Best Paper Organization Culture Efffectiveness

8/2/2019 Best Paper Organization Culture Efffectiveness

1/21

Toward a Theory of Organizational Culture and Effectiveness

Author(s): Daniel R. Denison and Aneil K. MishraReviewed work(s):Source: Organization Science, Vol. 6, No. 2 (Mar. - Apr., 1995), pp. 204-223Published by: INFORMSStable URL: http://www.jstor.org/stable/2635122 .

Accessed: 10/02/2012 03:08

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

INFORMS is collaborating with JSTOR to digitize, preserve and extend access to Organization Science.
http://www.jstor.org/action/showPublisher?publisherCode=informshttp://www.jstor.org/stable/2635122?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2635122?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=informs


2/21

Toward a T h e o r y o f OrganizationalC u l t u r e a n d Effectiveness

Daniel R. Denison * Aneil K. MishraSchool of Business Administration, Universityof Michigan, Ann Arbor, MI 48109Smeal College of Business, Pennsylvania State University,UniversityPark, PA 16802

T his paperuses case studiesand surveydata to explorethe relationshipbetweenorganizationalculture and effectiveness. The results provide evidence for the existence of four culturaltraits-involvement, consistency,adaptability, nd mission-and indicate that these characteristicsare positivelyrelatedto perceptionsof performance s well as to objectivemeasuressuch as returnon assets and sales growth.Organizationalulture is found to be measurableand to be relatedtoimportantorganizational utcomes. RichardL. DaftAbstractThis paper develops a model of organizational culture andeffectiveness based on four traits of organizational cultures;involvement, consistency, adaptability, and mission. Thesetraits are examined through two linked studies: In the first,qualitative case studies of five firms are used to identify thetraits and the nature of their linkage to effectiveness; In thesecond, a quantitative study provides an exploratory analysisof CEO perceptions of these four traits and their relation tosubjective and objective measures of effectiveness in a sampleof 764 organizations. The results show support for the predic-tive value of the traits, and help to illustrate the complemen-tarity of qualitative and quantitative methods for studyingorganizational cultures.Two of the traits, involvement and adaptability, are indica-tors of flexibility, openness, and responsiveness, and werestrong predictors of growth. The other two traits, consistencyand mission, are indicators of integration, direction, andvision, and were better predictors of profitability. Each of thefour traits were also significant predictors of other effective-ness criteria such as quality, employee satisfaction, and over-all performance. The results also showed that the four traitswere strong predictors of subjectively-rated effectiveness cri-teria for the total sample of firms, but were strong predictorsof objective criteria such as return-on-assets and sales growthonly for larger firms.This paper suggests that culture can be studied as anintegral part of the adaptation process of organizations andthat specific culture traits may be useful predictors of perfor-mance and effectiveness. The paper also illustrates how qual-itative case studies and inductive theory building can be

combinedwith quantitative omparisonsand theory-testingto makeprogresson specificaspectsof organizationalultureresearch.(Theory Building; CultureEffectiveness)

The relationship between the culture and functioningof social organizations has been a recurring theme inthe social sciences for over 50 years. Sociologists, socialanthropologists, and social psychologists have oftenpresented culture and ideology as integral features ofthe functioning of a society (Weber 1930, Mead 1934,Radcliffe-Brown 1952). Each of these authors focusedon culture as a critical aspect of the adaptation ofsocial organizations, and viewed culture as a system of"socially transmitted behavior patterns that serve torelate human communities to their ecological settings"(Keesing 1974).. This perspective has also been re-flected in the work of ethnographers such as Whyte(1949) and Rohlen (1974), and by psychologists such asSchein (1985, 1990) and Hofstede (Hofstede 1980, 1991;Hofstede et al. 1990).Organizational researchers have also addressed therelationship between culture and functioning (Wilkinsand Ouchi 1983, Barney 1986, Barley et al. 1988,Saffold 1988, Ott 1989), but have seldom developedexplicit theories of organizational culture and effec-

1047-7039/ 95/0602/0204/ $01.25Copyright ? 1995. Institute for Operations Research204 ORGANIZATION SCIENCE / Vol. 6, No. 2, March-April 1995 and the Management Sciences


3/21

DANIEL R. DENISON AND ANEIL K. MISHRA Organizational Culture and Effectiveness

tiveness or presented supporting evidence (Siehl andMartin 1990). Progress has been made in related re-search areas such as socialization (Van Maanen andSchein 1979, Chatman 1991), and change (Schein 1985,Kotter and Heskett 1992) but with few exceptions(e.g., O'Reilly 1989) little attention has been given tothe issue of organizational culture and effectiveness.The purpose of this paper is to develop and test amodel of the cultural traits that appear to characterizeeffective organizations. Our intention is to develop anexplicit theory about culture and effectiveness that canextend the implicit, but often unelaborated themes thatappear in many culture studies. This is accomplishedby presenting two linked studies: The first uses qualita-tive studies of five organizations to develop hypothesesabout four cultural traits and their linkage to effec-tiveness. The second study then presents a quantitativetest of these four hypotheses by comparing CEO per-ceptions of culture with subjective and objective mea-sures of effectiveness in 764 organizations.

Studying Organizational Cultureand EffectivenessThe topic of culture and effectiveness is of centralimportance in organizational studies, but progress inthe development of theory and research has beenslowed by a formidable set of research problems. Forexample, an integrative theory must encompass a broadrange of phenomena extending from core assumptionsto visible artifacts, and from social structures to indi-vidual meaning. Such a theory must also address cul-ture as a symbolic representation of past attempts atadaptation and survival as well as a set of limiting orenabling conditions for future adaptation. Several at-tempts at integration have been presented (Allaire andFirsirotu 1984; Ott 1989; Schein 1985, 1990) but thereis still limited consensus regarding a general theory oforganizational culture, and healthy skepticism aboutwhether culture can ever be "measured" in a way thatallows one organization to be compared with another.The &oncept of effectiveness also presents a chal-lenging set of problems. The multidimensional natureof the concept requires that effectiveness be defined bya complex of stakeholders, who may hold differing,incompatible, and changing criteria (Cameron andWhetten 1983, Lewin and Minton 1986). In recognitionof these difficulties, some researchers have even calledfor a moratorium on empirical research on effective-ness, suggesting that inquiries be limited to singleoutcome studies (Goodman et al. 1983). Joining the

two concepts of culture and effectiveness thus defines aresearch question which is important, but often prob-lematic.The dearth of empirical research on culture andeffectiveness can also be traced to the origins of cultureresearch within organizational studies. The develop-ment of the culture perspective and the critique of thepositivist approach to social science (Geertz 1973,Pettigrew 1979, Weick 1979, Martin et al. 1983, Frostet al. 1985) are closely linked. Most culture researchersadvocate a phenomenological approach, emphasizingthe qualitative understanding of meaning and interpre-tation. Several have argued that culture should not bestudied as a "variable" with "outcomes" (Trice andBeyer 1984, Meek 1988, Sackmann 1991, Siehl andMartin 1990), and have thus discouraged the integra-tion of culture research with other constructs of organi-zation theory (including effectiveness) which are pri-marily discussed in terms of variables, dimensions,causes, or effects. Thus, the literature carries a mes-sage about epistemology as well as substance.The culture perspective is also intertwined with thepopular literature on organizations that appeared dur-ing the 1980s (Ouchi 1981; Deal and Kennedy 1982;Peters and Waterman 1982; Kanter 1983; Kilman et al.1985; Tichy 1987; Schein 1985, 1990). This literature,while perhaps loosly connected with the epistemologi-cal critique of positivism, was directly linked to a farmore immediate concern: the rising tide of interna-tional competition. In contrast to the minor role thateffectiveness has played in the scholarly literature, herethe functional linkage between culture and effective-ness was a central issue. Unfortunately, the evidencepresented was often no more than selective anecdotesand normative assertions regarding the value of cul-ture. Thus, the popular literature, like the scholarlyliterature, has contributed little comparative evidenceregarding culture and effectiveness..Several authors have attempted to integrate thefunctionalist and phenomenological perspectives. On atheoretical level, the crux of the issue has been thetension between the functionalist perspective empha-sizing the predictable impacts of purposive, intentionalforms of social organization, and a phenomenologicalperspective emphasizing the emergent and epiphenom-enal nature of those same organizations. Giddens(1979) and Riley (1983), for example, have argued thatresearchers must seek a concurrent understanding ofthe impact of social structure on individuals as well asthe part that individuals play in creating those struc-tures. Each organizational actor must be understood as

ORGANIZATION SCIENCE / Vol. 6, No. 2, March-April 1995 205


4/21

DANIEL R. DENISON AND ANEIL K. MISHRA OrganizationalCultureand Effectiveness

both a subject and an agent of the system at the sametime (Mead 1934). To wit;Structures are both the medium and the outcome of interac-tion. They are the medium, because structures provide therules and resources individuals must draw upon to interactmeaningfully. They are its outcome, because rules and re-sources exist only through being applied and acknowledged ininteraction-they have no reality independent of the socialpractices they constitute. (Riley 1983, p. 415)

On an empirical level, there is also increasing attentionto the integration of approaches. Jermier et al. (1991),Hofstede et al. (1990), Siehl and Martin (1988), andDenison (1990) each present empirical research thatincorporates qualitative and quantitative data, and ac-knowledges the importance of both the functionalistand phenomenological perspectives.Other researchers have recently presented empiricalculture studies which have relied upon quantitativesurvey data to characterize and compare organizationalcultures. (Rousseau and Cooke 1988, Hofstede et al.1990, Chatman 1991, O'Reilly et al. 1991, Jermier et al.1991, Kotter and Heskett 1992). This research hascharacterized organizational cultures in terms of un-derlying traits and value dimensions, and has typicallyacknowledged that culture is being treated as a vari-able for a specific research purpose. These studiesillustrate the range of approaches that have been takenby culture researchers. They also reflect the approachtaken in this study; a combination of qualitative andquantitative methods coupled with an attempt to gainthe insights of both the contextualized meaning ofspecific settings and the generality offered by compari-son across multiple organizations.

Culture and Effectiveness:Empirical ResearchThe empirical literature on organizational culture andeffectiveness can be traced back to early studies ofculture and adaptation (Weber 1930, Buckley 1967),and to the work of classic organizational theorists suchas Likert (1961), Burns and Stalker (1961), or Lawrenceand Lorsch (1967). More recently Wilkins and Ouchi(1983) discussed the concept of a "clan" organizationand explored the hypothetical conditions under whichclans would be a more efficient organizational form.The evidence presented by Peters and Waterman (1982)identified cultural characteristics of successful compa-nies and built a "theory" of excellence that has causedmuch debate (Carroll 1983, Van de Ven 1983,

206

Aupperle et al. 1986). Kanter (1983), in a study oforganizational change, showed how companies withprogressive human resource practices outperformedthose with less progressive practices. Denison (1984),using survey-based culture measures, showed that per-ceived involvement and participation on the part oforganizational members predicted both current andfuture financial performance. In addition, Gordon(1985) has shown that high and low performing compa-nies in the banking and utilities industries had differentculture profiles.Kravetz (1988) amplified Kanter's earlier findings byshowing that management practices fostering participa-tion, autonomy, and creativity were closely correlatedwith objective indicators of organizational perfor-mance. Hansen and Wernerfelt (1989) contrasted theeffects of internal organization and market position onperformance. Their study showed that internal organi-zational factors were stronger predictors than marketposition and that the two predictors were largely unre-lated. A French study of managerial values andpractices also identified a set of cultural values andpractices associated with the growth patterns of fivesingle-business firms (Calori and Sarnin 1991). Finally,Kotter and Heskett (1992) presented an analysis of therelationship between strong cultures, adaptive cultures,and effectiveness.In addition, some theoretical development has alsooccurred. Barney (1986) and Camerer and Vepsalainen(1988) discussed culture from an economic perspectiveand defined conditions under which culture might in-fluence efficiency or effectiveness. Saffold (1988) hascritiqued the literature on cultural strength and norma-tive integration, and urged the development of a moresophisticated theory of culture and performance.Gordon and DiTomaso (1992) provide further empiri-cal support for the impact of cultural strength andvalue consensus on performance in a recent study ofthe insurance industry.This collection of empirical and conceptual studieshas established a modest precedent for the compara-tive study of culture and effectiveness. The emergingparadigm generally relies upon the identification of alimited set of underlying values and traits (often identi-fied through inductive analysis), which are then mea-sured through survey responses. The culture measuresare compared, in most cases, to measures of businessperformance, defined in financial terms.While this approach has provided a way to begin toaddress the underlying research question, it has severallimitations. These limitations suggest a general strategyfor research on culture and effectiveness that has in-

ORGANIZATION SCIENCE / Vol. 6, No. 2, March-April 995


5/21

DANIELR. DENISON AND ANEIL K. MISHRA OrganizationalCulture and Effectiveness

fluenced this paper: First, building a more generaltheory of the cultural traits and values associated witheffectiveness; and second, beginning to test the impactsthat these traits have on multiple dimensions of effec-tiveness.

Study #1: Building a Grounded TheoryThe first study begins by defining a set of five firmsselected to facilitate theory building. Brief case de-scriptions of the firms are then presented followed byan analysis of the cultural dynamics that appear to bemost closely related to effectiveness. The cases wereselected based on a prior study by the first authorshowing a close relationship between the level of in-volvement and performance (Dension 1984). The focuson this single dimension of culture, however, raised abroader question: What other characteristics of organi-zational cultures may be related to effectiveness? Whatare the processes by which these traits are linked toeffectiveness? What other aspects of effectiveness maybe influenced by cultural traits?To address these questions, the findings of this ear-lier study were used to select the set of firms shown inFigure 1 for case studies. Neither the survey instru-ment (Taylor and Bowers 1972) nor the traits opera-tionalized were ideal for culture research. Nonetheless,this data source was attractive because it was one offew databases that offered any possibility of comparingculture and effectiveness. The survey data were col-lected from 34 organizations at one point in timebetween 1968 and 1980 and matched with performancedata for the five years following the survey date. Theindex scores for each organization were derived byaggregating individual responses to the organizationallevel. The measure of performance shown in Figure 1is return on assets, expressed as the percentile rankingof each firm within its respective industry. The overallcorrelation between involvement and industry-adjustedperformance for the data reported in Figure 1 is 0.42.This scatterplot of involvement and performance wasused to define an "interesting" set of cases to study.Three of the cases, falling along the diagonal, appearto confirm the involvement-performance results, point-ing to a high involvement-high performance organiza-tion (Medtronic), a moderate involvement-moderateperformance organization (People Express), and a lowinvolvement-low performance organization (DetroitEdison). The remaining two cases, in contrast, bothseem to disconfirm the general finding: Procter &Gamble appears as a high involvement, medium per-

ORGANIZATION SCIENCE /Vol. 6, No. 2, March-April 995

Figure 1 Selection of Case Studies3.1 MedtronicProcter3.0 &2.9 Gamble2.8 -

Level of 2.7- PeopleInvolvement 2.6 Express TexasCommerce2.5 Detroit * Bancshares2.4 Edison2.32.220 30 40 50 60 70 80 90 100

Percentile Ranking in Industry (ROA)

formance firm; while Texas Commerce Bancshares ap-pears as a high performance, low involvement firm.As a set, these cases serve two purposes. First, theyallow for a qualitative examination of the hypothesisthat involvement is a cultural trait that has an impacton effectiveness. The diagonal cases support that no-tion and can be used to examine the underlying pro-cesses by which involvement is linked to performance.The off-diagonal cases appear to contradict the in-volvement hypothesis, and thus are particularly likelyto reveal additional cultural traits that may be linked toeffectiveness. Both diagonal and off-diagonal cases canbe useful in developing a more general theory. Finally,these cases can be used to explore the multidimen-sional nature of effectiveness and to attempt to linkspecific cultural traits to specific dimensions of effec-tiveness. Thus, as Eisenhardt (1989) has advocated, ourcase studies began with a point of view and wereselected in a specific manner designed to facilitatefurther theory building.Case StudyMethodsEach case study began with publicly available sourcessuch as annual reports, lOKs, the popular press, andbusiness histories written on three of the five firms.These preliminary sources of background data wereanalyzed before we began negotiating access. The sec-ond step was to interview knowledgeable outsiders,such as researchers, journalists, consultants, and pastemployees, who provided insights about the firm andsuggestions for gaining access. When direct access wasslow in coming, this stage continued for several monthsand may have involved interviews with 30 to 40 individ-uals. When direct access came more quickly, this stage

207


6/21

DANIEL R. DENISON AND ANEIL K. MISHRA Organizational Cultureand Effectiveness

overlapped with interviews of the current members ofthe organization.Primary data for the case studies came from directinterviews, conducted by the first author, with a sampleof organizational members. A minimum of 25 to 30individuals were interviewed in each firm, and in somecases as many as 100. Direct on-site contact with theorganizations varied from a minimum of one week toseveral months. Contact with three organizations con-tinued over several years, and in two of these organiza-tions the first author served as a consultant to sustainresearch access over this time period. Individuals inter-viewed in each firm represented a cross-section of theorganization, including members from different hierar-chical and functional groups. The CEO or Presidentwas interviewed in four of the five organizations, alongwith other top executives. Particularattention was givento interviewing both new and old members of each firmand to identifying "storytellers" or "historians" whohad particular insights or perspectives.Successive interviews developed and tested anemerging picture of the organization's culture. Thequestions were informal, but in all cases focused on theinterviewee's identity and career history; the core val-ues of the organization and its "uniqueness," the powerand prestige of different subgroups; the impact of theorganization's history on its culture, and the intervie-wee's perception of the linkage between the culture ofthe organization and its effectiveness over time.A specific set of questions guided the interviews, butthe interviewer resisted the approach of asking thesame questions each time and summarizing the inter-views by counting and aggregating responses. Instead,each successive interview was used to expand under-standing of the organization. For example, after thefirst set of five to ten interviews, clear areas of overlapand redundancy (or sometimes conflict and inconsis-tency) began to emerge. At this point, the results weresummarized and served as a basis for the developmentof a new set of questions intended to develop anunderstanding of the organization that was bothbroader and deeper. This process of summarizationand refocus often happened four to five times during acase study. This approach also resulted in some"topic-specific snowball sampling." That is, one inter-viewee would recommend that another individual beinterviewed, or that relevant organizational records,reports, or memoranda be reviewed in order to clarifythe issues that we had raised. Whenever feasible, theserecommendations were followed.The first part of each interview was nondirective.After a brief introduction to the project, questions

208

were asked to allow the interviewee to project theirown experience and perceptions of the organizationalculture. Later in the interview, the questions becamesomewhat more directive, to help clarify the emergingpicture of the culture and integrate the interviewee'sperceptions into that emerging picture. This usuallytook the form of asking the interviewee to comment onthe observations and perceptions that other organiza-tional members had expressed and on the interviewer'sown observations. This approach was particularly use-ful in elaborating inconsistencies that arose or address-ing specific issues of redundancy or agreement.When possible, interviews were recorded, and laterreviewed and reanalyzed. Extensive notes were takenduring each interview, and again when the tapes werereviewed. Transcriptions were done to capture directquotes and other specific data. The identification ofintegrative themes within each case and comparisonsbetween cases often required further analysis of thetapes and transcriptions. Through this ongoing itera-tive process, the summary themes emerged. Draft re-ports of each of the cases also received extensivereview and comment by members of each of the orga-nizations.More extensive discussions of each of these casestudies have been presented by Denison (1990). Thesummaries presented below give a brief description ofeach organization's culture, with a primary focus oneach firm's core values and their expression throughstructure and action. The case summaries describeeach organization by presenting an historical back-ground, an overview of the business, and a descriptionof critical events and quotes that provide insight intothe organization's culture. Conclusions about the con-nection between culture and effectiveness at the end ofeach case reflect the perspective of organizational ac-tors as well as the researcher's perspectives. The goalof these case studies was to identify a broad set oftraits that can enable a fuller understanding of theimpact that culture has on effectiveness.The Case Studies

Medtronic. Since its founding in the early 1960s,Medtronic has been the premier firm in the cardiacpacemaker industry. The firm has grown steadily to asize of 5,000 employees and $650 million in annualsales in 1990. Building on their perennial dominance ofthe pacemaker industry, they now have a diversifiedline of products in the bioelectronic health care area.Medtronic has been driven from the very beginningby an explicit humanistic vision that goes beyond a

ORGANIZATION SCIENCE / Vol. 6, No. 2, March-April995


7/21


solely economic rationale for the existence of the orga-nization. In the early days, the main structure was avalue-based consensus about applying technical knowl-edge to the needs of patients, surgeons, and cardiolo-gists. The Medtronic culture was described by severalorganizational members as a "scientific-technical club"which was dedicated to preserving human life. Otherscommented that in these early days, "Medtronic didn'thave a mission, it was a mission." This informal struc-ture was sustained by constant contact and risk sharingin the operating room, led by the inventor and CEO,Earl Bakken, who was also the company's most com-pelling salesman. Bakken also made a practice (whichcontinued at least through the time of our study) ofhaving lunch with each new employee to convey thecorporate mission.This value-based system worked very well for 10 to15 years in an industry that had little competition andoffered steady and predictable growth. By the mid1970s, however, when the organization had grown to2,000 employees and $150M in sales, the company hadseveral serious product problems and became the focusof increasing regulatory attention on the pacemakerindustry. At the same time, a new CEO succeeded thefounder and began to create explicit bureaucratic con-trol systems designed to regulate the organization. Theculture, which had always been seen as an implicithumanistic consensus central to the organization's suc-cess, came to be seen as insular, overly concerned withits own internal processes, and (at times) in shock overthe discovery of its own fallibility when product prob-lems emerged.These changes also coincided with the onset of newmethods of payment in the health care industry, creat-ing price-based competition for the first time. Duringthis period, organizational members often contrastedthe "old culture" characterized by a humanistic pater-nalism, a strong sense of entitlement and commitment,and a clear sense of purpose, with the "new culture"and its emphasis on cost control, bureaucracy andaccountability. This period saw heightened conflictwithin the organization and coincided with several in-terruptions in their otherwise steady growth in salesand profits.

Another change of leadership in 1985 helped toreintegrate the business goals and the mission. Ratherthan being viewed as contradictory logics, the newCEO emphasized these as complementary values thatserved to integrate the organization. In fact, the newleadership of the organization appears to have used aheightened sense of the implications of the mission asa means to raise expectations and commitment among


organizational members. This reconvergence of thebusiness goals and mission has coincided with theintroduction of a new and successful product, and withrestored growth and profitability. In addition, thesepositive changes coincided with a fundamental changein the industry: all major competitors have now beenacquired by major corporations, leaving Medtronic asthe only independent firm.The Medtronic case provides some support for theidea that involvement is a cultural trait that is relatedto effectiveness. The voluntarism, commitment, andidentification with the humanistic mission were seen bythe organization's members (and the researchers) ashaving a positive effect on the organization's function-ing, particularly in the early stages. It is equally clear,however, that at several points during the organization'slife cycle, involvement coupled with success has led toa sense of entitlement and a preoccupation with inter-nal processes rather than external adaptation. As oneemployee put it, "we spent most of our time meetingand eating." Thus, the case illustrates that involvementcan lapse into insularity and have a limited, or evennegative impact on effectiveness.Another interesting element of the Medtronic cul-ture is its fundamental humanistic mission. This trait ofthe culture sensitized us to what might be called the"teleological nature of cultures." Purpose and destinywere strongly prescribed in noneconomic terms thatwent far beyond Medtronic's goals as a business. Unityof purpose had a powerful positive effect by creatingmeaningful work for individuals and a mission andsense of direction for the organization as a whole.Medtronic also provides an interesting example ofthe ebb and flow of cultural strength and its relation-ship to effectiveness. Out of the early stages of thecompany's history grew an agreed upon set of systems,behaviors, and meanings. Nonetheless, a primary taskof the CEO who succeeded the founder in 1976 was toestablish an organizational system that was capable ofcoordinating a larger firm and to counter the percep-tion that the firm was "under-managed." In so doing,"bureaucracy"was placed at odds with "the mission,"and, in the words of organizational members, a conflictbetween "old and new sub-cultures." This introducedconflict, inconsistency, and lack of integration. Interest-ingly enough, the most recent CEO, since 1985, hasattempted to reintegrate the mission and the businessand appears to have recreated a high level of consis-tency.Finally, the Medtronic case provides a caveat regard-ing the attribution of an organization's effectiveness toits culture. Medtronic, after all, is a company that first

209


8/21

DANIEL R. DENISON AND ANEIL K. MISHRA OrganizationalCulture and Effectiveness

created, and then dominated an industry through in-cremental technological innovation. While it is plausi-ble that the culture influenced this success, it is alsoclear that the opposite is clear: the success has influ-enced the culture. One interpretation of the diver-gence and reconvergence that occurred between thebusiness goals and the mission, for example, is thatwhen the product technology was successful, the busi-ness goals and mission appeared consistent; when thetechnology was unsuccessful, however, the mission andbusiness goals appeared to diverge.People Express Airlines. This innovative, low-fareairline was formed in 1981, shortly after the deregula-tion of the industry, and grew to nearly $1 billion insales in 1986 before its collapse and acquisition byTexas Air Corporation in 1987. Ironically, the airlinehad been created in 1980 by a group of executives andmanagers from Texas Air who left the airline led byDonald Burr, founder of People Express, and formerTexas Air President.The story of the People Express culture is quite wellknown. The plan was to create a highly egalitarianorganizational culture and apply many innovative hu-man resource practices. The new members of the orga-nization were trained and socialized in teams, andwork was designed to be done by teams. "Cross-utiliza-tion," or the rotation of teams and individuals throughdifferent jobs on a regular basis, created variety andchallenge in an industry with many repetitive jobs. Thesystem was also based on "self-management," or the

responsible autonomy of each employee as a managerof the firm's resources. All employees held the title ofmanager, owned stock, and shared in profits. After1985, when a formal hierarchy began to emerge fromthis essentially flat organization, employees even beganto elect their own supervisors and team leaders.People Express was also based on a mission extend-ing far beyond the realm of economics. The organiza-tion was designed to "unleash the power of the individ-ual" and create "flying that was cheaper than driving,"bringing air travel to the masses as never before. Theairline's "precepts" also included the idea that theorganization should serve as a role model for industryand the world. By doing so they would demonstrate thecompatibility of the growth and achievement of individ-uals with the creation of an effective organization.Other aspects of the People Express culture are notas well known. Rather than being a homogeneousculture, for example, pilots and flight attendants formedsubcultures that had very different reactions to themanagerial ideology. Among the strongest examples of

210

these differences was the pilots' tendency to refer tothe organizational ideology as "kool-aid" (a sardonicreference to the mass suicide of the followers of Rev.Jim Jones in Guyana in 1978), and to contrast thePeople Express system with "a real airline." Pilotswere generally older, with much more airline and mili-tary experience, and were more likely to reject thecollective values of the organization in favor of theirown individual interests, and more likely to rebelagainst the "cult of charisma" that formed around DonBurr. Flight attendants, in contrast, were both maleand female and had little airline experience, and inmost cases no military experience. As a group, theytended to support the organizational ideology morethan the pilots. While significant socialization to theorganizational ideology occurred, these were nonethe-less strong occupational subcultures.The People Express system worked very well forthree to four years. By 1985, the airline had grown to asize of 2,500 employees and $500M in annual salesbased on this ideology and culture. After 1985, how-ever, the system began to show signs of strain andincreasing competition placed limits on the amount oftime and organizational slack that the airline coulddevote to making its unique system work. For example,since all managers were promoted from within, theinternal labor market could not produce experiencedmanagers at the rate that they were needed. The finalblow came when major airlines such as United andAmerican introduced variable pricing reservations sys-tems that allowed them to match People's low price onsome seats (with restrictions), while maximizing rev-enues by selling other seats on the same flight at muchhigher prices.In the beginning, People Express created a powerfulcultural system based on shared values transmittedthrough extensive socialization. The PEX culture wasstrong and pervasive, and held a powerful vision forthe future. Involvement and empowerment were criti-cally important to the implicit coordination and fiercecommitment that characterized the start-up. The ex-pectations for involvement were extremely high, andthis seems to be the best explanation of the fact thatPeople Express appears in Figure 1 as a mediuminvolvement firm. Even though "actual" involvementmight seem very high, it did not always meet theexpectations that the organization's members held.Trying to explain the ineffectiveness of the PeopleExpress culture also provides several insights. Oneinterpretation is that the leaders and members of theorganization placed more emphasis on their socialtechnology and internal integration than on external



9/21


adaptation. Symptoms of their internal focus includethe conflicts they experienced between their structuralevolution and remaining true to the organizationalideology. The limits of implicit coordination had beenreached, but the evolution of more complex coordinat-ing structures seem to be impeded by the organiza-tional ideology. Other examples of their relativeneglect of external adaptation include the assumptionthat they would be able to absorb and integrate newlyacquired Frontier Airlines, and their difficulty in re-sponding to the development of variable pricing reser-vations systems by their major competitors. Their orga-nizational innovations did not, in themselves, seem tofail, but they did direct the attention of leaders andmembers of the organization away from changes thatwere taking place in the business environment.Detroit Edison. With its origins in the consolidationof the utility industry in the early decades of thiscentury, Edison's history began with 50 years of steadygrowth; a predictable doubling in size every ten years.The company grew rapidly along with the region'seconomy and became one of the nation's leading utili-ties.During this long growth period, the company wasalmost entirely controlled by engineers who saw themission of a public utility as the continuous building of"bigger and better power plants." The organizationwas a "family"with a highly stable work force in whichline workers often socialized with the Chairman andtop executives. Although present-day workers still re-flect some of the "rough and ready" posture of an erawhen they moved from town to town stringing uppower lines in a rapid expansion of the system, todaythey are far more likely to hold their jobs because ofthe security they offer. Nonetheless, when the powergoes out, a sense of those traditional values stillemerges.Edison is an organization with well-defined author-ity. As several members of the organization said, "inthis organization, everyone knows who they report toand what their job is." The organization was oftendescribed as functioning "like a machine," and wasstaffed by a very stable population of employees, pro-moted from within the organization in almost all cases.Historically, the organization has also been very male,and until the early 1960s there was a rule that femaleemployees who got married had to quit their jobs.The changes of the 1970s created turbulence withinthe organization. The energy crisis created a decline insales and a decline in the growth of the region theyserved. Nuclear power brought a new regulatory pres-


ence and an increasing dependence on the expertise ofoutside contractors. A court ruling on affirmative ac-tion required that Edison adopt a strict quota system ofhiring and promoting minorities until the proportion ofminority employees in the company more closelymatched the proportion of minorities in the populationthat the company served. This forced integration ofEdison's work force was regarded by management assomething that was "done to" the organization, ratherthan a positive, proactive step on their part. The lackof ownership of this issue and the advent of integrationitself unfortunately undermined the sense of "family"that had been a part of the Edison culture for decades.Edison was hit by all of these changes at once, withseveral results: First the mission of the company gradu-ally changed from "building bigger and better powerplants'' to "creating safe and efficient options for con-sumers." Second, adaptation to these changes requiredthat the organization be led by a combination of engi-neers and politicians (who held radically different def-initions of "reality") and that the debate between thesetwo subcultures be carried out within the organization.Third, many employees within the organization resistedthe changes, preferring instead to go back to a simplerera.This set of circumstances makes Edison an interest-ing study of what happens when an organization'smission and identity are altered by changes in theirenvironment. Traditionally, the Edison culture has val-ued authority, predictability, technical skill, and themanagement of stable expansion. Involvement, to thedegree that it exists, is of secondary importance tostability and consistency. Redirecting the organizationto respond to politically defined stakeholders has beentraumatic, because it has meant a fundamental reexam-ination of the basic mission, and a shift from an inter-nal to an external locus of control.

Procter & Gamble. This 150 year-old consumerproducts giant is well-known as an innovator in suchareas as brand management, profit sharing, advertisingand promotion, and innovative work design. Concen-trating first on soap and then for decades on a range of"high quality consumer products found within everyhome," the corporation is now a $20 billion companywith 73,000 employees worldwide. The strong, method-ical culture is seen by many current and past employ-ees as a key factor in their steady doubling in size everydecade. While P&G has been quite profitable, perfor-mance is remarkable not so much for its level, but forits predictability.

211


10/21


At the core of the P& G culture is a highly rational,objective view of the world. Research, on both marketsand products, is central to all decisions. When makingor discussing a proposal, one must always "know thenumbers." A central goal has always been to develop atechnically superior product that will win in a blindtaste test. A classic example of this logic is Olestra, anew synthetic cooking oil and food ingredient free offats, calories, and cholesterol. It grew out of years ofbasic research, and the product will not only be solddirectly to consumers, but also will serve as a techni-cally superior base for other P&G food products.Heavy attention is paid to socialization by P&G.New employees in the brand management organiza-tion, for example, are socialized as a cohort, and manysee this peer competition and cooperation as the realsource of learning and motivation for new members.New assignments, which usually included a period ofstructured training, are alternated on a regular basis inshort cycles of 6 to 24 months. P&G careers, in gen-eral, are like a tournament in which the memberscontinually move up or out. Requirements for perfor-mance and conformity to the P&G way lead many newrecruits to leave the organization after a few years.Ironically, P&G often trains the key employees ofmany of their competitors, although some in the orga-nization will claim that they "never lost an employeethat they wanted to keep."Strong emphasis is also placed on written communi-cation. Lessons on writing a memo in the proper P&Gform are of central importance in a new recruit'ssocialization. This system results in efficient communi-cation in a common "language," and the creation of awritten corporate record of all significant events. Thesystem is intended to be independent of any one per-son, and everything important can always be quicklyreconstructed from the records. This is one of manysystems used to reinforce the idea that work at P&G isthe product of the organization, and not of any oneindividual.In manufacturing plants P&G is highly innovative, iflargely secretive, in their design of organizations. Prin-ciples of sociotechnical design are used to build pro-gressive, high commitment systems that place a highlevel of autonomy and responsibility on workers. Theywere among the first American corporations to seethese innovations as a source of competitive advantage,and to see them as an outgrowth of their organization'straditional assumption that the interests of the individ-ual and the organization overlap. This assumption ledto the use of innovative practices such as profit sharingas early as 1887.

212

A qualitative examination of P&G's culture supportsthe idea that it is a high-involvement organization, butthe context in which involvement occurs is far morestructured than in the Medtronic and People Expressexamples. Involvement seems to reflect the high over-lap of interest between the individual and the organiza-tion more than voluntarism and autonomy. Further-more, this involvement takes place within a structuredcompetition created by the organization, in a contextwith a high degree of normative integration, commit-ment, and a common language and symbolic system.P&G is a classic example of a "strong culture"system. As many past and current members of theorganization have noted, this characteristic of the cul-ture has both positive and negative impacts on effec-tiveness; P&G is an organization that seldom makesbig mistakes, but is often beaten to the punch bysmaller, faster companies. Their historical limitationhas been a difficulty in moving quickly primarily be-cause of their commitment to research, objectivity, andmethodical review. Their key capabilities are oftendescribed, as "science, not art." The P&G system isalso generally seen as being very well suited to theconsumer goods mass market where objectivity and amethodical approach pay off, but slow and ponderousin fast moving markets that are driven primarily bytaste and fashion. Recent acquisitions and expansionsinto food, health, and beauty products may test theadaptability of P&G culture. As one past employee putit, "the question is, does the 'corporate gene pool' havethe variety necessary for future adaptation?"TexasCommerceBancshares. With its origins in thefinancing of the Houston cotton trade in early 20thcentury, Texas Commerce Bancshares was formed bythe merger of two historic Houston banking institutionsthat led the Houston region through the Great Depres-sion. The bank was transformed in the 1960s by a newemphasis on marketing and managing by the numbers,and has become an elite, aggressive, and conservativemiddle market regional bank. It has been one of thepremier Texas banks through both boom and bust,often outperforming the larger money-center banks.More recently, in 1986, TCB was one of the first Texasbanks to take advantage of the change in U.S. inter-state banking laws and merge with Chemical Bank ofNew York.The bank has been led for the past 25 years by BenLove, a charismatic and demanding CEO who wasequally skilled at devising internal controls, motivatingmanagers and executives, and selling and marketing.Love placed ultimate priorityon "the numbers"(shared



11/21


throughout the organization in a monthly "blue book"),and has stated that "organizing human resources inpursuit of statistical objectives is uppermost in myphilosophy of management." Love's smooth but force-ful style was also a primary source of motivation andoften was exhibited in public settings such as loancommittee meetings and officers' meetings. Numerousstories describe his unrelenting style, formidable witand memory, and the dire consequences of presentinga proposal without knowing it inside out.The culture of the bank reflects an interesting para-dox: Leadership comes from the top down and respectfor that authority is high. It is virtually impossible totalk about the culture of the bank without the discus-sion focusing on Ben Love. Yet at the same time, thestructure of the bank is relatively flat (Love once had70 direct reports) and much of the work is done inlarge committees. Loan committees, for example, oftenhad 5 to 10 regular members, 5 to 10 who might bepresenting loans, and perhaps 10 to 15 observers in anouter ring around the conference table. Until recently,approval of any loan over $50,000 required unanimousapproval of the loan committee-a single dissentingvote could veto the proposal. These large committeesprovided a broad forum which served as a way tosocialize new members, communicate the organization'sculture, and disseminate best practices.The Texas Commerce culture is also influenced bythe strong regional culture of Texas. Rugged individu-alism, patriotism, and respect for authority run deepand tend to be projected onto Love as the symbolicleader. A 6'6" patrician figure, Love is an authentic"Texan hero" to which many of the organization'sstrengths are attributed. Perhaps the best statement ofLove's influence on the organization and the reactionof organizational members to that influence came fromone newly appointed loan officer who said, "I'm proudto walk in Ben Love's shadow."TCB's top down style, uniformity of practices, andstrict financial controls and objectives are all examplesof a high level of normative integration and consis-tency. The culture is pervasive and ensures that allmembers of the organization learn the "TCB way" andconform to it. At the same time, the system is alsooriented toward individual achievement and the ac-complishment of objective goals. The strength of thisculture and the ability of the holding company totransmit it to newly acquired member banks appears tohave been instrumental in the organization's success.The more difficult question to answer about TexasCommerce is how adaptable the system can be. Theirrecent merger with Chemical Bank of New York im-


plies that TCB must quickly reach beyond the middlemarket into consumer and investment banking by ex-tending an organizational system and underlying set ofassumptions that have been highly adapted to that onemarket. The strong culture of the bank seems to com-plicate that problem by instilling a belief in the existingsystem that may limit its future ability to adapt.Developing HypothesesThe case studies on the diagonal in Figure 1 provided astarting point for developing hypotheses about cultureand effectiveness. These three case studies generallyseemed to support the idea that organizations withhigher levels of involvement were more effective. Dur-ing the time period reflected in Figure 1, for example,involvement in Medtronic was very high and was seenby organizational members as highly instrumental tothe company's objectives. High involvement fostered astrong sense of psychological ownership and commit-ment to the organization and its goals. Detroit Edisonalso fits the involvement-performance pattern, but inthe opposite way. Their low-involvement bureaucracyseemed to be an obstacle in the organization's struggleto respond to fundamental environmental change. Thestable internal core prescribed by their history limitedEdison's ability to adapt.The third case on the diagonal, People ExpressAirlines, is more complicated to interpret. It appears inFigure 1 as a moderate-involvement, moderate-perfor-mance firm, but in the case studies it appears to haveexceptionally high involvement and was effective as astart-up organization. The best explanation for thesesomewhat inconsistent findings may be twofold: First,"actual" involvement or involvement compared to otherfirms may have been very high, but not nearly as highas expectations regarding involvement. Thus, the ques-tionnaire responses, in comparison to baseline expecta-tions, were only average. Second, the performancemeasures of profitability, also average, show that Peo-ple Express, when the data were collected, was abreak-even firm in a break-even industry. The organi-zation's primary achievement, growth and innovation,were not reflected in Figure 1. The People Expresscase study may thus provide some support for thehypothesis that involvement is linked to growth, ratherthan return on assets as suggested by Figure 1.Procter & Gamble, an "off-diagonal" firm, provides adifferent example of involvement. The P&G case por-trays an organization with a high level of involvement,primarily because of autonomy and a high overlapbetween individual and organizational goals. Involve-ment in P&G, however, takes place within a highly

213


12/21


structured setting, which is far less voluntary than ourearlier examples. As a result, there seemed to be less"ownership" of the system than in Medtronic or Peo-ple Express, even though most organizational memberswere highly involved in a system perceived as a positiveintegration of individual and organizational interests.Interestingly enough, the case also shows that P&Gwas not highly profitable, but its stable rate of growthover time was impressive. P&G thus appears to sup-port the idea that involvement may have greater impacton other aspects of effectiveness, such as long-termgrowth.Thus, our first hypothesis suggests that high levels ofinvolvement and participation create a sense of owner-ship and responsibility. Out of this ownership grows agreater commitment to the organization and a growingcapacity to operate under conditions of autonomy.Increasing the input of organizational members is alsoseen as increasing the quality of decisions and theirimplementation. This theme is a classic in organiza-tional theory, reflected in the work of McGregor (1960)and Likert (1961, 1967), as well as Argyris (1964), Ouchi(1981), Peters and Waterman (1982), Lawler (1986),and Walton (1986). Thus, our first hypothesis:

H1. Involvement is a cultural trait which will bepositively related to effectiveness.While P&G can be used to illustrate involvement, italso illustrates a different cultural trait, normative inte-gration, that may be related to effectiveness. This trait

is also apparent in the other off-diagonal firm, TexasCommerce Bancshares. Effectiveness in these firmsappears to spring from normative integration or consis-tency: the collective definition of behaviors, systems,and meanings in a integrated way that requires individ-ual conformity rather than voluntary participation.Texas Commerce Bancshares, for example, developedan internally consistent system that served to definemarkets, identify lending opportunities, and approveand monitor loans. The system was integrated around asmall number of key values; TCB as an elite middle-maiket bank, well-defined areas of personal responsi-bility, individual rewards, and management by thenumbers.This trait of consistency also appears in the otherorganizations. For example, a high level of sharedmeaning, and a strong sense of the "Medtronic" or the"TCB" way was typically a central source of integra-tion, coordination, and control. Each organization hada developed mind-set and set of organizational systemsthat could be projected on ambiguous situations. The

214

root of this trait is the concept.of normative integrationor the pervasiveness of a consensual system of behav-ioral control. This characteristic has been popularlyrecognized as a "strong" culture.This trait also helps in understanding subcultures.For example, in Medtronic, the "old" culture centeredaround the concept of organization as a humanisticmission, which eroded as the firm became more explic-itly managed as a business organization. A "new",bureaucratic subculture and complementary set of con-trol systems began to serve as the focus of integration.Only recently has the inconsistency between thesecompeting subcultures begun to reach some resolution.People Express and Detroit Edison also provide usefulexamples of the tensions between strong subculturesand overall integration. In People Express, pilots andcustomer service managers held different ideologiesand were often in conflict. In Detroit Edison, "en-gineers" and "politicians" struggled over which versionof "reality" they should respond to. The culture ofthese firms showed the tension of differences and theneed for integration.The concept of consistency also helps to address thecritique of "unitary" organizational cultures that hasappeared in the literature (Martin et al. 1983, Allaireand Firsirotu 1984, Rose 1988). Subcultures coexist inall organizations and an "organizational" culture mayexist only in the form of a set of assumptions, beliefs,and practices regarding the integration of the subpartsinto a whole. Sources of integration range from alimited set of rules about when and how to agree anddisagree, to a "unitary" culture with high conformityand little or no dissent. Nonetheless, in each of thesecases, the consistency and integration of the system is asalient trait of the organization's culture.The culture literature often tends to equate involve-ment and consistency, and does not emphasize thatthey are different routes to gaining the commitment oforganizational members. Equating democratic decisionprocesses and high levels of voluntary participationwith "Disney-like" systems of socialization designed toforge a collective mind-set overlooks a fundamentaldistinction. In high-involvement systems, the socialdefinition of reality is created by individuals in aninductive manner-each individual's stamp appears onat least some corner of social reality-and while theprocess may be instigated by a leader, it is not pre-scribed in detail. In contrast, in the highly consistentcultures, such as TCB or P&G, social reality comes"predefined." The existing managers, executives, andhigh-status individuals are the agents of socialization,and newer lower-status individuals are the subjects.



13/21

DANIELR. DENISON AND ANEIL K. MISHRA OrganizationalCulture and Effectiveness

The contrast between "agents" and "subjects" is highin these "strong-culture" organizations. In a high-involvement organization, this contrast between agentsand subjects is generally much lower.Thus, the concept underlying the second hypothesisis that an implicit control system, based upon internal-ized values, can be a more effective means of achievingcoordination and integration than external control sys-tems relying on explicit rules and regulations (Pascale1985, Weick 1987, Saffold 1988, O'Reilly 1989). Thebasis for an implicit control system is a normativesystem of regulation that can be projected by organiza-tion members on ambiguous or ill-defined situations.Interesting enough, normative integration appears tolie at the root of both the spontaneous emergence ofvalue-based cultures such as Medtronic or People Ex-press exhibited in their early stages and the inbredbureaucratic cultures shared by Detroit Edison and toa lesser degree, by Procter & Gamble and Texas Com-merce Bancshares. Thus, our second hypothesis:H2. Consistency, or the degree of normative integra-tion, is a cultural trait that will be positively related toeffectiveness.The next step in the theory building process grewfrom the observation that both involvement and consis-tency take as their primary focus the internal integra-tion of the organization. Thus, these two conceptsalone could not capture organizational members' fre-quent statements (and our frequent observations) aboutthe linkages between their organizations and their en-vironments. As cultural ecologists (Steward 1955, Ben-nett 1976, Rappaport 1979), and organizational schol-ars (Schein 1990) have noted, culture is one of theprimarymeans by which social organizations are linkedto their environments.The concept of consistency, in particular, made thispoint apparent because of its close relationship tochange and adaptation. The positive influence of con-sistency is that it provides integration and coordina-tion. The negative aspect, however, is that highly con-sistent cultures are often the most resistant to changeand adaptation. All of the cases showed this to somedegree.One of the best positive examples of adaptabilitycomes from the early stages of Medtronic's historywhen they worked closely with cardiologists, sharingthe risk of each new implant by being in the operatingroom and interacting directly with patients and sur-geons during operations. The organization rapidly re-acted to the needs of the physicians and patients, and


did so in a way that was totally compatible with theorganization's underlying character. Another exampleof more limited adaptability comes from P&G's abilityto apply the logic of brand management to new prod-uct settings without making fundamental changes inthe character of the system. A third example, DetroitEdison, illustrates how widespread changes in the busi-ness environment can outstrip an organization's capac-ity to adapt, and amplify the uncertainty over theunderlying character of the firm. Finally, People Ex-press had great difficulty overcoming the basic assump-tions of an entrepreneurial culture as it grew larger.The psychiatric term alloplastic-the capacity tochange in response to external conditions withoutabandoning one's underlying character-has been usedto describe this phenomenon at the organizational level(Nicholls 1985). Adaptability, in principle, could alsorefer to the capacity to redefine underlying character inresponse to large-scale change, but there were noexamples of this in the five organizations studied. De-troit Edison, for example, experienced fundamentalchange in its environment, but its traditional underly-ing character remained quite stable.Thus, the adaptability hypothesis asserts that aneffective organization must develop norms and beliefsthat support its capacity to receive and interpret signalsfrom its environment and translate these into internalcognitive, behavioral, and structural changes (Starbuck1971, Kanter 1983). Cases in which the foundations ofnormative integration become detached from the exter-nal environment often develop into insular bureaucra-cies, and are unlikely to be adaptable. Calori andSarnin (1991), for example, found that companies whichvalued adaptation were likely to hold ambitious objec-tives, give priority to the satisfaction of clients, and bewilling to try new ideas. These values and practiceswere closely related to the growth of the firms theystudies. Kotter and Heskett (1992) also found a closerelationship between adaptability and firm perfor-mance. Thus, our third hypothesis:H3. Adaptability, or the capacity for internalchangein response to external conditions, is a cultural trait thatwill be positively related to effectiveness.Our final hypothesis came from the observation thatseveral of the organizations were effective because theypursued a mission combining economic and noneco-nomic objectives, which provided meaning and direc-tion to organizational members. The mission hypothe-sis contrasts with the adaptability hypothesis, in that itemphasizes the stability of an organization's central

215


14/21


purpose and de-emphasizes its capacity for situationaladaptability and change.The importance of mission to culture and effective-ness was also supported by the observation that themost critical crises in each organization came when thebasic mission was questioned or altered. Each ofthe organizations provided a compelling example of theclose relationship between the overall purpose anddirection of the firm, and the meaning held by each ofthe organizational members. This loss of meaning anddirection seemed to coincide with significant losses ofmomentum and effectiveness.Several examples help to illustrate. Medtronic, forexample, began with an implicit humanistic missionthat served as a powerful integrating force. After about15 years of success, product problems, competition,and the leadership transition from the original founderto "professional" management all combined to bringgreat pressure on the firm to behave as a "business"organization, rather than a humanistic mission. Con-flict emerged between the old "humanistic culture"and the new "business culture" and seemed to threatenthe integrity of the organization as a whole.As another example, in Detroit Edison the missionchanged in the 1970's from "building bigger and betterpower plants" to "providing the public with safe, effi-cient energy choices." This fundamental realignmenthas taken years to translate throughout the organiza-tion as responsiveness to external stakeholders gradu-ally replaces a technical concern with producing elec-trical power.

Relatively few authors have written directly on thistopic (Selznick 1957, Torbert 1987, Bourgeois andEisenhardt 1988, Robbins and Duncan 1988, Hameland Prahalad 1989, Westley and Mintzberg 1989, West-ley 1992), but most have agreed that a sense of missionprovides two major influences on an organization'sfunctioning: First, a mission provides purpose andmeaning, and a host of noneconomic reasons why theorganization's work is important. Second, a sense ofmission defines the appropriate course of action for theorganization and its members. Both of these factorsreflect and exemplify the key values of the organiza-tion. Thus, our fourth and final hypothesis:H4. Sense of mission or long-termvision is a culturaltrait that will be positively related to effectiveness.

A Framework or Integratingthe HypothesesFigure 2 organizes these four traits into a frameworkdesigned to acknowledge two contrasts: the contrastbetween internal integration and external adaptation,

216

Figure 2 Theoretical Model of Culture Traits

Externalto Adaptability Mission

InternalIntegration Involvement ConsistencyChange StabilityFlexibility Direction

and the contrast between change and stability. Forexample, involvement and consistency have as theirfocus the dynamics of internal integration, while mis-sion and adaptability address the dynamics of externaladaptation. This focus is consistent with Schein's (1990,p. 111) observation that culture is developed as anorganization learns to cope with the dual problems ofexternal adaptation and internal integration. In addi-tion, involvement and adaptability describe traits re-lated to an organization's capacity to change, whileconsistency and mission are more likely to contributeto the organization's capacity to remain stable andpredictable over time.A second reason for integrating these four traits intoa framework is to qualify the apparent linear form ofthe four basic hypotheses. Taking each hypothesis in-dependently neglects important data revealed by thecase studies. For example, Medtronic showed clearlyhow involvement, over time, evolved into insularity andentitlement. Detroit Edison showed how internal con-sistency could be an obstacle to adaptation, or to aredefinition of their underlying mission. These exam-ples acknowledge that there may be many contradic-tions among the individual hypotheses. Well-integratedsystems are often the most difficult to change, andhighly adaptable systems may be less likely to achieve

the high rates of efficiency or common purpose associ-ated with consistency and integration. In a similarsense, mission and involvement may be contradictory:the meaning and direction established through an or-ganizational mission may limit the involvement of somemembers. As many organizations have discovered,declaring a new organizational mission does not neces-sarily imply the support and commitment of organiza-

ORGANIZATIONCIENCEVol. 6, No. 2, March-April 1995


15/21


tional members, and high involvement among anorganization's members does not necessarily imply aclear sense of direction.One way to acknowledge these observations wouldbe to pose each of the four hypotheses in a curvilinearform. This approach, however, would introduce theintractable problem of "how much is too much" toeach of the hypotheses. The alternative, and the ap-proach taken here, is to argue that each trait is "bal-anced" by the others. Thus, the excessive involvementthat led to insularity and entitlement in Medtronic isnot acknowledged by a curvilinear involvement hypoth-esis; but rather as an absence of one of the other traitssuch as mission or adaptability. Detroit Edison waslimited, not by the fact that it had too much consis-tency but rather because it had too little adaptabilityand involvement. While we are not able to provide adirect test of all of these ideas in the current study,representing them in the same framework recognizesthe inherent interdependence and contradiction ofthese traits.As such, this framework bears similarity to otherauthors' attempts to present general theories of func-tioning and effectiveness (Parsons 1951; Katz and Kahn1966, 1978), as well as more recent authors such asMitroff (1984), Lewin and Minton (1986), and Quinn(1988) who have recently argued for the importance ofunderstanding the paradoxes and contradictions thatare an integral part of organization life. As Quinn(1988) has noted, it is the balancing of competingdemands that distinguishes excellent managers and or-ganizations from their more mediocre counterparts.This theory suggests a similar logic for understandingorganizational culture.

Study #2: An Exploratory Testof the ModelThe second study conducts an exploratory test of thefour hypotheses, examining the linkage between thesefour cultural traits and effectiveness. Survey measuresof the traits were obtained from top executives in 764organizations and compared to a set of effectivenessmeasures including subjective and objective measuresof profitability, quality, sales growth, satisfaction, andoverall effectiveness. Compared to Study #1, this re-search design takes an opposite set of risks: richness ofdetail is traded off for the scope of generalization. Ifthe measures are too superficial, only represent a topmanagement viewpoint, or ignore the importance ofcontextualized meaning, then it is highly unlikely thatthe measures will be related to effectiveness. If theORGANIZATION SCIENCE / Vol. 6, No. 2, March-April 995

measures do show an interpretable relationship to ef-fectiveness, however, it signals a fundamental relation-ship strong enough that the impacts can be discernedthrough these limited measures.From the hypotheses described above, questionnaireitems were developed for each of the four traits. Next,we identified a population of firms in the industrialmidwest and drew a probability sample of firms fromthe files of the state employment security commission(Birley 1984, Denison and Hart 1987). The sample wasdrawn by SIC code from five major industries:(1) manufacturing, (2) business services, (3) finance,insurance, and real estate, (4) retail, and (5) wholesale.The probability of selecting an individual firm wasproportional to its size, such that larger firms weremore likely to be selected than smaller firms.After verifying the names and addresses of all firmsin the sample, a survey questionnaire, follow-up letter,second survey, and follow-up postcard were mailed out,and at one week intervals. Surveys were addressed tothe chief executive officer or top executive in eachorganization, and the respondents typically were thepresident CEO, COO, or CFO. After these four mail-ings, telephone calls were made to nonrespondents. Alldata collection was completed in 1989. From 3,625surveys sent out, 764 completed surveys were returnedfor an overall response rate of just over 21 percent.Since our intention was to conduct an exploratory testwith data from a large number of firms, this responserate was judged to be adequate. Furthermore, as Hen-derson (1990) has shown, a response rate of 20 to 30 %is fairly typical for a mail-out survey to a large sampleof firms.Response bias by industry sector and firm size wasminimal. Response rates were quite constant acrosscategories, generally varying no more than 2 to 3%from the overall response rate. There are two excep-tions to this pattern: The very smallest category offirms in the sample, those with 1 to 5 employees, hadonly a 16% response rate. In addition, wholesale firmsalso had a low response rate, with only 12% of thesurveys returned. Manufacturing firms and medium-to-large firms has the highest response rates.MeasurementFactor analysis and multidimensional scaling were usedto assess the validity of the questionnaire items devel-oped to measure the four cultural traits. Figure 3presents a multidemensional scaling analysis (Takaneet al. 1977) of eight items that shows good convergentand discriminant validity for the four traits, presentedin two dimensions as suggested by the theoretical

217


16/21


Figure 3 Multidimensional Scaling Analysis of Culture Items

2.1 -

I' \involvement1.0

-0.0 - - - - -

adaptability I consistency-1.0-- 8i

mission-2.1l I I I I. I I I I-2.5 -1.5 -0.5 0.5 1.5 2.5

Table 1 Correlation Matrix of Survey Items(N = 667)

Index Item 1 2 3 4 5 6 7 8Involvement 1. Input

(0.76) 2. Collaboration 0.62Consistency 3. Agreement 0.35 0.42

(0.79)4. Predictability 0.28 0.37 0.65

Adaptability 5. Change 0.24 0.25 0.26 0.24(0.63)

6. Responsiveness 0.29 0.30 0.33 0.20 0.46Mission 7. Direction 0.22 0.26 0.40 0.36 0.32 0.27(0.81) 8. Vision 0.28 0.26 0.40 0.35 0.25 0.30 0;68

model. Table 1 presents the correlation matrix of theitems along with alpha coefficients for the four indexes.A complete listing of the items is included in theappendix.The analyses presented here examine the relation-ship between the four culture measures and the effec-tiveness of the organizations. The effectiveness mea-sures fall in three categories: (1) a set of five question-naire items asking respondents to compare their ownfirm's performance to the performance of similar firmson the dimensions of profitability, quality, sales growth,employee satisfaction, and overall performance:(2) return on assets (ROA) for the most recent year(1989) as well as the average for the preceding threeyears (1987-1989); (3) reported sales growth for themost recent year (1989) and the average growth for thepreceding three years (1987-1989).ResultsTable 2 presents the correlations between the cultureindexes and the subjective effectiveness items. Theseresults are predominantly positive, with 15 of 20 corre-lations significant at the 0.001 level, and many falling inthe 0.25-0.35 range. The culture measures prove to beweak predictors of sales growth and profits, but strongerpredictors of quality, employee satisfaction, and overallperformance. Measurement error could potentially in-flate these correlations, but it is somewhat more dif-ficult to attribute all results to measurement errorwhen the correlations vary in magnitude from 0.01to 0.35.After reviewing these initial results, we began check-ing to see how robust they were with respect to poten-tial confounding effects such as industry, size, age,stage of development, and so on. In additions to check-ing for any bias resulting from the position of therespondent, we also analyzed the data for top executiverespondents (80% of the sample) separately from the

Table 2 Correlations Between Culture Traits and Subjective Effectiveness Measures(N = 674)

Sales Employee OverallGrowth Profits Quality Satisfaction Performance

Involvement 0.04 0.01 0.26*** 0.32*** 0.16***Consistency 0.03 0.13*** 0.29*** 0.33*** 0.29***Adaptability 0.08** 0.06 0.20*** 0.21*** 0.18***Mission 0.22*** 0.10*** 0.18*** 0.27*** 0.35***

*p < .05**p < .01

***p < .001

218 ORGANIZATION SCIENCE / Vol. 6, No. 2, March-April 1995


17/21

DANIEL R. DENISON AND ANEIL K. MISHRA Organizational Culture and EffectivenessTable 3a Correlations Between Culture Traits

and Return on AssetsTotal Sample (N = 220) Sub-Sample (N = 24)

ROA ROAROA 3-Year ROA 3-Year1989 Average 1989 Average

Involvement 0.00 0.09* 0.35** 0.23Consistency 0.11* 0.14** 0.36** 0.31*Adaptability 0.07 0.12** 0.25 0.22Mission 0.11* 0.10* 0.46*** 0.55***Table 3b Correlation Between Culture Traits

and Sales GrowthTotal Sample (N = 409) Sub-Sample (N = 50)

Sales SalesSales Growth: Sales Growth:

Growth: 3-Year Growth: 3-Year1989 Average 1989 AverageInvolvement 0.07* 0.05 0.25** 0.25**Consistency 0.05 0.02 0.18* 0.18*Adaptability 0.08** 0.08** 0.25** 0.24**Mission 0.08** 0.08** 0.22* 0.22*

*p < .10p < .05p < .01

rest of the sample. None of these tests, however, hadany significant impact on the results presented inTable 2.Table 3a presents the results for the first of the twoobjective measures of performance, return on assets(ROA). The table presents correlations between thecultural traits and 1989 ROA, and for average ROAfor the last three years. The correlations are presentedfor the entire sample, and for a subsample of largerfirms with top executive respondents.The correlations for the total sample are positive,but quite weak. None of the correlations with 1989ROA are significant at the 0.05 level, and only two ofthe four correlations with 3-Year Average ROA reachthat level of significance. More importantly, the overallmagnitude of the correlations is quite small, with thelargest correlation reaching 0.14.A quite different picture appeared, however, whenwe examined only those organizations with greaterthan 100 employees that have a top executive as re-spondent. For this subsample, all of the correlations


are greater than 0.20 and range from 0.22 to 0.55. Halfof the correlations are significant at the 0.05 level; twoof the eight correlations are significant at the 0.01.Strong correlations with all four traits are observed,and from these data it appears that mission is thestrongest predictor of performance.The analyses in Table 3 rely on a much smallernumber of cases than do the analyses presented inTable 2. Only about one-third of the firms reportedenough usable financial data to compute ROA forthree years, leaving only 220 firms total for the ROAanalyses. Since only about 15% of the firms in thesample have greater than 100 employees, and onlyabout 80% of the firms in the sample had the CEO orCFO as respondent, this left only 24 cases for the ROAsubsample analyses.The second objective performance measure, salesgrowth, offers slightly more data, since nearly 50% ofthe firms reported enough sales data to compute aver-age sales growth for three years. Table 3b presents thecorrelations between the four culture measures andsales growth. The results present a similar picture tothat presented in Table 3a. The results for the totalsample are quite modest, both in size and in signifi-cance, but the results for the subsample of larger firmswith top executive respondents shows much strongerresults. Overall, the four culture traits seem to beslightly stronger predictors of ROA than of SalesGrowth.The results for the objective data support the ideathat different cultural traits are related to differentcriteria of effectiveness (Pennings 1976). The data forthe sample of large firms shows that profitability crite-ria are best predicted by the stability traits, mission andconsistency, while the sales growth criteria are bestpredicted by the flexibility traits, involvement andadaptability. The analysis of the subjective effective-ness data presented in Table 2 also shows a similarpattern.Comparing the results for the subjective and objec-tive performance criteria also reveals an interestingfinding: the correlation between the four traits and theobjective criteria are size dependent, but the correla-tions between the four traits and subjective criteria arenot. These differences suggest at least two alternativeinterpretations: (1) that culture and effectiveness aremore closely linked in larger firms, (perhaps becausethe coordinating effects of culture are more importantin larger and more complex systems); or (2) that thesubjective measures of effectiveness are better suitedfor the comparison of a disparate set of firms than arethe objective measures of effectiveness.

219


18/21


Parallel regression analyses of these data presenteda similar picture when the four traits are used aspredictors of each of the effectiveness measures. Theculture traits are strongest as predictors of objectiveROA among large firms. Combining the four traits aspredictors of ROA in this sample resulted in a multipleR2 of 0.33.

DiscussionThis research has shown how an inductive, theory-building approach to organizational culture and effec-tiveness can be coupled with deductive, quantitativeresearch to provide an emergent theory and an ex-ploratory test of the theory. The convergence of thesetwo logics creates a broader understanding than eithercould alone.The results of these two studies suggest that culturemay indeed have an impact on effectiveness. Each ofthe four cultural traits showed significant positive asso-ciation with a wide range of both subjective and objec-tive measures of organizational effectiveness, as well asinterpretable linkages between specific traits and spe-cific criteria of effectiveness. These findings helpsupport one of the basic premises of many cultureresearchers: That the cultures of organizations have animportant influence on effectiveness.But these conclusions would be premature withoutcareful consideration of several limitations present inthis study. First, as a cross-sectional study, there is afamiliar set of limitations on infering causality. Thecase studies develop plausible explanations for thelinkage of culture and performance in the five firms,but provide little basis for inference. The quantitativestudy generates results that are consistent with thehypotheses, but those results are also consistent withthe hypothesis that effectiveness determines the cul-tural traits, or that the two are simply coincident intime. Clearly, the ideal research design would not onlyincorporate in-depth measures of cultural traits, butwould examine those with appropriate lag time, and abroad range of effectiveness measures. Although suchresearch is time-consuming and difficult, future studiesof organizational culture and effectiveness must de-velop research designs that incorporate both breadthand depth over time.A second limitation of this study is our reliance oncultural "traits" as a key concept in our analysis. Toaddress a specific comparative research problem wehave taken a narrow view of the relevant culturalcontent of each of the firms. To argue that the fourtraits in face are culture would be inappropriate. In-

220

stead, the four traits are summary characteristics of anorganization's culture and the processes by which cul-ture may have an impact on effectiveness. Without thisfocus on the comparative analysis of effectiveness, a farbroader model of culture would be required.Yet the reliance of this study on the concept of traitreflects a strategy that is becoming more widespreadamong culture researchers (Hofstede et al. 1990,O'Reilly et al. 1991, Chatman 1991, Gordon andDiTomaso 1992, Kotter and Heskett 1992, Denison1993). This approach appears to represent an interest-ing compromise between the phenomenological andfunctionalist perspectives on culture. The use of aconcept such as "trait" implicitly makes the assump-tion that comparison and generalization are warranted.The alternative assumption, that the linkage betweenvalues and behavior or between meaning and practiceis contextually defined, makes comparative analysisdifficult, if not impossible. In the extreme, organizationitself is epiphenomenal. Thus, the choice of the con-cepts used to describe culture is a critical one.At the crux of this issue of comparison and general-ization is the linkage between meaning and practice.This issue appears several times in the case studies;most notably in the People Express case. A comparisonbetween the survey data and the case clearly illustrateshow survey measures of traits are "expectation depen-dent" and how the meaning given to a particularpractice is contextually defined. This example is high-lighted to urge caution on a literal and unquestioningreliance on the survey measures. Ironically, however,the "innovative" nature of the People Express organi-zation is dependent on a rather constant relationshipbetween meaning and practice at the industry level.People Express practices were perceived as innovativebecause they disconfirmed the dominant link betweenmeaning and practice at the industry level.A related issue concerns our reliance on CEO per-ceptions of firm's cultural traits in the quantitativestudy. The second study took the risk that CEO per-ceptions of firm culture might have such low validitythat they would be insignificant as predictors of effec-tiveness. Fortunately, this was not the case. The ob-served association between the four cultural traits andthe objective measures of performance, for example,would be unlikely if the underlying culture measureshad not achieved at least some level of validity. Thus,trading the depth and detail of qualitative research forbreadth in the second part of this study met with somedegree of success. Nonetheless, the depth of theseperceptual measures and their reliance on a singlerespondent has obvious limitations. Now that this study



19/21


has suggested that the four cultural traits may have animportant link to effectiveness, future quantitative re-search must concentrate on in-depth measures from abroad range of organizational members to provide aricher test of the model.Conceptualizing and measuring organizational ef-fectiveness has also posed a host of difficult issues inthis study. Not only is cultural meaning contextualized,but so is effectiveness. Their interaction presents nearlyendless complexities. Nonetheless, this study does pro-pose a viable approach for future research: a focus onmultiple single measures grounded in a stakeholdermodel of effectiveness. This approach provides a recog-nition of the multidimensional and equivocal nature ofeffectiveness, but as this study has shown, also allowsfor some interesting results: This research providesevidence that the stability traits of mission and consis-tency are useful predictors of profitability, while theflexibility traits of involvement and adaptability aremore potent predictors of growth. These results sug-gest that future research might fruitfully focus onlinking specific cultural traits to related criteria ofeffectiveness.Finally, in our attempt to identify some core traits ina theory of organization culture and effectiveness, wehave generally paid only limited attention to thebroader cultural contexts within which the organiza-tions themselves exist. Societies, industries, occupa-tions, regulatory environments, and history, to nameonly a few, generate cultural contexts that influenceorganizations and their effectiveness. Further progresstoward a general theory of organizational culture andeffectiveness will clearly require that these factors aswell be incorporated.AcknowledgementsWe gratefully acknowledge the comments and suggestions made byJane Dutton, Thomas D'Aunno, Karl Weick, and other members ofthe Michigan Organization Studies Seminar on previous drafts ofthis paper. We also wish to thank the three anonymous OrganizationScience reviewers for their many helpful suggestions.

Appendix: Survey ItemsInvo

Date post:	06-Apr-2018
Category:	Documents
Upload:	princessshraddhateli
View:	221 times
Download:	0 times

Best Paper Organization Culture Efffectiveness

Documents