled students. The CSI is a tool kitof almost 500 indexed questions that can be used todraft surveys, questionnaires, and protocols for inter-views and focus groups.

Sample Flashlight questions and a more detaileddescription of the CSI are posted at: <http://www.wiche.edu/flshlght/flashinv.htm>.

materials. In addition, they liked the reports theyreceived documenting what portion of the class actuallyaccessed each reserve item and how many repeat uses ofitems were recorded.

As one faculty member stated, "I truly believe that mystudents had access to more timely and accurate literaturethrough using electronic reserves.... Electronic reserveshelps me do a better job of providing good readings to mystudents as well as monitoring their use of them." Anotherwrote, "[Electronic reserves] provided all students withinstant and continuing access to course materials.Electronic access far outstrips the traditional reserve systemfor providing access, especially in a high enrollment class.Also, some reserves were... needed for long-term access."

King's College, LondonKing's College is focusing on two issues of relevance tolibraries: electronic journals and the use of electronic vs.printed information. They are collecting data by electron-ic means when possible.

In exploring the topic of electronic journals, the King'steam is gathering data on use, cost per transaction/user,usage profiles by journal and by department, and systemavailability. They are also seeking qualitative data on rea-sons for use, user satisfaction, and ease of administration.To collect this information, they are examining system logsand administering web-based surveys.

Their analyses of the use of electronic vs. printedinformation also includes quantitative data on usage, costper item/user, use profiles by department, and documentavailability. Their qualitative assessment addresses rea-sons for choosing print or electronic information, userpreference, user satisfaction, and ease of administration.

The driving forces behind King's assessment projectare the need for accountability to users and funding bodiesand the desire to improve services where needed. As mostof the higher education institutions in the U.K. are makingheavy use of and investing in electronic resources as partof the eLib Programme and other efforts, the institutionswant to know if those investments are paying off to users

3-7 A R L 1 9 7 A P R I L 1 9 9 8

Continued

and in what ways users are satisfied or dissatisfied withthe networked information resources and services.

Challenges of ImplementingAssessment ProgramsIn their manual, McClure and Lopata state some criteriaby which they will judge the impact and success of theirpublication. They include whether:

campuses will experiment with assessment tech-niques;campuses will share information and insights onhow assessments can be done more effectively;evaluation research concepts and procedures willmove forward in this area;campus decision makers will be able to designand plan more effective networked environments;anddata generated will promote incorporation ofusers' viewpoints into the way the networkevolves.

Through the CM project, a small number of institu-tions have taken up this challenge. For most of the institu-tional teams, the road has not been easy. The time andresource investments have been significant and the ever-changing networked environment makes some techniquesproblematic. However, in some cases there is a stronginstitutional mandate for the development of assessmentmeasures throughout the university, and in others, there isstrong commitment by unit heads to work towardsimproving services using assessment as a diagnostic tool.In addition, the project provided a mechanism for individ-uals from many units on campus to coordinate assessmentefforts in relation to networks and networked services.

Project team members report that strong support forassessment from top campus and unit administrators hashad a positive impact on the amount of resources avail-able for assessment efforts. The surveys and data collec-tion efforts by the institutions involved have enabledthem to get a first look at the impact of electronic informa-tion services on users and to begin answering the ques-tion, "What difference do these electronic resources andservices make to users?" The institutional teams havefound that a variety of data collection techniques can beuseful, from log analysis to user surveys (both print andon the Web) to focus groups and individual interviews.The project has given the institutions experience with a setof tools and a start in establishing a baseline of data onelectronic resources and services use for their campuses.

CNI has received support for this project fromIndiana University; and Christopher Peebles and hisstaff have been instrumental in the project's implementa-tion. Charles McClure has been a guiding force in thisphase of the project and provided its initial inspiration.In addition, CM has received support from the Councilon Library and Information Resources (CLIR).

THE STATE OF PERFORMANCEMEASUREMENT IN LIBRARIES:A REPORT FROM THE 2NDNORTHUMBRIA INTERNATIONALCONFERENCEby Amos Lakos, Coordinator of Management InformationServices, University of Waterloo

In September 1997, Northumbria was the venue foran international conference to discuss various librarymeasurements and assessment issues and activities,

to exchange experiences, to increase awareness ofcurrent research, and to identify issues for furtherstudy and work. The conference, sponsored by theDepartment of Information and Library Management atthe University of Northumbria at Newcastle and by theBritish Library, focused on outcomes and impacts, mov-ing from research, definition, and standardization mat-ters to actions and implementation. Total attendancewas 141 persons from 24 countriesnine of which werefrom North Americaa truly international gathering.

The creation and draw of this conference is just oneindication of the greater role that performance measuresare playing in libraries today than they have in the past.Measures are being applied in a wide variety of settingsand in a number of locations. Further, the 46 papers ofthis conference, including five poster sessions and fivekeynote speakers, note the trend of increased coopera-tion between libraries across national boundaries indeveloping appropriate measures, especially amongthose supported by the European Union. Major themescovered during the conference were:

Performance MeasurementsGeneral Analysisand CriticismDefining Relevant Performance Indicators andDeveloping International Performance StandardsBenchmarkingQualitative Measurement ActivitiesComparability Across Sectors and NationalBoundariesService Level AgreementsManagement Information ServicesElectronic/Digital Library and Assessment ofLibrary Networks

A brief summary of some key papers presented follows.For more detailed information, as well as for informa-tion on papers not described here for lack of space,please see <http:/ /library.uwaterloo.ca/aalakos/North97/norsum.html>.

Performance MeasurementsGeneral Analysis and CriticismThe first keynote paper, "Does PerformanceMeasurement Improve Organizational Effectiveness?

A R L 1 9 7 A P R I L 1 9 9 8

A Post-Modern Analysis," presented by RowenaCullen of Victoria University in New Zealand, setboth the tone and the standard for the conference byasking: After all the research is done, the data gather-ing activities and analysis undertaken, and reportspresented, are libraries more effective? Cullen'smodel of organizational effectiveness used afocus/value/purpose matrix to reach the followingconclusions:

performance measurements are politicalactivities, taking place both on a macro and amicro level;how institutions act is dependent on theirrelative internal to external focus, the culture ofthe organization, and the resolve they bring totheir activities;the profession seems to be reactivedependenton rewards and incentives;performance measurements aremultidimensional in nature and each librarywill use the measurements that suit its socialconstruct, environment, and goals; andaction, leadership, and initiative are needed fororganizational effectiveness.

The question is, then, what kinds of measures willbest support a successful institution? Stephen Town's(Cranfield University) paper, "Performance orMeasurement?", took a critical look at the current stateof performance measurements and their impact, criti-cized the over-reliance on traditional library quantita-tive data series, and advocated that more attention bedirected to service quality indicators and to the devel-opment of assessment measures both in the digitalenvironment and for networked services. Town alsosuggested increasing attention to the diverse needs ofstakeholders, a concern Ian Winkworth (University ofNorthumbria on Newcastle) raised as well, in hispaper, "Making Performance MeasurementInfluential." Winkworth provides an excellentoverview of past and current performance measure-ment activities, primarily focusing on U.K. public andacademic libraries and the development of internation-al standards. Winkworth saw the need for outcomeand impact measurements that have real managementutility and feels that libraries may be collecting toomany quantitative indicator data without relatingthem closely enough to stakeholder's needs and con-cerns. However, Winkworth pointed out that, in spiteof deficiencies, libraries are well ahead of many otherpublic services in developing effective performancemeasures.

The role of stakeholders in performance measureswas, in fact, a theme that speakers came to time andtime again. The use of stakeholders in constructing,

39

conducting, evaluating, and, finally, acting on mea-sures (including making plans for the future) were alldiscussed. Speakers emphasized the need for morequalitative measures, and ones that serve more practi-cal purposes than performance measures have in thepast. The stakeholder focus in electronic environmentstudies has gained special attention, as this mediumoffers greater possibilities of further tailoring servicesto stakeholders' needs.

Qualitative Measurement ActivitiesThe growing emphasis on the use of qualitativemeasures to test performance were the focus in anumber of papers examining aspects of service quality.Elisabeth Hart, Patricia Gannon-Leary, and LorraineNoel's (University of Huddersfield) "The Use of FocusGroups in the Evaluation of Services" examined theuse of focus groups as an assessment tool and detailedthe value of this method in the assessment of lendingservices. Robert Bluck's (University of Northumbriaat Newcastle) paper, "Team PerformanceMeasurement of Mystery?", pointed out that,although teams are increasingly used in many aspectsof library work, performance measurement activitiestend to concentrate on the individual or the organiza-tion as a whole, rather than as part of an overallsystem. Of particular interest is a study done bythe CAVAL-Reference Interest Group Working Partyin Victoria, Australia. "Performance Indicators forReference and Information Services: A Study ofAcademic Libraries in Victoria, Australia," written byBarbara Paton, Eva Fisch, David Cunnington andRosemary Cotter from LaTrobe University in Victoria,describes CAVAL's activities in identifying measuresused to evaluate reference and information services inacademic libraries. Further, Marjorie Murfin (OhioState University) and Michael Havener's (Universityof Oklahoma) "Cronbach Revisited: Positive Bias:A Powerful Enemy to Validity in Library Surveys"tells of work done on the use of the ReferenceTransition Assessment Instrument (RTAI) in academiclibraries. This instrument has been used since 1983 inthe Wisconsin-Ohio Reference Evaluation Program toevaluate reference services in 109 academic and 121public libraries in the U.S. and Canada. Although theinstrument's statistical validity has been proven, only4% of U.S. academic libraries have chosen to partici-pate in the program. In their paper, Murfin andHavener make a strong case for use of the RTAI inmore research and academic libraries. -

The applicability of the SERVQUAL instrument inthe library environment was examined in two papers.The first, Danuta Nitecki's (Yale University)"Assessment of Service Quality in Academic Libraries:Focus on the Applicability of the SERVQUAL" is an

A R L 1 9 7 A P R I L 1 9 9 8 17

Continued

overview of the instrument and its applicability tothe study of service quality in academic libraries.According to Nitecki, in an environment whereacademic libraries want to be more responsive to theirclients, the SERVQUAL instrument offers a strategyfor defining and measuring quality of library services.Nitecki reviewed the results of eight academic librarystudies that used SERVQUAL to measure the qualityof their ILL and reference services, but pointed out thatthe usefulness of the instrument in improving servicemanagement in academic libraries is only beginning tobe discovered.

SERVQUAL was also the topic of, "Feedback froma Captive Audience: Reflections on the Results of aSERVQUAL Survey of Interlibrary Loan Services atCarnegie Mellon University Library." Presented byJoan Stein, this paper emphasizes the choice of theSERVQUAL instrument as a valid instrument tomeasure user perceptions. Stein reported that theCarnegie Mellon survey helped reshape staffperceptions of users' expectations, increased a senseof shared purpose, and helped staff set realisticpriorities for ILL services.

Service Level AgreementsAnother way to ensure that clients' needs are metare through Service Level Agreements (SLAs). SLAsmay be defined as a set of agreements, or a working"contract," that establishes the relationship betweenthe service provider and its clients, quantifying theminimum acceptable service to the customer.Malcolm Smith's discussion, "The Use of Service LevelAgreements at the British Library," focused on theLibrary's effort to establish SLAs between its variousdepartments as well as with external customers. Hedescribed the SLA format; the benefits, disadvantages,and obstacles to implementing SLA's; and how theseare to be used and implemented in the British Libraryin the future. Smith concluded that SLAs are valuabletools in achieving a number of objectives, such asincreasing the accountability of service providers,creating a customer-oriented institutional culture, anddeveloping an internal quality chain in addition tobetter relations with customers.

Jo Aitkins' poster session, "Setting Standardsand Monitoring Performance: The Experience ofInformation Services at the University of Sunderland,"was an example of a variation of the SLA, where,instead of a contract, the library decides to self-imposeunilateral standards focused on customer service.Among the cited benefits are: better customer service,heightened staff and client awareness of available andexpected services, improved monitoring of trends, andheightened staff morale.

Developing InternationalPerformance StandardsIt is no wonder that, with the concepts of internationalcooperation and digital libraries gaining in popularity,International Standard Organization (ISO) standardsand their implementation are given attention. Twopapers covered work undertaken by a number ofgroups in order to facilitate the creation and adoptionof the ISO Standard for Library PerformanceIndicators. Jacob Harnesk's (The Royal Library,Stockholm) paper, "The ISO Standard on LibraryPerformance Indicators" detailed the slow and difficultwork involved in developing international standardsand gives a description of the current situation. On asmaller scale was Reider Jan Zwart's (Delft Universityof Technology) paper, "Implementing the ISOStandard on Library Performance Indicators at DelftUniversity of Technology Library." He describedactivities undertaken to develop a document deliverysystem (DocUTrans) in partnership with KN/Minolta,with the goal of acquiring the ISO 9002 certificate forquality service--which they did achieve, in January1997.

Comparability Across Sectorsand National BoundariesA number of papers addressed the difficulty of com-paring libraries across sectors and national boundaries.Antje Cockrill and Judith Broady's (University ofWales) "Practical Issues of Performance Measurementin British and German University Libraries"; and "Useand Interest in Performance Measures: DifferencesBetween Library Sectors," by Sian Lambert, JonathanWilson and Tony Ou 1ton (Manchester MetropolitanUniversity) are two examples.

Management Information ServicesManagement Information Systems (MIS) and DecisionSupport Services (DSS) were topics in one session.John Blagden and Jane Barton's (Cranfield University)thought-provoking paper, "Can You Compare OneUniversity's Performance- with Another?" describes aproject whose aim was to develop a set of indicators tofacilitate comparisons of university library perfor-mances. It builds on existing work done in the FollettReport, The Effective Academic Library Report, andthe group of Concerted Action on ManagementInformation for Libraries in Europe (CAMILE) pro-jectsDECIDE, EQLIPSE, DECIMAL, and MIN-STREL. The project is unique in its goal to developindicators acceptable to university funding bodies,Vice Chancellors, and other senior staff within theU.K. educational sector. Issues studied are electroniclibrary or networked resources measures, access vs.ownership, document delivery, "stock" utilization and

A R L 1 9 7 A P R I L 1 9 9 8 10

availability, user population definitions, and how toassess and compare user satisfaction levels acrossinstitutional boundaries.

The CAMILE projects were discussed in moredetail in the poster session "European Union SupportSystems in Libraries Projects CAMILE." It is anticipat-ed that the project's concerted action will lead to agreater and more effective use of management infor-mation tools and techniques in libraries in Europe, andthat it will encourage the use of European-focused per-formance measures and contribute to greater commu-nication of information between European libraries.More information about CAMILE is available on theWeb: <http:/ /www2.echo.lu/libraries/en/projects/camile.html>.

The Electronic/Digital Library andAssessment of Library NetworksA number of papers examined various aspects ofassessment as they pertain to the digital library, toelectronic resources, and to issues of assessment oflibrary networked services. All papers acknowledgedtheir debt to the pioneering work of Charles McClureand Cynthia Lopata in the study Assessing the AcademicNetworked Environment: Strategies and Options(Washington, DC: CNI, 1996).

Peter Brophy and Peter Wynne (Central Universityof Lancashire), in their paper, "PerformanceMeasurement and Management Information for theElectronic Library (MIEL)," reviewed the MIELProgramme, whose aim was to examine the need fornew performance measures for the emerging electronicor digital library. Recommendations included: furthertesting of any proposed indicators in a live environ-ment, agreement on core indicators and internationalstandards, and the possibility of using qualitativeassessment instruments such as SERVQUAL.

Amos Lakos (University of Waterloo) sought tocreate a framework of discussion on the new networkedenvironment in his paper, "Identifying and AssessingLibrary Clients in a Networked Environment: Issuesand Possibilities." In a networked environment, espe-cially a web-based one, traditional ways to identifyand measure clients and their activities becomes muchmore complicated. Issues of identification and authen-tication of clients become essential to the delivery ofservices and to the development of service policies.Issues of privacy and confidentiality also must to beaddressed. In order to examine these concerns, vari-ous tools for tracking identities and activities onOPACs and websites are explored, as well as thepossibility of using new and changing businessintelligence tools for analysis.

F. W. Lancaster, University of Illinois at Urbana-Champaign, brought the conference to a close with his

keynote presentation, "Evaluating the Digital Library."Lancaster identified a number of concerns we have inthe move toward the electronic library, among themchanging clientele and a changing resource base. Healso raised the issue of a change in professional front-line activities, with the major concern that technologynot only forces but also enables users to use less libraryservices. Finally, Lancaster emphasized the importantrole librarians will play in the educational process.

Peter Brophy, in his conference summary, pointedout the many future-oriented sessions that had beenheard, the community's growing understanding of thelibrary as a social construct, and the growing focus onmeasures of outcome and user satisfaction in the pro-fession. He emphasized the need for constant changein order to shape effective roles for libraries in thefuture, and reminds us to keep in mind two key ques-tions as we face this future: What are libraries for?What is at the core of information management?

Those attending the conference were encouragedto continue to look internationally for new ideas toassess library performance, since the issues are ofinterest worldwide.

For more information, see the published proceedings:Proceedings of the 2nd Northumbria InternationalConference on Performance Measurement in Librariesand Information Services, University of Northumbria atNewcastle: Information North, 1998. Information North'saddress: Information North, Bolbec Hall, Westgate Road,Newcastle upon Tyne NE1 1SE, England. Tel: +44 (0)191232 0877. Fax: +44 (0)191 232 0804.

JUST RELEASED

Measuring the Performance of InterlibraryLoan Operations in North American Research

& College Libraries

The report of a two-year study funded by TheAndrew W. Mellon Foundation. By Mary E. Jackson,

ARL Access and Delivery Services Consultant.1998, 122 pp., $45

ISBN 0-918006-33-3

To order, contact <pubs@arLorg>.

A R L 1 9 7 A P R I L 1 9 9 8

4

'4k,a1

y---

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-296-2296 FAX 202-872-0884<http://www.arl.org/>Copyright: @ 1998 by the Association of Research Libraries

ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

ARL CALENDAR 1998-

1998June 25-July 2 American Library Association

Washington, DC

ARL Board MeetingWashington, DC

ARL/OLMS FacilitationSkills InstituteWashington, DC

ARL Board andMembership MeetingWashington, DC

Educom '98

Orlando, FL

ARL/OLMS Leading ChangeInstituteChicago, IL

ARL/OLMS LibraryManagement Skills Institute I:The ManagerWashington, DC

CNI Task Force MeetingSeattle, WA

CAUSE '98

Seattle, WA

July 27-28

October 7-9

October 13-16

October 13-16

October 27-28

November 16-19

December 7-8

December 8-11

Executive Director Duane E. WebsterEditor G. Jaia Barrett, Deputy Executive DirectorCopy Manager Karen A. WetzelDesigner Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscription; Nonmembers$50 per year.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arl.org>.

99

1999February 11-12

May 11-14

July 26-27

October 12-15

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingKansas City, MO

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingWashington, DC

BEST COPY AVAILABL

42

J

cf)

C3\

A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONS

LICENSES AND INFORMATION POLICY:AN UPDATE ON UCC ARTICLE 2Bby Laurel Jamtgaard

Att he October 1997 ARL MembershipMeeting, Robert Oakley, Director ofthe Georgetown Law Library, gave a

thoughtful and informative presentation about aproposed model law known as Article 2B of theUniform Commercial Code.' This law is poised toshape the legal landscape for transactions in infor-mation products, including copyrighted works,databases, and computer software. It is thereforelikely to directly impact the operations of alllibraries and academic institutions.

Some BackgroundIn the United States, Uniform State Laws aredrafted by committees of attorneys, reviewedby two organizationsthe American LawInstitute (ALI) and the National Conferenceof Commissioners on Uniform State Laws(NCCUSL)and then, if approved, sent aroundto the fifty states for adoption in whole or in part.Adopting a model law helps to facilitate interstatecommerce because market participants can thenbe confident that they are operating under asimilar body of law from one state to the next.Currently, areas of law encompassed in theUniform Commercial Code include contracts forgoods, leases, banking agreements, and securedlending transactions.

The intention of the drafters of Article 2B is toprovide standard rules for regulating licenses ofinformation products and intellectual propertyrights. The hope, as with all Uniform State Laws,is to codify existing case law.2 Current case lawvaries, however, across states and federal circuits,so this process is not trivial.

43

Although many agree that the idea of codifyingthe law of licenses is a good one, there has been anamazing amount of controversy involved in the effort.The drafting process for Article 2B has been underwayfor more than five years. In the past year, the draft ofArticle 2B, a 200-page document, has been updatedapproximately every eight weeks. The April 1998draft of Article 2B addresses, among other things:

contract formation (includes offer and acceptancerequirements for electronic agents and mass-marketlicensesa.k.a. "shrink-wrap" licenses);construction and interpretation of license terms;warranties (includes implied warranty ofmerchantability of a computer program);contract performance; and

remedies (including the right for licensors toemploy "self-help" measures to retrieve or disableinformation products if licensees breach thelicense).

Article 2B is in many ways a moving target and enter-ing the debate over specific provisions involves a steeplearning curve, but more entrants are important to theprocess.

Concerns .

In March of this year, Geoffrey Hazard, Chair of theAmerican Law Institute (ALI), distributed a letteroutlining some reasons why ALI would not submit 2Bto a member vote at its May Annual Meeting inWashington, D.C.3 The first concern expressed wasthat the scope of 2B was so broad that there werelikely many voices still to be heard from and manywho may be unaware that they will be affected.

CURRENT ISSUESContinued

The example he gave hits close to homehe queriedwhether a library card constituted an "access contract"under 2B and whether the library community hadvoiced an opinion one way or another.4

ScopeArticle 2B began as an effort to standardize softwarelicenses, but its scope has expanded (and contracted)at various times to include motion picture, broadcast,publishing, banking,and other industries.Representatives from severalof these sectors have onlyrecently chimed in. Manyparties who will be dramati-cally affected have yet toeven put Article 2B on theirradar screens.

The April 1998 draft ofArticle 2B covers "licensesand software contracts,"defining a license as "anagreement that authorizesaccess to or use of informa-tion or of informationalproperty rights....""Information" is defined as "data, text, images, sounds,mask works, or works of authorship."

A motion by Stephen Chow, an attorney with Smith& Sohen in Massachusetts, to reduce the scope of Article2B was discussed at the ALI Annual Meeting in May.The motion had sought to limit the scope to:

1) Software contracts;

2) access contracts; and

3) such other transactions that are included expresslyand defined with sufficient clarity to avoid surpriseto affected parties.5

However, Mr. Chow withdrew the motion withoutputting it to a member vote on the understanding thatthe Drafting Committee had plans to revisit the questionof scope during the summer.

Intersection with Federal LawAnother concern of Mr. Hazard, the relationship ofArticle 2B to federal law, was the focus of a three-dayconference at the University of California at Berkeley inApril.6 Some argue that if non-negotiated licenses(a.k.a. "shrinkwrap licenses") are honored as providedin Article 2B, licensors will use licenses to strip awayfair use rights from consumers and thus jeopardize thedelicate balance struck by copyright law. Conferenceparticipants also raised questions about the interactionof Article 2B with trade secrecy law, patent law,competition policy, and the First Amendment.

To what extent does federal copyright law preemptlicense terms that interfere with copyright limitationssuch as fair use and first-sale rights? This is a toughquestion and one courts have struggled with on a case-by-case basis.' Article 2B has been drafted with the statedintent of "staying neutral" with regard to the issue offederal preemption. This stance was criticized at theBerkeley Conference by David Nimmer, co-author of

the Nimmer on CopyrightTreatise (and no relationto Raymond Nimmer,Reporter for Article 2B).He sees Article 2B's"neutrality" as disingenu-ous because it fails "toprotect even.., obvioususer rights" but blesses"as presumptively validprovisions that would robusers of those rights."5

Although Article 2Badmittedly creates federalpreemption problems, thedrafters are looking to

The April 1998 draft of Article 2Bcovers "licenses and software

contracts," defining a license as "anagreement that authorizes access to oruse of information or of informationalproperty rights...." "Information" is

defined as "data, text, images, sounds,mask works, or works of authorship."

Congress and the courts to solve them. At this time,attempts to pass clarifying language in Congress havenot been successful.9

Impact on LibrariesAt the Berkeley Conference, Peter Lyman of theUniversity of California at Berkeley spoke of the poten-tial impact of Article 2B on research libraries and thebroader research community. He described the alreadyspiraling costs of periodicals and online database feesand expressed frustration at what he perceived tobe Article 2B's contribution to increasing thebargaining power of information publishers incontract negotiations.

For example, as Bob Oakley mentioned in his talk ayear ago, 2B-502 declares that license terms that restrictthe transfer of informational property rights are enforce-able. Does this affirm a position that licenses can castaside first-sale rights granted under copyright law?1°I believe the answer is an oblique "yes."

The explanatory notes to the April 1998 Draft statethat 2B does not apply to the sale of books, but only tolicenses of information products. As Professor CharlesMcManis pointed out in Berkeley, most printed booksalready contain onerous language prohibiting the copy-ing of any portion for any purpose, but such statementshave not been found to trump "fair use."11 Well, it istime to take another look because 141 that is missingfrom books is the snap [to seal them closed[and a law

A R L 1 9 8 J U N E 1 9 9 8

saying that a snap creates an enforceable contract."12With a snap and Article 2B, what in the past has been asale of a book could, with the blink of an eye, become alicense (could there be a more dramatic reason forlibraries to pay attention to Article 2B?).

Access contracts as defined in 2B-615 will beextremely important to libraries because these contractsare dealt with daily. 2B-615 provides that a license canplace use restrictions on theinformation accessed. Andwhere does fair use fit inhere? Well, it doesn't unlessyou put the rights in thecontract explicitly orCongress acts to resolve thepreemption question infavor of fair use.

It is crucial that thelibrary community enter intothe Article 2B conversation.The statute, when/if adopt-ed, will provide default rulesthat can be overcome by lan-guage in the contract itself sofamiliarity with Article 2B

information about Article 2B.I3 It includes html versionsof most comments submitted to ALI, NCCUSL, and theDrafting Committee, as well as links to online news arti-cles on the topic. In addition, Ed Foster of InfoWorld hasrecently started hosting a discussion page on Article 2B.14

More voices from the library and educational com-munities need to be heard. One way to do so is to submitcomments to the NCCUSL Commissioner from your

home state. Letters can also

Although Article 2B is a state-basedlaw, it is a critical front in the battle

because it strengthens contentproviders' ability to use "shrinkwrap"

and "clickwrap" licenses to controltheir information products'

post-distribution.

may therefore enable librarians to overcome the mostproblematic defaults during license negotiations. Butbe forewarnedwith the affirmation of mass-marketlicenses, Article 2B will transfer considerable bargainingpower to the publisher's side of the table.

Where Does Article 2B Stand?ALI has announced plans to review the progress onArticle 2B at its November Council meeting and may atthat time recommend the Article for a final vote at its1999 Annual meeting.

As late as April, NCCUSL had planned to submit 2Bto its membership for a final vote at its Annual Meeting,July 24, 1998. It has stepped back somewhat from thatposition and has announced plans to read "non-contro-versial" aspects of 2B into the record at the July meetingand to entertain motions and debate on specific sections.NCCUSL will not submit Article 2B to a final state-by-state member vote until its 1999 Annual Meeting atthe earliest.

The Drafting Committee had planned that nofurther drafting committee meetings would be neededafter May, but with the concerns raised by ALI andothers more meetings have been scheduled.

There is an increasing awareness of Article 2B anddiscussion forums are popping up where librarians andother members of the research community may enterthe debate. For example, attorney Carol Kunze's "2BGuide" site offers the most comprehensive online

4 5

be submitted directly to theleadership of NCCUSL, theArticle 2B DraftingCommittee, and ALI.15

The Big$er InformationPolicy PictureThe significance of Article 2Bshould be considered in con-junction with the variouscopyright and database billscurrently working their waythrough Congress. The battlewaging over the appropriatedirection of domestic andinternational information

policy has reached a frenzy this summer.Some content providers are reacting to the prolifera-

tion of digital technology with a push for stronger legalprotection of information products. Not satisfied withadvances in existing legal and emerging technologicalmeans to protect information, these media, Hollywood,and publishing interests are driving bills throughCongress to extend the term of copyright,16 provide for anew right in databases wholly outside of copyright law,I7and create a "traveling right of trespass" (my phrase, nottheirs) for technologically protected works! Each ofthese efforts is a truckload of cement primed to pave theroad to a "pay-per-view" digital age.

Others, including the library community, seek toprotect the balance between information as property andinformation as a public resource. There has been supportfrom this group for a digital age copyright bill sponsoredby Representatives Boucher (D-VA) and Campbell (R-CA)19 that seeks to preserve the fair use and first-sale doc-trines and to provide a clear signal to courts about federalcopyright preemption of non-negotiated license terms.The challenge in this effort, besides facing the incrediblefinancial resources of the media, motion picture, andpublishing industries, is how best to translate thesebalancing principles into the digital world.

Although Article 2B is a state-based law, it is acritical front in the battle because it strengthenscontent providers' ability to use "shrinkwrap" and"clickwrap" licenses to control their information

A R L 1 9 8 TUNE 1 9 9 8

Continued

products' post-distribution. As currently drafted, it isyet another avenue enabling increased private "fencing"of information.

Try to Imagine...At the Berkeley Conference, Peter Lyman painted agrim picture of a university environment in which allinformation products are licensed. He shuddered as hedescribed the revelation that nothing in Article 2B pro-hibits the application of the proposed license regime toprinted books and periodicals. With that in mind, thecrystal ball view became one in which archiving was toocostly to continue; all information was available only ona pay-per-view basis and only to university studentsand professors; students were unable to re-sell textbooks; something akin to "fair use rights" were avail-able only for the least valuable resources thanks to aloss of bargaining power; and university librarieswould no longer be the catalyst for community-wideentrepreneurial research efforts. Innovation suffers.The economy suffers. Etc.

(You get the idea.)IS* P./ eI

About the AuthorLaurel Jamtgaard is an attorney and member of theState Bar of California. She has just completed a six-month research fellowship at UC-Berkeley withProfessor Pamela Samuelson focused upon intellectualproperty policy issues. She is currently working withARL and the Special Libraries Association on informa-tion policy and scholarly communication issues. Shewill return to California in October to practice law.

2

Robert L. Oaldey, UCC Article 2B: Some PreliminaryComments on a New Issue for the Library Communihy, October16, 1997. Available online at <http://www.arl.org/arlIproceedings/131 /oakley.html>. All references to the text ofArticle 2B in this article are to the April 1998 Draft, availableat <http://www.law.upenn.edu/bll/ulc/ulc.htm>.

Today when judges decide disputes involving licensingtransactions (such as publishing contracts, database accesscontracts, and contracts for motion picture rights) they basetheir decisions upon applicable case law precedents. If Article2B is adopted by the states, then judges will look to thestatute for initial direction on how to decide similardisputes.

3Geoffrey C. Hazard Jr., Letter to Gene N. Lebrun, Presidentof NCCUSL, and Charles Alan Wright, President, TheAmerican Law Institute, March 26, 1998. Available online at<http://www.softwareindustry.org/issues/guide/docs/ghmar98.html>.

4 An "access contract" is defined in 2B-102(a)(1) to involveaccess to electronic resources, but if a library card is used to

enable access to electronic resources in the library, it mayqualify.

5Revised Motion of Stephen Y. Chow, American Law Institute1998 Annual Meeting, May 14, 1998.

6 See Intellectual Property and Contract Law in the Information Age:The Impact of Article 2B of the Uniform Commercial Code on theFuture of Transactions in Information and Electronic Commerce,April 23-25, 1998, Berkeley Center for Law & Technology,<http: / / sims.berkeley.edu / BCLT/ events /ucc2b/>.

7 Compare Vault Corp. v. Quaid Software Ltd., 847 F. 2d 255 (5thCir. 1988) and ProCD v. Zeidenberg, 86 F. 3d 1447 (7th Cir.1997).

8David Nimmer, Elliot Brown, and Gary N. Frischling, TheMetamorphosis of Contract Into Expand (DRAFT), ConferenceProceedings, Berkeley Conference supra note 7.

9See e.g., House Bill 3048, 105th Congress, the Digital EraCopyright Act (1998), introduced by Representative RickBoucher of Virginia.

10The First Sale doctrine, codified in §109 of the Copyright Act,allows libraries or other owners of a copy of a copyrightedwork to transfer ownership or possession of the copy. Thisenables library lending practices as well as donations ofmaterials to libraries.

" See Bobbs-Merrill v. Isidor Straus, 210 U.S. 339 (1908).12

Charles McManis, The Privatization (or "Shrink-wrapping") ofAmerican Copyright Law (DRAFT), Proceedings of the April1998 Berkeley Conference on Article 2B, supra note 7.

13See <http: / /www.webcom.com/ software /issues/ guide/index.html>.

14 See <http:/ /forums.infoworld.com/threads/get.cgi?56033>.

18 Comments on the draft are encouraged by ALI and NCCUSLand the Drafting Committee itself. Send letters to NCCUSL at211 E. Ontario Street, Suite 1300, Chicago, Illinois 60611; to theArticle 2B Drafting Committee care of Reporter, Ray Nimmer,Law Center, University of Houston, 4800 Calhoun, Houston,TX 77004; and/or to the American Law Institute at 4025Chestnut Street, Philadelphia, PA 19104-3099.

16 H.R. 2589, 105th Congress, Copyright Term Extension Act(1998).

17H.R. 2652, 105th Congress, The Collections of InformationAnti-Piracy Act (1998). This bill, which creates broad newprotection for compilations, states that its provisions willnotpreempt the state law of contract and thus increases thepotential power of a licensing regime.

18 H.R. 2281, §1201(a), 105th Congress, WIPO Copyright TreatiesImplementation Act (1998). The anti-circumvention provisionof this bill creates a new right for information providers tocontrol post-distribution access to their works.

19H.R. 3048, supra note 10.

46A R L 198 JUNE 1998

ACADEMIC COMMUNITY SETS AGENDATO RECLAIM SCHOLARLY PUBLISHINGby Mary M. Case, Director, ARL Office of ScholarlyCommunication

Research universities have it within their powerto work with each other and scholarly societiesto transform scholarly communication into "a

system of electronically mediated publications that willprovide enhanced access to scholarly information andrelief from the escalating prices of commercial publish-ers." So conclude the participantsuniversity presi-dents, provosts, faculty, librarians, counsels, and repre-sentatives of scholarly societies and university pressesin a special Roundtable on Managing IntellectualProperty in Higher Education. The findings of thisgroup are reported in the essay "To Publish andPerish," featured in the March 1998 issue of PolicyPerspectives, the publication of the Pew HigherEducation Roundtable.

Noting that the rising cost of scholarly publicationsis not a "library problem," but a symptom of the deeperconflict between the sociology and economics of acade-mic publishing, the essay contrasts the expectation of anopen exchange of information within the academy tothe pricing and copyright practices of some commercialpublishers that control many of the major scholarly pub-lishing venues. In an effort to regain some control overthe research and scholarship generated by the academiccommunity, the Roundtable participants proposed a setof five strategies to address the problem. They recom-mended that:

promotion and tenure committees disentangle thenotions of quality and quantity in the work of thefaculty;libraries leverage their resources by creating a morecoherent market for scholarly publications;universities, led by their national associations, helpfaculty understand the implications of signing awaytheir intellectual property rights;universities and scholarly societies invest inelectronic forms of peer-reviewed scholarlycommunication; anduniversities and scholarly societies decouplepublication and faculty peer reviewed evaluationof the merit of scholarly work.

The participants stated that "The outcome we seek is aset of specific arrangementslinking institutions, theirfaculty, and their scholarly organizationsthat protectsthe rights of faculty and secures for their appointinginstitutions a more assured ability to provide access toresearch and scholarly information." While the chal-lenges are not insignificant, the group concluded that:

the risks of doing nothing substantially outweigh thedifficulty of doing somethingand doing it now!

A moment of opportunity is at hand, occasioned bythe potential for peer-reviewed electronic publishingand a sense of desperation spawned by runawayacquisition costs. Missing this opportunity willmean more rapidly accelerating costs, greatercommercial control, and, in the end, less access toscholarly communications.

The Roundtable was convened in Baltimore inNovember 1997 to confront the challenges facing theacademic community in maintaining access to significantresearch and scholarship at a time when both the volumeand price of information have increased nearly three-foldin the last decade alone. The Roundtable was hostedby the Johns Hopkins University and sponsored by theAssociation of American Universities (AAU), theAssociation of Research Libraries (ARL), and the PewHigher Education Roundtable. Funding was providedby the Gladys Krieble Delmas Foundation, the W. K.Kellogg Foundation, and The Pew Charitable Trusts.

ARL is pursuing a number of the recommendationsin the report, including the continued development ofSPARC, the Scholarly Publishing & Academic ResourcesCoalition, as an investment in electronic forms of scholarlycommunication that stimulate competition in the market-place. In addition, ARL will seek to partner with the AAUto develop an organized campaign to inform faculty of thecurrent economics of scholarly publication and to exploreoptions for faculty to retain rights to their own work forthe purposes of research and education.

In the meantime, ARL member institutions are find-ing "To Publish and Perish" a useful document for open-ing discussions on their campuses. Several directors ofARL libraries have distributed the report to and heldmeetings with state legislators, Boards of Regents, cam-pus administrators, university budget officers, deans,faculty senates, library committees, journal editors, juniorfaculty, graduate students, campus copyright commit-tees, and departmental faculty. One institution reportedsending the essay to all faculty and students. ARL isdeveloping a website to track progress on the recommen-dations that resulted from the Pew Roundtable and toshare experiences that ARL members have as they use thereport in campus and broader community discussions.Watch <http://www.arl.org/scomm/pew/> for furtherinformation.Individual copies of the March 1998 issue of PolicyPerspectives can be obtained from the Institute forResearch on Higher Education, University of Pennsylvania,4200 Pine Street, 5A, Philadelphia, PA 19104-4090; phone1-800-437-9799. The issue is available on the Web at<httplIwww.arLorglscomm/pew/pewrept.html> or in PDF at<http:11www.irhe.upenn.edulcgi-binIpp-cat.pl#V7N4>.Member libraries may contact ARL at <pubs@arLorg> forinformation about ordering copies of the report.

47 A R L 1 9 8 J U N E 1 9 9 8

Continued

A SPOTLIGHT ON HIGH-PERFORMINGILL/DD OPERATIONS IN RESEARCHLIBRARIESby Mary Jackson, ARL Access & Delivery ServicesConsultant

Over the past two years ARL collected a wealthof data on the 1996 performance of interlibraryloan/document delivery (ILL/DD) services of

North American research and college libraries. Withfunding from The Andrew W. Mellon Foundation andadvice from the Council on Library and InformationResources, ARL set out to identify the attributes oflow-cost, high-performing ILL/DD operations againstwhich other libraries could compare themselves andfind ways to improve their local performance. High-performing borrowing or lending operations weredefined as the top ten research libraries performingbest in one or more categories of the three borrowingperformance measures (cost, fill rate, or turnaroundtime), or in one or both of the lending measures (costand fill rate). Averages were calculated for the tenlibraries with the lowest borrowing unit costs, the tenlibraries with the highest borrowing fill rates, the tenwith the fastest borrowing turnaround times, and forthe ten with the lowest lending unit costs and the tenlibraries with the highest lending fill rates. The accom-panying chart presents these averages and comparesthem with the mean and median performance of all 97research library participants. The level of user satisfac-tion with ILL/DD services, as measured in a samplesurvey in this study, did not vary significantly acrossall research library participants and thus is not includ-ed in this analysis.

What are the characteristics of ILL/DD operationsin the ten libraries with the lowest unit costs, the tenwith the highest fill rates, or the ten with the fastestturnaround times? How do these high-performingoperations differ from the other research libraryparticipants?

High-Performing Borrowing OperationsThree measures were used to track the performance ofILL borrowing: unit cost, fill rate, and turnaroundtime. Of the 97 research library participants, ColoradoState University was the only research library thatappeared in the top ten for all three borrowing mea-sures. In addition, the University of Chicago wasamong the ten libraries with very low unit costs andvery high fill rates. Four additional research librarieshad very low unit costs and very fast turnaroundtimes: the University of Cincinnati, Ohio University,the University of Illinois at Chicago (UI-C) MainLibrary's LCS unit, and the LCS unit of UI-C's Libraryof the Health Sciences. (LCS, or Library Computer

System, is the unit that initiates loan requests on theILLINET Online system, the union catalog of Illinoislibraries.)

Nineteen other research libraries ranked in thetop ten for a single borrowing performance measure.Thus, a total of 25 research libraries ranked in the topten for one or more borrowing performance measures.These high-performing borrowing operations share anumber of characteristics.

User-Initiated OrderingA majority of the top-performing borrowing opera-tions offer electronic user-initiated ordering, eithervia statewide systems such as OhioLINK or throughlocally-developed electronic order forms. Systems thatpass electronic patron request forms directly to thepotential lender or into national messaging systemseliminate the need for ILL staff involvement in theseinitial steps of the borrowing process. Staff costsrepresent two-thirds of the borrowing unit cost, soby increasing user-initiated ordering staff costs arelessened and the borrowing unit cost is thus lowered.

Maximizing TechnologyMaximizing use of technology is a second characteris-tic shared by high-performing ILL borrowing opera-tions. Nine of the ten research libraries with the veryfastest turnaround times encourage or even requirepatrons to submit requests electronically. Fourlibraries use OCLC's ILL Prism Transfer orFirst Search-ILL link to enable patrons to identify itemsin online databases and place electronic ILL requests.These system features are designed to reduce delays ininternal processing because patron requests are trans-ferred directly into the OCLC ILL system without staffre-keying requests. Eliminating the need to re-keyrequests speeds up the process and saves staff time,thereby freeing staff to work on other requests.

Many of the high-performing borrowing opera-tions use a single ILL messaging system. Workflowand procedures are streamlined when only one systemis used, resulting in faster turnaround time and lowerunit costs.

The ILL operation at Colorado State may be theultimate example of maximizing technology, as theirborrowing operation is truly paperless. Patrons arerequired to submit requests electronically; in sevenkeystrokes staff transfer those requests into the OCLCILL system. OCLC ILL requests are subsequentlycopied into the CLIO management software for follow-up tracking and report generation. Colorado State hasdeveloped a series of "custom holding paths" andOCLC macros (keystroke shortcuts) that permit stu-dent assistants to process the majority of their borrow-ing requests. Using available software and a modest

A R L 1 9 8 J U N E 1 9 9 8

amount of local programming permits ILL requests tobe received from the patron and to be sent to the firstpotential lender within hours, rather than the averageof four days it takes most research libraries to com-plete the same segment of the process.

Many high-performing borrowing operationsalso use OCLC's ILL Fee Management system to payinvoices, another example of software that can signifi-cantly minimize the time spent by staff in handlinginvoices, thus reducing the staff unit costs.

Delivery TechnologiesResearch libraries with the fastest borrowing turn-around times use delivery services such as Ariel, fax,state/regional couriers, FedEx, or UPS to a greaterextent than do libraries with the very slowest turn-around times. Libraries with fast turnaround time askother libraries to supply all articles via Ariel and/orfax, and some even choose lenders based on Arielavailability. These borrowers are also willing to payfor next-day delivery for most requests, rather thanjust for "super, super rush" requests.

Staffing Levels and Staff ExpertiseHigh-performing borrowing operations employ differ-ent levels of staff than the aggregate of all researchlibraries. In particular, these operations have fewerprofessional non-supervisors and are more often man-aged by support staff supervisors than the borrowingoperations of other research libraries that participatedin the study. The ten libraries with the very lowestborrowing unit costs also rely slightly more on stu-

dents and slightly less on support staff than do eitherall research libraries or the subset of the 25 librarieswith high-performing borrowing operations.

Staff in these operations also monitor actualperformance of specific lenders and choose only thosewith fast turnaround time and liberal lending policies,both of which translate into higher fill rates.

Libraries with high borrowing fill rates have staffwho know unique collections in other libraries nation-ally and internationally, are aware of a variety ofcommercial suppliers, have command of complexcataloging rules, and search print bibliographies ratherthan assuming that "if the item isn't found on OCLC,it doesn't exist." Thus, staff are able to obtain difficultmaterials without expending significant additionaltime on each request.

Local PoliciesLocal policies also contribute to a library's high bor-rowing performance. Nearly three-quarters of thehigh-performing borrowing operations do not limit' thenumber of requests patrons are permitted to submit atone time. Half do not charge patrons for loans orcopies, eliminating staff involvement in collecting feesand/or maintaining accounts. Both of these policiesreduce staff mediation, and therefore reduce theircosts. Two-thirds of the high-performing borrowersmail photocopies to patrons, again eliminating anadditional patron contact with ILL staff at the end ofthe process. Finally, these libraries do not limit theamount they are willing to pay to obtain an itemthey

RESEARCH LIBRARY ILL/IDD PERFORMANCE MEASURES

Borrowing

Unit Cost

Fill rate

Turnaround time

Lending

Unit Cost

Fill rate

High-Performance

$9.76

93%

10.2 calendar days

Mean Performance Median Performance

$18.35

85%

15.6 calendar days

$16.63

86%

14.9 calendar days

$4.87 $9.48 $8.14

78% 58% 57%

4 9 A R L 1 9 8 J U N E 1 9 9 8

Continued

choose libraries that provide reliable service (even ifthere is a charge) rather than trying to find librariesthat do not have a lending fee, thereby saving stafftime, reducing staff costs, and minimizing overallturnaround time.

High-Performing Lending OperationsTwo measures were used to track the performance ofILL lending operations: unit cost and fill rate. Of the97 research library participants, only the University ofAlberta and the University of Wisconsin-Madisonrecorded very low unit costs and very high fill rates.An additional 17 libraries recorded either very lowunit costs or very high fill rates, but not both. Thestudy confirmed that high-performing lendingoperations in research libraries are comparable toor slightly better than the average performance oflending operations in the college libraries thatparticipated in the study.

Expenditures on TechnologyOn average, high-performing lending operations inresearch libraries spend (on a unit cost basis) just overtwice as much on equipment (hardware and software)than do most research libraries. Although the studydid not confirm this, it is reasonable to speculate thatincreased use of technology may reduce unit cost iftechnology is deployed in ways that minimizes staffinvolvement in some of the steps of the lendingprocedures.

Innovative Use of StaffLike most research libraries, Alberta and Wisconsin-Madison have centralized lending operations, requir-ing retrieval from multiple branch libraries. Bothlibraries train staff (student assistants and/or full-timestaff) to retrieve from one or two branch or departmen-tal libraries rather than assigning staff to pull materialsfrom all branch or departmental libraries on campus.Staff become familiar with a specific collection (suchas where materials may be located if not properlyshelved), thus increasing the lending fill rates.

Alberta has taken this specialization a step furtherand encourages libraries to contract with them for pri-ority lending services. A borrowing library pays thesalary of an Alberta staff member to fill their requests;in doing so, the borrowing library receives preferentiallending service from the Alberta collection. This is aninteresting example of outsourcing a library's borrow-ing operation to a library with a collection that can filla large portion of requests. (One of the college libraryparticipants in the study also relies on a neighboringlibrary to verify and fill its borrowing requests, result-ing in a very low borrowing unit cost for that collegelibrary.) These innovative lending services confirm theonly statistically significant finding in the study: as

lending volume increases, the lending unit costdecreases. Libraries wishing to lower their lendingunit costs could do so by increasing the number oflending transactions they fill.

Staff CompositionLike high-performing borrowing operations, high-performing lending operations rely on a different mixof staff categories than most research libraries. Theten libraries with the lowest lending unit costs use noprofessional staff in non-supervisor positions in theirlending operations. Libraries with the lowest lendingunit costs use student assistants slightly more andsupport staff slightly less than other research libraries.However, the 19 high-performing lending operationsuse support staff slightly more and student assistantsslightly less than all research libraries. Given this vari-ability in use of student assistants and support staff, itis reasonable to speculate that the low unit costs forboth groups of high-performers are driven more bythe absence of professional staff in non-supervisoryroles than in the mix of support staff and studentassistants used.

ConclusionThe ILL/DD Performance Measures Study identifieda handful of operations that chart new and innovativeways to manage interlibrary loan operations. Thechallenge is to spread awareness of the potential forimprovements and develop strategies for replicatingthe best practices in more libraries.

To encourage this process, ARL is planning aseries of workshops to assist attendees in evaluatingand adapting these performance enhancing proceduresand tools for improving local services. The workshops,aimed at ILL/DD managers and their supervisors,will provide additional details on characteristics ofhigh-performing borrowing and lending operations.The initial workshops are tentatively planned for thesecond half of 1998. Specifics will be posted on the ARLwebsite <http:/ /www.arl.org/> when available.

Measuring the Performance of Interlibrary LoanOperations in North American Research and CollegeLibraries was published in May by ARL. This finalreport details the findings of the ARL ILL/DDPerformance Measures Study. The Executive Summaryof the report and ordering information may be found at<http:/ /www.arl.org/access/illdd/pr.shtml>. Alsoonthe Web is a previous newsletter article summarizingthe wide range of ILL/DD performance documentedin research and college libraries, "Measuring thePerformance of Interlibrary Loan and DocumentDelivery Services" (ARL 195, Dec. 1997); it isavailable at <http:/ /www.arl.org/access/illdd/illdd-measperf9712.shtml>.

f_A R L 1 9 8 J U N E 1 9 9 8

Clifford Lynch, Executive Director

QUANTTTIES OF INFORMATIONby loan K. Lippincott, Associate Executive Director, CNI

The Coalition for Networked Information's Spring Task ForceMeeting was held in Crystal City, Virginia on April 14-15,1998. A wide variety of topics related to networked informa-tion were covered in a plenary talk by Michael Lesk, a panel ondigital preservation, and project briefing sessions. A summaryof Lesk's talk is provided here; the complete meeting report andmaterials from the project briefing sessions are available onCNI's website at <http:11www.cni.orgltfms/1998a.springl>.

The opening plenary session at CNI's Spring TaskForce Meeting featured a presentation by MichaelLesk, on leave from Bellcore and currently serving

as Division Director of Information and IntelligentSystems at the National Science Foundation, where heoversees the Digital Libraries 2 initiative. Lesk has writ-ten extensively about information and digital libraries,and recently produced the valuable book, Practical DigitalLibraries: Books, Bytes, and Bucks (Morgan KaufmannPublishers, 1997). CNI's Executive Director CliffordLynch observed in his introduction of the speaker Lesk'sreputation for provocative comments and for insightfuland pithy observations that open up new areas of specu-lation. Lesk's theme was the amount, worth, and usabili-ty of the information available in the world.

Managing the sheer quantity of information availablein digital form is a daunting task. Lesk gave examplesand presented charts to demonstrate the great amountand variety of information that exist. His examplesranged from photograph collections by professionals andamateurs to digitized phone conversations to satellitedata. He referred to economists' discussions about sup-ply and demand and its relationship to value and thenasked whether we will have so much informationincluding freely available informationthat it willbecome worthless.

A related issue is how much information will actuallybe seen or used by humans. Lesk cited work by a psy-chologist who theorizes that the human memory holdsabout 200 megabytes and that people can take in onebyte/second of information. An average Americanspends 300 hours/year with some kind of media. In theworld of the future, then, there will be more informationthan human memory can cope with, leading Lesk to con-clude, "[I]n a short time there will be so much informa-tion that only a small fraction of it will be seen by ahuman being." This is true already of some largedatasets, such as those collected by NASA.

Lesk's next question, then, was: If there's so muchinformation, is it worth anything? He noted that sincelibraries don't charge for information, no one truly knowswhat it isor they areworth. Universities spend anaverage of three percent of their budget on libraries, but

do they know if it is well spent? Lesk believes, however,that having a lot of information is, in fact, valuable. Theinformation sector of our economy is growing, and in moreand more industries a larger portion of costs will be infor-mation-related instead of materials- and/or labor-related.

Lesk had some specific suggestions for the universitycommunity. He encouraged universities to develop high-prestige sections of their websites, stimulated by cashpayments to authors, that would become desirable venuesfor scholarly publications. Universities could control theeconomic system of scholarly publication and the preser-vation of that information. He also stated that, while in afew years it will be possible to record literally all informa-tion, we need to decide what we can do with that informa-tion; therefore, we need more research on information-seeking and information use.

Lesk concluded with his observations on what thesetrends in information mean to society and to informationprofessionals:

Summarizing is the key problem for us to work on.We need to be able to take quantities of information andabstract the useful parts in all formatstext, audio,video. There has been relatively little research on thisbut people are starting to attend to it.

Librarians will be worth more, and libraries may beworth less. People good at managing memory might bethe ones who matter.

Attention is the scarce resource, not information.Organizing information and helping people find theirway through it is a 'good thing.'

Information professionals will become more valuableas people increasingly rely on an information specialistto help them deal with the quantity of information outthere. The focus of information specialists must be onhelping people.

Paul Evan Peters Award and Scholarship FundThe selection process is underway for the first recipientof the Paul Evan Peters Award. The winner will beannounced at CNI's Fall 1998 meeting. The award issponsored by ARL, EDUCAUSE, Microsoft, and Xerox.Additional information is available at <http://www.educause.edu/awards/awards.html>.

Duane Webster, Executive Director of ARL, encour-ages additional contributions to the Paul Evan PetersScholarship Fund. A committee headed by Charles Henryof Rice University is soliciting gifts. Information on thescholarship fund is available on CNI's website at <http:/ /www.cni.org/docs/scholarship/pep-scholarship.html>. I

Fall Task Force MeetingThe Fall 1998 Task Force meeting will be held at theSheraton Seattle Hotel and Towers in Seattle, Washingtonon December 7 and 8, immediately preceding CAUSE98.

51 A R L 1 9 8 J U N E 1 9 9 8

EVALUATING LIBRARY DIRECTORS:A CHECKLISTprepared by George Soete, ARLIOLMS OrganizationalDevelopment Consultant

To assist those who are responsible for or who willparticipate in the performance evaluations of aca-demic library directors, the ARL Committee on

Leadership and Management worked with ARL/OLMSOrganizational Development Consultant George Soeteto collect and analyze information about library directorroles and responsibilities. The project resulted in achecklist of recommendations whose aim is to promotegreater knowledge among those who participate in thedirector evaluation process, to foster both effectivenessand fairness in director evaluations, and to strengthenlibrary leadership as part of the overall process ofimproving libraries. The checklist is also available at<http://www.arl.org/oms/checklist.html>.

The Director's ResponsibilitiesToday's academic library directors often hold positionsof broad responsibility and high visibility within theirinstitutions. They may be responsible for informationsystems beyond the walls of the library; education andoutreach programs; archives, museums, and galleries;and computing and telecommuthcations operations ofvast proportions. Virtually everyone on campus uses orbenefits from the library and its allied operations.

Clearly, the director's leadership is a key factor inthe success or failure of the library. Directors have lead-ership responsibility for envisioning the future of theirorganizations and setting goals to achieve that future,choosing and leading the best possible staff, managingresources wisely, fund raising, seeing that exciting newprograms get implemented, and assuring that the tech-nology is there to support those programs. Becauselibraries are a critical aspect of the educational process,designing and maintaining systems for evaluating theperformance of library directors is now more importantthan ever before.

Ideally, participants in the director review processshould be well informed about what library directors doand what should be expected of the director. Moreimportant, participants need to be aware of the direc-tor's goals and expectations as well as her or his accom-plishments during the period of review. Participantsshould also work to make sure that the evaluationprocess is one that will promote the growth anddevelopment of the director and the improvement oflibrary services and collections.

Key Leadership Roles' In order to assess the library director's performance, the

full range of her or his leadership roles should be con-sidered. Though the director may delegate portions

(0

of these responsibilities to others, he or she hasultimate leadership responsibility for their effectiveimplementation.

1. Chief Representative and Spokesperson. Directorsact as the chief external representatives of theirlibraries; they present and explain the library toothers; they distribute information to people(especially influential stakeholders) outside thelibrary; they inform outsiders of progress withinthe library; and they promote the library toexternal constituents.

2. Campus Administrator. Directors actively partici-pate in the governance of the university or collegethrough membership on committees, standingadministrative groups, and task forces; they helpdevelop policy on information issues, but alsoserve the larger community in tasks that may havelittle or no direct bearing on the library or on infor-mation policy; and they may administeror part-ner with others who administeroperations out-side the library (e.g., information technology).

3. Liaison. Directors maintain contacts outside thelibrary with key stakeholders in the parent institu-tion (e.g., faculty and other constituent groups,advisory groups, other administrators), as well aswith stakeholders outside the parent institution(e.g., community advisory and advocacy groups);they build external information networks; theyserve as a significant contact point for those whowish to influence the library's goals; and theyattempt through interactions with outside organi-zations to influence the environment in ways thatare beneficial to the library (e.g., legislation).

4. Monitor. Successful directors remain informedabout critical developments in the external envi-ronment, including changes in how library usersuse the library and what users need in terms ofinformation services; they are aware of currentdevelopments in other libraries and in the libraryprofession; they use that knowledge to solveproblems and to develop new services; and theyeducate the parent institution and the internalorganization about information and communica-tion technology issues.

5. Negotiator and Advocate. Directors negotiate withorganizations and individuals outside the libraryto secure funding, reach agreement on key issues,and safeguard the interests of the library.

(Continued on page 12)

ARL 198 JUNE 1998

CHECKLIST FOR ASSESSING THE PERFORMANCE EVALUATION PROCESS

The checklist is divided into three general areas:review process guidelines, review processcriteria, and review process participants.

There are at least two possible uses of the checklist:1) as a framework for discussion between directorsand those who are responsible for evaluating them;and 2) as a tool for educating those who participate inthe director's review. Although this list is not meantto be conclusive or restrictive, it presents importantquestions that should be addressed during theperformance evaluation process.

I. Review Process Guidelines

A. Is there a formal process for performance reviewof the library director, or are there other effectivemeans for monitoring her or his performance andproviding feedback?

B. Are there documented procedural guidelines forthe review process? If not, are the ad hoc guide-lines mutually satisfactory to both the directorand her or his supervisor?

C. Does the review have a clear purpose? Is it:a decision tool (used to decide whether thedirector will be reappointed or given a salaryadjustment);a developmental tool (used to assist thedirector in performance improvement);a communication tool (used to shareinformation about campus and library goals,problems, etc.);or any combination of the above?

D. Is the frequency of the review satisfactory to keyparticipants in the process?

E. Does the process result in specific, candid feed-back that is behavioral and recognizable and thatwill help the director build on performancestrengths and work on performance weaknesses?

F. Are discussions of institutional and library priori-ties, goals, and objectives between director andparticipants part of the review process?

G. Is there an opportunity for the director to providedocumentation or context, which may includeplanning goals, accomplishments, constraints,and other contexhial information that will helpparticipants in the review to make informedjudgments?

II. Review Process Criteria

A. Are there documented criteria for the review thatspecifically refer to the director's position andresponsibilities? Alternatively, are there genericcriteria that are customized as part of the reviewprocess?

B. A principal measure of the director's performanceis her or his success in achieving negotiated expec-tations as documented in library and institutionalplanning documents. Are the expectations dearlyoutlined and understood by participants? Doesthe review initiator provide an overall context forthe review participants, especially regardingfactors outside the director's control (e.g., acampus-mandated budget cut)?

C. Are the changing and evolving roles of the director(e.g., fund raising) sufficiently recognized in thereview process?

D. Is there a distinction between the performanceof the library and the performance of its director?Is it clear to those involved in the process whichaspects of the library's performance might beattributable to the director's leadership andwhich might not?

III. Review Process Participants

A. Is there an opportunity for a variety of participantsto have input into the performance review process,which may include library staff, library users,institutional stakeholders with whom the directorworks, and external persons with knowledge ofthe director's work?

B. Is the input of those who know the director'sperformance first-hand given more weight thanothers? If anonymous input is offered, is itevaluated as such?

C. Does the review committee include anotheradministrator comparable to the director whocan provide assistance to the group during theprocess?

D. Is the director's supervisor actively involved in theprocess even though she or he might not be thereview initiator? Does the supervisor give directfeedback to the director and provide participantswith contextual information that will help themevaluate the director's performance?

53A RL 198 JUNE 1998 11

-OFFICEOFLEADERSHIP-&-M--ANAGEMENTSERVICES1-;Continued

6. Fund Raiser. Directors lead the effort to identifyneeds that cannot be adequately supported by theparent institution, set priorities, and garner exter-nal funding through grants, gifts, endowments,and other development activities.

7. Leader of Planning and Operations. Directors leadmembers of the library organization in developingvalue systems, visions, and goals for the library;they promote high-quality services by involvinglibrary staff appropriately in planning and deci-sion making; they assure that there are perfor-mance measures and accountability systems andhold staff to them; they seek to understand inter-nal library issues, problems, and operations insufficient depth to make informed decisions; andthey plan, coordinate, and oversee major multi-year capital projects, as well as smaller facilitiesprojects.

8. Leader of Staff. Directors create and support acontinuous learning environment within thelibrary and encourage staff to participate activelyin virtually all the leadership roles noted in thislist; they monitor the human relations side of theoperation, insuring high-quality hiring, placement,retention, training, motivation, performance evalu-ation, and reward systems; they practice "facilita-tional leadership," especially in leading a diversework force; they handle conflicts and crises withinthe library; they take corrective actions whenunexpected disturbances occur; and some direc-tors may serve as deans with responsibility forlibrarians who have faculty appointments.

9. Communicator. Directors share and distributeinformation within the library through staff meet-ings, personal contacts, and other means; theyinvite input from individuals and groups withinthe library, listen attentively to that input, and acton it for the good of the library and its users.

10. Change Agent and Entrepreneur. Directors introducechange within the library by identifying problems,recognizing and seizing opportunities,and imple-menting new systems and programs; they promoteexperimentation and risk taking within the library(e.g., through the application of new technologiesand innovative uses of networks); and theyencourage staff to develop entrepreneurial skills.

11. Resource Allocator. Directors develop priorities forresource allocation and design the organizationalstructure to achieve those priorities; they allocate

funds, time, staff, materials, and equipment toassure that the library is a successful one; and theyauthorize major resource-related decisions madewithin the library.

This list of leadership roles is based on the work ofHenry Mintzberg1 and Michael Ann Moskowitz2, withinput from several ARL directors.

2

Henry Mintzberg, The Structuring of Organizations,Englewood Cliffs, NJ: Prentice-Hall, 1979.Michael Ann Moskowitz, "The Managerial Roles ofAcademic Library Directors: The Mintzberg Model,"College and Research Libraries 47 (Sept. 1986): 452-59.

RECENTLY PUBLISHED BY ARL

OLMS SPEC Kit #229, Evaluating Academic LibraryDirectors, May 1998.

OLMS Occasional Paper #21, Evaluating LibraryDirectors: A Study of Current Practice and a Checklist of

Recommendations, May 1998.

5 4

A R L 1 9 8 J U N E 1 9 9 8

Julia Blixrud, Senior Program Officer

THE STATE OF SALARIES INRESEARCH LIBRARIES

The median salary figure for professional staff inARL university libraries increased in FY1997-98to $44,534 from $43,170 in FY1996-97, representing

a slight increase in the purchasing power of ARLlibrarians when inflation is taken into account. Fornonuniversity libraries, the median salary figureincreased even more: from $51,150 to $55,055. Themedian beginning professional salary in universitylibraries is $28,500; in nonuniversity libraries it is$28,724. Both numbers are slightly higher than thosereported last year. The figures come from the recentlypublished ARL Annual Salary Survey 1997-98, a reporton ARL professional staff salaries.

Librarians at private U.S. university libraries stillearn more than their counterparts at publicly fundedU.S. university libraries. The differential in FY1997-98 is$2,220, which is 4.6% more for the average position in aprivate library. The overall average salary at privateU.S. institutions is $49,778; at public U.S. institutions itis $47,558; and at Canadian institutions it is $45,236 (U.S.dollars). The overall average salary in all ARLuniversity libraries, excluding medical and law libraries,is $48,090. ARL university libraries in the Pacific region($55,325) and New England ($52,014) again have thehighest average salaries, while those in the East SouthCentral ($42,382) and West South Central regions($42,782) continue to have the lowest.

The 848 minority librarians in the 97 U.S. memberuniversity libraries account for 11.03% of ARL's U.S.library professionals. Fewer minority professionals areworking in the New England, East North Central, WestNorth Central, East South Central, West South Central,

and Mountain regions, while more are to be found inthe Middle Atlantic, South Atlantic, and Pacific regions.The salary differential between minority averagesalaries from overall average salaries is $2,054, which is4.5% lower for minority staff, a slight improvementfrom last year.

For the second consecutive year, the average salaryfor female directors in university libraries is higher thanthat of male directors, 3.6% higher. However, atmedical libraries female directors earn 6.1% less thanmale directors, while at law libraries they earn 4.5% less.Overall, women still earn less in ARL universitylibraries although they represent 65% of all professionalstaff: the average salary for women is $46,888, which is93% that of men's, $50,171.

The newly released ARL Annual Salary Survey1997-98 is a compilation of detailed tables of salaries foralmost 12,200 ARL professional librarians. These tablesare arranged by job category, years of experience, sex,minority status, size of library, and geographic region.The Survey, an important research and managementresource, is based on data collected from the 110university and 11 nonuniversity ARL libraries, andcontinues to be the most comprehensive and thoroughguide to current salaries in major U.S. and Canadianacademic and research libraries.

The ARL Annual Salary Survey 1997-98 is availablefor $35 to member libraries and $70 for nonmembers(plus $6 shipping and handling per publication), and isavailable on standing order. Please contact ARLPublications, Department #0692, Washington, DC20073-0692; tel.: (202) 296-2296; e-mail: <pubs@arLorg>.Ken Rodriguez, Co-compiler and Co-editor, ARL AnnualSalary Survey 1997-98.

ARL ACADEMIC LIBRARIANS, FY 1997-98 *

Women Men CombinedAverage Salary $46,888 $50,171 $48,090

Average Years of Experience 16.6 17.0 16.8

Total # of Filled Positions 4,438 2,561 6,999

Minority Librarian's Avg. Salary(U.S. only) $44,896 $49,002 $46,036

Total # of Minority Librarians 510 196 706

Total # of Directors (filled positions) 49 59 108

*Excludes medical and law libraries.Source: ARL Annual Salary Survey 1997-98

55 A R L 1 9 8 J U N E 1 9 9 8

A-ertvirtEs-Lee Anne George, ARL Program Planning Officer

WINNERS ANNOUNCED IN 1997-98LC/AMERITECH NATIONAL DIGITALLIBRARY COMPETITION

The following seven projects were funded in thesecond round of the Library of Congress/Ameritech program to fund the creation of

cultural digital library collections as part of its"National Digital Library Program."

With a gift from Ameritech, the Library of Congressis sponsoring a three-year competition to enable public,research, and academic libraries, museums, historicalsocieties, and archival institutions (except federalinstitutions) to create digital collections of primaryresources. Eleven institutions received seven awards inthe second year of the competition. The postmark dead-line for the third competition year is November 2, 1998.Specific Guidelines for the last round will be available inJuly 1998. For further information, see the competitionhomepage at <http:/ /memory.loc.gov/ammem/award / >.

Chicago Historical SocietyHaymarket Affair: Chicago Anarchists on TrialApproximately 5,500 pages/images including the com-plete original transcripts of the proceedings of the his-toric Haymarket trial; the evidence books; the originaldraft autobiographies written by two of the Haymarketdefendants; and numerous pamphlets, accounts, andpictorial portrayals of the events in the popularpress of the time. Contact person: Bernard Reilly,(312) 642-5035, <reilly@chicagohistory.org>.

Duke UniversityThe Emergence of Advertising in America, 1850-1920Eight thousand five hundred images relating to thehistory of advertising including Eastman Kodak ads,tobacco related posters and insert cards, and ephemerarepresenting ads for bicycles, patent medicines, andfood. Contact person: Ellen Gartrell, (919) 660-5836,<gartrell@duke.edu>.

Nebraska State Historical SocietyPrairie Settlement: A Story of DeterminationTwo thousand five hundred glass plate negatives ofimages recording the process of settling Nebraskabetween 1886 and 1912, and approximately 3,000 pagesfrom diaries and letters written by the Oblinger familyas they moved from Indiana to settle in a sod house onthe Nebraska prairie. Contact person: Jill MarieKoelling, (402) 471-4409, <koelling@nebraskahistory.org>.

Northwestern UniversityNorth American Indian Photographs by Edward S. CurtisOne of the best recognized and most significant recordsever produced of the culhire and daily life of about 80

4 A R L 1 9 8 J U N E 1 9 9 8

Native American tribes, captured in 2,222 plates. Contactperson: Richard Frieder, (847) 491-4672, <frieder@nwu.edu>.

University of IowaTraveling Culture:Circuit Chautauqua in the Twentieth CenturyThirty-eight thousand talent flyers and promotional pam-phlets representing text and images from performers andpublic speakers, including educational, cultural, and reli-gious lecturers; politicians; and vaudeville and varietyacts. Contact person: Carol Hughes, (319) 335-5900,<carol-hughes@uiowa.edu>.

University of MiamiReclaiming the Everglades:South Florida's Natural History, 1884-1934A consortial collection that includes materials from theUniversity of Miami, Florida International University,and the Historical Museum of South Florida. The collec-tion documents the history of South Florida, especiallythe Everglades, a unique subtropical ecosystem that has arich, but troubled history. Contact person: William E.Brown, Jr., (305) 284-3247, <wbrown@umiami.ir.miami.edu>.

University of WashingtonAmerican Indians of the Pacific NorthwestA consortial collection that includes materials from theUniversity of Washington, the Eastern Washington StateHistorical Society in Spokane, and the Museum ofHistory and Industry in Seattle. The collection consistsof 2,350 pictorial images and 6,000 pages of selectionsfrom manuscripts, printed ephemera, and journal articlesconcerning Native Americans of the Pacific Northwestfrom 1763 to 1920. Contact person: Gary Menges,(206) 543-1929, <menges@u.washington.edu>.

HONORSArnold Hirshon, Vice-Provost for Information Resourcesat Lehigh University, was named the 1998 recipient of theHugh C. Atkinson Memorial Award.

Allen B. Veaner, former Library Director at UC-SantaBarbara, was named ACRL 1998 Academic/ResearchLibrarian of the Year. He is currently adjunct AssistantProfessor in the School of Library Science at theUniversity of Arizona.

UT-Austin: The Peny-Castaqeda Library Map Collectionon UT Library Online <http://www.lib.utexas.edu/Libs/PCL/Map_collection/Map_collection.html> wasselected by Library Journal as one of the ten Best ReferenceWeb Sites for 1997. The PCL Map Collection is the onlyeducational ".edu" site chosen for the yearall of theother sites were ".org," ".gov" or ".com" sites.

56

TRANSITIONSUC-Berkeley: Penny Abell was appointed the InterimUniversity Librarian, effective July 1.

UC-Irvine: Gerald J. Munoff was appointed UniversityLibrarian at the University of California-Irvine, effectiveJuly 1. He was previously Deputy Director of theUniversity Library, University of Chicago.

UC-Santa Barbara: Joe Boissé, University Librarian,will be unable to return from disability leave and hasfinalized his plans for retirement.

CISTI: Bernard Dumouchel was appointed DirectorGeneral of the Canada Institute for Scientific andTechnical Information, effective May 11.

Illinois at Urbana-Champaign: University LibrarianRobert Wedgeworth has announced his plan to retire asof August 1999.

Iowa State: Olivia M. A. Madison has accepted theappointment as Dean of Library Services. Her appoint-ment begins July 1.

NYU: Carlton Roche 11, Dean of Libraries, hasannounced his intention to retire as soon as a successorcan be appointed. Following the appointment of a newdean, he will serve as Executive Publisher of the NYUPress for a short time before his formal retirement fromthe University.

Syracuse: Peter S. Graham will become UniversityLibrarian, effective September 1. He is currently theAssociate University Librarian for Technical andNetworked Information Services at Rutgers University.

Western Ontario: Joyce Garnett was appointed Directorof Libraries, effective July 1. She was previouslyExecutive Director of University Advancement atLaurentian.

ARL StaffLee Anne George, formerly Librarian for Informationand Document Delivery Services at Harvard CollegeLibrary, joined the ARL staff in May as ARL ProgramPlanning Officer for a two-year appointment. Lee Annewill be assuming some of Jaia Barrett's responsibilitiesduring her time in Kiev, including assisting in planningARL membership meetings, developing the periodicARL program activity reports, and preparing the annualARL program plan.

Richard K. Johnson was named SPARC EnterpriseDirector, effective June 15. Rick will lead the businessand strategic development of SPARC, the ScholarlyPublishing & Academic Resources Coalition. Rickcomes to ARL from Congressional Information Service,

Inc. (CIS), a subsidiary of LEXIS-NEXIS, a ReedElsevier company, where he had been Vice-Presidentof Planning and Development.

Other TransitionsEDUCAUSE: Brian L. Hawkins was selected to be thefirst CEO for EDUCAUSE, the new higher educationinformation technology association formed by theconsolidation of Educom and CAUSE, as of June 1.Hawkins was previously Senior Vice-President forAcademic Planning and Administrative Affairs atBrown University.

Dr. Carole A. Barone was named a Vice-Presidentat EDUCAUSE, where she will focus on the NationalLearning Infrastructure Initiative. She has served asAssociate Vice-Chancellor for Information Technologyat the University of California-Davis.

NEA: On May 21, the full Senate confirmed theappointment of William J. Ivey as Chairman of theNational Endowment for the Arts. He was Director ofthe Country Music Foundation in Nashville, andreplaces NEA former chair Jane Alexander.

OCLC: Robert L. "Jay" Jordon was named Presidentand Chief Operating Officer of OCLC, Inc. becomingOCLC's fourth president in its 31-year history. Hebegan in May on the retirement of K. Wayne Smith,who stepped down after nearly 10 years. He was pre-viously President, New Business Development,Information Handling Services Group.

US National Commission on Libraries andInformation Science: Robert S. Willard was namedExecutive Director by Jeanne Hurley Simon, Chair ofNCLIS. Willard's career includes experience as a mar-keting executive for legal publishing companies and asstaff for the Information Industry Association. JaneWilliams, Acting Executive Director of NCLIS fromMay 1997 to February 1998 resigned to take a positionat the University of Maryland Library in College Park.

The Senate confirmed the reappointment of JeanneHurley Simon as NCLIS Chair.

ARTHUR CURLEY, 1938 -1998Arthur Curley, Director Emertius of the BostonPublic Library, died March 31, 1998. He hadserved as Director of the Boston Public Libraryfrom 1985 to 1996 and as a member of the ARLBoard of Directors from 1989 until 1992. He wasARL President in 1991.

5 7 A R L 1 9 8 J U N E 1 9 9 8

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-296-2296 FAX 202-872-0884<http://www.arl.org/>Copyright: © 1998 by the Association of Research Libraries- - -ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

- - .

Executive Director Duane E. WebsterEditor: G. Jaia Barrett, Deputy Executive DirectorCo-editor. Lee Anne George, Program Planning OfficerCopy Manager Karen A. WetzelDesigner Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscription; Nonmembers$50 per year.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arLorg>.

ARL CALENDAR 1998 99

1998July 27-28 ARL Board Meeting

Washington, DC

October 7-9 ARL/OLMS Facilitation SkillsInstituteWashington, DC

October 13-46 ARL Board andMembership MeetingWashington, DC

October 13-16 Educom'98Orlando, FL

October 27-28 ARL/OLMS Leading ChangeInstituteChicago, IL

November 2-4 OLMS/DORAL ManagementInstitute for DevelopmentOfficersBaltimore, MD

November 5-6 Workshop on LicensingElectronic InformationResourcesKansas City, MO

November 16-19 ARL/OLMS LibraryManagement Skills Institute I:The ManagerWashington, DC

December 7-8

December 8-11

1999February 11-12

May 11-14

July 26-27

October 12-15

58

CNI Task Force MeetingSeattle, WA

CAUSE98Seattle, WA

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingKansas City, MO

ARL Board MeetingWashington, DC

ARL Board andMembership MeetingWashington, DC

0

`Mk

A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONS

SPARC AND CHEMISTS TO COLLABORATEON NEW REDUCED-COST JOURNALS

The American Chemical Society (ACS), theworld's largest scientific society, is the firstScholarly Publishing & Academic Resources

Coalition (SPARC) publishing partner in an innov-ative collaboration aimed at distributing researchresults faster and at significantly less cost to librarysubscribers. ACSa non-profit scientific societythat currently publishes 26 peer-reviewed researchjournalswill collaborate with SPARC to publishat least one new scientific journal each year for thenext three years. SPARC is a newly formed alliancelaunched with support from ARL member libraries.

On June 30, ARL signed a memorandum ofunderstanding with ACS to set out the terms bywhich SPARC and ACS will proceed. The agree-ment frames an ongoing working relationship.At a press conference announcing the partnership,Duane E. Webster, ARL Executive Director,declared it "a major step forward in resolvingthe debate between libraries and publishers overacademic journal costs and access, especially asthe benefits and challenges of electronic publishingbecome more evident." He noted that theCoalition's aim is to substantially reduce the costof scholarly publication while shortening the timerequired to deliver information, especially in theform of journals.

"Today we replace debate with dialogue andaction, moving forward together to make cutting-edge scientific research available to the academiccommunity in the most efficient and cost-effectiveway possible," said Robert D. Bovenschulte, ACSDirector of Publications. "We're especially pleasedto be the first scientific publisher to enter into a

5 9

partnership with SPARC, a creative, forceful, and wel-come initiative by ARL. We're also confident that thiscollaboration will result in publishers being able to offerhigher quality products than are currently available."

The first new journal will publish letters in organicchemistry both online and in print. Under the agree-ment, SPARC members advise ACS by identifying thesubject area of interest, and, in return, the participatinglibraries will help provide instant market acceptance byendorsing the project and ensuring purchase of thenew publication.

Richard K. Johnson serves as Enterprise Directorfor SPARC, which is seeking additional publishingpartners. "One of the key advantages SPARC offers toa publishing partner," said Johnson, "is the commit-ment of Coalition members to subscribe to SPARCproducts. This reduces the time to market acceptanceand cost-recovery of a new title."

Currently comprised of 98 member libraries with apurchasing power of nearly $500 million, SPARC wascreated as a result of the growing concerns amonglibrarians and researchers over the rising cost of acade-mic publications, particularly scientific journals. Librarymaterials budgets, which have increased over the pastdecade at almost seven percent a year, have not beenable to keep pace with the 12-percent annual increasein the average price of science journals. Both the highprices and steep annual increases charged by somemajor scientific publishers have forced libraries to cancelthousands of journal subscriptions, prompting publish-ers to raise prices even higher to make up for the loss.

The advent of electronic publishing has also been aconcern for both publishers and librarians, despite itsdistinct advantages over print in terms of delivery time,

2

4Continued

Pictured at the June 30, 1998 signing of the publishing agreementare (clockwise from left): Duane Webster, ARL Executive Director;

Richard Johnson, SPARC Enterprise Director; Kenneth Frazier,Director of the University of Wisconsin Libraries and Chair of the

SPARC Working Group; and Robert Bovenschulte,Director of the ACS Publications Division.

additional features, and flexibility. Higher subscriptioncosts, licensing agreements for access, and the uncertaintyof archiving all have been sources of increasing tensionbetween academic publishers and their customers.

The new organic letters title will include the enhancedfeatures that ACS developed for all of its 26 journals thatit began offering on the World Wide Web last year. Themost important of these is "Articles ASAP" (As Soon AsPublishable), which releases journal articles on the Webas soon as they are finished, accelerating publications bytwo to 11 weeks over print.

ACS announced additional services this year thatwill help improve the cost efficiency to the subscriber ornon-subscriber, such as free access to tables of contentssince January 1996; supporting information on experi-mental details, also at no charge; and the option topurchase single articles without subscribing. ACSalso altered its licensing arrangements with academiclibraries, eliminating restrictions on use of the webeditions of ACS journals in interlibrary loan programs.

"We are very pleased to have the prestige and powerof ACS in this first partnership," said Kenneth Frazier,Director of the University of Wisconsin Libraries and Chairof the SPARC Working Group. He continued, "ACS publi-cations are generally higher quality and more heavily citedby researchers, and they are considered among the morereasonably priced scientific journals. We hope that otherscientific publishersboth non-profit and for-profitwillrecognize the significance of this collaboration and join us inthis effort to fundamentally change academic publishing."

SPARC FOUNDING MEMBERSas of August 3,1998

University of AlbertaUniversity of ArizonaArizona State UniversityAuburn UniversityBrigham Young UniversityUniversity of British ColumbiaBrown UniversityUniversity of California-

BerkeleyUniversity of California-

DavisUniversity of California-

IrvineUniversity of California-

Los AngelesUniversity of California-

RiversideUniversity of California-

San DiegoCalifornia Digital LibraryCanada Institute for Scientific

and Technical InformationUniversity of ChicagoUniversity of CincinnatiUniversity of ColoradoColorado State UniversityColumbia UniversityUniversity of ConnecticutCornell UniversityDartmouth CollegeUniversity of DelawareDuke UniversityEmory UniversityUniversity of FloridaGeorgetown UniversityUniversity of GeorgiaGeorgia Institute

of TechnologyUniversity of GuelphHarvard College LibraryUniversity of HoustonUniversity of Illinois-ChicagoUniversity of Illinois-UrbanaIndiana UniversityIowa State UniversityJohns Hopkins UniversityUniversity of KansasKent State UniversityLaval UniversityLinda Hall LibraryLouisiana State UniversityMcGill UniversityMcMaster UniversityUniversity of ManitobaUniversity of MarylandUniversity of MassachusettsMassachusetts Institute of

TechnologyUniversity of MiamiUniversity of Michigan

6 0

Michigan State UniversityUniversity of MinnesotaUniversity of MissouriUniversity of NebraskaUniversity of New MexicoNew York UniversityUniversity of North CarolinaNorth Carolina State UniversityNorthwestern UniversityUniversity of Notre DameOhio UniversityOhio State UniversityUniversity of OklahomaOklahoma State UniversityUniversity of OregonUniversity of PennsylvaniaPennsylvania State UniversityUniversity of PittsburghPrinceton UniversityPurdue UniversityRice UniversityRutgers UniversityUniversity of SaskatchewanUniversity of Southern

CaliforniaSouthern Illinois UniversityState University of New York-

AlbanyState University of New York-

BuffaloState University of New York-

Stony BrookSyracuse UniversityTemple UniversityUniversity of TennesseeUniversity of TexasTexas A&M UniversityTexas Tech UniversityUniversity of TorontoTulane UniversityUniversity of UtahVanderbilt UniversityVirginia TechUniversity of WashingtonWashington State UniversityWashington UniversityUniversity of WaterlooWayne State UniversityUniversity of Western OntarioUniversity of WisconsinYork University

AffiliatesAssociation of College and

Research LibrariesCanadian Association of

Research Libraries(Association desbibliotheques derecherche du Canada)

A R L 1 9 9 A U G U S T 1 9 9 8

SPARCTHE SCHOLARLY PUBLISHING & ACADEMIC RESOURCES COALITION

WHAT IS SPARC?SPARC, the Scholarly Publishing & Academic ResourcesCoalition, is a newly formed affiance of libraries thataims to foster expanded competition in scholarlycommunication.

Launched with support from membership of theAssociation of Research Libraries (ARL), SPARC hasbegun creating "partnerships" with publishers whoare developing high-quality but economicalalternatives to existing high-price publications.

By partnering with publishers, SPARC seeks to:create a more competitive marketplace;reduce the prices of journals;ensure fair use and educational and library usesof electronic resources; and

apply new technologies to improve the processof scholarly communication and to reduce thecosts of production and distribution.

HOW WILL SPARC ACCOMPLISH THIS?SPARC will influence the marketplace positively byencouraging publishers to enter markets where theprices are highest and competition is needed mostprimarily in the science, technical, and medical areas.

Through its activities, SPARC will reduce the riskto publisher-partners of entering the marketplace,while providing faculty with prestigious andresponsive alternatives to current publishing vehicles.

To accomplish this, SPARC will:

solicit and encourage the introduction of newpublications of high quality and fair price;guarantee a subscription base and market newproducts to potential subscribers;provide start-up capital (in selected cases); andgenerate support for SPARC projects fromdistinguished faculty, educationalorganizations, professional societies,and scholarly publishers.

LIBRARY SUPPORT IS NEEDEDSPARC will finance its efforts through coalitionmember fees that support operating expenses and help

build a capital fund to provide start-up money forSPARC projects. SPARC will also seek grant moneyto augment the capital fund.

The key to SPARC's success, however, will bethe commitment of coalition members to supportSPARC-endorsed products.

Although the impetus for SPARC grew out ofARL member initiatives, SPARC membershiprecently was opened to other libraries.

WANTED: PUBLISHER PARTNERSSPARC is taking action to create a competitivemarket for scholarly publishing. It will concentrateits resources on supporting ventures in which thegain from introducing price and service competitionis great.

SPARC seeks partners that are committed to fairpricing, intellectual property management policiesthat emphasize broad and easy distribution andreuse of material, and the ethical use of scholarlyresources.

Priority will be given to innovative applicationsof information technology that enrich and expandresearch and scholarship; assure that new channelsof scholarly communication sustain quality require-ments and contribute to promotion and tenureprocesses; and enable the permanent archiving ofresearch publications and scholarly communication.

Potential partners in SPARC include:

Professional societies and universitypresses interested in launching newpublishing initiatives.Start-up electronic publishers that havealready created publications in subject fieldsdominated by commercial publishers.For-profit enterprises that offer newstrategies for controlling prices andimproving access to research information."Visionary" enterprises, including bothdiscipline and institution-based servermodels, seeking to create entirely neweconomic models for scholarlycommunication.

(Continued on next page)

REST COPY AVAILABLi-ARL 199 AUGUST 1998

CtiRRENT---ISSUES---Continued

SPARCTHE SCHOLARLY PUBLISHING & ACADEMIC RESOURCES COALITION

(Continued from previous page)

WHO BENEFITS?SPARC aims to stimulate creation of better, faster,and more economically sustainable systems fordistributing new knowledge. These advanceswill benefit:

Researchers. SPARC will encouragedevelopment of high-quality, lower-costdelivery channels, which will drive expandedaccess.

Society. Research conducted by faculty atuniversitiesoften with the aid of governmentfundspropels key advancements benefitingthe economy. SPARC aims to facilitateimproved and expanded communication of thisresearch.

Publisher-partners. Working with a variety ofpublishers whose business goals are compatiblewith SPARC, SPARC will support developmentof a vibrant and economically viable new modelof scholarly communication.

Libraries. By fostering development of newalternatives to established high-price journals,SPARC encourages the kind of competition thatwill bring skyrocketing journal prices back toearth. Libraries will get more for their money.

CURRENT STATUSEndorsementsThe concept of SPARC has been endorsed by theseleading academic organizations:

Association of American Universities (AAU)

National Association of State Universities andLand Grant Colleges (NASULGC)

Association of American University Presses(AAUP)

Big 12 Provosts

OversightThe development of SPARC is being guided by aformal working group and steering committee, chaired

1111111

by Ken Frazier, University of Wisconsin at Madison.Formation of an advisory committee of leaders inhigher education, scientific disciplines, scholarlypublishing, and business is targeted for fall 1998.

MembershipSPARC now has 98 Founding Members from amongthe 121 ARL member libraries. Founding Membershave each made a financial commitment to the start-upof the SPARC enterprise. On July 28, the ARL Boardendorsed a plan to open SPARC membership beyondARL to other libraries and organizations. (For mem-bership information, see the SPARC website.)

AffiliatesThese organizations have supported SPARC bybecoming affiliates:

Association of College & Research LibrariesCanadian Association of Research Libraries(Association des bibliotheques de recherche duCanada)

Business DevelopmentSPARC development efforts already are off to apromising start:

A full-time SPARC director was hired.A first partnership agreement has beennegotiated with the American Chemical Societythat calls for the creation of several new journalsthat challenge established high-price titles.Negotiation of additional partnerships isunderway.

MORE INFORMATIONFor more information about SPARC, contact:

Richard K. JohnsonSPARC Enterprise Director21 Dupont Circle, NW, Suite 800Washington, DC 20036

Tel: 202-296-2296Fax: 202-872-0884E-mail: rick@arl.orgURL: www.arl.org/sparc/

A R L 1 9 9 A U G U S T 1 9 9 8 62

STATISTICS-Sr ME-AMAZEMENT-Julia Blixrud, Senior Program Officer

STILL PAYING MORE;STILL GETTING LESS

The most recent edition of ARL Statistics docu-ments the continuing struggle by researchlibraries to keep up with serial and monograph

costs. Specifically, ARL data for 1996/97 show that,while ARL libraries more than doubled expendituresfor serials from 1986 to 1997, they purchased 6% fewerserial subscriptions. During the last decade, librariesshifted expenditures from monographs to serials tomeet some of the demands of increasing serial prices,reducing the number of monographs purchased by14%, while the unit cost for monographs increased by62%. Since 1986, the annual average increase for theserial unit cost has been 9.4% and for the monographunit cost 4.5%, both higher than the general inflationtrends in North America during the same period.

Annual statistics for university libraries have beencollected since 1908 and represent the oldest and mostcomprehensive continuing library statistical series inNorth America. Machine readable datafiles datingfrom 1908 are available on the ARL ftp server at<http://www.arl.org/stats/arlstat/mrstat.htm>.The 1996/97 publication describes the collections,staffing, expenditures, and service activities for the121 ARL member libraries. To order, contact<pubs@arLorg>.

MONOGRAPH AND SERIAL COSTSIN ARL LIBRARIES, 1986-1997

200%

150%

s.c)cocr,e-,a) 100%

Ia) tir?lr Crs rhbt,

1(+62%)

U /50%e ... ronoirph( +TO;

Serial Unit Cost(+169%)

Serial

(Ex+061)tures

0%

-50%

1986 1988 1990 1992 1994 1996

Fiscal Year

SerialsPurchased(-6%)

rmougzels

(-14%)

MEDIAN VALUES FOR TIME-SERIES TRENDS

Year

(# of Libraries)

SerialUnit Cost

(43)

SerialExpenditures

(103)

MonographUnit Cost

(63)

MonographExpenditures

(99)

Serials MonographsPurchased Purchased

(43) (63)

1986 $88.81 $1,517,724 $28.65 $1,120,645 16,198 33,2101987 $104.30 $1,770,567 $31.76 $1,064,484 16,518 27,2141988 $117.25 $1,979,604 $35.63 $1,141,226 16,443 26,5411989 $128.47 $2,130,162 $37.74 $1,241,133 16,015 27,2681990 $130.81 $2,304,744 $40.26 $1,330,747 16,182 27,9991991 $152.43 $2,578,309 $42.04 $1,400,738 16,149 28,0271992 $162.72 $2,630,827 $43.31 $1,353,865 15,846 27,1581993 $184.71 $2,919,756 $41.78 $1,295,807 15,463 25,5831994 $191.13 $2,932,091 $44.51 $1,309,807 15,583 25,8031995 $211.29 $3,133,885 $45.27 $1,365,575 14,942 25,7191996 $219.46 $3,393,307 $46.73 $1,444,015 15,069 26,2621997 $238.69 $3,674,368 $46.42 $1,460,234 15,297 28,658

Annual averagepercent change 9.4% 8.4% 4.5% 2.4% -0.5% -1.3%

, 6 3 A R L 1 9 9 A U G U S T 1 9 9 8

1STICS & MEASUREMENTContinued

4

. 444.uwaw4.444

UNIVERSITY FUNDING FORRESEARCH LIBRARIESby Julia Blixrud, Senior Program Officer

In January 1994 (ARL 172), using a decade of dataabout total university educational and generalexpenditures (E&G) and ARL academic library

expenditures, Kendon Stubbs, University of Virginia,provided a detailed analysis of the trends of the libraryexpenditures as a percent of university E&G expendi-tures. This analysis revealed a "widespread decreasein the proportion of university funding directed toresearch libraries, averaging over half a percent of E&Gexpenditures from1982 through 1992."

Data from themost recent E&Greport to the ARLacademic member-ship shows that thedownward trendcontinues, althoughthe rate of descenthas been much lessover the last fouryears than previously.

For the 102 ARLuniversity librariesproviding data onlibrary expendituresas a percent of uni-versity E&G expendi-tures in 1995/96, the mean figure is 3.23% and the medianis 2.90%. For U.S. private university libraries (26 institu-tions reporting), the mean is 2.96% and the medianis 2.77%, while for the U.S. state university libraries(65 institutions reporting) the mean is 2.95% and themedian is 2.82%. For the 11 Canadian institutionsreporting, the mean is 5.47% and the median is 4.86%.

Caution should be used in interpreting the E&Gfigures because the method for calculating universityexpenditures is not entirely consistent among institu-tions, nor is the university figure necessarily calculatedon the same basis as that for the library. Universityfigures are as reported to the U.S. National Center forEducational Statistics, Integrated PostsecondaryEducation Data Systems (IPEDS) Finance form, SectionI, Part B, line 12, column 3. There is a variation in howthe IPEDS definitions are applied on individual cam-puses. University figures typically include fringe bene-fits, buildings, and maintenance. Library expendituresare as defined and reported in the ARL Statistics andusually do not include maintenance, fringe benefits,and information resources purchased through other

accounts (for example, the computing center).Canadian institutions attempt to provide university

data comparable to the U.S. IPEDS categories.However, the financial accounting systems in the twocountries may be fundamentally incompatible. Forexample, Canadian universities may not count as directexpenditures certain allocated federal monies, becausethose grants accrue to individual researchers rather thanbeing transferred or reimbursed to the university via theinstitutional budget. Thus, Canadian university expen-ditures, as reported, may not include all funds spent insupport of the institution. This lower "total" may inturn create a misleading figure for the library's percent-

age of the institu-tion's funds.(Canadianexpenditures areexpressed in U.S.dollars at the rateof 1.3613 Canadiandollars to one U.S.dollar, the averagemonthly noonexchange rategiven in the Bankof Canada Reviewfor the period July1995June 1996.)

The 1996 report1992 1994 1996 will complete this

trend series due tochanges in the

accounting and financial reporting standards issued bythe Financial Accounting Standards Board (FASB). WithFY 1996-97, FASB has changed the IPEDS Finance Surveyforms, the source for E&G data. There are currently twoforms, one for private institutions (Form IPEDS-F1-A)that incorporates the accounting changes, and anotherform for public institutions (Form WEDS-F-1), which is,for now, the same as in previous years. ARL has alreadynoted from reports received from private institutions thatthe changes have generally resulted in an increased uni-versity expenditure figure and thus have a negativeimpact on the library percentage. Some public institu-tions will be testing their new form this year and nextyear both public and private institutions are expected tobe completing the same form. Since the 88 librarieswhose data is used to create the chart are a mix of publicand private institutions, the ARL trend series cannot becalculated in this interim period.

Electronic data for this time series and a copy of thesummary chart that illustrates the trend line are avail-able also on the ARL website at <http:/ /www.arl.org/stats/eg/>.

ARL LIBRARY PERCEXPENDITURES FOR 88 L

ENTAGE OF E&GIBRARIES, 1982-1996

4.0%

3.9%

3.8%

3.7%

"o' 3.6%

3.5%

3.2%

3.1%

3.0%1982 1984 1986 1988 1990

1 YearFisca

A R L 1 9 9 A U G U S T 1 9 9 8 6 4

?RESERVATIONMary Case, Director, Office of Scholarly Communication

LIBRARY OF CONGRESS MASSDEACIDIFICATION PROGRAMby Kenneth E. Harris, Preservation Projects Director,Preservation Directorate, Library of Congress

In 1996, for the first time since the Library of Congressbegan its search for a mass deacidification process20 years ago, LC moved beyond evaluation and testing

and began to deacidify books from the Library's collec-tions. In a contract with Preservation Technologies, theLibrary committed itself to deacidify 72,000 books over an18-month period. By late May 1997, the U.S. Congressapproved a new MassDeacidification Action Planthat enabled the Library to(1) incorporate deacidificationinto the Library's existingarsenal of preservationoptions, (2) extend the exist-ing contract while negotiatinga multi-year contract todeacidify significant quanti-ties of important endangeredbooks from the Library'scollections, and (3) continueto offer companies the oppor-tunity to have the Libraryevaluate and test other viable,operational mass deacidifica-tion processes. Following are

prolong the useful life of invaluable library collections andarchival holdings through mass deacidification, either bynegotiating separate agreements or by forming with otherinstitutions partnerships that could achieve economies ofscale through the treatment of large quantities of materials.

Selection Criteria and ProceduresDeacidification treatment is reserved for books that areacidic and at risk of loss if no action it taken. Due to itsrole as the national library and the official library of theU.S. Congress, the Library is focusing primarily on selec-tion of "Americana" for early treatment under the mass

deacidification program,emphasizing the selection ofendangered volumes fromcollections that are central tothe Library's mission.

Quality ControlsThe deacidification process,utilizing magnesium oxide(Mg0) to neutralize acid inthe paper, takes two hoursfrom the time books areplaced in the Bookkeeper IIItreatment cylinders until thevolumes are ready to bepacked for return to theirhome library. All steps in theprocess, from selection to

reshelving, are monitored to ensure that the intendedresults are achieved. The Bookkeeper process meets theLibrary's basic preservation requirements by:

raising the pH level of treated paper to the acceptablerange of 6.8 to 10.4 pH;achieving a minimum alkaline reserve of 1.5% or more;andextending the useful life of paper (measured by foldendurance after accelerated aging) by over 300%.

Manuscript Deacidification EquipmentPreservation Technologies has designed new equipmentthat it intends to use to offer deacidification services forloose manuscript and archival materials. The Libraryhopes to perform tests to evaluate the effectiveness of thisnew equipment in treating manuscripts and other paper-based materials in non-book formats and will be pleasedto share the results with other libraries and archives.

Additional InformationMore information and data about the Library's massdeacidification program is available at the followingInternet address: <http:/ /lcweb.loc.gov/preserv/>.At that site, select "Research and Testing Online" toaccess a number of reproduced publications aboutdeacidification, including an informative, illustratedarticle that appeared in the April 1997 Library of CongressInformation Bulletin.

LIBRARY OF CONGRESS"DEMONSTRATION SITE

ft

Given the effective operation of its mass deacidificationprogram for books over the past two years, the Library

of Congress is serving as a demonstration site formanagers and technical staffs from other libraries,

archives, and cultural institutions. Anyone interestedin learning firsthand about administrative and workflow procedures required for mass deacidification

programs should contact Kenneth E. Harris,Preservation Projects Director, PreservationDirectorate, Library of Congress, LM-G21,

Washington, DC 20540-4500.Phone: (202) 707-1054;Fax: (202) 707- 3434;

E-mail: <khar@loc.gov>.

highlights of the LC Mass Deacidification Program in 1998.

Library Awards Book Preservation ContractWith strong support from the U.S. Congress, the Libraryof Congress has provided leadership in the developmentand evaluation of deacidification processes on a massscale and their application to increasingly larger volumesbf books and other paper-based items to achieve economiesof scale. Through a competitive process, the Library hasnegotiated a second contract for mass deacidification.The contract was awarded to Preservation Technologies,Limited Partnership (PTLP) of Cranberry Township, asuburb of Pittsburgh, Pennsylvania. The company willprovide book preservation services to the Library usingthe firm's Bookkeeper mass deacidification process. Thiscontract runs through October 2001. It is an outgrowth ofa two-year, limited production contract that enabled theLibrary to deacidify about 100,000 books and to resolvemany book selection, shipping, and quality control detailsof the deacidification program.

Production Incentives: LC EncouragesOthers to Mass Deacidify Library &Archival MaterialsWithin four years, the Library will deacidify up to 275,000books. Under an incentive plan, the per-book treatmentcost will decrease as the number of books deacidifiedincreases. The Library encourages other institutions to

6 5 A R L 1 9 9 A U G U S T 1 9 9 8

Prudence S. Adler, Assistant Executive Director-Federal Relations and Information Policy

W.I.P.O. LEGISLATION UPDATEJULY 21, 1998

On Friday, July 17, the House Committee onCommerce approved the Digital MillenniumCopyright Act of 1998 (DMCA), H.R. 2281. This

is the legislation to implement the World IntellectualProperty Organization (W.I.P.O.) treaties by updatingU.S. copyright law for the digital environment. The billas crafted by the Committee on Commerce is signifi-cantly different than H.R. 2281, as approved by theHouse Committee on the Judiciary as well as the Senate-passed S. 2037. In addition to provisions included inS. 2037 and the House Committee on the Judiciary bill,this legislative package reflects a broader set of interests,such as those relating to fair use, privacy, and encryp-tion research.

Changes to the bill followed well over a weekof intensive negotiations by members of the library,education, copyright proprietor interests, and commit-tee staff. The library community was represented bythe Shared Legal Capability attorney, Arnold Lutzker ofFish & Richardson. The library and education commu-nities have agreed not to oppose the compromise bill asdrafted by the Committee on Commerce with theunderstanding that copyright proprietor interests willactively support DMCA, other bills such as H.R. 2652(database protection) will not be included in this pack-age, and this agreement is binding through the Senateconference. It is not clear that these conditions will hold.

In announcing the fair use amendment, CommerceCommittee Chairman Bliley (R-VA) stated, "The agree-ment we have today gives consumers a reliable and reg-ular process that ensures they will have 'fair use' accessto information and copyrighted workswithout stiflinggrowth- of electronic commerce. Digital technology hasthe potentialand let me emphasize the word 'poten-tial'to lock up information and works that are other-wise widely available to consumers today. The fact thatthis is only 'speculative' or 'hypothetical' does not con-vince me that we should do nothing. Quite the oppo-site, it raises the possibility that it will inhibit growth inelectronic commerce."

An amendment offered by Rep. Klug (R-WI) wasincorporated in the DMCA; it has several key elements,including:

The new access provision which would have prohib-ited circumvention of a technological protection mea-sure for any purpose, including lawful purposes suchas fair use, was dropped from the Commerce-draftedbill.In lieu of this statutory "access" provision, a regula-tory regime will be in place. The Secretary ofCommerce, in consultation with the Register ofCopyrights and others, is directed to adopt a

regulation no later than two years following enact-ment of the DMCA that would prohibit drcumven-tion once a formal rulemaking process is completed.This rulemaking would consider "whether users ofcopyrighted works have been, or are likely to be . . .

adversely affected by the implementation of techno-logical protection measures that effectively controlaccess to copyrighted works."

In conducting the rulemaking, the Secretary shallexamine several factors, including the availabilityfor use of copyrighted works; their availability forarchival, preservation, and educational purposes;the impact of technological protection measureson traditional fair uses, such as scholarship, teach-ing, and research; the effects of circumvention oftechnological measures on the market for or valueof copyrighted works; and such factors as theSecretary and others consider appropriate.If the Secretary finds that an adverse impact isdemonstrated or is "likely" on any particular classof copyrighted works, such as journal articles, thisclass would be exempt from the prohibition oncircumvention for the following two years to per-mit "lawful uses."The Secretary, in consultation with the Registerand others, shall conduct a rulemaking every twoyears and evaluate the waivers of certain classesof works, if applicable.

In addition to the Klug amendment, there are sever-al other key provisions included in the Committee onCommerce bill. Many of the provisions originated inthe Senate-passed bill and include:

a framework for online service provider liability;the ability to use digital technologies for preservationpurposes;the ability to circumvent in order to engage inencryption research;permission to circumvent to protect personal privacy;a study to consider distance education opportunitiesand the impact of such programs on copyrightedworks in the networked environment; anda study to examine the implications of online serviceprovider liability for educational institutions.Since the House Committee on the Judiciary and the

Committee on Commerce bills differ, there will need tobe consultation between the Committees to resolvethese differences prior to the bill reaching the Housefloor. In addition, there will need to be a similar confer-ence to sort through the differences between the Houseand Senate W.I.P.O. legislation. The House bill couldreach the floor prior to the Congressional recess in earlyAugust. Given the highly charged political atmospheresurrounding this legislation, it will be important toretain a legislative package that is balanced.

A R L 1 9 9 A U G U S T 1 9 9 8 6 6

011, ,M1 MI

De Etta Jones, Senior Program Officer for Diversity

ARL APPLAUDS THE 1997-98 CLASSOF THE LEADERSHIP AND CAREERDEVELOPMENT PROGRAM

The pilot experience of ARL's Leadership andCareer Development (LCD) Program has beena tremendous success. The Program exposes

participants, mentors, and ARL faculty to a breadthof knowledge with significant opportunities forin-depth exploration.The LCD Programconsists of severalcomponents: twoone-week Institutes,a research project,a mentoringrelationship, andfinally, a closingceremony. The closingceremony for the1997-98 LCD Programclass was held July 25in Washington, D.C.,just before the ALAConference. Directors,mentors, supporters,and ARL staff joinedtogether as LCDProgram participantsand faculty sharedtheir impressions ofthe Program andcelebrated its success.

The 21 LCDProgram participantshave and continue toattest to the long-term value of having participatedin this Program. Participants have also committed tosharing their insights and their enhanced leadershipskills as mentors to ALA's Spectrum InitiativeScholars, a program established to encourage andassist minorities pursuing library science degrees.This first group of LCD Program participants helpsthe library community to appreciate the future oflibrarianship, more fully explore synergisticlearning, and truly understand the value ofprograms for developing and giving professionalvisibility to research librarians from diversebackgrounds.

As we applaud the efforts of the 1997-98 class ofparticipants, we also applaud the efforts of theleaders of academic and research libraries,

information science schools, and associations thatcontributed significantly to the Program's developmentand success. In addition, we thank the Department ofEducation and Institute of Museum and LibraryServices for their invaluable financial support.

Next StepsARL is now working with the LCD Programparticipants to complete their research projects and

publish orotherwise givevisibility to theirproject results.Leading Ideas, a newARL publicationseries that focuseson diversity,leadership, andcareer development,was created throughthe DiversityProgram as avehicle forhighlighting theparticipants'research; it isavailable at<http:/ /www.arl.org/diversity/leading/index.html>. Inorder to assess theProgram's impact,ARL maintainscontact with

participants, regularly monitoring and tracking theircareer paths. Furthermore, the participants continue tonetwork through an electronic discussion list and aremaintaining relationships with their LCD Programmentors.

There is a clear need for and interest in continuingthe Leadership and Career Development Program.Strategies for deploying resources and securingfunding for future classes are currently beingaddressed by the ARL membership. This opportunityis expected to be available again for 1999-2000.

For more information about the ARL Leadershipand Career Development Program, visit the website at:<http:/ /www.arl.org/diversity/lcdplist.html>.

LCD PROGRAM CLASS OF 1997-98From left to right, Back row: Poping Lin, Jon Cawthorne, Gloria Rhodes,

Valerie Wheat, Tracey Joel Hunter, Carlos Rodriguez, Mark Winston, JuliaBlixrud (ARL); Center row: Thura Mack, Barbara Lehn, Vicki Coleman,

Maria de Jesus Ayala-Schueneman, Patrick Dawson, Neville Prendergast;Front row: Janice Simmons-Welburn, De Etta Jones (ARL), Elayne

Walstedter, Johnnie Love, Denise Stephens, Janet Lee-Smeltzer, NereaLlamas, Glendora Johnson-Cooper, Teresa Neely, Kathryn Deiss (ARL).

b 7 A R L 1 9 9 A U G U S T 1 9 9 8

Kathryn Deiss, OLMS Program Manager

NEW FROM THE OLMSINFORMATION SERVICES PROGRAMby Patricia Brennan, ARL Publications Officer

To order copies of any of these titles, contact theARL Publications department at <pubs@arLorg>.

SPEC Kits

Use of Teams in ARL LibrariesSPEC Kit 232, July 1998compiled by George J. Soete, ARLIOLMSA notable trend in academic library management in recentyears is the movement toward use of teams. Eighty-threelibraries responded to a February 1998 SPEC survey thatsought to determine the extent to which teams are being utilizedand/or whether ARL libraries are becoming team-based orga-nizations. While only five libraries described themselves asteam-based, the majority of respondents indicated that theyhave both permanent and project teams in place. In addition todescribing the characteristic features of teams, the survey alsoaddressed training and assessing team effectiveness. The Kitincludes both a listing of teams currently in place in ARLlibraries and representative documents, such as team defini-tions, a team-building handbook, and team training manuals.

Customer Services Programs in ARL LibrariesSPEC Kit 231, May 1998compiled by George J. Soete, ARL/OLMSThis SPEC Kit reports on the extent to which ARL librariesengage in formal or organized customer service programs,programs that focus, in highly specific ways, on the peoplelibraries serve. The most common characteristics of suchprograms in ARL libraries are staff training efforts focused oneffective customer service, regular data-gathering aboutcustomers and their needs, and written customer servicepolicies. Included in the Kit are the survey results;documentation from member libraries, such as patron servicemanuals, case studies, principles for service statements, andneeds assessments surveys; and a selected reading list.

Affirmative Action in ARL LibrariesSPEC Kit 230, June 1998compiled by Joseph Diaz and Jennalyn Tellman, University ofArizona Library, in collaboration with DeEtta Jones, ARLAn examination of the extent to which ARL Libraries haveaffirmative action policies or programs in place, this SPEC Kitincludes documentation on employment policies, instructionsfor hiring and interviewing staff, guidelines for search commit-tees, appropriate interview questions, sample statisticalreports, and recruitment plans. Included also is an extensivelist of resources, such as newsletters, articles, electronic lists,and other publications.

Evaluating Academic Library DirectorsSPEC Kit 229, May 1998compiled by George J. Soete, ARL/OLMSPerformance evaluation is a widely acknowledged tool formonitoring and promoting effective leadership, an essentialingredient for success in any library organization. This SPECsurvey examined factual attributes of director review process-es, including frequency of review cycles, participants, and

outcomes as well as level of satisfaction with the processes.Included in the Kit are the survey results and documentationfrom member libraries such as sample review forms andprocedures from responding institutions. In addition, afeature of this Kit is "A Checklist of Recommendations."See also Occasional Paper #21, described below.

Organization of DocumentCollections and ServicesSPEC Kit 227, April 1998compiled by Cynthia Clark and Judy Horn, University ofCalifornia at IrvineThis survey examined how collections, technical services,and public services for government documents are currentlyorganized in ARL libraries; the extent to which documentcollections are being reorganized; and the impact of suchreorganizations. Survey results indicated that the majormotivations for change in documents collections is the desireto improve workflow and services, the increased availabilityof bibliographic records, and the growth of electronic publica-tions. Included in the Kit are the survey results and documen-tation from member libraries, including organizational charts,sample websites, and planning and policy documents.

SPEC Kits: ISSN 0160-3582, $40 ($25 ARL members).SPEC Flyers summarizing SPEC Kits are available via theARL web at: <httpl/www.arLorgIspecIspeclist.html>.

OLMS Occasional PapersEducational Background of Systems LibrariansOLMS Occasional Paper #20by Jean Dorian, April 1998, ISBN 0-918006-32-5, 49 pp.,$25 ($18 ARL members)Although the impact of automation on library staff as a wholehas attracted a great deal of attention during the past twodecades, the emergence of a new specialization within theprofessionsystems librarianshiphas remained relativelyunexamined. In December 1996, a survey of systems librariansin ARL libraries was conducted to provide more informationabout the manner in which these positions have developed;the educational qualifications and training required; theirplace within the organizational structure of the library; andthe long-term prospects of systems librarians.

Evaluating Library Directors: A Study of CurrentPractice and a Checklist of RecommendationsOLMS Occasional Paper #21by George J. Soete, May 1998, ISBN 0-918006-34-1,48 pp., $25 ($18 ARL members)The project to study the performance evaluation processes forARL directors began in 1997 and included a SPEC Survey,available as SPEC Kit 229. This paper expands on the surveyto include additional analysis of survey results as well as thefindings from telephone interviews with ARL library directors.The interviews report on directors' perceptions of their reviewprocesses as well as suggestions for how these processes maybe improved. In addition, the current roles and responsibilitiesof library directors are examined. The paper also includes areading list and a checklist of recommendations, availableonline at: <http:/ /www.arl.org/oms/checklist.html>.

A R L 1 9 9 A U G U S T 1 9 9 8 6 s

ARL---Acrw-rnEsLee Anne George, ARL Program Planning Officer

ARL HONORS K. WAYNE SMITHOn June 29, ARL presented K. Wayne Smith,

retiring President and CEO of OCLC OnlineComputer Library Center, Inc., with an award

honoring his many distinguished contributions toresearch libraries and the scholarly community.The award was presented by ARL President JamesNeal, Johns Hopkins University, and ARL ExecutiveDirector Duane Webster during the OCLC President'sLuncheon in Washington, D.C. The award acknowl-edged that during his years at OCLC Dr. Smith was aninnovator, a coalition-builder, and a catalyst for change.His visionary leadership resulted in improved access toinformation and in reduced costs for managing thatinformation.

Dr. Smith forged successful partnerships withresearch libraries to build a cooperative base for deliv-ering technological innovations for integrating localsystems and OCLC services. Of great importance toNorth American research libraries are the powerfulenhancements in the OCLC Interlibrary Loan Systemand groundbreaking initiatives in the publishing andarchiving of electronic information resources.

In the words of the award, "K. Wayne Smithastutely confronted the challenges of managing thelargest library computer network and guided thetransformation of OCLC into a global organizationand the world's foremost bibliographic database."

Dr. Smith was the third president of OCLC.He served the corporation for nearly ten years.

6 9

HONORSMiriam A. Drake, Georgia Institute of Technology, wasawarded an honorary Doctor of Library Science degreeby Simmons College. The award was presented at aceremony in Boston on May 17, and recognized herprofessional vision, scholarship, and leadership as wellas the impact she has made on higher education and thefield of library and information science.

David Kohl, University of Cincinnati, and WilliamPotter, University of Georgia, received the 1998Blackwell's Scholarship Award from the Association forLibrary Collections & Technical Services for their pair ofarticles on library consortia which appeared in theWinter 1997 issue of Library Trends.

Maureen Sullivan, OLMS Organizational DevelopmentConsultant, became the 61st President of the Associationof College and Research Libraries on June 30. AsPresident of ACRL, Maureen will promote her theme ofLeadership of Learning.

Winston Tabb, Library of Congress, received theMelvil Dewey medal from the ALA at its annualconference. The Dewey medal is given annually torecognize creative professional achievement of highorder and was awarded to Winston in recognition of hiswork with the LC Program for Cooperative Cataloging.

Wayne State University: The Shiffman Medical Librarywas awarded the 1998 Innovation in Instruction Awardby ACRL for the program Health Sciences InformationTools 2000.

TRANSITIONSAssociation of American Universities: Nils Hasselmobegan as President July 1. Previously the President ofthe University of Minnesota, he replaces CorneliusPings, who announced his retirement last year. Pingshad been AAU president since February 1993.

Association of Research Libraries: Deborah Jakubswas appointed Director, Collections Services at DukeUniversity, effective September. Since 1996 she hasbeen a Visiting Program Officer at the Association ofResearch Libraries, leading the Global ResourcesProgram funded by the Mellon Foundation. In her newcapacity at Duke, Deborah will continue as head of theARL Global Resources Program but with a somewhatreduced time commitment.

Trish Rosseel, Librarian/Analyst for the ElectronicLibrary Network in British Columbia, joins the OLMSas Visiting Program Officer August 17. She willinvestigate uses of technology for deliveringtraining electronically.

A R L 1 9 9 A U G U S T 1 9 9 8

A

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-296-2296 FAX 202-872-0884<http://www.arl.org/>Copyright: © 1998 by the Association of Research Libraries

ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

-

L

Executive Director: Duane E. WebsterEditor: G. Jaia Barrett, Deputy Executive DirectorCo-editor: Lee Anne George, Program Planning OfficerCopy Manager: Karen A. VVetzelDesigner: Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscription; Nonmembers$50 per year.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arl.org>.

ARL CALENDAR 1998October 7-9 ARL/OLMS Facilitation

Skills InstituteWashington, DC

October 12-14 Workshop on ElectronicPublishing of Data Sets onthe WWWCharlottesville, VA

October 13-16 ARL Board andMembership MeetingWashington, DC

October 13-16 Educom'98Orlando, FL

October 19-20 From Data to Action: An ARLWorkshop on Strategies toRedesign ILL/DD ServicesEdmonton, Alberta

October 22-23 From Data to Action: An ARLWorkshop on Strategies toRedesign ILL/DD ServicesDenver, CO

October 27-28 ARL/OLMS LeadingChange InstituteChicago, IL

November 2-3

November 2-4

November 5-6

November 12-13

November 16-19

December 7-8

December 8-11

70

From Data to Action: An ARLWorkshop on Strategies toRedesign ILL/DD ServicesOttawa, Ontario

OLMS/DORAL ManagementInstitute for DevelopmentOfficersBaltimore, MD

Workshop on LicensingElectronic InformationResourcesKansas City, MO

From Data to Action: An ARLWorkshop on Strategies toRedesign ILL/DD ServicesChapel Hill, NC

ARL/OLMS LibraryManagement Skills Institute I:The ManagerWashington, DC

CM Task Force MeetingSeattle, WA

CAUSE98Seattle, WA

\C

r-A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONS

VIEWS OF THE CURRENT MARKETPLACEFOR SCHOLARLY JOURNALSby Mary M. Case, Director of the ARL Office of Scholarly Communication

Scholarly research is the life-blood of theadvancement of knowledge. Faculty, univer-sities, societies, federal agencies, research insti-

tutions, publishers, and libraries all play a role insupporting the creation and dissemination of thatresearch. Broad, easy, and early access to researchfindings facilitates the work of the practicing scien-tist, while access to quality-controlled, permanentlyavailable articles provides the basis for rewardsystems, cross-disciplinary investigation, graduatestudy, and the historical record. However, theformal and informal systems of scholarly commu-nication that have grown up to support theadvancement of knowledge are in jeopardy as aresult of market forcesmost visible in scientificjournals publishingthat are reducing accessto the published literature in all disciplines.

Librarians are acutely aware of the dynamicsin the market for academic journals. Price increasesfor a serial subscription for a research library haveaveraged 9.5% a year over the past decade. Despiteinfusions of funds into serials budgets, librarieshave been unable to keep pace with the steep priceincreases. As a result, libraries have had no choicebut to cancel significant numbers of journal sub-scriptions and to reduce monographic purchasing,decimating their collections. Many believe theexploitative practices of a few large commercialcompanies operating in a near monopolisticmarket are the fundamental cause of high prices.

In 1988, ARL contracted with the EconomicConsulting Services (ECS) to conduct a serials pricesstudy. Using best estimates available, ECS conclud-ed that the increasing costs of production and the

71

growing size of the journal titles could not fully accountfor a large portion of the prices charged by four commer-cial publishers. In 1997, a 20 -year study of scientificjournals by Tenopir and King (Journal of ScholarlyPublishing 28.3: 171-79) came to a similar conclusionand attributed an unknown but significant proportionof price increases to the pricing practices of commercialpublishers. Studies within individual disciplines havetime and again demonstrated higher costs per characterand per page of some commercially published journalswhen compared to journal titles published by societiesor other non-profit groups.

Library responses to this growing body of researchabout the extent and nature of price increases for schol-arly journals have been multifaceted. National andinternational library conferences have drawn attentionto the research findings and explored possible solutions.On the local level, librarians have combined campusexperience with the literature published about theresearch to help inform faculty and students about thefinancial challenges and choices inherent in managingserial collections. Over time, this process of informingand educating research library users, boards, and fund-ing agencies has contributed to a broadening circle ofawareness about dysfunctions in this marketplace.This awareness has finally reached into the offices offaculty, provosts, university presidents, and antitrustauthorities in the U.S. and Europe.

With this awareness has emerged a sense that theentire system of scholarly communication is in dangerof collapsing unless there is concerted effort by theacademic community to promote less expensive chan-nels for publication, dissemination, and archiving ofscholarly research.

Continued

This special issue of the ARL newsletter exploresthe nature of the journals market in depth and presentsstrategies for challenging the status quo. Earlier issuesof this newsletter have reported on ARL's commitment toSPARC (the Scholarly Publishing & Academic ResourcesCoalition) as a strategy to create partnerships with pub-lishers who are developing high-quality and yet economi-cal alternatives to existing high-priced publications andto foster expanded competition in scholarly communica-tion. ARL views SPARC as one strategy among manyfor addressing the complex issues involved in the currentscholarly publishing system. The four articles that followin this issue of the newsletter propose additional, comple-mentary strategies that libraries, faculty, universities, andthe academic community as a whole may pursue to keepchallenging the current marketplace.

We begin the issue with a report by Mark McCabe,an economist formerly with the U.S. Department of Justice(DOD. McCabe presents the preliminary findings of theDOJ's work to develop a new model for understandingthe competitive impact of publisher mergers on the acade-mic journals market. Librarians' outcry at the proposedReed Elsevier/Wolters Kluwer merger encouraged theDOJ to look beyond its established criteria for evaluatingpublishing mergers and to explore whether other dynam-ics were operating that would permit mergers of relativelymodest size to cause competitive harm and higher prices.Librarians across the country spent hours on the phonewith McCabe and his legal colleagues at the DOJ as theyattempted to understand how libraries behave as a mar-ket. Subsequently, staff in over 50 ARL libraries providedDOJ with data on the holdings of over 3,000 journal titlesfor a ten-year period. Some of these staffs worked over-time for two weeks to gather the information from manualrecords. These data, combined with that collected else-where by the DOJ, allowed McCabe and his colleagues todevelop and test models by comparing projected effectswith the actual effects of previous mergers and acquisi-tions. Based on this testing, McCabe outlines here his newportfolio theory of journal pricing that suggests that "pub-lisher mergers of relatively modest size can cause compet-itive harm." While the Reed Elsevier/Wolters Kluwerdeal was ultimately abandoned by the companies,several other publisher mergers have come before theDOJ. McCabe will therefore continue his research for theDOJ even as he takes on a new position as an assistantprofessor at the Georgia Institute of Technology.

McCabe's work helps us to understand the dynamicsof the marketplace in scholarly journals publishing.It may lead the DOJ to some future action that will pro-tect the academic community from the anti-competitivepractices of some publishers. Most importantly, it willreplace some of our speculations with fact, allowingus to develop even more effective strategies.

In the second article, Brendan Wyly, a librarian inthe Johnson Graduate School of Management Library atCornell University, uses publicly available data to analyzethe financial health of four major publicly traded compa-nies that have significant scholarly publishing operations.Wyly describes the measures of profitability found in thesecompanies' annual reports and concludes that the profitsof some of the major commercial publishers of scholarlyjournals are, in a word, exceptional. He also concludesthat these profits confirm a lack of competition in themarketplace. His response is to call for a new system ofscholarly communication, created primarily by universi-ties, that provides the kinds of innovations that will lureauthors away from commercial publishers and endreliance on journal purchasing in a non-competitive market.

Regardless of what we may think about their price,commercial publications, according to Louisiana StateUniversity Library's Stanley Wilder, are of less value tofaculty than society publications when measured againstthe revenue they generate. Wilder presents the specificcase of chemistry, where he finds that commercially pro-duced journals account for 74% of the revenue generatedby a core set of chemistry titles but contribute onlybetween 22-35% of the value. The response? He calls fora change in philosophy of those academic librarians whocurrently pursue comprehensiveness in collecting scientificand technical journal literature. He recommends insteadplacing value at the heart of scientific and technicalcollecting and relegating lower value literature to morecost-effective document-on-demand acquisition.

Concluding this issue is a report from Mike Sosteric,a faculty member of Athabasca University's Centre forGlobal and Social Analysis and founding director of thenew International Consortium for Alternative AcademicPublication (ICAAP). Sosteric writes in practical termsabout the very real barriers to independent scholarly publi-cation and describes how the recently formed ICAAP plansto tackle these barriers in order to develop an internationalalternative scholarly communication system outside of thecommercial mainstream. Most noteworthy about ICAAPis that it is a faculty-initiated project intended to facilitatecultural change in faculty editors and authors while pro-moting the acceptance of independent publishing outlets.Sosteric calls on research libraries and their institutions tojoin this Canadian-based, international initiative.

ARL's Office of Scholarly Communication welcomescommunity response to these proposals for influencing thecurrent marketplace for academic journals. We realize thatthere are many other worthy initiatives, which we willfeature in upcoming OSC newsletter columns. Moreover,we will continue to promote and facilitate actions such asthese that complement and reinforce the agenda of SPARC.

A R L 2 0 0 O C T O B E R 1 9 9 8

THE IMPACT OF PUBLISHER MERGERSON JOURNAL PRICES:A PRELIMINARY REPORTby Mark J. McCabe, Assistant Professor of Economics,Georgia Institute of Technology

When I was asked by the Department of Justice(DOD to consider the potential competitiveimpact of a number of publishing mergers

on the market for academic journals, my initial reactionwas, franldy, one of skepticism. I thought to myself,Hundreds of unique titles exist and the number of publishersis immense. The first step in antitrust analysis is definingthe market in order to determine whether a monopolistcould in fact wield power in this market. In scholarlypublishing, doesn't each unique journal title constitute adistinct market for the purposes of antitrust analysis?Certainly no one would argue that articles in BrainResearch could be easily substituted for ones in theNew England Journal of Medicine, much less those in theAmerican Economic Review. If each title corresponds toan antitrust market, then owners of individual titlesalready have the capacity to achieve monopoly returns;a corollary is that mergers don't matter. Furthermore,even if markets are defined somewhat more broadly,say, to include titles whose content overlaps, the likeli-hood that two publishers (among dozens) wouldtogether control sufficient content in enough subjectareas to warrant antitrust scrutiny seemed small.

I conveyed these first impressions to my legal col-leagues at the DOJ, and, well, they were not particularlyenamored with this point of view. Their initial discus-sions with university librarians across the country asthey began their investigations into the proposed ReedElsevier/Wolters Kluwer merger had revealed a certain(high) level of agitation over persistent journal priceinflation during the past decade or more. Furthermore,the librarians' ire was focused on a small number oflarge publishing houses (large in the sense that theyhad large portfolios of titles) whose price "leadership"had helped decimate their budgets and imperil scholar-ship at their institutions.

So, what was to be done? The Antitrust Divisionhad recently evaluated a number of publishing mergers(Thomson/West, for example). As a result of theseinvestigations, several of my economist colleagues atDOJ had concluded that measuring book or journalcontent differentiation was an appropriate method forevaluating the impact of increased concentration inthese types of markets; that is, only if two or more jour-nals were close substitutes might there be any potentialconcern.' In other words, their approach paralleled myown initial impressions! Obviously, if companies werebehaving anti-competitively in this market, we would

have to break new ground to determine how and why.But first, what about alternative explanations for

the persistent price increases of journals? For example,perhaps the price inflation reflects increased costs.However, an extensive review of the relevant literature(published largely in the library science field) revealedthat the actual costs of journal editing and printingseemed not to have suffered any unusual run-up duringthe past decade. Another potential and more compellingexplanation relies on a simple application of supply anddemand analysis. Our interviews with libraries indicatethat university budgets for periodicals are relatively fixedin a given year so that "large" increases in the populationof available titles might induce librarians to cancel sometitles as they add new, more desirable journals to theircollections (this "population" invariably increases overtime). And since each issue of a journal involves certainfixed editorial costs in addition to variable printing costs,a persistent decline in a title's circulation will eventuallyforce firms to raise prices as they attempt to cover thefixed costs of publication. In other words, everythingelse equal, a smaller subscriber base necessitates higherprices. Of course, the interesting feature of this explana-tion is that entry by new journals is a source of priceinflation.2 "Demand" for new titles eventually resultsin higher prices across the board.

Roger Noll, Morris M. Doyle Centennial Professor ofEconomics at Stanford University, is an advocate of thishypothesis. To test the hypothesis, he collected price,quality, and circulation data for a sample of journals thatwere included in the Institute for Scientific Informationindexes in a number of fields (for the period 1978-88).After estimating supply and demand equations, he foundthat supply was downward-sloping, i.e., price and circu-lation were inversely related. Since entry by new journalsreduces demand for existing journals, this estimated "sys-tem" predicts increased prices with entry. It should benoted that Noll's analysis does not determine what por-tion of journal price inflation is due to entry (presumablythis could be assessed by measuring the change in journalpopulation during his sample period). Furthermore,his system estimation indicates that significant priceinflation has occurred independent of changes in circula-tion. That is, even after accounting for the effect ofcirculation on prices, there remains a large unexplainedincrease in prices.

The existence of this "unexplained" residual inflationopens the door for additional and complementary theo-ries. Over the past six or seven months, I began initialdevelopment and testing (with some assistance from aDOJ colleague, David Reitman) of a portfolio theory ofjournal pricing that suggests that publisher mergers ofa relatively modest size can cause competitive harm.The remainder of this brief report summarizes this work.

A R L 2 0 0 O C T O B E R 1 9 9 8

-SPECIAL ISSUE ON-jOURN-AL-SContinued

Some Views of the Marketplacefor Academic JournalsThe Demand SideFirst, some useful facts regarding what the DOJinquiry found about the demand for academic journals.Purchase of academic journals by libraries is generallybased on two factorsprice and expected usage. Toassemble and maintain their collections, most librariesappear to construct a cost per use ratio for each title.3Given a budget for a relevant academic field, e.g.,medicine, they then proceed to rank journals from low-est to highest in that field according to this ratio, andidentify a cutoff above which titles need to be canceled.Conversely, if their holdings in the relevant field do notexhaust the budget, additional titles can be purchaseduntil the budget constraint is met. From year to year, asbudgets and titles' usage change, collections are adjust-ed accordingly. Over the past decade or so the generaltrend is for increases in library budgets to lag journalprice inflation; a consequence is that many librarycollections have suffered a fair amount of attrition.

The most interesting aspect of library journal acqui-sition, of course, is that individual titles within a givenfield are considered simultaneously. So, for example,medical libraries group titles from various sub-fields,e.g., neurology, biochemistry, clinical medicine, etc.,into a single "portfolio" and broadly apply the cost peruse criterion. Thus, titles compete with each other forbudget dollars across an entire field rather than across anarrow sub-field, as intuition might otherwise suggest(ar} intuition based on the perspective of the typical userof journal materials, and a basis for my initial skepti-cism). Furthermore, since journals are highly differenti-ated even within sub-fields, libraries try to purchase asmany titles as possible (except for the most widely usedjournals, libraries purchase no duplicate subscriptions).

Publishers' Pricing StrategiesGiven this demand structure, how do publishing firmsprice their journals? First, there exist at least two typesof journal companiescommercial and non-profitthathave different strategies. In general, the latter are most-ly intent on disseminating knowledge, whereas the for-mer are primarily interested in profits. Here I assumethat the non-profit firms set prices to cover costs (andare thus ignored in the analysis that follows). DavidReitman and I have both considered how this questionmight be answered for commercial firms. Although ouranalyses differ in certain details, they should be seen ascomplementary.

David's modelDavid assumes that a continuous distribution of librarybudgets exists, i.e., libraries have unique budgets thatcan be ordered from smallest to largest. He shows that

each journal company (owning a single journal title)will set prices so that higher quality journals will exhibitlower cost per use ratios. Given n journals, these can beranked from lowest to highest in terms of cost per useby libraries, with the lower quality journals (that havehigher cost per use) purchased by relatively high-budgetlibraries. Conversely, higher quality journals (that havelower cost per use) are purchased by most libraries.4

The intuition for this particular ordering is thathigher quality imparts a "cost advantage" that makes itmore profitable to price low and sell widely. Given thisstrategy, lower quality, or "high-cost," journals find itmost profitable to price high and sell to fewer, relativelyhigh-budget libraries. Note that although the latterfirms could match the "low-cost" firms' prices, thisstrategy is less profitable than targeting the smallerbase of well-endowed customers.5

Mark's modelMy approach differs from David's primarily in theassumed distribution of budgets. I consider a finite setof discrete library budget classes; each class is populat-ed by one or more libraries. The number of journalsexceeds that of budget classes (one motivation for thisapproach is that, in the case of medicine, for example,the number of journals is much larger than the popula-tion of academic libraries in the U.S.). Given theseassumptions, my model shows that single-title journalfirms targeting the same budget class set prices so thattheir cost per use ratios are identical to each other; forhigher budget classes this ratio increases. Furthermore,given some set of journals of varying quality, higherquality journals will generally target the lower budgetclasses, resulting in lower cost per use ratios (as inDavid's approach). The intuition for. these results is thesame as before: higher (lower) quality implies a cost(dis) advantage, which dictates a lower (higher) priceand therefore broader (narrower) circulation.6

While both approaches to the question of how pub-lishing firms set their prices are based on single-titlecommercial journal firms, they set the stage for futureanalysis of larger portfolio firms.

Implications for MergersUnder David's scenario, mergers between firms control-ling two adjacent journalsi.e., journals with adjacentcost per use ratiosresult in higher prices post-merger,and therefore a change in the holdings of libraries.Mergers between firms with non-adjacent journals (atleast in the few simulations that David performed)appear to have no impact on prices. This is probablydue to the fact that a given journal's price is constrainedby the next higher journal, in terms of cost per use.Relaxing this constraint benefits the lower cost peruse journal but hurts the higher cost per use journal.

ARL 200 OCTOBER 1998 7 4

Mergers of firms with non-adjacent journals are lessprofitable since the lower cost per use journal's pricingis still constrained by an adjacent journal(s) that is notinvolved in the merger.

Under Mark's scenario, mergers between firms thatcontrol two journals that target the same budget classor adjacent budget classes may be profitable. In otherwords, mergers do not have to involve journals of adja-cent quality. Note that with this approach some pricesincrease while others decrease, depending on whetheror not a journal switchesbudget class post-merger.For example, consider amerger between firms withtwo journals in the samebudget class. The price ofthe lower quality journal ofthe two involved in themerger is raised enough sothat the libraries in thisbudget class would cancelthe journal and replace itwith one of lower quality that is now more attractive interms of price. Using the low-cost analogy, the intro-duction of this relatively "high-cost" journal allows therest of the journals in this class (including the high qual-ity journal owned by the merged firm) to raise prices.Depending on the parameters, the correspondingincrease in profits for the merged firm may or maynot outweigh the losses incurred by its lower qualityjournal.

Much work remains to be done here, including theanalysis of larger portfolio firms and the interplaybetween price inflation and budgets that was alludedto earlier.

of internal growth versus that due to acquisition. Wespeculate that internal growth due to entry may producebenefits (such as coverage of an otherwise uncoveredfield) that help offset any intentional competitive harm;any harm associated with acquisitions, on the otherhand, is less likely tO be balanced by substantial bene-fitstitles are simply reshuffled, and any fixed costsavings seem to be small.8

We adopted a reduced form or hedonic method ofestimation, where prices are regressed on several fac-

tors: firm portfolio size,journal quality (measuredby ISI citation level), aBrandon-Hill dummy (theBrandon-Hill journal list9indicates essentiallywhether a journal is a gen-eral or specialized title andis thus an exogenous mea-sure of circulation "poten-tial"), and a number ofvariables that capture the

effects of residual price inflation and idiosyncratic firm-specific effects. For further technical details see theshort empirical appendix that follows.

Briefly, our results for journals sold by commercialpublishers indicate that prices are indeed positivelyrelated to firm portfolio size, and that mergers result insignificant price increases. For one specific transactionwe investigatedWolters Kluwer and Waverlytheresults predicted an average price increase of between20-30% for the affected medical titles. What is particu-larly noteworthy are the modest portfolio sizesinvolved. Based on the total number of ISI journalssold by commercial publishers (with portfolios of tentitles or greater), Wolters Kluwer and Waverly haveshares equal to about 11% and 5%, respectively.Contrary to the current rule of thumb in anfitrustthecombined shares of the merging firms typically need toapproach 35-40% before measurable effects arisetheseshares are relatively modest.1° An explanation for thisanomalous result is that unlike most markets, wherebuyers purchase a product from one of several sellers, inthe market for journals, buyers purchase titles from asmany sellers as possible. This strategy delivers moremarket power to "small" firms, hence the large impactfrom modest-sized mergers.

Finally, even after controlling for the effects of firmsize and other variables, there remains significant resid-ual price inflation. One explanation is that our portfolioapproach and Noll's entry story are complementaryboth factors contribute to the observed price inflationseparately.11 And the residual price inflation in eachestimation reflects the impact of the omitted factor,

Briefly, our results for journals sold by

commercial publishers indicate thatprices are indeed positively related to

firm portfolio size, and that mergersresult in significant price increases.

Testing the Portfolio TheoryDuring the course of our investigation at DOJ we col-lected a range of data in order to test the portfolio theoryof journal pricing. We collected price data for some3,000 journals, holdings information from variouslibraries, and quality information from the Institutefor Scientific Information (ISI) (for the period 1988-98)across a number of academic fields for ISI-rankedjournals. Since our progress was greatest for medicaljournals,' my discussion is focused in that subject area.In 1998, the number of ISI-ranked medical journalspublished by commercial firms exceeds 1,000.

The sample period, 1988-1998, is useful in at leasttwo respects. First, it is sufficiently long to assesswhether inflation continues to plague the journalmarket (and it dovetails nicely with Noll's sample).Second, the period contains a number of natural experi-ments, i.e., publishing mergers, that help us distinguishbetween different "types" of portfolio effects: the effect

7 5A R L 2 0 0 O C T O B E R 1 9 9 8

*

Continued

whether it be firm portfolio size or declining circula-tion. Clearly, one important avenue to pursue in thefuture is how these two factors should be incorporatedinto a single estimation (the medical libraries holdingsdata should be useful in this regard since it providessnapshots over time of how journal circulationchanged for a representative sample of libraries).

Implications and Future DirectionsWhen the proposed merger between Reed Elsevier andWolters Kluwer foundered earlier this year, oppositionfrom antitrust authorities in Europe and the U.S. wascited as a primary cause. Although no formal com-plaints were filed by agencies on either side of theAtlantic, regulators had sent a variety of signals indi-cating their serious concerns. Negotiations with theEuropean Union had progressed the farthest and itappeared that the proposed deal would proceed onlyif the parties agreed to significant divestitures. It waswidely reported at the time that the EU's preferred setof divestitures upset the financial logic of the mergerand resulted in its demise.

What is interesting here is that the EU's mainfocus was not on academic journals, but rather onlegal publishing (in Europe), and that its theory ofanti-competitive harm was based on traditionalantitrust principles, i.e., excessive overlap in content(and therefore similar to the DOJ's approach to theThomson/West merger).

The U.S. focus, of course, was far different, in partbecause European legal publishing was not germaneand because the model of harm relied upon was novel.Though one can only speculate on how a U.S. antitrustcase might have proceeded, it is clear that the com-bined Reed Elsevier/Wolters Kluwer entity wouldhave controlled large journal portfolios in a numberof broad fields, including biomedicine. Assuming thatthese broad fields constituted antitrust markets, someof these portfolios would have crossed the U.S.Government's threshold with shares in excess of 35%.Based on the preliminary results discussed here, sucha merger would have resulted in substantial priceincreases over time. If the U.S. had filed a complaintand had been successful with this market definition,an important legal precedent would have been set, onethat would have made it easier to employ a portfoliotheory in mergers involving combined market sharesless than the threshold andlor a large firm buying a rel-atively small portfolio of journals (see endnote 10).Unfortunately, since any future deals are likely to berelatively small in scope, opposition to journal mergerswill need to be successful in both dimensions: marketdefinition and market shares.12

This increased burden is why further refinement ofthe work started over the last eight months is necessary.

Using the price, portfolio and holdings data collectedduring the course of our investigation, I hope tofurther develop our portfolio theory and to test therobustness of our initial empirical results. Althoughantitrust policies in the U.S. and Europe havechanged considerably over the past two decadesin response to a better understanding of marketdynamics, this type of reform requires the develop-ment of new and persuasive evidence that existingpolicies are inadequate.

Empirical AppendixHedonic estimation does not provide estimates ofthe "structural parameters" from a supply/demandsystem. This can limit the quality of inference. Forexample, suppose one regressed auto prices on a setof car characteristics, e.g., horsepower, passengercapacity, transmission type, etc. If we found that carprices were positively associated with horsepower,should we conclude that the cost of additional horse-power is positive or that demand for larger horse-power engines is greater? In the present case, thisinference problem is not likely to affect the firmportfolio size parameter. Since a merger should notincrease costs, a merger-related price increase in thehedonic framework is demand-related, i.e., a resultof enhanced market power.

Another concern whenever "firm size" is speci-fied as an explanatory factor in a price or profitsequation is the issue of endogeneity. What deter-mines firm size? If firm size is merely a proxy forsome other X-factor (say, lower costs, or skill at intro-ducing new journals) it is possible that any conclu-sions regarding the negative effects of firm size per semight change if this additional factor(s) were includ-ed in the analysis. This problem is most acute withcross-sectional data, since firm-fixed effects are omit-ted and no changes in the relevant variables can beobserved over time (see the Handbook of IndustrialOrganization chapter by Richard Schmalensee on thistopic13). However, panel data (cross-sectional dataover time) of the sort collected here is ideally suitedto address these endogeneity concerns. A panelallows us to observe the impact of changes in portfo-lio size over time, holding firm attributes constant.And to address any lingering concerns that thesechanges in firm size are masking some X-factor (asnoted above), the data allows us to distinguishbetween increases in firm size due to internal growthversus growth from merger. Since we are able tocontrol for quality changes, a significant (in thestatistical sense) post-merger price increase is agood indication of an anti-competitive effect.

ARL 200 OCTOBER 1998 76

Endnotes1For example, as a condition of DOJ approval of theThompson/West merger, the companies were required todivest a number of law journals in order to reduce the con-centration of titles with similar content that would now beowned by the single company.

2How much of this "brand proliferation" and its impact onprices is due to anti-competitive intent, rather than morelegitimate reasons, is a question that remains to be addressed.

3This claim is literally true for medical libraries; though othertypes of academic libraries may not be as precise in their pro-cesses, they appear to behave in similar fashion. In any case,this is an empirical question that can be tested using holdingsdata (if one examines how a given library's holdings changeover time, it should be possible to test whether the changes inholdings correspond to the predictions of a cost per use port-folio model that takes into account journal quality and price).

4We used price divided by number of citations as a proxy forcost per use.

5For the more technically inclined, David's approach employsa "constrained" Nash equilibrium concept where each firmchooses a price so that journals with higher cost per use ratioshave no incentive to undercut.

6Again, for the more technically inclined, it should be notedthat to demonstrate existence of equilibria in this case it isnecessary to think of pricing as part of a two-stage game, i.e.,in the first stage firms select the budget class in which theywill competethis might reflect marketing costs and soforthand in the second stage they choose prices.

'When the Reed Elsevier/Wolters Kluwer merger was calledoff in March 1998, the DOJ refocused their analysis on thepending merger of Kluwer with Waverly Press, a publisherof biomedical journals.

5Furthermore, if publishing mergers do result in cost savings,economic theory implies that post-merger prices shoulddecline, everything else equal.

9Alfred N. Brandon and Dorothy R. Hill, "Selected List ofBooks and Journals for the Small Medical Library," Bulletin ofthe Medical Library Association 85.2 (Apr. 1997): 111-35.

w This rule of thumb is suggested by the government's antitrustguidelines, and helps to explain the recent reluctance of theAntitrust Division to oppose mergers in the publishingindustry. Note that even if two merging companies havecombined shares in excess of this threshold, one company'smarket share cannot be trivial, e.g., one or two percent.

11These are likely to be "separate" effects since the effects ofentry are not firm specific, whereas those of portfolio size are.

12Meanwhile, to help avoid future antitrust scrutiny theElseviers of the publishing world will likely grow byadding relatively small numbers of journals at frequentintervals. If pursued diligently, this stealth strategy can bejust as successful as any megadeal.

13RichardSchmalensee and Robert D. Willig, eds., Handbook

of Industrial Organization, Handbooks in Economics 10,Amsterdam; New York: North Holland, 1989.

© 1998 Mark J. McCabe. The author grants blanketpermission to reprint this article for educational use as longas the author and source are acknowledged. For commercialuse, a reprint request should be sent to the author<mark.mccabe@econ.gatech.edu>.

COMPETITION IN SCHOLARLYPUBLISHING? WHAT PUBLISHERPROFITS REVEALby Brendan J. Wyly, Johnson Graduate School of ManagementLibrary, Cornell University

Abs oth customers and critics of commercial scholarlypublishers, librarians might find useful a sum-mary of the recent finances of the publicly traded

companies that have significant scholarly publishingoperations. A financial analysis can help us determineif our trust in maintaining long-term relationships withthese companies is warranted and at the same time sug-gest whether our concerns that an anti-competitivemarket is operating are supported by the data.

Recent Profits of Publicly TradedScholarly PublishersThe publicly traded companies with significant scholarlypublishing operations focused on here are WoltersKluwer, Reed Elsevier, John Wiley & Sons, and PlenumPublishing. In March 1998, Wolters Kluwer and ReedElsevier canceled plans to merge. In early June 1998,Wolters Kluwer announced plans to acquire Plenum.

Finding meaning in the measurements of publishinghouses requires an understanding of what existing datarepresent (and/or exclude) and a context for interpretingthe data. This summary presents three tables to examinethe four companies. The first table looks at the overallsize of the companies. The second presents financialratios to indicate profitability of overall operations.And the third table presents a hypothetical summary ofsavings for the customers of these companies' scholarlypublishing segments had the companies operated theirscholarly publishing segments at the median measureof profitability for the periodical publishing industry.Taken together, these figures give a reasonable pictureof the profits achieved by these companies from theirsale of scholarly materials.

To illustrate the relative size of the companies, Table 1reports sales, operating income, net income, net incomeavailable for common shareholders, and common equityfor the four publishers under review as well as for twoother companiesThomson and Microsoft. Thomson isnot a major publisher of scholarly journals, but it is amajor vendor of other resources to the research librarycommunity (for example, Thomson owns Gale Research,Information Access, the Institute for Scientific Information,and West), so it is included for informational and compari-son purposes. Microsoft is included as well to provide

1some basis for comparison as a well-known, highly prof-itable company operating in a somewhat monopolistic I

market. (In the March 30, 1998 issue, Business Week rankedMicrosoft as the most profitable company in the S&P 500in 1997 based on a variety of measures.)

A R L 2 0 0 O C T O B E R 1 9 9 8

Continued

In addition to the raw financial figures in Table 1, thefinancial ratios in Table 2 are presented as they are moreinformative about the market in which commercial schol-arly publishers operate. For comparison, periodical pub-lishing industry ratios are also included in Table 2 forperiodical publishing in general (not limited to scholarlypublishing). Ratios for the banking and consumer prod-ucts segments of the S&P 500 are included because theywere the most profitable segments in 1997, as measuredby net margin and return on equity, respectively.

There is no single ideal measure of profitability,and all of the financial ratios presented in this summarymust be considered relative to other companies andindustries. Three ratios are particularly useful tocustomers of these companies:

Operating margin measures how much of each salesdollar customers pay to all the beneficiaries of anenterprise over and above the costs incurred in provid-ing the goods or service. Those beneficiaries includestockholders, lenders, and members of society benefit-ing from taxes on the enterprise. A high operatingmargin may be needed to cover taxes and provide areturn that will attract the necessary capital throughlending or stockholding, given the particular risks ofthe enterprise and alternative investments. Operatingmargin excludes interest costs/profits that are theresult of decisions about how to finance an operationand how to use the income it generates rather thanthe result of the direct costs and benefits of carryingon an enterprise.Net margin takes interest and taxes out of the profitconsideration so that only common and preferred

dividend stockholders are considered as beneficiaries.A high net margin may be necessary if the enterprise ishigh-risk and requires a considerable capital invest-ment. However, a net margin of profits on sales that ishigher than the market norm is a substantial indicatorof "high" profits and raises questions about whethera competitive market is operating and, if so, why thatparticular market supports a higher margin thanother markets.

Both operating and net margins are reportedexcluding extraordinary financial events that are notexpected to recur, such as the one-time costs and bene-fits of acquisitions, dispositions, legal actions, etc.Return on equity is the ratio of net income (afterpayment of preferred dividends) to the net worth orcommon equity in the enterprise. Return on equitymeasures income generated for each dollar of capitalinvested by the ordinary owners who have no specialfirst claim on profits in the form of preferred dividends.

As measured by net profit margin, PlenumPublishing was more profitable in 1997 than 491 compa-nies in the S&P 500, according to Business Week. ReedElsevier's net margin was higher than 473 of the compa-nies in the S&P 500 for 1997. Both Plenum and ReedElsevier had extraordinary net profit margins comparedto the periodical publishing industry as a whole, andWolters Kluwer was near the upper quartile. However,as noted, high margins on sales may be necessary to earna sufficient return on the equity that must be attracted.Wolters Kluwer provided a higher return on equity in1997 than 482 of the S&P 500 companies, Reed Elsevierhigher than 448, Plenum higher than 361, and Wiley higher

TABLE 1: SALES, INCOME, AND COIVIMON STOCK EQUITYFOR THE LATEST FISCAL YEAR REPORTED

Net Sales Operating Income(before interest &taxes & excluding

extraordinary items)

Net Income(excluding

extraordinaryitems)

Net IncomeAvailable for

CommonShareholders

CommonShareholder

Equity(prior year)

Wolters Kluwer $2,569,808,000 $548,101,000 $285,644,000 $285,644,000 $685,235,000Dec. 31, 1997

Reed Elsevier $5,603,880,000 $1,451,400,000 $998,760,000 $997,120,000 $3,534,570,000Dec. 31, 1997

J. Wiley & Sons $431,974,000 $34,797,000 $20,340,000 $20,340,000 $117,982,000April 30, 1997

Plenum Publishing $52,634,000 $17,626,000 $12,824,000 $12,824,000 $63,399,000Dec. 31, 1997

Thomson Corp. $8,766,000,000 $959,000,000 $579,000,000 $550,000,000 $4,946,000,000Dec. 31, 1997

Microsoft $11,358,000,000 $5,130,000,000 $3,454,000,000 $3,439,000,000 $6,908,000,000June 30, 1997

Source: Respective annual reports and SEC filings.

A R L 2 0 0 O C T O B E R 1 9 9 8

than 302. Even though Wiley's net margin was relativelymodest (4.7%), it excelled in providing a strong returnon equity (17.2%). This ability to turn a net margin intoa high return on equity is characteristic of the publish-ing industry in which either a small amount of equitycan be put to work to generate considerable sales or theindustry leverages equity by borrowing to produce highreturns on small equity stakes. In simple terms, thismeans that publishers either do not need to raise a largeamount of capital to produce publications or that pub-lishers borrow what money they need rather than sell-ing equity stakes. A detailed analysis of the origins ofthese returns on equity is beyond the scope of this arti-cle, but a high return on equity is at least a potentialindicator that equity holders are benefiting from invest-ing in activities not subject to competitive forces. Incombination with other evidence, the return on equityfigures help bear out this assessment.

The periodical publishing industry as a whole, atthe lower quartile, median, and upper quartile levels,converts margins into return on equity at multiples ofthree or four times. Wolters Kluwer and Wiley attainedthese industry levels of performance as measured byreturn on equity as multiples of their margins. However,because Wiley started from a lower margin, only

Wolters Kluwer approached the extremely high returnon equity of the upper quartile of the industry. Comparedto most other industries, Reed Elsevier fairly efficientlyused capital to earn a high return on equity from a veryhigh margin, but it was not nearly as effective with itsmargin as other top-performing publishers. Plenumdid a poor job of converting a very high margin intoreturn on equity; this gives some indication of likelyinefficiencies on the part of Plenum relative to the othertwo companies and relative to periodical publishing ingeneral. However, we cannot definitively state theorigins of these return on equity figures without a verydetailed analysis of the debt and equity structures ofthese companies.

If the net margins of Wolters Kluwer, Reed Elsevier,Wiley, and Plenum had been 5.0% (the median in theoverall periodical publishing industry as measured byDun & Bradstreet's Industry Norms & Key Business Ratios1997-98) the customers of the three companies wouldhave saved approximately $884,653,000 in 1997.Alternatively, if the three companies had earned a returnon equity of 18.8%, the customers would have savedapproximately $490,145,000. Of course, those customerswere not all purchasing only scholarly publications fromthese companies. That raises the issues of how much

TABLE 2: PROFIT RATIOS FOR THE LATEST FISCAL YEAR

Operating Margin Net Margin Return on EquityWolters Kluwer* 21.3% 11.1% 41.7%Dec. 31, 1997

Reed Elsevier* 25.9% 17.8% 28.2%Dec. 31, 1997

J. Wiley & Sons* 8.1% 4.7% 17.2%April 30, 1997

Plenum Publishing* 33.5% 24.4% 20.2%Dec. 31, 1997

Thomson Corp* 10.9% 6.6% 11.1%Dec. 31, 1997

Microsoft* 45.2% 30.4% 49.8%June 30, 1997

Periodical Publ. (SIC 2721)* unavailable 11.5% (upper quartile) 54.3% (upper quartile)5.0% (median) 18.8% (median)

1.7% (lower quartile) 5.6% (lower quartile)Consumer/S&P 500** 9.3% 28.2%Dec. 31, 1997

Banks/S&P 500**Dec. 31, 1997 14.5% 15.8%

Sources: *Respective annual reports and SEC filings.lIndustry Norms & Key Business Ratios 1997-98, Dun & Bradstreet, data for 280 establishments.

'Business Week's Industry Rankings of the S&P 500," Business Week (30 Mar. 1998): 123.

A R L 2 0 0 O C T O B E R 1 9 9 8

Continued

revenue these companies derive from scholarly publish-ing (as opposed to their other product lines) and howthese companies regard their own scholarly publishingsegments.

Company Views on theScholarly Publishing MarketThe annual reports contain some information about thecompanies' views of the scholarly publishing marketand their intentions to focus on this profitable business.

Wiley reports that 47% of revenue for 1997 wasderived from science, technology, and medical publish-ing, consisting of journals, encyclopedias, newsletters,and books: "A fundamental strategy is to developjournals, newsletters, subscription books, multivolumereference works, and other continuity products. Suchproducts serve customers well by keeping them up todate in fast-moving fields of science and technology,and they benefit Wiley by generating predictableincome and substantial cash flow." Educationalpublishing, mainly of undergraduate texts, accountedfor 25% of 1997 revenue for Wiley, and professionalpublications mainly focused on business professionalsgenerated 28% of revenue in 1997.

Plenum reported that in 1997 subscription journalsaccounted for 63% of sales, books for 26%, and databaseproducts for 9%. No particular emphasis is reported ongrowing one segment relative to another.

Reed Elsevier reports, "The Scientific segment ofReed Elsevier comprises worldwide scientific and med-ical publishing businesses.... The Scientific segmentrepresented approximately 17% of Reed Elsevier's totalnet sales and 26% of Reed Elsevier's total operatingincome before exceptional items for the year endedDecember 31, 1997. Within the Scientific segment,Elsevier Science contributed approximately 83% of thetotal net sales in the year ended 31 December, 1997."The fact that Reed Elsevier made only 17% of its totalsales in the Scientific segment yet made 26% of its oper-ating income from that segment may in large partexplain Reed Elsevier's enthusiastic disposal of most ofits operations in the more competitive and less prof-itable consumer publishing market in favor of focusingon lucrative scholarly publishing.' The three profitablesectors, Scientific, Professional, and Business, grew netsales and operating income (at constant exchange ratesand before extraordinary items and acquisitions anddispositions) as follows in 1997: Scientific net sales +8%and operating income +12%; Professional net sales +7%and operating income +8%; Business net sales +5% andoperating income +11%. Reed Elsevier is clearly goodat converting sales in both the Scientific and Businesssegments into income.

Wolters Kluwer reports that 14% of its sales and13% of its operating income were derived from the

Scientific and Medical publishing segment. The largestand most profitable Wolters Kluwer operations are inTax and Legal publishing, including not only the CCHdivisions but also a very substantial amount of scholarlylegal publishing. The profitability of the large, some-what monopolistic Tax/Legal segment partly accountsfor the difference between Scientific/Medical sales at14% and operating income at 13% of overall operations.The Scientific/Medical segment is very profitable forWolters Kluwer, and they have been concentrating ongrowing it through acquisitions, including the plannedacquisition of Plenum. The other quite profitable sectoris Business publishing, which includes some scholarlypublishing. The Educational and Trade publishingsectors are much less profitable in terms of operatingmargins. In the following calculations of scholarlypublishing profit ratios (Table 3), only the Scientific/Medical sector finances are included for WoltersKluwer, although they clearly derive income fromscholarly publishing in other sectors that is not specifi-cally quantified in their reporting. Wolters Kluwer's1997 annual report includes a section on "Developmentsin the Markets" that states, "We are furthermore facingsustained competition from 'non-publishers' in the fieldof information provision." There is no elaboration on thisinteresting point that seems to have significance especial-ly for the Tax/Legal and Scientific/Medical segments.

Since indications are that scholarly publishing isamong the most profitable segments for these compa-nies, a very conservative assumption is that their overallnet incomes derive from scholarly publishing in propor-tion to their sales or operating margins in that segment.Because the capital investment for scholarly publishingis relatively low, it is also conservative to estimate thatequity is invested in the scholarly publishing segmentsin proportion to the sales that derive from those seg-ments. By recalculating scholarly publishing sales,assuming the periodical publishing industry median of5.0% net margin or an industry median of 18.8% returnon equity, it is possible to calculate the savings or coststo scholarly publishing customers of these companies'profits beyond or below industry medians. These hypo-thetical savings and costs to scholarly publishing cus-tomers are shown in Table 3. (Complete tables fromwhich these calculations were derived are available inthe online version of this article <http:/ /www.arl.org/newsltr/200/wyly.html>.) Where segment data is notavailable, the conservative assumptions outlined aboveare used so that these calculations are based on thescholarly publishing sales figures for Wiley at 47% oftotal sales, Plenum at 63%, Wolters Kluwer at 14%, andReed Elsevier at 17%; and the scholarly publishing oper-ating and net incomes for Wiley at 47% of the totals,Plenum at 63%, Wolters Kluwer at 13%, and Reed

A R L 2 0 0 O C T O B E R 1 9 9 8 S

Elsevier at 26%. Reed Elsevier dominates these figuresin sheer costs to customers because of its size; therefore,to make clear the proportional savings within each com-pany's customer base, the hypothetical percentage sav-ings on scholarly sales is also shown for each company.By these proportional measures, Reed Elsevier stillclearly dominates these measures of profits beyond theperiodical publishing industry medians, and the ReedElsevier profits are on such large sales that they raisethe overall effect when the totals of all four companiesare combined.

It is noteworthy that Wiley's profits in scholarlypublishing are essentially similar to the periodicalpublishing industry overall. Reed Elsevier in particularis probably profiting from the structural differencesbetween the market for scholarly publishing versusthe market for periodical publications more generally.Plenum's relatively inefficient use of capital is illustrat-ed again here by the dramatic contrast between thesavings as calculated on industry-standard net marginversus industry-standard return on equity. WoltersKluwer is considerably more profitable in its scholarlypublishing businesses than the periodical publishingindustry median ratios. However, Plenum's highmargins and inefficient use of high margins and ReedElsevier's high margins and high return on equitystand out as worthy of our attention as customers.

DiscussionThe financial performance of vendors is of interest to(potential) customers for several reasons. On the posi-tive side, a financially sound company can be more safe-ly relied upon if the intention is to build a long-termrelationship. The profitability of a vendor may alsoindicate the superiority of its products and services or abetter understanding of the market served. If a compet-itive market is operating, the very definition of a freemarket means that successful companies succeed for

these reasons. Therefore, the financial performance ofa vendor can help to guide purchasing decisions byproviding information about the collective judgmentsof the marketplace of customers.

However, if a competitive market is not operating,the finances of vendors can show the absence of compe-tition. More importantly, the finances of the vendorswill provide some measure of the economic conse-quences for the customers of the lack of a competitivemarket, as in Table 3. When these consequences areclear, customers may come to understand that theymust seek to change the situation, unless the costs ofbeing a buyer in a non-competitive market are relativelysmall in terms of the overall enterprises of the customer.

The costs of acquiring scholarly information are rel-atively substantial in research libraries, and even interms of entire universities' budgets. If we find our-selves being customers in a non-competitive scholarlycommunication market that is too expensive for our par-ticipation in the breadth and depth needed for a healthyscholarly system, what is to be done? When regulationis as unlikely as it is in today's current political environ-ment, and when the product or service is as vital asscholarly information is to research communities, thenthe customers can only pursue innovations that will endthe reliance on buying in a non-competitive market.

SPARC (the Scholarly Publishing & AcademicResources Coalition <http://www.arl.org/spardindex.html>) seeks to create a competitive environmentfor scholarly communication by just such innovationboth within and beyond the current scholarly publish-ing system.

In the short term, this financial information might beuseful in negotiations or debates with commercial pub-lishers. However, the fundamental problem of a lack ofcompetition cannot be addressed through negotiations.Ultimately, research institutions must help to create a

TABLE 3: HYPOTHETICAL SAVINGS OR (COSTS)TO CUSTOMERS BASED ON INDUSTRY-STANDARD PROFITS

(Considering only Selected Scholarly Publishing Operations of Each Company)

Savings (Costs)Relative to

5.0% Net Margin

Wolters Kluwer $19,145,000

Reed Elsevier $212,045,000

J. Wiley & Sons ($592,000)

Plenum Publishing $6,421,000

Total $237,019,000

% Savings(Costs) fromActual Net

Margin

Savings (Costs)Relative to18.8% ROE

% Sayings(Costs) fromActual ROE

5.3% $19,098,000 5.3%

22.3% $146,286,000 15.4%

(0.3%) ($865,000) (0.4%)

19.4% $570,000 1.7%

15.3% $165,089,000 10.7%

811 A R L 2 0 0 O C T O B E R 1 9 9 8

Continued

12' A R L 2 0 0 OCTOBER 1 9 9 8

system of scholarly communication that is either strong-ly competitive and commercially based or a system thatis not predicated upon the generation of profits from thecommunication process itself. This summary of pub-lisher finances illustrates in an inexact way the premiumpaid for relying on the current inherently non-competitiveeconomic model of commercial scholarly publishing.The lack of competition is fairly obvious in the financesof the companies, but the lack of competition can beconfirmed by an analysis of how all the participantsauthors, readers, and pub-lishersoperate within thecommercial scholarly pub-lishing process.

Free markets encourageinnovation in the deploy-ment of resources preciselyby tolerating extraordinaryprofits for new enterpriseswhen risks are high andbefore a competitive marketis established. Thus, com-panies regularly makeclaims to investors aboutplaying in markets wherecompetition is sparse ornon-existent for themoment due to the innovations of the company.However, sustained non-competitive markets in amature industry are generally an indication of a struc-tural problem in the market that prevents the entry ofcompetitors over time.

In the case of scholarly publishing, the majorimpediment to competition is that all of the direct incen-tives for authors lead them to publish in well knownchannels to reach the largest number of readers withoutregard to the cost of those channels to the readers.However, the reader is not free to choose new channelsbecause those channels are typically not filled with thesame content early in the development of a publication.This system of incentives makes it very difficult todevelop competitive communications channels. Thecustomer, i.e., the reader in this economic model, isdisempowered within the system because the productis "must-have" and "single source."

Reed Elsevier comes close to stating this inElsevier's 1997 filing with the U.S. Securities andExchange Commission (20-F), when it notes forinvestors that the company is increasingly "concentrat-ing on high value-added areas of 'must-have' informa-tion and significantly reducing its exposure to the con-sumer markets." In the consumer publishing market,the publisher produces the content and other publisherscan compete by producing similar content themselves.In the scholarly market, the content is produced by

authors external to the publisher, and authors are reluc-tant to move to another publishing channel for any rea-son other than broader awareness or prestige, which areinherently difficult to produce early in the developmentof a channel. This is a very positive market situation fora publisher because it avoids competition so long as theeconomic model is not disrupted by innovation. New pub-lishing channels that include technological or structural-ly innovative features, reach broad audiences, and offerprestige early in the life-cycle might lure authors away

from established channels.But perhaps even moreimportantly, we must keepin mind that the innovationmust ultimately disrupt thenon-competitive economicmodel at work in commer-cial publishing. Improvingscholarly publishing func-tionality with technologicalinnovation will not in andof itself introduce competi-tion. The innovation tochange the economicmodel must focus ondeveloping a structure toalign authors' and readers'

shared interest in broad dissemination. This is preciselythe challenge to the customer community for scholarlycommunication. We must innovate to produce competi-tive communication systems over time or continue suf-fering under a system of shrinking access due to the lackof competition in the present system.

Universities have excellent reasons for pursuingsuch innovations. The most important reason is not thatthe current system is too expensive and cannot handle asteadily increasing volume of scholarly communication.Both of these are serious problems, and even minimiz-ing all non-beneficial "publishing for tenure" is only ashort-term mitigating approach in the face of increasingcollective knowledge. However, overcoming theseproblems of cost and volume has a much more impor-tant benefit for universities than any potential directcost savings on library acquisitions. The most impor-tant benefit of a changed economic model for scholarlycommunication falls to universities as producers ratherthan as consumers: scholarly publishing makes publiclyknown the value of research of all kinds. Only by mak-ing that research known can it be valued and financiallysupported by the various supporters of scholarly enter-prise, ranging from corporations to foundations to legis-lators to other scholars. The worst disservice to theresearch community by the current system of commer-cial scholarly publishing is that it dooms scholarlyresearch to reach a shrinking audience as commercial

The worst disservice to the researchcommunity by the current system of

commercial scholarly publishing is that itdooms scholarly research to reach a shrinkingaudience as commercial publishers profit fromthe artificial scarcity enforced by high prices.The greatest rewards for universities in creating anew system of scholarly communication are in the

potential for increased support based onincreased awareness of the intellectual

products of the universities.

82

publishers profit from the artificial scarcity enforced byhigh prices. The greatest rewards for universities in creatinga new system of scholarly communication are in the potentialfor increased support based on increased awareness of theintellectual products of the universities. If any point bearsspecial emphasis in our thinking about the value ofinnovations in electronic scholarly communication,the value of increasing readership should.

Ease of access to quality information is the funda-mental goal of both authors and readers"producersand consumers," to put it simplistically. When aresearcher takes on the maze of literature databases,the frustration of shrinking local library collections,the long-delayed document delivery services, and theshocking copyright fees, it rapidly becomes clear thatthere has to be a better way, given that the author andthe reader share the desire to communicate. Readersand authors are already finding ways around the frus-tration imposed by the current scholarly publishing sys-tem. SPARC and other initiatives give us the opportuni-ty to facilitate and institutionalize that process.

Finally, it should be noted that this article is notintended as an indictment of any particular commercialpublisher or all commercial publishers as a group.Indeed, we are finding many for-profit publishers areadopting business philosophies and practices that mayultimately contribute to a more competitive and viablemarketplace for scholarly communication.

/V

Endnote1According to Reed Elsevier's annual report, the operatingmargin of the Scientific segment ran at 40.28% (1997),41.77% (1996), and 39.66% (1995) as a percentage of sales.As a point of comparison, Microsoft's operating incomeas a percentage of sales was 45.17% for 1997, 35.50% in 1996,and 34.33% in 1995.

© 1998 Brendan J. Wyly. The author grants blanketpermission to reprint this article for educational use aslong as the author and source are acknowledged. Forcommercial use, a reprint request should be sent to theauthor <bjw4@cornelLedu>.

8 3

COMPARING VALUE AND ESTIMATEDREVENUE OF SCITECH JOURNALSby Stanley J. Wilder, Assistant Dean for Technicaland Financial Services, LSLI Libraries

Aca.demic libraries currently devote about 74% oftheir chemistry journal budgets to titles pro-duced by commercial publishers. Relative to

U.S. association titles, however, commercial titles aregenerally much lower in value to the academic scientiststhey serve. The imbalance between the low value ofcommercial publications and the significant revenuethey generate from academic library subscriptions is atthe heart of the current crisis in scientific/technical (ST)journal costs.

The StudyThese claims are the result of a study of three measuresof value (faculty ratings, Science Citation Index [SCI] cita-tion counts, and library use) and estimated revenue fora core group of chemistry titles. The list of chemistrytitles, the measures of faculty ratings, and the SCI cita-tion counts were derived from a dataset originallydeveloped for Stephen J. Bensman's (1996) "TheStructure of the Library Market for Scientific Journals:The Case of Chemistry." The primary component ofthis set was faculty survey data collected by theLouisiana State University (LSU) Libraries for its 1993Serials Redesign Pilot Project. The LSU surVey askedfaculty in the Chemistry Department to name thosetitles necessary to support curriculum and research,and to indicate whether each title was needed on sub-scription or if access through a subsidized, unmediateddocument delivery service would suffice (Hamaker1994). The second dataset was provided by Tina E.Chrzastowski of the University of Illinois' ChemistryLibrary. Chrzastowski had collected library use data fora study of chemistry journals she conducted at Illinois.She defined use as the number of times an item wasreshelved or circulated. The two datasets were com-bined using only those titles from the Illinois use studythat were also included in the LSU data. Estimates ofrevenue generated by the titles were then added toallow for the exploration of the relationship between theallocation of resources on a national level to chemistrytitles and these three measures of value.

Underlying this study of chemistry journals are thefindings of recent research conducted by Bensman andWilder. The Bensman/Wilder analysis,"Scientific/Technical Serials Holdings Optimization inan Inefficient Market: A LSU Serials Redesign ProjectExercise" (July 1998), establishes that throughout eachof the ST journal literatures, faculty value highly a rela-tively few titles, most of which are U.S. association pub-lications. They also establish that a relatively few titles

ARL 200 OCTOBER 1998 13

Continued

account for a substantial portion of the cost of sub-scribing to these literatures. Most of these titles arecommercial publications.

Bensman and Wilder use three measures of valuein their analysis of ST literature: faculty ratings, SCIcitation counts, and use. One of the surprising resultsof their analysis is that, once appropriate subjectgroupings have been established, each of these mea-sures of journal value produce virtually equivalentresults in terms of journal rankings. In particular,equivalency among these value measures is evidencethat researchers in each ST discipline operate withina similar system. This system is marked by broadconsensus on what is important research, whichacademic programs tend to produce it, and which jour-nals publish it. This consensus is so pervasive that thesame journals rise to the top regardless of which of thethree value measures one uses. Given that the basicrelationships between value, cost, and publisher typewere found to be identical throughout the ST disci-plines, the results of the present study of chemistry

journals can be said to be illustrative of general patternsin other ST journal literatures.

Measuring RevenueThere is no way of knowing with certainty the amountof subscription revenue a journal generates, but a rea-sonable estimate can be obtained by multiplying itsprice by the number of libraries holding it as reflectedin the OCLC database. This approach misses subscrip-tions held by individuals or by institutions that do notplace their holdings in OCLC, and it may overestimatethe number of subscriptions to titles that have beencanceled by libraries but are still listed in the OCLCholdings statement.

Nonetheless, even publishers with substantialindividual membership income rely primarily onlibrary subscriptions for their revenue. For instance,Lorrin Garson (1997) of the American ChemicalSociety (ACS) recently stated that 90% of ACS' sub-scription revenue comes from institutions. Given thatan overwhelming majority of journal revenue stems

TABLE 1: REVENUE ESTIMATES WITH FACULTY SCORE AND TOTAL CITATION*

Numberof Titles

EstimatedRevenue

LSU FacultyScore

TotalCitation

U.S. Association 29 19% $13,165,270 18% 5,502 40% 872,141

U.S. Commercial 38 25% $11,250,297 16% 2,038 15% 248,589

Foreign Association 5 3% $5,355,872 7% 501 4% 87,257

Foreign Commercial 81 53% $42,522,015 59% 5,862 42% 727,026

Combined Commercial 119 78% $53,772,312 74% 7,900 57% 975,615

Total 153 100% $72,293,454 100% 13,903 100% 1,935,013

*Figures in table are rounded.

TABLE 2: REVENUE ESTIMATES WITH ILLINOIS USE*

Number of Titles

U.S. Association 25 21%

U.S. Commercial 27 23%

Foreign Association 5 4%

Foreign Commercial 63 53%

Combined Commercial 90 75%

Total 120 100%

Estimated Revenue % Illinois Library Use

$12,798,017 20% 36,004

$9,776,615 15% 4,724

$5,355,872 8% 4,700

$36,728,964 57% 24,644

$46,505,579 72% 29,368

$64,659,468 100% 70,072

*Figures in table are rounded.

%

45%

13%

5%

38%

50%

100%

51%

7%

7%

35%

42%

100%

A R L 2 0 0 OCTOBER 1998 8 4

from sales to libraries, and that a large number of U.S.academic libraries participate in OCLC, countingOCLC holdings is sure to capture a substantial portionof both subscriptions and revenue from U.S. academiclibraries.

Publisher TypesBensman and Wilder separated the journal titles intofour publisher types: U.S. commercial, U.S. association,foreign commercial, and foreign association. Both theLSU and Illinois datasets under-represent commercialtitles and over-represent U.S. association titles inrespect to the universe of published research in chem-istry. Many commercial titles were excluded from thesets either because they were not selected by LSU facul-ty or because they are not indexed in the print versionof the Science Citation Index (SCI). Both cases are anindication of low value to academic users.

This under-representation of commercial titles mayresult in an understatement of the actual value of theassociation titles.

TABLE 3: REVENUE LEVERAGE*

LSUFacultyScore

TotalCitation

IllinoisLibrary

Use

U.S. Association 2.17 2.47 2.60

U.S. Commercial 0.94 0.83 0.45

Foreign Association 0.49 0.61 0.81

Foreign Commercial 0.72 0.64 0.62

Combined Commercial 0.76 0.68 0.58

*Figures in table are rounded.

TABLE 4: TITLE LEVERAGE*

LSUFacultyScore

TotalCitation

IllinoisLibrary

Use

U.S. Association 2.09 2.38 2.47

U.S. Commercial 0.59 0.52 0.30

Foreign Association 1.10 1.38 1.61

Foreign Commercial 0.80 0.71 0.67

Combined Commercial 0.73 0.65 0.56

*Figures in table are rounded.

85

ResultsRevenue SkewWithin each publisher type, a small number of titlesaccounts for a significant proportion of the revenue.A look at the top five revenue-generating titles in eachcategory, however, reveals wide variation. Of the 29chemistry titles from U.S. associations, the top fivejournals are 17% of that set's titles, and 51% of the totalrevenue. For U.S. commercials, the top five are 13% ofthe 38 titles, and 38% of revenue. For foreign commer-cials, the top five of the 81 journals are only 6% of thetitles, but 33% of revenue.1 This concentration ofrevenue in the foreign commercial set is particularlynoteworthy, indicating that a very small core offoreign commercial titles subsidizes a long list oflower revenue titles in this set.

Relationship of Value to Revenue and TitlesTable 1 presents the revenue and value measure data(faculty ratings and SCI citation counts) for chemistryfrom the LSU data set. U.S. association publishersproduce value in excess of the revenue they receive:they account for 19% of titles and 18% of revenue, but40% of faculty ratings and 45% of total citations.Commercial publications, on the other hand, receivemore revenue than the value they produce: theyaccount for 78% of titles and 74% of total revenue, butonly 57% of faculty ratings and 50% of total citations.

The Illinois dataset on library use produces similarresults, with disproportionately high value among theU.S. association publishers, and disproportionatelylow value among commercial publishers (Table 2).

One might conclude from these data that commer-cial publications supply over half of faculty value asreflected in faculty ratings (57%). However, this con-clusion does not account for the much larger numberof commercial titles present in both data sets. To cor-rect for this imbalance, Bensman/Wilder created aconcept called leverage, a ratio that normalizes thevalues. In the present analysis, two types of leverageare considered: revenue and title leverage.

Revenue leverage expresses the value of journals inrelation to the revenue they generate, and is calculatedby dividing the percentage of each value measure bythe corresponding percent of total revenue (Table 3).For example, the percentage of faculty ratings for U.S.association titles (40%) is divided by the percentage oftotal revenue for those same titles (18%), resulting in arevenue leverage of 2.17 (variations due to rounding).Using the revenue leverage scores in Table 3, one con-cludes that U.S. association journals have 2.9 times thevalue of commercial publications (U.S. and foreigncombined) as measured by faculty ratings (2.17 dividedby .76), 3.6 times the value as measured by total citations

A R L 2 0 0 O C T O B E R 1 9 9 8

Continued

(2.47 divided by .68), and 4.5 times the value asmeasured by library use (2.60 divided by .58).

Title leverage expresses value obtained per title, andis calculated by dividing the percentage of each valuemeasure by the correspondin§ percent of total titles foreach publisher type (Table 4). Comparing title leverageas for revenue leverage above, one concludes that U.S.association journals have 2.9 times the value as measuredby faculty ratings, 3.7 times the value as measured bytotal citations, and 4.4 times the value as measured bylibrary use. The revenue and title leverage measuresreveal a more accurate picture of the relationshipbetween publisher types: commercial publicationsmanage to contribute over half of total value asmeasured by faculty ratings only because thereare so many more of them.

Revenue per TitleFor the set of chemistry titles examined, there is nostatistically significant difference between the averagerevenue per title generated by commercial publishersand that generated by U.S. association publishers fromU.S. academic subscribers. How can this be, given theunusually high cost of commercial publications? Oneexplanation lies in the concept of "first copy costs,"expenses attributable to the preparation of a journal forprinting, including peer review, technical editing, andmarketing, for example.

While estimates of first copy costs vary from60-80%,3 first copy costs are an important considera-tion for all publishers. When first copy costs must becovered over a smaller subscription base, a higher priceis required. On average, the commercial publications inthe chemistry dataset have only half of the subscriptionbase of association journals as measured by OCLC hold-ings, but charge about 80% more (Bensman 1996). Inother words, the commercial publishers' high pricesmake up for smaller subscription bases, and their risingprices may be in part a function of cancellations. If so,commercial publishers are caught in the unenviableposition of charging higher prices for products of lowervalue to a declining subscriber base.

ConclusionWhatever the components of commercial ST journalpricing, the value to academic scientists is small relativeto the revenue they receive. Whereas commercial jour-nals in chemistry account for 74% of the revenue gener-ated by ST publications, on a per-title basis they con-tribute only between 22-35% of the value of their U.S.association counterparts. This pattern detailed in chem-istry has been demonstrated by Bensman/Wilder toexist in the ST literatures in general.

Addressing the imbalance between value and rev-enue will require a change of philosophy for academic

libraries that currently pursue comprehensiveness inregard to ST journal literatures. It is the nature ofprogress in ST disciplines to require a high degree ofconsensus among researchers, and one reflection of thisconsensus is the relatively small number of ST journalsdeemed important by researchers themselves. Use ofjournal literatures within the conscribed sphere of high-value titles is so heavy that subscription is easily themost cost-effective form of access. For titles outside thissphere, use is so infrequent that document deliveryfrom commercial document suppliers is currently muchless expensive. This is true for large and small institu-tions alike, meaning that cooperative collection develop-ment in ST journal literatures is not cost-effective.

By placing value at the heart of their ST collectiondevelopment philosophies, academic libraries can freethemselves from the high-cost, low-value commercialtitles that currently consume such a large proportion ofST journal budgets. The resulting re-allocation ofresources will reverse the present imbalance betweenvalue created and revenue received by funding onlyhigh-value titles, and relegating lower value titles tomore cost-effective on-demand acquisition strategies.

Endnotes1Foreign association journals are not listed here because onlyfive of them were in the LSU data.

2In considering the results presented in Table 4, it should benoted that the high title leverage of foreign association jour-nals is partially due to the necessity of combining the fivebibliographically separate sections of the British Journal of theChemical Society into a single entity.

3 Lorrin Garson, in his presentation at the National Meetingof the American Chemical Society (8 September 1997),"Economics of Scientific Publishing," stated that first copycosts for the ACS are 84% of the journal publishing expens-es. John Cox of Carfax Publishing estimated in his paper atthe 17th annual Charleston Conference (November 1997)that first copy costs of scholarly journals range from 60-70%.

References

Bensman, S. J. (1996). "The Structure of the Library Marketfor Scientific Journals: The Case of Chemistry." LibranyResources and Technical Services 40: 145-70.

Bensman, S. J., and S. J. Wilder (1998). "Scientific/TechnicalSerials Holdings Optimization in an Inefficient Market:A LSU Serials Redesign Project Exercise." LibraryResources and Technical Services 42.

Hamaker, C. A. (1994). "Re-Designing Serials Collections."Journal of Library Administration 20(1): 37-47.

©1998 Stanley J. Wilder. The author grants blanketpermission to reprint this article for educational use as longas the author and source are acknowledged. For commercialuse, a reprint request should be sent to the author<swilder@lsu.edu>.

A R L 2 0 0 O C T O B E R 1 9 9 8 86

0 t.

AT THE SPEED OF THOUGHT:1PURSUING NON-COMMERCIALALTERNATIVES TO SCHOLARLYCOMMUNICATIONby Mike Sosteric, Assistant Professor, Centre for Globaland Social Analysis, Athabasca University, and Director,International Consortium for Alternative AcademicPublication (ICAAP)

Scholars, information specialists, and academiclibraries have warned of a crisis in scholarly com-

munication for years (Sosteric 1996). For a time, itwas hoped that electronic publication would bring somemuch needed relief by reducing the cost of distributionof the primary journal literature. However, despitehopeful statements in the early years, it now appearsincreasingly unlikely that electronic publication willbring relief. Some commercial publishers, rather thanseeing electronic publishing as an opportunity to reducecosts and bring needed relief to the academic librariesthat they serve, have chosen to exploit the opportunitiesfor increasing profit presented by the new technologies.From our 1998 perspective, the 1996 vision of a futurecommunication system where scholarly skywriting isconducted in a low-cost, collectivist manner by the schol-ars themselves now seems largely a hopeless fantasy.

A recent statement issued by The InternationalCoalition of Library Consortia (ICOLC, Statement) con-firms our worst fears. Using their growing monopolyposition and control over the scholarly communicationsystem, commercial publishing houses are forcinglibraries to purchase both paper and electronic versionsof their journals at rates that are already higher than thestandard print cost and at rates that the ICOLC fearswill eventually add as much as 40% or more to the costof scholarly material in journals. This is certainly notthe dream of independent scholarly publishing; it isthe nightmare of unregulated monopoly control.

In a press release introducing the statement, theICOLC notes:

The explosion in electronic licensing, the wide variance inpublisher practices, rapidly escalating prices, and a con-cern about the reduction in the number of independentscholarly information providers all served as the impetusfor the statement. The Statement calls for developing mul-tiple pricing models, separating charges for electroniclicenses from those of paper subscriptions, and loweringthe cost for the electronic information below that of printsubscriptions. ICOLC expresses its concern over thegrowing practice of publishers that levy initial surchargeson electronic information, which is compounded by signifi-cant multi-year inflation surcharges and prohibitionsagainst libraries canceling print versions of journal titles.As a result, while libraries may receive access to a larger

87

array of titles by paying the 'print price plus electronicsubscription cost plus inflation,' the total base price forelectronic access over the print subscription could increaseby 40% or more within as little as three or four years.A similar attempt to consolidate the strength of

libraries and perhaps win support from the scholarlycommunity took place in Europe in late 1997. A coali-tion of 15 Dutch scientific research libraries came togeth-er to express their concerns over an announcement byReed Elsevier of its intent to merge with WoltersKluwer (both major scientific publishers) and to pursue"a strategy of increasing its focus on 'must-have' infor-mation." Seeing the anti-competitive nature of theproposed merger, the Dutch librarians adopted a setof principles aimed at bolstering their position innegotiations with publishers and reducing the possibleimpact of skyrocketing electronic journal prices.'

Unfortunately, in the long-run, libraries are virtuallypowerless by themselves to offset the practices of commer-cial publishing houses. When it comes right down to it,an academic library simply must have the key journaltitles. They serve an academic market, after all, and it isthis market that largely determines which journalsshould be made available. If the academic market con-tinues to demand the high-prestige titles, and if we, thefaculty authors, continue to submit our work to thesejournals, there is little libraries will be able to do but copewith our needs and purchase the high price journals.

Barriers to Independent Scholarly PublishingSo what is to be done? At one point, scholars like SteveHamad, Andrew Odylzko, and Bernard Hibbits madecalls for a new future of independent scholarly publica-tion outside of the commercial mainstream. In the con-text of the crises in the scholarly communication system,proselytizing a future for independent publication wasa laudable project. However, the vision of independentcommunication has not been realized primarily becausescholars have been reluctant to take up the call to inde-pendent publication. Despite the powerful potential ofinformation technologies to wrest control of the scholar-ly communication system from the big commercial pub-lishing houses, relatively few individual scholars haveinitiated journal projects. It is perhaps this failure onthe part of the academic community to move quickly toindependent publication that has encouraged commer-cial houses to grow bolder in their demands.

But why the reluctance? What are the barriers thathave slowed the revolution in scholarly publicationwithin the academic environment? Probably one of themost significant barriers to independent scholarly publi-cation is simply the amount of work involved. Therecan be no doubt that editing journals takes a lot of work.And it takes even more work when you add typesettingand production duties to the job requirements. Add to

A R L 2 0 0 O C T O B E R 1 9 9 8 17

Continued

this the necessity of developing a high degree of com-puter literacy and you have a daunting barrier to par-ticipation in the electronic revolution.

A second barrier, closely related to the first, is thateditorial duties are not as highly regarded as publish-ing scholarly articles or books when it comes to tenureand advancement decisions. As a result, even thoseindividuals who may be willing to take on some of thework of publishing a journal in their field may bereluctant to do so because of the time it takes awayfrom other, more important, activities. In a decisionbetween putting time into an activity that moves youtowards tenure and editing an electronic journal, thecurrent reward structure virtually guarantees thatscholars will pass up opportunities to start publicationprojects.

Third, access to the system of scholarly communi-cation is largely closed to all but the most establishedscholars. Legitimate restrictions on access, which fol-low an almost guild-like structure, ensure that qualityis maintained. However, there is a significant contra-diction and irony in the way control is maintained.Mechanisms for apprenticing scholars in even the oldtechnologies seem undefined and there are no effortsto develop a craft structure for electronic publication.Because of this guild-like structure, and because of therelatively inadequate uptake mechanisms, most upand coming scholars would not have sufficient confi-dence to cross the line into the scholarly communica-tion system without proper guidance and encourage-ment (and we would probably be safe in assuming thatmost new scholars would be discouraged from initiat-ing journal projects). This is, of course, the rub, since itis young scholars who are most likely to take the risksassociated with initiating an electronic journal project.Perhaps it is this weakness in the current system ofquality control that is preventing a more rapid adop-tion of new technologies by scholars. And while somemay argue that it is necessary to move slowly, failureto utilize initiative at any level supports commercialpenetration and control almost by default.

Fourth, independent scholarly publication hasbeen largely seen as an anomaly and there are stillquestions about the quality of independent publicationon the Net. These concerns seem largely undeserved.Publications like the Electronic Journal of Sociology (EJS)<http:/ /www.sociology.org/> get,tens of thousandsof hits a month, draw papers from all levels of theacademy (including established authors), and haveachieved international recognition. And the EJS is notalone in this. So the basic problem seems to be con-vincing "the establishment" of the acceptability ofalternative publication. The general scholarly worldseems to have already reached its own conclusions.

Fifth, there hasn't been a deep sense of urgency tospur scholars into actionup until now. At this point,however, it is hard to deny that serious and damagingchanges are occurring in the academy. Higher educa-tion, once at least marginally dedicated to liberalideals, is slowly and inexorably being turned into avenue for profit generation. The long-term implica-tions have been explored elsewhere,' but it is clear thatnot only access, but also our own livelihoods are beingthreatened.

Sixth, while individual projects have abounded,there have been few attempts to develop the level oforganization required to create a broad, multi-discipli-nary coalition of educators, researchers, and adminis-trators dedicated to revolutionizing the scholarlycommunication system. Such an organization is sorelyrequired. Not only because a coalition of journaleditors and institutions will raise the standards andacceptability of "craft based" journal publication, butalso because such a coalition will be able to exert con-siderable market power in the highly monopolisticworld of academic journals.

International Consortium forAlternative Academic PublicationThis is a call for participation in a broad-based initia-tive dedicated to the development of an internationalalternative scholarly communication system outside ofthe commercial mainstream. Founded in Canada andhoused at Athabasca University, it is named theInternational Consortium for Alternative AcademicPublication <http:/ /www.icaap.org/>. TheConsortium's mission is to reduce the barriers toindependent scholarly publication by bringing togeth-er scholars and institutions from all countries and alldisciplines who are interested in bringing economichealth back to the scholarly communication system.ICAAP resolves to work towards overcoming blocksthat have prevented a shift away from commercialdependence and will pursue the following agenda:

Providing Editorial AssistanceIn order to increase the chance that individual scholarswill publish their own journals, ICAAP will devote thebulk of its revenue to hosting journals and providingfinal production assistance (HTML markup, copy-editing, etc.). ICAAP will also develop onlineresources and tutorials in order to help facilitate thedevelopment of alternative outlets for scholarly work.

Enhancing the Prestige of Editorial WorkIn order to encourage scholars to develop their ownjournal projects, editorial work will have to be suffi-ciently rewarded. We believe it is possible to raise thestatus of editorial work in the academy simply byexposing the difficulties, challenges, rewards, and

ARL 200 OCTOBER 1998 88

0

contributions that editors make to the progress ofscience. ICAAP will tirelessly proselytize this causeto the scholarly world.

Enhancing the Prestige ofIndependent Electronic PublicationThere should be no reason that commercial publishersare seen as any more prestigious than independentpublication efforts. After all, it is we who provide thecritical editorial and review work. Commercial housessimply provide production services and editorial stan-dards. To remedy this imbalance, ICAAP will developcross-disciplinary publication standards. In develop-ing these standards, ICAAP will focus on developinghigh-quality publication without the expensive frillsthat commercial publications most often use to justifyhigher prices.

Providing a Route For Apprenticing YoungScholars in the Craft of Scholarly CommunicationRecognizing that part of the problem is a lack ofappropriate training, ICAAP will, through its officialorgan The Craft <http: / / www.icaap.org /journals /hosted/TheCraft/>, provide a venue for apprenticingyoung scholars in the art and science of scholarly com-munication. It is hoped that eventually these scholarswill go on to initiate their own independent journalprojects outside of the commercial mainstream.

Providing Technical Expertise and Standards toMove Independent Efforts Toward4 GreaterStandardizationIt is imperative, if independent scholarly communica-tion is to advance beyond its current uncoordinatedstate, that technical standards be developed to ensureefforts are not duplicated or wasted and that the costto the system is not raised through this duplication ofeffort. ICAAP proposes to develop, in consultationwith stakeholders, technical standards for the produc-tion and distribution of scholarly information. Onesuch effort already underway is our web indexingrobot <http: / /www.icaap.org/JournalsIndex/>. Thisrobot is unique in the realm of web robots not onlybecause it targets only scholarly resources (thus elimi-nating the problems associated with the many less dis-cerning robots on the Web), but also because it is capa-ble of structuring the indexing of journals on familiarsearch fields.

ICAAP already has affiliations in Australia,Canada, Denmark, Mexico, and Russia. We are cur-rently seeking to expand our board of directors withinterested scholars from all disciplines, and we areseeking affiliations with stakeholder organizationswith an international scope. Discussions to explore anappropriate relationship with SPARC are currentlyunderway. We at ICAAP hope that you and your

organization will join us in our mission to revolutionizethe scholarly communication system.

EndnotesiThe author would like to acknowledge the work of SteveHamad as inspiration for the title of this article and theICAAP catch phrase, At the Speed of Thought, from his work,"Post-Gutenberg Galaxy: The Fourth Revolution in theMeans of Production of Knowledge," 1991, <ftp:/ /ftp.princeton.edu/pub/harnad/Harnad/HTML/harnad91.postgutenberg.html>.

2 These guidelines state that libraries that subscribe to a printversion of a journal should not have to pay more than anadditional 7.5% for electronic access to that same journal,and that libraries should not pay more than 80% of the printrate to subscribe exclusively to the electronic version."We've been talking about a 'journal crisis' for years," saysone of the Dutch librarians. "It looks like it's finally arrived.We're fed up." Quoted in the International Federation ofLibrary Associations mailing list <IFLA-L@INFOSERV.NLC-BNC.CA> from Terry Kuny <Terry.Kuny@xist.com>,in the article "Libraries Join Forces on Journal Prices"(Science 278.5343 [28 Nov 1997]: 1558).

3Mike Sosteric, ed. (1998), Pedagogy and the Colonization of theAcademy: Business and Consumerism in the Classroom, Spec.issue of Electronic journal of Sociology 3.3. Available at<http:/ /www.sociology.org/vol003.003/>.

ReferencesEgan, Timothy (1998). "Microsoft Loses some of its Swagger."

Globe and Mail Report on Business (19 Jan.): B10.

International Coalition of Library Consortia (ICOLC) (25 Mar.1998). Press Release. <http:/ /www.library.yale.edu/consortia /icolcpr.htm>.(25 Mar. 1998). Statement of Current Perspective and Preferred

Practices for the Selection and Purchase of ElectronicInformation. <http://www.library.yale.edu/consortia/statement.htmb.

Noble, David F. (1997). "Digital Diploma Mills: TheAutomation of Higher Education." <http://www.journet.com/twu/deplomamills.html>.

Reed-Elsevier (27 Oct. 1997). "Possible Divestment of IPCMagazines." Press release. <http:/ /www.reed-elsevier.com/Reed-Elsevier/newsreleases/nr32.asp>.

Sosteric, Mike (1996). "Electronic Journals: The GrandInformation Future?" The Electronic Journal of Sociology 2.<http:/ /www.sociology.org/vol002.002/Sosteric.article.1996.html>.

@ 1998 Mike Sosteric. The author grants blanket permissionto reprint this article for educational use as long as the authorand source are acknowledged. For commercial use, a reprintrequest should be sent to the author <mikes@athabascau.ca>.

89

ARL 200 OCTOBER 1998

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-296-2296 FAX 202-872-0884<http://www.arl.org/>Copyright: © 1998 by the Association of Research Libraries

ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

Executive Director Duane E. WebsterEditor G. Jaia Barrett, Deputy Executive DirectorCo-editor Lee Anne George, Program Planning OfficerCopy Manager Karen A. WetzelDesigner Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscri2tion; Nonmembers$50 peryear.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arLorg>.

TABLE OF CONTENTS AND HIGHLIGHTS

The Impact of Publisher Mergers onJournal Prices: A Preliminary Report .. .. 3

by Mark J. McCabe, Assistant Professorof Economics, Georgia Institute ofTechnology

" ...our results for journals sold by commercialpublishers indicate that prices are indeedpositively related to firm portfolio size, and thatmergers result in significant price increases."

Competition in Scholarly Publishing?What Publisher Profits Reveal 7

by Brendan J. Wyly, Johnson GraduateSchool of Management Library, CornellUniversity

"The worst disservice to the research communityby the current system of commercial scholarlypublishing is that it dooms scholarly researchto reach a shrinking audience as commercialpublishers profit from the artificial scarcityenforced by high prices."

Comparing Value and EstimatedRevenue of SciTech Journals 13

by Stanley J. Wilder, Assistant Deanfor Technical and Financial Services,LSU Libraries

"Whatever the components of commercial STjournal pricing, the value to the academicscientist that is created by such journals issmall relative to the revenue they receive."

At the Speed of Thought: PursuingNon-Commercial Alternatives toScholarly Communication 17

by Mike Sosteric, Assistant Professor,Centre for Global and Social Analysis,Athabasca University, and Director,International Consortium for AlternativeAcademic Publication (ICAAP)

"This is a call for participation in a broadbased initiative dedicated to the developmentof an international alternative scholarlycommunication system outside of thecommercial mainstream."

90

(f)

A

C.)

4.1

A BIMONTHLY NEWSLETTER OF RESEARCH LIBRARY ISSUES AND ACTIONS

THE DIGITAL MILLENNIUMCOPYRIGHT ACT

Following years of oftentimes contentious debate, onOctober 28 the President signed into law extremelycomplex legislation that updates selected aspects ofU.S. copyright law to address the digital environment.The changes to U.S. copyright law included in the DigitalMillennium Copyright Act (DMCA), PL 105-304, aresignificant and place new requirements on libraries andeducational institutions. As a consequence, there is aneed for broad awareness, indeed, deep understanding,about the major points of this law. In addition,implementation of the DMCA will require activeengagement within all educational institutions andlibraries so that they may comply with the newrequirements as well as take advantage of theopportunities presented in the law.

To contribute to our communities' understanding ofthe new law, ARL and its partners in the Shared LegalCapability are developing several papers that reviewand evaluate pertinent titles and/or provisions in theDMCA. Many of these pieces include next steps andrecommended actions for libraries and educationalinstitutions to take in order to benefit from selectedaspects of the law. This issue of the ARL newsletterincludes excerpts from two of these papers. The first is

bk) from an analysis of DMCA that highlights time-sensitiveactions that libraries and educational institutions must

E-1 address to take advantage of a new exemption for onlineservice providers. The second report is excerpted froma paper on distance education. Both papers wereprepared by Arnold Lutzker, Lutzker & Lutzker,legal representative for the Shared Legal Capability.Additional resources on other aspects of the DMCA,including preservation and fair use and access issues,are forthcoming.

rCD

9 1

HIGHLIGHTS OF NEW COPYRIGHTPROVISION ESTABLISHINGLIMITATION OF LIABILITY FORONLINE SERVICE PROVIDERSExecutive Summary

of the principal provisions of the DigitalMillennium Copyright Act ("DMCA") is aimitation on the potential money damages that

Online Service Providers ("OSPs"), including librariesand educational institutions, could face when they func-tion like a common carrier, allowing online users accessto copyrighted material placed there by someone else.Rather than confront huge financial claims if the thirdparty material infringes someone's copyright, OSPs canescape liability provided they comply with these newrules. Since the statute takes effect immediately, it is urgentthat all institutions act promptly to ensure that their systemsare in compliance with the terms. Note that the limitationdoes not apply to copyrighted material the OSP mayplace online itself, such as on its home page. Standardcopyright rules, including proper clearance and fair use,apply to that material.

The statute defines a "service provider" as an entitythat transmits, routes and connects users to online com-munications or provides online or network services,such as storing digital material, caching, or providinglocation tools (directories, hyperlinks, etc.). When deal-ing with copyrighted material available through its net-work, an OSP must be passive. It cannot place materialonline, modify content, store it longer than necessaryor know that it infringes someone else's copyright.Its systems must operate automatically and it cannotchoose recipients of transmissions. Finally, it mustnot directly profit from an infringement.

41 -41

Continued

The statute requires that in order for an OSP to qualify,it must implement several novel requirements immedi-ately. Institutions would be well advised to turn thesematters over to an established committee that managescopyright policies, or to create a new group for that pur-pose. In light of the fact that the statute calls for takingprompt action and making informed decisions, such abody could find itself involved in important policyquestions. Among the things an institution needs to doright away to qualify for the limitation are the following:

Designate an agent to receive statutory notices fromcopyright owners about infringements and to sendstatutory notices to affected subscribers.Advise the Copyright Office of the agent's name andaddress and post that information on the OSP's website.Develop and post a policy for termination of repeatoffenders and provide network users withinformation about copyright laws.Comply with "take down" and "put back" noticerequirements.Ensure that the system accommodates industry-standardtechnical measures used by owners to protect theirworks from unlawful access and copyrightinfringement.

A special exception has been created for public andnonprofit institutions of higher education, which allowsthem to qualify for the limitation even when the offend-ing user is a member of the faculty or a research graduatestudent. The law also gives immunity from third partyuser claims, provided there is a good faith compliancewith the statutory rules. It should also be borne in mindthat it is not necessary to actively monitor material on theInternet. The limitation requires an OSP to take action whenit has "actual knowledge" of an infringement (by facts broughtto its attention or by notice from the copyright owner), but itdoes not impose the burden on the OSP to monitor or discoverinfringing behavior.

In all, the limitation on liability gives library andeducational service providers a critical legal exemption at atime when their exposure to online copyright infringementis growing, not only because of the increased volume ofmaterial on the Internet, but also because of several adversecourt rulings. To make full and effective use of the limita-tion, each institution should take the time now to carefullyreview the details of the Act set forth in this memo.

$.0.000.,

See the Library of Congress OSP registration site<http://lcweb.loc.gov/copyright/onlinesp/> for the OSPInterim Regulations.

STUDY ON DISTANCE EDUCATIONAND DIGITAL TECHNOLOGIES

Although not originally part of the WIPO Implemen-tation Act debate, distance education became atopic of the negotiations because libraries and

educational interests raised it. The point was that it

would be inappropriate to expand legal protection forcommercial owners of digital works without remedyingsome of the legitimate legal concerns of librarians andeducators who use copyrighted works and technologiesin education. With a commitment from the Chairmanand Ranking Minority Members of the Senate JudiciaryCommittee, Senators Orin Hatch (R-Ut.) and PatrickLeahy (D-Vt.), the Senate included a Copyright Officestudy of distance education in the bill. The studyremained part of the Digital Millennium CopyrightAct (DMCA) as finally enacted.

The provision requires the Register of Copyrightsnot later than six (6) months of passage of DMCA(April 28, 1999) to provide Congress with its recom-mendations on how to promote distance educationthrough digital technologies, including interactivedigital networks, while maintaining an appropriatebalance between the rights of copyright owners andthe needs of users of copyrighted works.

The study requires the Register to consult withrepresentatives of copyright owners, nonprofit educa-tional institutions, libraries and archives and to focuson the following factors:

The need for an exemption from exclusive rights ofcopyright owners for distance education throughdigital networks.The categories of works to be included in anyexemption. The extent of appropriate quantitativelimitations on the portions of works that may be usedunder any distance education exemption.The parties who should be entitled to the benefits ofany exemption.The parties who should be eligible to receive distanceeducation materials under the exemption.Whether and what types of technological measurescan or should be employed to safeguard againstunauthorized access to and use or retention ofcopyrighted materials as a condition of eligibility.The extent to which the availability of licenses shouldbe considered.Other factors that the Register deems appropriate.

On November 16, 1998 the Copyright Office releaseda preliminary notice soliciting identification of interestedparties and a statement of their concerns by December 7,1998 (see <http: / /frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=1998_register&docid=frl6no98-111/>). The Copyright Office will also invite writtensubmissions over the next few months, and plans tomeet with interested parties and hold public meetings.

The papers excerpted above are available in their completeform on the ARL website, as are other reports about theDMCA and related copyright issues <http://www.arl.org/info/frn/copy/dmca.html>. ARL is also preparing a printcompilation of DMCA resources.

A R L 2 0 1 D E C E M B E R 1 9 9 8

Clifford Lynch, Executive Director

ACCESS MANAGEMENT FORNETWORKED INFORMATIONRESOURCES

IntroductionOver the past year the Coalition for NetworkedInformation (CNI) has been leading a broad-based exploration of access management and

authentication issues that are emerging as institutionsshare resources and license access to published materi-als through the Web. Scores of institutions and contentproviders participated in the discussion and con-tributed to the development of a lengthy white paperthat is now being cast into final form; both the draftfrom the spring of 1998 and, when completed, the finalversion, can be found at CNI's website <http:/ /www.cni.org/>. The interested reader is urged to consultthis white paper for a much more in-depth discussionof the issues summarized here.

This article is intended to briefly describe the cruxof the access management problem and the two majorarchitectural approaches that are now corning into useto address it. It will also devote considerable emphasisto important policy issues that emerge as part of theconsideration of approaches to access management;these issues are real, immanent, and potentially contro-versial. Research libraries have both an opportunityand an obligation to provide their patron communitieswith leadership in understanding the issues, and towork with these communities to formulate appropriatepolicies proactively and thoughtfully, rather than inresponse to a crisis or scandal. Many universities arenow implementing institutional authentication andaccess management systems to support specific appli-cation and security requirements; typically this work isbeing overseen by the institution's information tech-nologists and it is often motivated by policy impera-tives and priorities that are very different from those ofthe library. Because authentication and access man-agement represents a common infrastructure that willultimately need to serve many purposes, it is essentialthat libraries and information technology groupsestablish a dialog now to ensure that the systemswhich are ultimately deployed honor the completeset of campus community requirements.

This article does not discuss the technical detailsinvolved in authentication and access managementthey are complex and the state of commercially avail-able technology is continually changing. Suffice it tosay that, as of this writing (November 1998), there doesnot seem to be any simple, inexpensive, ready-to-deploy comprehensive solution for authentication and

93

access management; while there are many promisingcomponents available, management complexity, usertraining and acceptance, system integration, and costare still major issues. Organizational readiness withinthe content provider community is also a barrier to theuse of advanced technology-based approaches; contentsite operators are no more able than libraries to obtaincommercial turn-key solutions. But as libraries offer agrowing portfolio of network-based informationresources, they face immediate and pragmatic needsfor access management. In developing strategies, it isimportant to recognize that technical choices aboutaccess management have policy implications.

The Access Management Problem DefinedA library negotiates a license agreement on behalf ofits patron communityfor example, the faculty, staff,and students of a university. This license agreementgives members of the community the right to use somenetwork-based resourceperhaps a commerciallyoffered electronic encyclopedia or scholarly journal ata publisher website, or a specialized research databaseat another university that is part of a resource-sharingagreement. Users connect to the site using webbrowsers running on their personal workstations or onpublic workstations. As the site hosting the licensedresource processes these requests for web pages, itmust determine whether the requesting user is in fact amember of the appropriate user community. If so, heor she is given access; if not, access is refused.

Obviously, this basic scenario can be elaborateduponfor example, there is a need for finer-grainedsystems that can be used to allow only registrants in aspecific multi-institutional course access to electronicreserve materials provided by the library at one of theinstitutions. These more elaborate situations raiseadditional problems both in terms of scale and systemdesign and are useful to have in mind to understandthe extent of the requirements; however, we willconcentrate on the basic scenario here.

Access management needs to be routine and easyto implement; once a contract is signed, lengthy techni-cal negotiations between institution and content sup-plier should not be necessary before users can haveaccess. In a world of networked informationresources, access management needs to be a basicpart of the infrastructure, and must not become abarrier to institutional decisions to change or addresource providers.

While it may be a complex administrative questionto determine the membership of the patron communi-ty in a world of extension students, affiliated medicalprofessionals, visiting scholars, and others who may

A R L 2 0 1 D E C E M B E R 1 9 9 8 3

Continued

blur the boundaries, libraries have a great deal ofexperience in devising these policies. The problemhere is a technical one between two computers: howdoes the resource host machine actually determine thatthe user at the requesting workstation is really a mem-ber of the relevant community? In other words, givena policy for determining community membership, howis this policy implemented technically? In the physicalworld of circulating materials, the "technical" problemis dealt with through the presentation of a library cardindicating that the holder is indeed an authorized bor-rower; this library card is issued upon the presentationof policy-specified documents that identify the userand prove that he or she meets the policy criteria forauthorized community membership (a faculty ID, forexample, plus a picture ID).

Some Approaches That Don't Work WellHistorically, many computer systems employed userIDs and passwords. When a user ID was first defined,it would have certain rights and privileges associatedwith it. The user would then supply a password upondemand to demonstrate his or her rights to use a givenID. Leaving aside technical problems associated withthe use of passwords on an insecure network and thepossible technical remedies, there is a more basic archi-tectural issue. We are in a world where users may rou-tinely want access to many different network resourcescontrolled by many different organizations, such aspublishers; but with the ability to follow links on theWeb it may not even be clear to the user which pub-lisher owns what resource. Users will not be able toremember and manage a large number of differentuser IDs and passwords issued by different publishers;further, each publisher will need some cumbersomeprocedure to validate the user as a community mem-ber prior to issuing a user ID and initial password (forexample, involving mediation by local library staff).This does not scale up in a practical way to a worldfilled with electronic information resources.

The library might issue the user a single ID andpassword for all licensed external network resources,and then transmit lists of these IDs and passwords toeach content supplier so that the supplier can use thelist to validate users. However, there are architecturalproblems here, as well: a very large number of pub-lishers would need to be notified every time the list ofuser IDs and passwords changes, with inevitable time-liness and synchronization problems. Also, each pub-lisher takes on an enormous responsibility for protect-ing the security of the ID/password list; a securitybreach at any publisher will mean a security breach atevery publisher doing business with the library for

networked resource access, an unacceptable liability.Here the networked information scale-up means thattoo many independent parties must rely on each otherto maintain security, and that the cost of accuratelymaintaining a synchronized common-access manage-ment database will be very high (to say nothing of thestandards that would need to be put in place to makesuch a shared database a reality).

As a stopgap measure, many institutions are cur-rently using electronic "place"the user's source TPnetwork addressas a substitute for other methodsof demonstrating proof of community membership.If the user's connection request came from a networkthat belonged to the university, it is assumed to befrom a community member. Network ownership doesnot change too frequently, so maintaining a list ofvalid network numbers and making this available topublishers is a tractable administrative burden.Further, since the list of network numbers, unlike IDsand passwords, doesn't need to be kept secret, there islittle interdependence. This approach works well and,indeed, has the virtues of simplicity and transparencyto the user, as long as all users come to the resourceproviders through the campus or library network.However, in an era when many institutions are discon-tinuing dial-up modem pools in favor of commercialInternet service providers, and when new access tech-nologies to the home, such as cable television modemsfor Internet access, are beginning to deploy, a greatdeal of legitimate user access now takes place fromsources other than the campus or library network.With growing needs to support distance learners, part-time students who may do academic work from theirplace of business, and people who want to exploit the"anytime, anywhere" promise of networked informa-tion resources from their homes, limiting access bysource IP network address disenfranchises more usersevery day. For most universities, for example, it isclearly no longer acceptable to tell community mem-bers they can only access networked informationresources from on-campus workstations.

The Two Emerging Architecturesfor Access ManagementTwo general approaches are emerging to address theaccess management problem. The first is the use ofproxies. Here, an institution develops an internalauthentication system and uses it to validate useraccess to a special machine on the institutional net-work, called a proxy. Once a user is validated, he orshe can ask the proxy to contact an external resource;in some variations, the proxy mediates the user's entireinteraction with the external resource, while in other

ARL 201 DECEMBER 1998 9 4

(perhaps slightly less secure and user-transparent)variations the proxy drops out of the interaction aftermaking the introduction. From the resource operator'spoint of view, it is only necessary to know that theinstitution's proxy server can be trusted to pre-validateall users before contacting the external resource host.Proxies have a number of advantages: authentication isan internal matter to the institution and external serviceproviders need not be concerned with the details ofhow this is accomplished; at least in theory, the institu-tion has complete flexibility in deciding what resourcesa user can have access to through the proxy and when;and the proxy can act as acentral control point for theinstitution in access manage-ment. On the negative side,the proxy is a high-impactcentral point of vulnerabilityfor outages or capacity prob-lems (though, of course, it'spossible to have multipleproxy machines), and config-uration management in theproxy can become extremelycomplex and labor-intensive,particularly if not every valid

The simplest credentials-based approach would beto have the institution just issue the user a user ID andpassword for external resources, and to have externalresource providers validate the ID/password pair withthe institution (as opposed to having the institutiondistribute lists, as discussed earlier). This reduces, butdoesn't eliminate, the interdependence among externalresource providers in the maintenance of security;further, standards don't exist for such a validationprocess and no off-the-shelf software supports it.The industry is moving towards a technology basedon public key cryptographic certificates (X.509) in off-

the-shelf software, and, atleast in theory, this shouldwork well for personalmachines (but not forshared or public worksta-tions, which would have tobe handled some other way,such as IP source addresschecking, at least until amuch more complex hard-ware-based infrastructurebecomes widely deployed).X.509 removes the interde-pendent security issue

because credentials are computed for each use frominformation that is held only by the user and neverdirectly transferred to the content provider.

The problems with this approach include integra-tion with browsers, the mechanics of issuing and dis-tributing these electronic cryptographic certificates tousers, the cost and complexity of acquiring and operat-ing the infrastructure for managing and validatingpublic key certificates, and government regulation ofthe export and use of cryptography in various coun-tries, which causes problems in an increasingly inter-national world of scholarly resources. All of theseproblems with cryptographic certificates are slowlybut steadily getting better (except, perhaps, for thegovernment regulation issues), but, at least today,implementation of such an approach is an enormouschallenge. Finally, it is important tO note that X.509was really designed to support applications such aselectronic commerce, and there are some significantproblems in relation to the policy problems discussedin the next section.

Policy Issues in Authenticationand Access ManagementThe development of any access management strategyraises policy issues in areas such as privacy, account-ability, and the collection of management data. It is

Patron privacy has been such an

important value to libraries that theyhave, by and large, used a dual

technical/legal strategy to providetheir patrons with the strongest

possible protection.

proxy user has access to all resources. Further, proxiesdo not eliminate the need for an authentication system;they only isolate its scope to the proxy and the mem-bers of the institutional community.

The other approach is based on credentials. Thebasic idea here is that the institution issues each userthe electronic analog of a community ID card. Theuser, or the user's browser, presents these credentialsupon demand to any resource provider that requeststhem; the resource provider can then, through a fastelectronic transaction, validate the credentials with theissuing institution. The validation process shares withproxies the vulnerability to outages or capacity prob-lems on the part of institutional systems that verifycredentials, though these vulnerabilities are morecircumscribed. A compromise of the credential-verification system may be more serious thancompromise of a proxy: proxy compromise usuallymeans that unauthorized users get access to resourcesfor the period during which the proxy is compro-mised, while a compromise in the credentialingsystem may well mean that new credentials need to beissued to the entire authorized user community. Themajor practical difficulties with the credentials-basedapproaches, however, involve technical problems,standards, cost, and software integration.

95 A R L 2 0 1 D E C E M B E R 1 9 9 8

Continued

important to recognize that libraries must decidewhether to address these issues through legal means(that is, by negotiating contractual obligations on theresource supplier as part of the license agreement),through technical means (for example, by making itimpossible to collect personal data), or by a combina-tion of the two.

Library experience in other contexts offers someinsights. For example, libraries have aggressivelychampioned and defended the privacy of theirpatrons. They have done this through both legalmeansby developing privacy policies and by requir-ing a legal subpoena as a condition for divulgingrecordsand also by technical means by not keepinghistorical circulation data on an individual basis,which then limits the amount of information that theycan be compelled to disclose under any circumstances.Patron privacy has been such an important value tolibraries that they have, by and large, used a dualtechnical/legal strategy to provide their patronswith the strongest possible protection.

In the electronic environment, it is easy for a pub-lisher to track the use of content in great detailwhatmaterial is being viewed, for how long, and how often.Depending on how the access management system isstructured, particularly when credential-basedapproaches (and specifically, routine implementationsof X.509 certificates) are employed, the publisher maybe able to correlate this usage information to a specificindividual by name, to a long-term "pseudonymous"identity that the publisher can link to an institution butnot to an actual individual within that institution'scommunity, or simply to a transient and anonymousmember of the institution's user community. Clearly,the license contract between institution and publishercan speak to the collection, retention, use, and disclo-sure of such usage data on a policy basis, but librariesand patrons may find it desirable to limit the ability ofthe publisher to collect information by the design of theaccess management system on a technical basis, as well.

One important point needs to be made about userprivacy that underscores the need for contractual con-straints even in conjunction with an access manage-

, ment system that provides some level of anonymity:often users will make their actual identities known tocontent providers for other reasons, independent ofthe access management system, such as to take advan-tage of email-based current awareness services or per-sonalization options in a user interface. The accessmanagement system is not the only way in which pri-vacy can be compromised, or bargained away forincreased function or convenience.

A license agreement represents a commitment onthe part of the licensee institution to honor the termsof the license, and to educate members of its commu-nity about their obligations under the license. Thepublisher and library share a need for some level ofaccountability by community members: if a singleuser accesses publisher content hundreds of times aday from three continents, it's likely that somethingis wrong; perhaps that user doesn't understand his orher obligations, or perhaps credentials have beencompromised. There is a need for the publisher andthe library, acting together, to be able to investigatesuch situations effectively, and, if need be, to blockaccess by specific individuals who seem to be violat-ing the terms of a license agreement. But, in order todo this effectively, a publisher needs to be able to atleast provide an institution with enough information(such as a pseudonym) to permit the individual inquestion to be identified; note that this does not nec-essarily mean that the publisher can directly identifythe individual in question, but only that the publishercan provide the institution with enough informationto identify the user. The need for accountability con-tradicts, to some extent, the mandate to designanonymity into an access management system, andargues for a pseudonymous approach. This can beachieved with credential-based approaches (though itis not the standard model for X.509 certificates, forexample), but is more difficult with proxy-basedarchitectures.

Finally, there is the issue of management data.Electronic information resources promise librariesmuch more accurate and detailed data about whatcontent is actually being used and how oftenthough even at this level libraries may want to makecontractual stipulations to protect patron privacy; forinstance, it is not clear how many universities wouldbe comfortable having a list of all the articles read bymembers of their community (with frequency counts)posted on the Web every week. But greater problemsarise when libraries want to have these usage statis-tics, at whatever level of aggregation, demographical-ly facetedfor example, to drive internal cost alloca-tion processes within the library's institution. Thesimplest solutions are often to pass demographicattributes to the publisher along with identities orpseudonyms and to get the publisher to do the workof generating management data for the librarybutthis path can rapidly compromise the privacy ofpseudonymous users by making them more identifi-able and, if actual identities are used, it makes theprivacy problem even more acute by raising the

A R L 2 0 1 D E C E M B E R 1 9 9 8 9 6

stakes on the amount of information disclosed.In discussing this issue as part of the CNI authentica-tion work, it appears that many libraries, because ofthe privacy considerations, are backing away fromthe continuous collection of demographically facetedusage data, recognizing that they can always dospecific studies in order to obtain "snapshot"information, much as has been done with journalusage historically.

Demographically faceted usage data offers a goodillustration of the scaling issues that we face in themove to network-based electronic information. If alibrary were dealing with just a single publisher, itwould be reasonable to have the publisher returndetailed (transactional) usage data with pseudonymsto the library. The library could then look up thepseudonyms to obtain demographic data and summa-rize the transaction data into management informa-tion. Here privacy is protected by library policies anddoes not depend on the publisher, who knows onlypseudonyms. But this is intractable with hundreds ofpublishers, each supplying transaction data at differ-ent time intervals and in different formats (and evenbased on different models of the transactional eventsbeing tracked and logged). Without very detailedstandards, which don't exist today, and very rigorousadherence to such standards by the publisher commu-nity, libraries needing demographically faceted usagedata will be forced to export demographic informa-tion to each publisher.

Conclusion: Towards PracticalShort Term SolutionsCredential-based solutions require substantial workby publishers; at least a few of the larger publishershave a sophisticated understanding of the access man-agement issues and are, I believe, prepared to workwith libraries and universities to put credential-basedsolutions in place. However, many content providersoffer only IP source, network-based authenticationand have a limited understanding of the broaderissues. One of the great advantages of proxy-basedapproaches is that, from the content provider perspec-tive, they appear to be a simple extension of IP sourcenetwork authentication: they do not require any addi-tional work by the content provider and all of thecomplexity of an authentication system is hiddenwithin the institution's proxy servers. This suggeststo me that most institutions will be forced to continueto support a proxy-based approach for the foreseeablefuture, if they are to manage access to a wide range ofpublishers. Credential-based approaches may be use-ful in authenticating users to the proxy, and can also

be exposed to those content providers that understandand support them where it is useful, but user needs forconsistency of mechanism may well mandate that allaccess be managed by proxy except for specializedinter-institutional resource sharing arrangements (suchas electronic reserves).

Finally, it is worth noting that institutional authen-tication systems are not driven solely by demands toaccess network-based information resources; they areneeded as a means of managing access to institutionalresources and services, and, in a time when many net-works within higher education institutions are underconstant attack, they are also being viewed by someinstitutions as a means of improving security. Someinstitutions are now seeking to establish both policyand technical infrastructure that forbids unauthenticat-ed access to the institutional network and the servicesavailable on it. To the extent that these local authenti-cation mechanisms are exposed externally and reducethe privacy of library patrons in interacting with exter-nal resources, I believe that they require very carefulexamination and discussion. At many institutions, thefocus of the information technology managers is pri-marily on the management of local resources, whilethe focus of the library is on access to resourcesbeyond the boundaries of the institution, and it isessential that these two perspectives be balancedin the development of institutional infrastructureand policy.

@ 1998 by Clifford Lynch. This article will also be pub-lished in CAUSE/EFFECT, an EDUCAUSE publication.The author grants blanket permission to reprint this articlefor educational use as long as the author and source areacknowledged. For commercial use, a reprint request shouldbe sent to the author <clifford@cni.org>.

CNT's 1998-99 PROGRAM PLANCNI's 1998-99 Program Plan has recently been issued.Specific projects for the year are detailed under thebroad themes of:

Developing Networked Information ContentTransforming Organizations, Professions, andIndividualsBuilding Technology, Standards, andInfrastructure

The full text of the plan is available at: <http://www.cni.org/program/>.

A R L 2 0 1 D E C E M B E R 1 9 9 8

Richard K. Johnson, Enterprise Director, Scholarly Publishing and Academic Resources Coalition

ARIZONA BIOLOGIST/EDITORBREAKS WITH PUBLISHER TOESTABLISH LOW-COST ALTERNATIVE

Biologist Michael Rosenzweig is abandoning thethriving scholarly journal he founded a dozen yearsago because he believes the publisher has made it

so expensive that many libraries and colleagues no longercan afford it. He says price increases on his journal, aver-aging almost 19% annually, have harmed the scientificcommunity, the very group that supplies articles to thejournal for free.

With endorsement from SPARC (the ScholarlyPublishing & Academic Resource Coalition), the Universityof Arizona professor is staking a considerable amount of hisown money on independently launching aneW journal focusing on the conjunction ofevolution and ecology. Evolutionary EcologyResearch (EER), the comparatively low-pricedalternative to his original project, will beginpublication in January.

"I conceived the original journal as ascholarly resource. In addition to its basicscience value, its articles help to identify thelong-range consequences of environmentalchange for ecosystems. I wanted that infor-mation to be affordable so it would bebroadly available," said Rosenzweig, whofounded and serves in the University ofArizona's Department of Ecology andEvolutionary Biology in Tucson. "But soonafter the project was launched in 1987, ouroriginal publisher, Chapman and Hall, wasbought by International ThomsonCorporation. Then, this year, they sold thejournal to Wolters-Kluwer. At every changeof hands, the price increased significantly. Ithas become obvious over the years that thejournal's editors and publishers have incom-patible goals. We want to disseminate knowledge; theywant to maximize profits."

Because the skills, prestige, and connections of theeditor and editorial board are key to the success of a sci-entific journal, EER is well positioned to compete forreaders and manuscripts. Rosenzweig is an acknowl-edged force in his field and his entire editorial board istaking the plunge with him, having told Wolters-Kluwerthey will no longer edit the existing journal. Because oftheir reputation, EER already has attracted an amplepipeline of manuscripts for the first year of publication,with solid potential for more. Moreover, Rosenzweigsays his experience starting a successful academic journaland managing it for 12 years will also serve EER well.

While scholars commonly pursue commercial ventures,entrepreneurship is the last thing on Rosenzweig's mind.

He is publishing EER as a resource for essential scientificknowledge, not as a cash cow. Rosenzweig shares theopinion of many researchers and librarians around theworld that commercial publishers of scholarly literaturein recent years have become too profit-oriented, thuscrippling the authors' and editors' hunger to advanceknowledge.

The mechanics of publishing EER are simple toRosenzweig: he will outsource the journal's production,just as many large publishers do. But the initial chal-lenge lies in the publication's rapid market acceptance,especially when another journal covering the same sub-ject is already established. After exploring variousoptions for publishing EER, Rosenzweig turned toSPARC because its members share his sentiment that

the spiraling price of information threatensscientific discovery.

While Rosenzweig cares little about thenew journal's profitability, he recognizes itmust be financially viable. "I've run thenumbers every which way, and I'm confi-dent this will be a going concern," saidRosenzweig. "I'm prepared to lose moneyfor a whileit's the cost of starting a busi-ness. But over the longer haul, we've gotexcellent prospects." Whatever may come,he says he will stick to his promise of keep-ing the journal's price down by basing iton the actual production costs. EER websubscriptions will be offered to libraries in1999 for just $272one-third the currentsubscription price of Wolters-Kluwer'sprint publication. Moreover, individualsassociated with each subscribing institutioncan purchase personal print subscriptionsfor $33 in the U.S. and $41 elsewhereabout one-quarter of Wolters-Kluwer'sprice for individual subscriptions.

As is the case with so many scholarlyjournals, Rosenzweig nurtured his original project per-sonally in return for minimal compensation by his pub-lishers. He expects to do the same with EER, but sayshis real compensation will be the gratification thatimportant research is being made readily available to hiscolleagues.

"Scholars rarely publish for profit," saidRosenzweig. "If I'm not motivated by profit, thenthere's no sense in pricing this journal far above the costof production. But getting EER quickly recognizedwithin the marketplace indeed is essential. Our associa-tion with SPARC will go a long way toward making thata reality."

More information on Evolutionary Ecology Research isavailable on the World Wide Web at <http://www.evolutionary-ecology.com/>.

It has become obviousover the years that thejournal's editors and

publishers have incom-patible goals. We wantto disseminate knowl-

edge; they want tomaximize profits.

Michael Rosenzweig

A R L 2 0 1 D E C E M B E R 1 9 9 8

SPARC NOTESSPARC Notes is a new occasional feature of the ARLnewsletter that provides a rundown on activities of theScholarly Publishing and Academic Resources Coalition(SPARC) and its publishing partners.

Europeans Join Effort to ExpandCompetition in Journal PublishingEuropean chemists and libraries have joined NorthAmerican efforts to expand competition, drive downprices, and enhance timeliness in scientific journal publish-ing. SPARC and the Royal Society of Chemistry (RSC),based in the United Kingdom, have agreed to collaborateon a series of new high-quality, peer-reviewed electronicjournals that will be offered at far below the prices of com-peting commercially published journals.

PhysChemComni, the first product of the SPARC-RSCpartnership, provides rapid communication of articlesin physical chemistry. With a 1999 price tag of $353(200 Pounds Sterling) for site-wide institution-al online access, it competes head-to-headwith a commercial title priced at over $8,000.The RSC title is an all-electronic journal pro-viding double refereeing, an esteemed inter-national team of editors, searchable full-text,multi-media presentation features, and webavailability.

SPARC also announced it will be repre-sented in the UK and Ireland by the StandingConference of National & UniversityLibraries (SCONUL), which includes 129leading libraries in its membership, and hasestablished an affiliation with Denmark'sConference of Directors of ResearchLibraries, an organization of 12 major libraries. Thesealliances are the first expansion of the program beyondNorth America.

ACS Appoints Editor To HeadNew Organic Letters JournalThe world's largest scientific society, the AmericanChemical Society (ACS), has announced the appointmentof Amos B. Smith, III as Editor-in-Chief of Organic Letters,a new peer-reviewed chemistry journal slated to begin pub-lication in July 1999. The journal is the first of three to bepublished by ACS as part of a collaboration with SPARCaimed at developing lower-cost alternatives to high-pricedjournals and encouraging rapid dissemination of researchresults. The new journal will be priced at $2,300 per year,about one-quarter the price of its direct compefitor.

A leading researcher in synthetic organic chemistry,Smith is the Rhodes-Thompson Professor of Chemistry atthe University of Pennsylvania and served as chairman ofits Chemistry Department from 1992 to 1996. He also is amember of two of its interdisciplinary institutes: theLaboratory for Research on the Structure of Matter andthe Monell Chemical Senses Center.

"Organic Letters will offer rapid communication oforiginal, concise, and timely reports of significantresearch," said Smith. "The speed of access offered by theSociety's innovative web publishing features will makethis journal an essential tool in organic chemistry."

Smith named six associate editors to serve on thejournal. They are: Peter Beak and Scott Denmark, both ofthe University of Illinois, Urbana-Champaign; C. DalePoulter and Cynthia J. Burrows, both of the University ofUtah; Daniel H. Rich, University of Wisconsin, Madison;and Jeffrey D. Winlder, University of Pennsylvania.

SPARC Invites Expanded MembershipMembership in SPARC is now open beyond the ranks ofARL. A newly released SPARC membership plan invitesinternational support from academic and research institu-tions that share an interest in creating a more diverse mar-ketplace for scholarly communication by encouraging thedevelopment of high-quality, economical journals.

"Founding Member" status is available toqualifying members that join by December31, 1998. Several membership categoriesallow individual institutions to provide sup-port ranging from $1,500 to $5,000 annually.Consortium memberships also are available,utilizing a dues structure based on the aggre-gate library materials spending of participat-ing libraries.

SPARC is a new but already formidablecoalition of libraries that aims to create amore competitive scholarly communicationmarketplace. Its initial focus is on sciencejournals, where the problems of high andescalating prices are most serious. Created

as an initiative of ARL, SPARC membership has grownmore than 50% since June 1998 to include 122 memberlibraries and library consortia. In addition to SCONULand the Conference of Directors of Research Libraries(above), SPARC affiliate organizations include theAssociation of College & Research Libraries and theCanadian Association of Research Libraries.

The SPARC membership plan and related documentscan be viewed at the SPARC web page <http://www.arl.org/sparc/index.html>.

SCHOLARLYPUBLISHING AND

ACADEMIC RESOURCESCOALITION

More information on the organizations and productsdescribed above may be found at the following LIRLs on theWorld Wide Web:

Conference of Directors of Research Libraries:<http:/ /www.dlh.dk/dpb/fc/fc_eng.html>Royal Society of Chemistry: <http://www.rsc.org/>PhysChemComm: <http://www.rsc.org/physcc/>SCONIJL: <http://www.sconuLac.uk/>ACS Publications: <http://pubs.acs.org/>Organic Letters: <http://pubs.acs.org/journals-or1ef7/index.htm1>Evolutionary Ecology Research: <http://www.evolutionary-ecology.com/>

9 9 A R L 2 0 1 D E C E M B E R 1 9 9 8

Mary M. Case, Director

ELECTRONIC THESESAND DISSERTATIONS:THE INEVITABLE FUTUREIssue #7 of Transforming Libraries' summarizes thecurrent state of electronic theses and dissertations and reportson current developments in 13 institutions. The followingis a excerpt of the introduction to this issue that was written byGeorge I. Soete with technical assistance from Gail McMillanof Virginia Tech.

perhaps no technological development in recentyears has so energized the academic community aselectronic theses and dissertations (ETDs). Though

only one university is currently in full production mode,with a second close behind, many others are runningpilot projects or discussing the issues on their campuses.Some in academia are very strong advocates, pointing toclear benefits for all concerned, but others are not so sure,suggesting that ETDs may compromise the publicationpotential of student work.

ETDs: The Current SituationETDs are digitally produced, archived, and accessed thesesand dissertations. Typically, students produce their workusing standard word processing systems, then convertthem into PDF or SGML formats. They may insert multi-media files: still pictures, maps, or video and audio clips.Students have a menu of choices for distribution; typicallythese choices include 1) unrestricted distribution via theWeb; 2) distribution restricted to the university communi-ty; and 3) no distribution allowed for a certain period.

One institution, Virginia Tech, is in full productionmode with more than 1,100 ETDs approved so far. WestVirginia University is the second institution to make elec-tronic submission of theses and dissertations mandatory,effective August 1998. Several other institutions are inpilot mode, and many others are in the process of orient-ing their campus communities and/or investigating anddiscussing ETDs.

Virginia Tech is the clear leader in the field; they areviewed by many as setting standards for ETD programs.With grant support, they are providing training, soft-ware, and other kinds of assistance to anyone who asksfor it. The Networked Digital Library of Theses andDissertations (NDLTD), based at Virginia Tech, is anumbrella organization of 35 institutions (as of July 1998)that are at various stages in the development of ETDprograms. NDLTD acts as a clearinghouse for informa-tion about ETDs, holds discussions about ETDs withpublishers, and promotes inquiry into ETD issues.

Benefits of ETDsThe benefits of ETDs are many. In general, they correlatewith the benefits of most electronic services andresources:

Accessibility from anywhere at any time.Quick availability after submission (paperdissertations may take up to a year to be processed).

Searchability, indexing.Inclusion of multimedia.Potential savings in storage space, circulation,interlibrary loan, and processing costs.Simplification of cataloging processes.Education of students in use of electronic technologiesin scholarly inquiry and publication.

Key IssuesWhile the benefits of ETDs are vividly apparent, move-ment toward widespread adoption of fully developed ETDprograms has been relatively slow. Libraries are grapplingwith several issues. As usual, the purely technical issuesappear to be solvable, while the human issues look some-what more daunting.

Readiness. The issue of readiness encompasses a greatdeal. For most, it means, "How ready is the institution toembrace ETDs such that electronic submission should bea requirement of the process?" Most institutions areanswering, "Not yet ready!" In general, though it is notinvariably true, faculty and students in the sciences aremore enthusiastic about Ems than their colleagues in thesocial sciences and humanities. Humanists and social sci-entists tend to be more concerned about loss of potentialpublication of their work and plagiarism. In some cases,student skills are not sufficient or the technical infrastruc-ture is considered inadequate.

Publication Potential. One of the thorniest issues for manystudents is the possibility that their chances of getting theirwork accepted by a publisher is compromised by the elec-tronic availability of their work. Several publishers havedeclared flatly that Ems are considered previous publica-tions and will not be accepted. But others (Elsevier, forexample) have declared no blanket prohibition againstETDs, suggesting that individual editors would factor ETDavailability in their publication decisions. It is likely thatpublishers will be less skeptical once there is wider use ofETDs and more data on trends available.

Intellectual Property. As with any electronic publication,Ems present property rights issues, even though owner-ship by the student and sponsorship by the university mayseem to simplify the problems. Some are concerned thatwider availability on the Web will lead to plagiarism andirresponsible redistribution of their work. Some facultyhave expressed concern that their own work, which issometimes the framework for the student's work, will beprematurely released and not appropriately credited.

Orientation and Training. Institutions running ETD pro-grams, whether in pilot or routine mode, have discoveredthat orientation and training need a great deal of attentionand may represent significant additional cost. Studentsneed to be trained in systems such as PDF. Sometimes theyeven need basic training in word processing. The decisionconcerning who will provide training for students is a keyone for planners. Peer trainers have been found by someinstitutions to be very effective. In general, interviewees

A R L 2 0 1 D E C E M B E R 1 9 9 8 100

L.

spoke of greater success with informal training methods,and spoke of the importance of having staff available toassist students when they needed special help.

Standards. Choice of formats can be troublesome, espe-cially when students are creating their work in every-thing from state-of-the-art systems to antique wordprocessing systems. Planners need to decide what stan-dards they will set for electronic submission and howmuch responsibility they will place on the student formeeting standards. Standards determine how computerlabs will be equipped, what systems trainers have tomaster, and how difficult it will be to manage files.

Costs. Opinions about costs are very much divided. Somefeel that ETDs, though a tremendous improvement in ser-vice, will not result in appreciable overall cost-savings;others feel that considerable savings will result. Still otherspoint to significant start-up costs. It seems undeniablethat there will be savings in some areas for those institu-tions that migrate to required electronic submission; forlibraries, such savings should be found in storage, process-ing, and interlibrary lending, as well as other areas. Andyet, until the top of the learning curve is reached, studenttraining might represent significant additional costs.

Archiving, Preservation. Archiving and preservation aremajor issues for most interviewees. Some institutions aremoving forward, trusting that technological solutionswill be found, while others remain cautious and are notcommitting to archiving. Many others are committed torunning parallel paper/microform systems or to relyingon UMI as an archival holding.

Restrictions. Many institutions will provide studentswith a menu of distribution choices, including the optionof restricting access to the campus or denying accessaltogether. Yet some have begun to question restrictionsand to work toward refinements. A critical question forstate institutions is whether they can, by law, keep thecitizens of their states from accessing the products ofuniversity research.

ETDs: The FutureVirtually all interviewees felt that electronic theses anddissertations are an inevitable development. Thoughmany institutions are taking it slow, their ultimate goal,clearly, is required submission of ETDs and, ultimately,a paperless thesis and dissertation program.

Much like the larger world of electronic publication,ETDs represent major changes and major challenges toestablished ways of doing things. In this case, the bene-fits seem so heavily to outweigh any negative aspects thatthe widespread implementation of ETD programs doesindeed seem inevitable.

1Issues and Innovations in Electronic Theses and Dissertations,Transforming Libraries #7 (also issued as SPEC Kit 236),Washington, DC: ARL, October 1998. See <http:/ /www.arLorg/transform/index.html>.

1 1f

SPRING 1999 CONFERENCESARL/SLA Licensing VideoconferenceMarch 4, 1999"De-Mystifying the Licensing of ElectronicResources," a videoconference co-sponsored by ARLand the Special Libraries Association and funded inpart by Lexis-Nexis, will be held on March 4, 1999,from 1:00-4:00 p.m. EST. The videoconference willcover the basic elements of a license agreement; thelegal foundations of a license, including copyrightand other relevant areas of the law; and user, access,and legal terms. Panelists include Molly H. Sherden,Attorney with Peabody & Arnold, Boston; PamelaClark, Vice President of American InternationalGroup, London and New York; and Trisha Davis,Head of Continuation Acquisition Division, OhioState University Libraries.

Site licenses for the live satellite broadcastare available to ARL libraries for $350 (beforeFebruary 9). There is no limit to the number ofparticipants at a site. This is an excellent opportunityto provide the basics of licensing for those library staffand university personnel who may not need or havethe time for the full two-day workshop offered byARL. Details are available on the ARL website<http:/ /www.arl.org/scomm/licensing/videoconf.html>.

New Challenges for Scholarly Communication inthe Digital Era: Changing Roles and Expectationsin the Academic CommunityMarch 26-27, 1999A conference sponsored by the American Associationof University Professors, American Council ofLearned Societies, Association of AmericanUniversity Presses, Association of Research Libraries,and the Coalition for Networked Information will beheld on March 26-27, 1999, in the Sheraton CityCenter Hotel, Washington, DC. Building on thesuccess of last year's conference on the scholarlymonograph, this conference will look at a number ofkey issues confronting scholarly communication inthe digital age. Topics include: Getting Ahead in theDigital World, Distance Education; What Does itMean to Publish?; Economics of Scholarly Publishing;and Preservation and Access. Detailed programand registration information is available at<http:/ /www.arl.org/scomm/conf.html>.

A R L 2 0 1 D E C E M B E R 1 9 9 8

Kathryn J. Deiss, OLMS Program Manager

STEAL A BASE OR STAY SAFE?TAKING RISKS TO GROW

Progress always involves risk; you can't steal second baseand keep your foot on first.

Frederick Wilcox'

Continued success in any organization, includinglibraries, depends on its individuals to learn atleast as fast as the rate of organizational change.

This learning often involves taking risks: risks in tryingnew behaviors, risks in abandoning what we do well toexplore what we know less well, and risks in developingnew models to deliver on our missions, just to name afew areas. In order for this to take place, however,we must first overcome the barriers to establishing arisk-taking environment.

Easier Said Than DoneAs consultants, we frequently hear the hope that peoplewill take more risks to help the organization breakthrough to new concepts, solutions, and paradigms.However, we also hear the individual and collectivevoice of fear saying that risk is too dangerousperson-ally and organizationally. Trying something new is likestepping out to steal a base: if we succeed, the success isabsorbed into our daily practice, routinized, and forgot-ten, but if we are unsuccessful, the failure is remem-bered and carried over into assumptions about our abili-ties, our leaders, and our safety. A common fear is ofretribution for failure. The concept of "permission tofail" is a phrase incapable of being meaningful in thepractices we exhibit. After all, do we have examples inour everyday work worlds where we applaud or evenpermit failure resultant from risk-taking? Or is this ararity in our organizational cultures?

The main question to ask, then, is: Why is risk-taking so difficult and what can organizations do toaddress that issue?

The "Real" Problem with Risk-TakingWhile fear of retribution if a risk turns to failure is oneinhibitor of risk-taking, it is not the only concern that isvoiced. One of most often-heard and cogent concerns isof the possible failure to provide the expected perfor-mance. People want to deliver work in the way that hasbeen comfortable, reasonably effective, and rewarded.Although those who support the concepts behindincreased risk-taking indicate that a higher degree ofeffectiveness is likely to be the direct result of a higherdegree of risk-taking, it is not before first passingthrough a forbidding level of diminished effectiveness,commonly known as the "learning curve." There is a

, definite sense that taking risks can lead to feeling, seem-ing, or even being "less" competent than one is used to

beingin fact, one might even go so far as to say we areasking people to be "consciously incompetent" whenwe ask them to take risks. Everything in their experi-ence tells them that risk-taking behaviors generallymean being in a state of less than total controlof beingin a state of exploration, puzzlement, and discovery, anuncomfortable or even unacceptablestate for many peo-ple in the workplace.

These worries point to the observation that fear andreticence to risk are consequences of the fact that wehave, apparently, only one arena in which to take risks:our "performance arena," where we actually do thework we are there to do. We feel that what we areengaged in is entirely too important to us to chance fail-ure or the appearance of failure. Obviously, we don'twant to risk not being able to perform our workforlibraries, the organization and delivery of informationserviceswork we are rewarded for performing in aconsistent and predictable manner. Therefore, we needto create an arena that will safely allow for risk-takingwithout sacrificing job performance when learning totake those risks.

Where To Learn?Creating Practice ArenasObviously, one can learn even in performing tasks asone has always performed them. However, in an areaof less competence, risk is required to beget learningand new practices. What is needed to encourage morerisk-taking? Where can this activity best take place? Ifour performance arena calls on us to deliver our workwith a fairly consistent and high degree of quality,where will we find the "practice arena" in which tolearn the new behaviors associated with risk-taking?One answer is to turn to areas of our work that aren'texpressly connected to our delivery of service. In thelibrary environment, such areas may include data gath-ering, performance evaluation, or designing responsesto our user communities. I would suggest that these areareas in which we could create "practice arenas" to trulyencourage greater creativity and risk-taking withoutinvoking those concerns that cause reticence to risk.

Where We Feel ComfortableThere are a finite number of areas where risk can takeplace: the physical, the intellectual, the financial, andthe relationship areas. In at least three of these areas(the latter three) we are being called upon to risk inorder to meet and respond to the changing environ-ment. Libraries, faced with a rapidly changing andgrowing intellectual environment, are clearly eager fornew ideas, concepts, models, and theories with which tofashion responses to this environment. In the financialarea, we might benefit from risk and innovation in theways we allocate funds, seek new funding, and assess

ARL 201 DECEMBER 1998102

fiscal performance. And the relationship area couldbenefit from true risk-taking that may result in newcommunication behaviors, facilitative and sharedleadership, the ability to confront organizationaldefensive mechanisms, and the building of a newsense of community.

The Leaders' RoleFinally, leadership support is necessary in order tocreate an environment that encourages risk-taking.To do so, leaders at all levels of the organization needto reflect on their own risk-taking and facilitativebehaviors; only through self-understanding can oneclearly explain the desired behaviors and designstructures and systems that support those behaviors.However, modeling risk-taking is not enoughin fact,this can create an "organizational nervousness" where-by others see no safe and stable place. Therefore, lead-ers' risk-taking needs to be accompanied by genuinesupport of others' risks, both psychologically and interms of rewards.

ConclusionDisorder, challenge to espoused tradition and practice,and play are natural elements of the risk-takingprocess. As such they need to be encoliraged and,indeed, protected in order to make risk-taking a morenatural behavior. David Campbell, Smith RichardsonSenior Fellow of the Center for Creative Leadership,identifies a number of organizational blocks to creativi-ty.' Two of these, a "preoccupation with order and tra-dition" and a "reluctance to play," seem to be especial-ly difficult for us in libraries. Creating "practice are-nas" where freedom to play with ideas and conceptsexists without high risk to immediate performancemight help groups and individuals "learn" to risk as amore natural behavior. But playfulness is only one partof risk; commitment to risk is just as important. AsDaniel Goleman, author of The Creative Spirit, has writ-ten, "Risk, like beauty, is in the eye of the beholder...[something that] may seem risky or dangerous from theoutside can seem entirely different to the person in itsmidst. The hidden variable is commitment."'

To gain commitment to and expand our capacityfor risk-taking, our organizations must be ingeniousin creating reward systems that encourage the behav-iors we want to develop. We must have credible lead-ers who themselves model and support this behavior,and we must formpractice or learning arenas thatallow people to develop confidence in themselvesand in the benefits of temporary puzzlement anddiscomfort. By doing so, we will create an atmosphereof safe change that allows for innovation, better jobresults and satisfaction, and, ultimately, a moresuccessful organization.

103

The Office of Leadership and Management Services is offeringtwo programs that will help libraries to develop "can do"cultures: the Edgework Institute: Stimulating Innovation inLibrary and Information Services, in November 1999; and thefour-part Institute on Organizational Learning for LibraryPerformance, which will open in October 1999.

1 Quoted in Sharon Gilchrest O'Neill's Lfir'nxing: 147Inspiring Thoughts for Learning on the Job (Berkeley, CA:Ten Speed Press, 1993).

2 David P. Campbell, Take the Road to Creativity and Get OffYour Dead End (Greensboro, NC: Center for CreativeLeadership, 1985).

3 Daniel Goleman, The Creative Spirit (New York: Dutton,1992).

NEW SPEC KITS FROM THE OLMSINFORMATION SERVICES PROGRAMby Patricia Brennan, ARL Publications Officer

To order copies of any of these titles, contact the ARLPublications department at <pubs@arl.org>.

The Role of ARL Libraries in Extension/OutreachSPEC Kit 233, August 1998Compiled by Claudine Jenda and Tamera Lee,Auburn University

Issues and Innovations in Distance LearningSPEC Kit 234, TL 6, September 1998Compiled by William G. Jones, University of Illinoisat Chicago; Joan K. Lippincott, CM, Editorial Advisor

Collaborative Collections Management Programsin ARL LibrariesSPEC Kit 235, September 1998Compiled by George Soete, ARL/OLMS

Issues and Innovations in Electronic Thesesand DissertationsSPEC Kit 236, TL 7, October 1998Compiled by George Soete, ARL/OLMS;Gall McMillan, Virginia Tech, Editorial Advisor

Managing Food and Drink in ARL LibrariesSPEC Kit 237, September 1998Compiled by George Soete, ARL/OLMS

SPEC Kits: ISSN 0160-3582, $40 ($25 ARL members).SPEC Flyers summarizing SPEC Kits are available at<http://wzvw.arl.org/spec/complete.html>.

A R L 2 0 1 DECEMBER

>

---STATISTICS-ANEt-M-EA-StTREMEM'Julia Blixrud, Senior Program Officer

PRESERVATION EXPENDITURES LEVEL;MICROFILMING, STAFFING DECLINE

The ARL Preservation Statistics provides annual statis-tical information on the current level of preservationefforts in research libraries and on the key organiza-

tional, functional, and fiscal components that characterizepreservation programs. The 1996-97 survey was revisedto streamline current data collection activities and to addtwo optional questions regarding digitizing: "number ofbound volumes/pamphlets digitized," and "number ofsingle, unbound sheets (manuscripts, maps, photographs)digitized." Although digitizing for preservation purposeswas defined broadly in the instructions, only 19 librariesresponded to the optional questions on bound volumesdigitized. The reports varied widely from one bound vol-ume digitized (two institutions) to 335,191 items reporteddigitized by one institution.

Preservation expenditures have been level since 1993.The total $80,772,236 reported for 115 reporting memberlibraries reflects a modest increase of about $3,000,000 for1996-97. Total preservation staff sharply declined to 1,742I- i Es in 1996-97 from 1,879 in 1995-96, falling back to thestaffing levels reported in 1989-90. Level 1 conservationtreatment has increased and Level 2 treatment hasdeclined, but Level 3 treatment has remained at approxi-mately the same level as last year.

Microfilming activity, measured in volumes, hasdeclined for the second year in a row. Availability ofexternal funds continues to play a critical role in thatpreservation activity. In 1988, the National Endowmentfor the Humanities (NEH) began a multi-year, expanded

VOLUMES FOR PRESERVATIONMICROFILM MASTERS

180,000

160,000

140,000

120,000

100,000

80,000

60,0001989 1990

VolumesPreserved

1991 1992 1993 1994 1995

Year

1996 1997

cooperative preservation microfilming program. ARLlibraries participated extensively in that program, and datashow that cuts in the NEH budget negatively affected theavailability of external funding for preservation activities byresearch libraries. External funding has been reduced con-stantly from a high of $11,090,547 in 1992-93 to a low of$7,364,841 in 1996-97. University libraries in the aggregatereported that special grants from external sources account-ed for about 9% of the total preservation expenditures, twopercentage points lower than last year. Grant funds wereexpended predominantly on preservation microfilmingprojects. The accompanying chart shows that ARL mem-ber libraries preserved 109,526 volumes on microfilm in1996-97, a 29% decline in volumes since last year.

The financial support for preservation activities inARL university libraries shows a substantial range fromapproximately $116,354 to $3.5 million during fiscal year1996-97. As a corollary, ARL university libraries spentfrom as little as 1% to as much as 7% of total library bud-gets for preservation.

The chart below displays the allocation of preservationexpenditures based on data from all reporting ARLlibraries. Local needs and capabilities will determine theexact allocation of budgetary resources to various activi-ties, but, as in past years, the largest category is salariesand wages, followed by binding expenditures.

ARL Preservation Statistics 1996-97 presents data from115 U.S. and Canadian research libraries. Copies are avail-able for $35 to member libraries and $65 to nonmembers(plus $6 shipping and handling per publication). Pleasecontact ARL Publications, Department #0692, Washington,DC 20073-0692; (202) 296-2296; or email <pubs@arl.org>.

PRESERVATION EXPENDITURES1996-97

ContractBinding30.82%

ContractMicrofilming6.91%

Supplies4.42% Equipment

1 7%

ContractConservation1.04%

Other ContractExpenditures1.58% Contract

Preservation Photocopying.82%

Salaries& Wages52.7%

A R L 2 0 1 D E C E M B E R 1 9 9 81 4

Lee Anne George, Program Planning Officer

GEORGE WASHINGTONUNIVERSITY JOINS ARL

Ats Membership Meeting held in October, themembership of ARL voted to invite The GeorgeWashington University Libraries to join as the

122nd member. Located in Washington, DC, theUniversity is classified as a Carnegie Research UniversityII institution. Jack A. Siggins, University Librarian,accepted the invitation.

The University Libraries serve as a resource in sup-port of the research and instruction conducted by theapproximately 1,175 faculty and 16,000 students. TheLibraries include a distinctive Washingtoniana collection,a collection of post-Cold War materials from the formerSoviet Union, Eastern Europe, and East Asia, as well as animportant Judaica collection donated by the Kiev family.

More information about The George WashingtonUniversity Libraries may be found at: <http:/ /www.gwu.edu/gelman/>.

ARL CONFRONTS CHALLENGESOF THE DIGITAL ERA

The 133rd Membership Meeting of ARL, heldOctober 14-16, 1998 in Washington, DC, was con-vened by ARL President James Neal, Johns Hopkins

University, to engage the theme of Confronting theChallenges of the Digital Era. University of MarylandPresident Dan Mote, Jr. opened the program with akeynote address on strategies for raising funds for thedigital library. The program included an economist'sassessment of the impact of journal publisher mergers onpricing practices, a panel of experts looking at staffingissues in the digital era, and perspectives from provostsDavid Shulenburger (University of Kansas) and CharlesPhelps (University of Rochester). Papers presented at theOctober meeting are made available electronically onARL's web server as they are received, at <http:/ /www.arl.org/arl/proceedings/133/index.html>.

ElectionsAt the ARL Business Meeting, membership elected threenew representatives to the ARL Board of Directors:Meredith Butler (SUNY at Albany), Joseph Hewitt(University of North Carolina), and Carolynne Presser(University of Manitoba). Each will serve a three-year termon the Board, October 1998 to October 2001.

ARL President Jim Neal also acknowledged and salutedthe contributions of Board members whose terms expiredthis October: Bill Crowe (University of Kansas), CaroleMoore (University of Toronto), and Gloria Werner (UCLA).

Kenneth Frazier (University of Wisconsin) was elect-ed Vice President and President-Elect of ARL by the ARLBoard of Directors. He serves as Vice President for a yearbefore becoming President in October 1999.

At the conclusion of the Business Meeting, Mr. Nealhanded the gavel to Betty Bengtson (University ofWashington), who began her term as ARL President.

HONORSKendon L. Stubbs: On October 30, Kendon L. Stubbsbecame the 45th recipient of the Thomas Jefferson Award,the University of Virginia's highest award. The award hon-ors an individual who exemplifies in character, work, andinfluence the principles and ideals of the University'sfounder. Stubbs was recognized for his innovations in mak-ing the University Library's vast resources more accessibleto the University and scholarly communities via the Internet.

TRANSITIONSUCBerkeley: Gerald R. Lowell was appointedUniversity Librarian, effective December 1, 1998.He was previously University Librarian and AssociateVice Chancellor-Academic Information Technologyat UCSan Diego.UCSan Diego: Phyllis S. Mirsky, previously DeputyUniversity Librarian, was appointed Interim UniversityLibrarian, effective November 7.Michigan: William A. Gosling was named Director of theLibraries. Serving as Michigan's Interim Director since1997, he came to the University in 1986 as AssistantDirector for Technical Services and Library Systems.Tulane: Philip E. Leinbach announced his resignation asUniversity Librarian, effective July 1, 1999.

International Federation of Library Associations: LeoVoogt has announced his resignation as lFLA SecretaryGeneral to accept the position of Executive Director of theNetherlands Royal Association for the Book Trade (KVB).ICANN: Mike Roberts was named Interim President ofThe Internet Corporation for Assigned Names andNumbers, the new non-profit group that is likely to takeover the Internet's address registration system. He wasformerly Vice President of EDUCOM.

ARL Staff TransitionsPatricia Brennan announced her resignation as ARLProgram Officer for Publications to accept the position ofPublications Manager in the Harvard University Library,effective January 25.

ARL/AAU Global Resources Program TransitionsRoger Brisson, Digital Access Librarian and Selector,German Language & Literature at Pennsylvania StateUniversity's Pattee Library, assumed responsibility asthe Coordinator, AAU/ARL German Resources Project,effective July 1, 1998.Mary E. Jackson, ARL Senior Program Officer for AccessServices at ARL, was appointed Director of the NationalCoordinating Committee (NCC) on Japanese LibraryResources. Ms. Jackson will also assume responsibilityfor coordinating the AAU/ARL/NCC Japanese JournalAccess Project.Eudora Loh, Latin American Bibliographer at UCLA,has taken over the leadership of the AdvisoryCommittee of the AMY ARL Latin AmericanistResearch Resources Project.

105 A R L 2 0 1 D E C E M B E R 1 9 9 8

ARL: A Bimonthly Newsletter of Research Library Issuesand Actions (US ISSN 1050-6098) is published six timesa year by the Association of Research Libraries,21 Dupont Circle, Washington, DC 20036.202-26-2296 FAX 202-872-0884<http://www.arl.org/newsltr/newsltr.html>Copyright: © 1998 by the Association of Research Libraries

ARL policy is to grant blanket permission to reprint any article inthe newsletter for educational use as long as the source, author, issue,and page numbers are acknowledged. Exceptions to this policy may

_Executive Director: Duane E. WebsterEditor: G. Jaia Barrett, Deputy Executive DirectorCo-editor: Lee Anne George, Program Planning OfficerCopy Manager: Karen A. WetzelDesigner: Kevin Osborn, Research & Design, Ltd., Arlington, VASubscriptions: Members$25 per year for additionalsubscription; Nonmembers$50 per year.

be noted for certain articles. For commercial use, a reprintrequest should be sent to ARL Publications <pubs@arl.org>.

ARL CALENDAR 1999January 7-8 From Data to Action: An ARL

Workshop on Strategies toRedesign ILL/DD ServicesChicago, IL

January 29 OCLC/ARL ResourceSharing SymposiumPhiladelphia, PA

February 11-12 ARL Board MeetingWashington, DC

March 4 De-Mystifying the Licensingof Electronic ResourcesARL/SLA Videoconference1:00-4:00 p.m. EST

March 10-12 Library Management SkillsInstitute I: The ManagerDenver, CO

March 26-27 New Challenges for ScholarlyCommunication in the DigitalEra: Changing Roles andExpectations in theAcademic CommunityWashington, DC

May 4-6 Facilitation Skills InstituteSeattle, WA

May 11-14 ARL Board andMembership MeetingKansas Cifty, MO

May 19-21

June 3-4

July 26-27

October 7-8

October 12-15 ARL Board andMembership MeetingWashington, DC

October 26-29 Library ManagementSkills Institute TheManagement ProcessEvanston, IL

November 3-5 Assistant/AssociateLibrarian InstituteCharleston, SC

November 10-12 Edgework Institute: StimulatingInnovation in Libraries andInformation ServicesWashington, DC

November 17-19 Library Management SkillsInstitute I: The ManagerAtlanta, GA

Training Skills Institute:Managing the Learning ProcessSan Diego, CA

Project Management Institute:Getting Things Done or Gettingthe Outcomes You WantEvanston, IL

ARL Board MeetingWashington, DC

Leading Change InstituteSan Antonio, TX

106

1:1

U.S. Department of EducationOffice of Educational Research and Improvement (OERI)

National Library of Education (NLE)Educational Resources Information Center (ERIC)

NOTICE

REPRODUCTION BASIS

IC

This document is covered by a signed "Reproduction Release(Blanket) form (on file within the ERIC system), encompassing allor classes of documents from its source organization and, therefore,does not require a "Specific Document" Release form.

This document is Federally-funded, or carries its own permission toreproduce, or is otherwise in the public domain and, therefore, maybe reproduced by ERIC without a signed Reproduction Release form(either "Specific Document" or "Blanket").

EFF-089 (9/97)