Puzzle ofThe Puzzle of Representation: Specifying Components of Responsiveness Author(s): Heinz...

8/12/2019 Puzzle ofThe Puzzle of Representation: Specifying Components of Responsiveness Author(s): Heinz Eulau and Paul D. Karps Represntation

1/23

The Puzzle of Representation: Specifying Components of ResponsivenessAuthor(s): Heinz Eulau and Paul D. KarpsSource: Legislative Studies Quarterly, Vol. 2, No. 3 (Aug., 1977), pp. 233-254Published by: Comparative Legislative Research Center

Stable URL: http://www.jstor.org/stable/439340Accessed: 27/09/2009 15:42

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at

http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=clrc.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Comparative Legislative Research Centeris collaborating with JSTOR to digitize, preserve and extend access

toLegislative Studies Quarterly.

http://www.jstor.org
http://www.jstor.org/stable/439340?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/action/showPublisher?publisherCode=clrchttp://www.jstor.org/action/showPublisher?publisherCode=clrchttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/439340?origin=JSTOR-pdf


2/23

HEINZ EULAUPAUL D. KARPSStanfordUniversity

ThePuzzle of Representation:Specifying ComponentsofResponsivenessThis study examines the conceptualization of representation, particularly the

problems resulting from conceiving of it simply in terms of congruence between theattitudes of constituents and of representatives on policy questions. It examines criti-cally some of the work that followed the innovative study of Miller and Stokes. Re-garding representation as responsiveness, it identifies four components of this concept:policy, service, allocation, and symbolic responsiveness.

Thepuzzle: Wehaverepresentativenstitutions,but like theGreekswe do not know what they areabout (Eulau,1967).With hepublicationn 1963 of Constituency nfluence in Congressby Millerand Stokes, the direction was set for a novel approach o the studyof politicalrepresentation.lThevirtueof this originalstudynotwithstanding,the approach had some quite unexpected consequences for subsequenttheoretical developmentand empiricalresearch. Much of this developmentand researchwas due less to the impact of Millerand Stokes' innovativeapproachas such than to its vulgarization.The questions addressed n thispaper are two: first, we propose to unravelthe continuing puzzle of repre-sentation which was probablymade even more puzzlingby the thoughtlessuse of the concept of congruence which Miller andStokeshad introducedinto discourse about representation;and second, we proposeto explicatetheconcept of responsiveness y decomposing t into four componentswhichseem to correspond o four targetsof representation.

The Miller-StokesModelMiller and Stokes (1963) themselveswere well awareof the broadercontext of theory and researchon representation,2but the focus of theirparticular analysis was a more limited one than representation. Theywere interested in the degree to which constituency control, rather thanpartyvoting, determinedcongressionalroll call behavior: 'The fact that

LEGISLATIVE TUDIESQUARTERLY, I, 3, August 1977 233


3/23

Heinz Eulau and Paul D. Karpsour House of Representatives . . has irregular arty votingdoes not of itselfindicate that Congressmendeviate from party in response to local pres-sures (p. 45). The analysisaddressedan old question:which factor,partyorconstituency, contributesmore to variance n roll-callvoting(all other thingsbeing equal)? The question had been previouslyasked in numerousstudiesrelying, of necessity, on aggregatesurrogate ndicatorsof presumeddistrictpredispositions,most of them demographicor ecological.Millerand Stokes' researchwas a giant stride in the study of repre-sentation because it freed analysis from dependence on surrogatevariablesas indicators of constituency attitudes or predispositions.Miller and Stokesinterviewed a sample of congressionalconstituents (voters and non-voters)and their respective congressmen (as well as non-incumbent candidates)whose attitudes in three broad ssue domainsthey comparedwith each other,with congressmen'sperceptions of constituency attitudes, and with cor-respondingroll call votes. Their tool of analysiswas the product momentcorrelation coefficient and their mode of treatment was causalanalysis,which was then being introduced into political science. Miller and Stokesfound the relationshipsamong the variables of their model to vary a gooddeal from issue area to issue area,being strongestin the case of civil rights,weaker in the case of social welfare, and weakest in the case of foreigninvolvement.They concluded:

The findingsof this analysisheavilyunderscore he fact that no singletradi-tion of representation ully accords with the realities of American egislativepolitics.The Americansystem is a mixture, to which the Burkean,instructed-delegate, ndresponsible-partymodels all can be said to havecontributedelements.Moreover, aria-tions in the representativeelation aremost likely to occur as we movefrom one policydomain o another p. 56).We have no quarrelwith this general conclusion concerning theAmericansystem. We are bothered by the definition of what MillerandStokes call the representativerelation and its operational expression.This relation is the similarity or, as it is also called, the congruencebetween the four variables of the causal model that serves the purposesofanalysis. This specification of congruenceas the expression of the repre-sentative relation has had great influence on later researchers,both thoseworkingin the traditionof, or with the datamadeavailableby, the Michigangroup and those working independently with fresh data of their own.5The concernhere is not this influence as such but rather he gradual rosion

of alternativetheoretical assumptionsabout representationof which Millerand Stokes themselves are fully cognizant.As a result of this erosion, whatfor Millerand Stokes (1963, p. 49) was only a startingpoint for a widerange of analyses became an exclusive definition of representation:high

234


4/23

The Puzzle of Representationcongruencewas interpreted as evidence of the presence of representation,and low congruencewastaken asproof of its absence.Whatevercongruence may be symbolizing,it is not a self-evidentmeasure of representation. Later researchers,poorly tutored in theoriesand practices of representation,tended to ignore this. Miller and Stokes,in order to use congruence as a measure,had stipulated three conditionsfor constituency influence or control. First, control in the representationalrelationshipcan be exercisedthroughrecruitment-constituentschoose thatrepresentativewho shares their views so that, by followinghis own convic-tions, the representative doeshis constituents' will. Second, control canbe obtainedthroughdepriving he representative f his office-the representa-tive follows his (at least tolerablyaccurate) perceptionsof districtattitudein order to win re-election. And third, the constituency must in somemeasure take the policy views of candidates into account in choosing aRepresentative pp. 50-51).The electoral connection is of course only one of the linksbetweenrepresentativeand represented.And it should by no means be taken forgrantedthat it is the most critical,the most important,or the most effectivemeans to insure constituency influence on or control over public policiesand the conduct of representatives.t is so only if one or all of the conditionsfor constituency control specified by Millerand Stokes are satisfied. Thisis also precisely the reason why attitudinal or perceptual congruence s notan exclusive measure of representation; t is simply the starting point,as Miller and Stokes knew, in the puzzle of representation.Anyone whohas the least sensitivity to the representativeprocess recognizesthat repre-sentatives are influenced in their conduct by many forces or pressuresorlinkages other than those arisingout of the electoral connection and shouldrealize that restricting he study of representation o the electoralconnectionproduces a very limited vision of the representationalprocess. MillerandStokes themselves were eminently aware of this, as their Conclusionindicated. Yet, only three years after publication of their analysis,whentwo other analysts (Cnudde and McCrone, 1966), subjecting the Miller-Stokes data to an alternativecausal analysis, found no support for recruit-ment as a condition of representation,constituency control was reduced toa purely psychologicalfunction in the representative'smind, and the dangerof limiting the representative elation to attitudinal and perceptualcon-gruence was demonstrated. Moreover, these analysts altogether ignoredMiller and Stokes' important third condition for constituency influencethrough the electoralconnection: constituents'takingaccount of the candi-date'spolicy views in choosingthe representative.Indeed, Miller and Stokes themselves had the most trouble with

235


5/23

Heinz Eulau and Paul D. Karpsthis last condition. The overwhelming vidence of their researchand that ofothers denies the condition: most citizens are not competent to performthefunction which the model assumes-that elections are in fact effective sanc-tioning mechanisms in the representationalrelationship.Miller and Stokesgave a number of reasons or why representatives eem to be so sensitiveabout their voting records-for if voters do not know the record, this sensi-tivity is surelypuzzling. They suggested hat the votingrecordmay be knownto the few voterswho, in close contests, make the differencebetweenvictoryor defeat, and that the Congressmans a dealer n incrementsandmargins.They also speculatedthat the votingrecordmay be known to opinionleadersin the district who serve as gatekeepersand conveyors of evaluation in atwo-step flow of communication.But there is no evidence for this in theirown research.

The Crisis n RepresentationalTheoryIt would not yield further theoretical dividendsto review in anydetail the empiricalstudies of representationthat, in one way or another,are predicatedon the attitudinal-perceptualormulation of congruencethathad served Millerand Stokes as a startingpoint but that, for most of theirsuccessors,became a terminalpoint. Most of these studies are distinguishedby lack of historical-theoreticalknowledge of representationand of inde-pendent theoretical creativity. In particular,they are cavalierin regardtoa number of dilemmas that, by the middle sixties, had forced themselveson the attention of scholars interested in theoretical understandingof theproblem of representation.That these dilemmas were articulated by dif-ferent scholarsat about the same time was probably coincidental,but thecoincidenceis importantbecauseit emphasizedthe possibility of alternativeresearchdirections.First, representational theory made assumptions about citizenbehavior that were negated by the empiricalevidence. Wahlke,examiningthe role of the represented n the representationalrelationship,concludedthat the evidence did not justify treating citizens as significantsources ofpolicy demands,positions or even broadorientationsthat could be somehowrepresented n the policy-makingprocess. Citizenssimply lack the neces-sary information for effective policy choices to be communicated to theirrepresentatives, ven if they were to make the effort to communicate. Thisbeing the case, Wahlkeconcluded that the simple demand-inputmodelof representationwas deficient. This is of course precisely the model thatMiller-Stokeshad in fact constructed in orderto organizeand explain theirdata. Wahlkesuggestedthat a support-inputmodel might be more appro-priate.

236


6/23

The Puzzle of RepresentationSecond, given the limited capacity of the represented o formulate

policy, a viable theory could no longer ignorethe asymmetryof the repre-sentational relationship. Eulau suggested, therefore, that research shouldproceed from the structuralassumptionof a built-in status difference be-tween representativeand represented n which the former rather than thelatter give direction to the relationship.Representational heory would haveto deal with the tensions arisingout of status differentiation rather thandeny their existence (Eulau, 1967). Once status is introducedas a variableinto the representationalequation, the model of the representationalrela-tionship can be recursive,and the causal orderingof the relevantvariablesis likely to be reversed.Finally, in a linguistic study of the concept of representation,Pitkin (1967) found the traditional theories of representationflawed. Sheadvancedthe proposition that representation,referringto a social relation-ship rather than to an attributeof the individualperson, could be meaning-fully conceptualized only as a systemic property. Representationmight ormight not emergeat the level of the collectivity, the criterion of emergencebeing the collectivity'spotentialfor responsiveness. oliticalrepresentationis primarilya public, institutionalizedarrangementnvolving many peopleand groups,and operating n the complex ways of large-scale ocial arrange-ments. What makes it representationis not any single action by any oneparticipant,but the over-all structure and functioning of the system, thepatterns emerging from the multiple activities of many people (pp. 221-222). Moreover,after consideringevery conceivabledefinition, Pitkin con-cluded that political representationmeans acting in the interest of therepresented,in a mannerresponsiveto them (p. 209). However, there isalso the stipulation that the representative mustnot be found persistentlyat odds with the wishes of the representedwithout good reason in termsof their interest, without a good explanationof why their views arenot inaccordwith theirinterests pp. 209-210).Pitkin's formulation creates many measurement problems forempirical research.Concepts like wishes, ' good reason, interest, orviews, are difficult to operationalize. She provides no clues as to howresponsiveness s a systemic property of the political collectivity can beascertainedand how, indeed, it can be measured n ways enablingthe scienti-fic observer o concludethat representationhas in fact emergedat the level ofthe political system. Pitkin'streatmentseemsto stressthe conditionin whichthe representative tands ready to be responsivewhen the constituents dohave somethingto say. A legislaturemay, therefore,be responsivewhetherornot there are specific instances of response. In other words, Pitkin empha-sized a potential for response rather than an act of response. There are

237


7/23

Heinz Eulau and Paul D. Karpsconsiderabledifficulties in empirically workingwith a concept stressingthepossibility of an act rather than the act itself. Moreover,the formulationignoresWahlke'snjunctionto jettison the demand-inputmodel. Nevertheless,Pitkin'swork had an almost immediate and profoundeffect on subsequentempirical research. (See Prewitt and Eulau, 1969; Muller, 1970; Peterson,1970.) Research on representationfollowing the watershedyear of 1967has taken two majorinnovative routes. First, takingtheir cue fromWahlke'scritiqueof the demand-inputmodel, Patterson,Hedlund,andBoynton (1975)have used a support-inputmodel that makes fewer requirementson thecapacity of the representedto play a role in the representationalprocess.However, their model continues to be based on congruence assumptions.Their analysis,conducted at the level of the individual,largely consists ofcomparisonof the representedandrepresentational lites in termsof relevantattitudes,perceptionsand behaviorpatterns.Second, taking a cue from Pitkin, Eulau and Prewitt (1973) trans-formed data collected at the level of individuals into grouped data, andconducted theiranalysisof representationat the macrolevel of smalldecision-making groups (city councils). In contrast to Patterson and his associates,Eulau and Prewitt stressedactualrather than potential responseto constitu-ent inputs, whether of the demand or support variety. In retrospect, itappears, they were harnessing reactive behavior rather than responsivebehaviorin Pitkin's sense, for they ignored the direction of the response-whether it was in fact in the interest of the constituents at the focus ofrepresentation.But these retrospectivemusingsonly suggestthat the problemof conceptualizing representation n terms of responsiveness emainson theagendaof theory and research.As Loewenberg(1972, p. 12) has summedup the situation morerecently:Representation . . is an ill-definedconcept that has acquired conflictingmeaningsthroughlong use. It may be employed to denote any relationshipbetweenrulersandruled or it may connote responsiveness, uthorization,egitimation,or accountability.It maybe used so broadlythatanypolitical nstitutionperforms epresentativeunctionsor so narrowlythat only an elected legislaturecan do so. To a surprisingxtent, theBurkeanconceptualizationof the representative unction is still in use, and Eulau'scall for a concept adequate o modern concernsabout the relationshipbetweenlegisla-torsandtheirconstituencieshasnot beenanswered.

Responsiveness sCongruenceAlthoughthe expectations or behavioralpatternsto which the termresponsiveness refers were implicit in the concept of representativegovernment, 8the term as such had not been used by Miller and Stokes or

238


8/23

The Puzzle of Representationothers as the defining characteristicsof representation. By 1967, whenPitkin's work was published, the term struck an attractive chord as theideals of participatorydemocracy were once more being revived in neo-populist movements that had intellectual spokesmenin the social sciences.Even though one should not expect a close affinity between the vocabularyof participationand the vocabularyof representationon logical-theoreticalgrounds, a term like responsivenessstemmingfrom considerationsof repre-sentative democracy could easily blend in with considerations of partici-patory democracy.Whenanalysts of political participationlike Verba andNie (1972) came to pay attention to empiricalwork on representation, heyhad little trouble in linking, by way of an adaptationof the assumptionofcongruence, the concept of responsiveness o their work on participation.Interestingly,althoughthey did not cite or referto Pitkin's inguistic analysis,Verbaand Nie found, on the one hand, that responsiveness, s faraswe cantell, rarelyhas been defined precisely,almost never has been measured,andnever has been related to participation (p. 300). On the other hand, theyacknowledgedMillerand Stokes, who had not used the term: MillerandStokes in their analysis of the relationshipbetween constituency attitudesand Congressmen,do deal with responsiveness n ways similar to ours(p. 300, ft. 3).Indeed, in examiningand seekingto explain the effects of differentdegreesof citizen participationon the responsivenessof community leaders,Verba and Nie presenta rechristenedversion of the congruenceassumptionof representationwhichthey call concurrence :Our measureof congruencedependson how well the prioritiesof the citizensand theleadersmatch. Severaltypes of concurrenceare possible . . . our measureof the con-currencebetween citizensand community eadersmeasures he extent to whichcitizensand leaders in the community choose the same agenda of community priorities(p. 302).But they immediately raise the critical problem of causality: whetherwehave the warrantto considerour measure of concurrenceto be a measureofresponsiveness.Just because leaders agree with citizens and that agreementincreases as citizens become more active, can we be sure that it is citizenactivity that is causingleaders to respond by adoptingthe prioritiesof thecitizen? (p. 304).In orderto test for the causalrelationship,Verbaand Nie comparedthe correlation coefficients obtained for the relationshipbetween citizenactiveness and concurrence,on the one hand, and between leaderactive-ness and concurrence,on the other hand. Findingthat the correlationsforcitizens are muchstronger hanthose for leaders,VerbaandNie concludedthat their measureof concurrence seemsto be a validmeasureof responsive-

239


9/23

Heinz Eulau and Paul D. Karpsness to leaders (pp. 331-332). But this mechanicalcomparison s not a testof causality at all in regardto the direction of responsiveness.In fact, itamounts to a false interpretationof the data. The correlationsfor citizenssimplymean that more active citizens see things (priorities o be done in thecommunity) more like leaders do than is the case with less active citizens;the correlations for leaders simply mean that the more active leaders seethings more like citizens do than is the case with less active leaders. Thestrength of the coefficients, all of which are positive for both citizens andleaders, does not prove anything about the direction of causality-whethercitizens influence leaders or leaders influence citizens, or whether citizensare responsiveto leadersor leadersto citizens.It cannotbe otherwisebecauseVerba and Nie's measureof concurrence, ike Millerand Stokes' measureofcongruence,is neutral as to direction and requiresthat the directionof therelationships nvolved in the model be theoretically stipulated. Thereis nosuch stipulationin the Verba-Nieapplicationof the concurrencemeasuretothe questionof linkagebetweenleadersandled.Causal analysis, then, does not free the analyst from defining histerms-be they power and influence, or be they responsiveness-in advanceand stipulating the direction of expected relationshipsin advance.9 Themechanicalapplicationof statisticaltests of a possible causalstructuredoesnot necessarilymodel real-worldrelationships f the operationaldefinitionsof the model's components make no theoretical sense. Verba and Nie'stwo-edged use of the responsiveness,operationalized n terms of the direc-tionless concept of concurrence,is intrinsicallycharacterizedby ambiguity.If concurrence s a measureof responsivenessof leadersto citizens, it cannotbe a measureof responsiveness f citizensto leaders.If one were to take theircomparison of the correlations between participationand concurrenceforcitizens and leaders as an indication of anything, it would have to be thatleaders are responsive to citizens and citizens are responsive to leaders,varying n degreewith degreeof participation.Pitkin, it was noted, had raised the importanceof responsivenessas the critical characteristic of representation,but she had left the termundefined. Representatives,in order to represent, were to be responsiveto their constituents, but Pitkin did not specify the content or target ofresponsiveness.Verbaand Nie had taken a step forwardby specifyingpublicpolicy issues as the target of responsiveness.In focusing exclusively oncongruenceor concurrence n regard o policy attitudes or preferences, heyignoredother possible targetsin the relationshipbetweenrepresentatives ndrepresentedwhich may also give content to the notion of responsiveness.By emphasizing only one component of responsivenessas a substantiveconcept, they reduced a complex phenomenon like representationto one

240


10/23

The Puzzle of Representationof its components and substituted the component for the whole. But ifresponsivenesss limited to one component, it cannot capturethe complexi-ties of the realworld of politics. It is necessary,therefore,to view responsive-ness as a complex, compositional phenomenon that entails a variety ofpossible targetsin the relationshipbetween representativesand represented.How else could one explain that representativesmanageto stay in office inspite of the fact that they are not necessarilyor always responsiveto therepresentedas the conceptionof representationas congruenceor concurrenceof policy preferencesrequires?It deservesmention that Millerand Stokes (1963) had themselvesrealized that there are possible targets of responsivenessother than policyissues. They emphasizedthe necessity of specifying the acts with respectto which one actor has power or influence or control over another. Theirtarget, they conceded, was only the set of issueslyingwithin the threepolicyareasof civil rights,social welfare and foreigninvolvement. But significantlythey added, Wearenot able to sayhow much control the local constituencymay or may not have over all actions of its Representative,and there maywell be pork-barrelssues or other publicmatters of peculiarrelevance o thedistrict on which the relation of Congressman o constituency is quite dis-tinctive (p. 48). Millerand Stokes did not specify what they referred toas other public matters. It is the task of the rest of this paperto suggestwhat some of these other targetsof responsivenessmightbe.

Componentsof ResponsivenessThere are four possible components of responsivenesswhich, as awhole, constitute representation. While each component can be treatedas an independent target of responsiveness,all four must be considered

together in the configurativetype of analysis which, it seems to us, thecomplexity of the representationalnexus requires. The first componentis, of course,policy responsivenesswhere the targetis the greatpublicissuesthat agitate the political process. Second, there is service responsivenesswhich involves the efforts of the representativeto secure particularizedbenefits for individualsor groupsin his constituency. Third,there is alloca-tion responsivenesswhich refers to the representative'sefforts to obtainbenefits for his constituency through pork-barrel xchangesin the appropri-ations process or throughadministrativenterventions.Finally, there is whatwe shall call symbolic responsivenesswhich involves public gestures of asort that create a sense of trust and support in the relationshipbetweenrepresentativeand represented.It is possible that there are other targetsof responsive conduct which, in composition with the four here tapped,

241


11/23

Heinz Eulau and Paul D. Karpsconstitute the matrix of representationalrelationships.But the main pointwe are trying to make is this: responsivenessrefers not just to this orthat target of political activity on the part of the representativebut toa numberof targets.Only when responsivenesss viewed as a compositionalphenomenon can the approach to representation-as-responsivenessecom-mended by Pitkin be useful. It is the configurationof the componentaspectsof responsiveness hat might yield a viable theory of representativegovern-ment undermoder conditionsof societalcomplexity.Policy Responsiveness

How the representativeand the represented nteract with respectto the makingof publicpolicy lies at the heart of most discussionsof respon-siveness.Responsivenessn this sense refers to the structure n which districtpositions on policy issues, specified as some measure of central tendency ordispersion,are related to the policy orientation of the representative-atti-tudinal or perceptual-and to his subsequent decision-makingconduct in agivenfield of policy.The premise underlying the specification of policy responsivenessis the presence of a meaningful connection between constituent policypreferences or demands and the representative'sofficial behavior. This iswhat Miller and Stokes called congruence and what Verba and Nie calledconcurrence. Whatever he term, the operationaldefinition is the same:if the representativeand his constituency agree on a particularpolicy, nomatter how the agreementhas come about, then the representatives respon-sive. Thereare, as has been noted, severalproblemswith the model of repre-sentation built on the operationalizationof responsivenessas congruence,notably the problem that congruence s neither a necessarynor a sufficientcondition for responsiveness.The representativemay react to constituencyopinion, and hence evince congruent attitudes or behavior,yet not act inwhat is in the best interest of the constituency as he might wish to definethat interest, thereby being in fact unresponsive.Further,the representativemay make policy in responseto groupsand interestsother thanhis constitu-ents, includingexecutive and bureaucraticagencies.Whethersuch conductis also in the interest of his district as he sees it is an empiricalquestion.But whatever the formulation and findings,it cannot be denied that policyresponsivenesss an importantcomponentof representation.

The notion of policy responsivenessis implicit in some of theclassic theories of representation.First of all, the controversy over man-date versus independence, whether the representativeis a delegate or atrustee, though considered obsolete by Eulau (1967, pp. 78-79) and in

242


12/23

The Puzzle of Representationmany respectsresolvedby Pitkin(1967, ch. 7, pp. 144-167), is still intriguingand relevant to the present discussion. For the debate is over whether therepresentative hould act according o whathe thinks is in the bestinterestof the constituency, regardlessof constituency wants, or whether heshould follow the expressedwishes of the district, regardlessof how hepersonallyfeels. The debate really turns on the competence of the citizenryin mattersof public policy. For while the citizenrymay know what it wants,it may not know what it needs. Secondly, therefore,an appropriatedefini-tion of policy responsivenesswill be related to the classic issue of districtinterest as against district will. There is no denying that the notion ofpolicy responsivenesspervades empirical research on legislative decision-making, even when the issue of representationas a theoretical one is notraised.(For recent research,see Turnerand Schneier, 1970; Kingdon, 1973;Clausen, 1973; Jackson, 1974; Matthews and Stimson, 1975.) However,precisely because this is the case, it is importantnot to ignore other com-ponents of responsivenessin the representational relationship. Exclusiveemphasison the policy aspectsof responsivenessmay give a one-sidedviewandmaynot help in solvingthe puzzleof representation.ServiceResponsiveness

A second target for responsivenessto define the representationalrelationshipconcerns the non-legislative ervices hat a representative ctuallyperforms for individualsor groups in his district. Service responsiveness,then, refersto the advantagesandbenefits which the representatives abletoobtain for particularconstituents. There are a number of services that con-stituents may expect and that the representative onsidersan intrinsicpartof his role. Some of them involve only modest, if time consuming,requests,such as responding to written inquiries involving constituents' personalconcerns,orfacilitatingmeetingsandtoursfor visitorsfrom the home district.Newsletters or columnsin local newspapersmay be used to informconstitu-ents of legislation that may be of interest and use to them. Much of thiswork is routineandcarriedout in regular ashion.Another link in the chain of serviceresponsivenesss often referredto as case work. (See Clapp, 1963.) Given his official position andpresumedinfluence, the representative s in a position to solve particularproblemsfor members of his constituency. The representative ntervenes betweenconstituentsand bureaucrats n suchmattersas difficulties with a tax agency,delays in welfare payments, securinga job in government,and so on. Pro-viding constituent servicesand doing case work constitute for many repre-sentatives more significant aspects of their representationalrole than does

243


13/23

Heinz Eulau and Paul D. Karpslegislative work like bill-draftingor attending committee hearings. Theseerrandboy functions deserve more theoretical attention than they havebeen given in contemporary research. In some important situations therepresentativemay actuallyserve as an advocateand even lobbyist for specialinterestsin his district vis-a-visthe legislature,departmentalbureaucracies rregulatory agencies. This type of responsiveness s indeed crucial in tryingto understandmodem representative overnment.This notion of serviceresponsiveness eemed to underlieEulauandPrewitt's(1973, pp. 424427, 649-650) operationaldefinition of responsive-ness. In their study of San FranciscoBay Areacity councils, they initiallydivided these small representativebodies into those which seemed to besomehow responsiveto constituent needs or wants and those which did notseem to be responsive.They then distinguished amongthe formercouncilsthose which were responsive to important standing interests in the com-munity or attentive publics, and those which more often were responsiveonly to temporary alliances having a particulargrievanceor request. Thisconception of responsiveness, hen, is based on the kind of groupor indivi-duals whom the representativeperceives as being primarilyserved by hisactivities.ZeiglerandJennings 1974, pp. 77-94), in a study of school boards,present a similar conception of responsiveness,conceptually distinguishingmore sharplybetween groupresponsiveness nd individualized esponsive-ness. Both of these research teams, then, defined responsiveness n termsof the significantrecipientsof representationalervices.Thatserviceresponsivenesss an importantelementin representationshould be apparent. Moreover,there is every reason to believe that it isincreasingrather than declining. Until the middle sixties, it was generallyassumedthat case work andthe advocacyof specialinterestsbringadvantagesand benefits only to those who take the initiative in soliciting the repre-sentative'shelp. But as Fiorina (1977, p. 180) has recently pointed out, atleast with reference to the federal level, increasedbureaucraticactivity inthe wake of increased federal largesse to all kinds of population groupshas also motivated congressmento undoubtedlystimulatethe demand fortheir bureaucratic ixit services. The representativedoes not just respondto demandsfor his good offices andservices;he has become a kind of hustlerwho advertisesand offers them on his own initiative.l0This explication of serviceresponsivenesshas been entirely focusedon the relationshipbetween the representativeand particularconstituents.The representativecan also be responsive n his unique role as a middlemanin the allocation of more generalizedbenefits. We refer here to what hasbeen traditionallycalled pork-barrel olitics and to what we shall refer,for lack of a better term,as allocationresponsiveness. Both servicerespon-

244


14/23

The Puzzle of Representation 245siveness,whether nitiatedby the representative rnot, and allocationrespon-siveness,which is always initiatedby him, are importantelements of repre-sentational behavior and important pillarsin the representationalrelation-ship.AllocationResponsiveness

It has long been recognizedthat pork-barrelpolitics in legislativeallocations of public projects involves advantagesand benefits presumablyaccruingto a representative'sdistrictas a whole. Althoughtraditionally heseallocationswere seen as publicgoods, with the expansionof the govern-ment's role in all sectors of society-industry, agriculture, ommerce,health,education, welfare, internal security, and so on-the distinction betweenpublic and privatebenefits is difficult to maintain.Again,as Fiorina(1977,p. 180) has felicitously put it in connectionwith federalpolitics, Thepork-barrelerneed not limit himself to dams and post offices. There is LEAAmoney for the local police; urban renewal and housing money for localofficials;and educationalprogramgrants or the local educationbureaucracy.The congressmancan stimulate applicationsfor federal assistance,put in agood word during consideration, and announce favorable decisions amidgreat fanfare. Such allocationsmay benefit the district as a whole, or theymay benefit some constituents more than others because they make moreuse of the benefits. The criticalpoint to be made is that in being responsiveas an allocator, whether in the legislativeor bureaucraticprocesses, therepresentativeseeks to anticipate the needs of his clients and, in fact, canstimulatetheirwants.

Legislators'committee membershipssometimes serve as indicatorsof allocation responsiveness,as revealedin Fenno's (1973) studies of legis-lative conduct in committees of the U. S. House of Representatives.A repre-sentative from a district that has a particular take in a committee's urisdic-tion will often seek a post on a parentcommittee but also on a particularlysuitable sub-committee;such membershippresumablyenables him to actin a mannerresponsiveto the best interestsof his district and some or all ofhis constituents.

However, one cannot automaticallyassume that a legislatorservingon a committee not relevant to his districtis necessarilyunresponsive ndnot interested in securingallocations. Legislatorsoften seek prefermentonimportant committees like Rules, Appropriations,or Ways and Meansnotbecause these committees are directly relevant to the interests of theirconstituents, but because they place members in positions of power andinfluence vis-a-vis administrativeagencies which distribute benefits, such


15/23

Heinz Eulau and Paul D. Karpsas the Army Corpsof Engineers, he ParkService,or the VeteransAdminis-tration. These secondarybonds are probably as criticalin securingbenefitsfor the districtas are the primarybondsresulting rom relevant ommitteeassignments.However, the secondarybonds have less symbolic value thando the primarybonds. And symbolic pay-offs, we shallsee, arean importantfourthcomponentof representational esponsiveness.Symbolic Responsiveness

The fourth component of responsiveness s more psychologicallybased than the others. The first three components all somehow tap a be-havioral aspect of representation:policy responsiveness s oriented towardthe decision-makingbehavior of the representativein matters of publiccontroversy;while serviceand allocation responsivenessare oriented towardparticularizedor collective benefits obtained through the acts of the repre-sentative. The representational elationship s not, however,just one of suchconcrete transactions,but also one that is built on trust and confidenceexpressed in the supportthat the representedgive to the representativeandto which he responds by symbolic, significant gestures,in orderto, in turn,generateand maintaincontinuingsupport.The notion of symbolic responsivenesshas been alluded to byWahlke(1971) in examiningthe role of the constituency in the representa-tional relationship.He found little evidence for presumingthat a districtmakes specific policy demands on its representative.Rather, he suggestedthe relevance of Easton's concept of diffuse support (1965, pp. 247-340)as a key component in the relationshipbetween the representedand theirrepresentative.He states that the symbolic satisfaction with the processof government s probablymore importantthan specific, instrumentalsatis-faction with the policy output of the process (Wahlke, 1971, p. 288).The important question then becomes, . .. how do representativebodiescontributeto the generationand maintenanceof support? p. 290).In an era of cynicismabout the functioningof representativensti-tutions, the ways in which representatives manipulate political symbolsin order to generate and maintain trust or support become criticalaspectsof responsiveness.Edelman (1964, 1971), following the earlier work ofMerriam,Lasswell,and Smith (1950), has emphasized the importance ofsymbolic action in politics. The need for giving symbolic reassurance sbeing demonstratedby the reachout efforts of the new President of theUnited States-walking down PennsylvaniaAvenue after his inauguration,fire-sidechats, telephonic call-a-thons,visits to economicallystrickenareas,being Jimmy Carter,and so on. The purpose of all of these symbolic

246


16/23

The Puzzle of Representationacts is to project an image that the President s trulythe people's representa-tive and ready to be responsiveto them. By mobilizingtrustandconfidenceit is presumablyeasier to go about the job of representationthan wouldotherwisebe the case.Fenno (1975), in a paperon Congressmenn theirConstituencies,emphasizesthe importanceof political supportin the representational ela-tionship. The representative's home style -how he behaves in his con-stituency-is designednot just to secure constituent supportand re-electionbut also to give the representativemore freedomin his legislativeactivitieswhen he is away from home. Symbolic politics has the purposeof buildingup credit to be drawn on in future contingencies. Although Fenno doesnot cite Wahlke at all, it is significantthat his analysis approximatesthesupport-inputmodel :. . . congressmen eek and voters may give supporton a non-policybasis.They maysupporta goodman on the basisof hispresentation asa person and trust himto bea good representative.So, we might consider the possibility that constituent trust,togetherwith electoralaccountability,may also providea measureof good representa-tion. The point is not that policy preferencesare not a crucialbasis for the representa-tionalrelationship.They are. The point is -thatwe should not startourstudiesof repre-sentationby assuming hat they are the only basis for a representationalelationship.Theyare not (p. 51).

Fenno's comments are all the moregermane o the argumentof thispaper because it is interestingto note that this most eminent of legislativescholars deflates the prevailingobsession with policy responsivenessas thesine qua non of representation.In fact, much of what may appearto bepolicy responsivenessis largely symbolic responsiveness.From session tosession, legislators on all levels of government-federal, state, and local-introducethousands of billswhichhave not the slightestchanceof everbeingpassed and, more often than not, are not intended to be passed.Yet repre-sentatives introduce these bills to please some constituents and to demon-strate their own responsiveness.1ResponsivenessandFocus of Representation

Once the concept of representation-as-responsivenesss decomposed,policy responsivenessappears as only one component of representationand, perhaps, as by no means the dominant link between representativeand represented. There is no intrinsic reason why responsiveness n onecomponent of representation cannot go together with unresponsivenessin another. An individual or group may disagreewith the representative'sposition and behavioron an issue of public policy and, as a result,may beunrepresented n this sense; yet, the same individualor group may be well

247


17/23

Heinz Eulau and Paul D. Karpsrepresentedby a person who is responsiveby attending to their particularrequestsfor some type of service.Similarly, t is possiblefor a representativeto be responsivewith regardto securing public goods for his constituency,while simultaneously being quite unresponsive with respect to issues ofpublic policy. Finally, what matters in symbolic responsiveness s that theconstituents feel represented,quite regardlessof whether the representativeis responsive n his policy stands or the services or public goods he providesfor his constituency.Moreover,even if attention is given only to policy responsiveness,researchcannot simplyneglect some of the classicalquestions of representa-tional theory, such as the issue of representingthe district'swill as againstits interest, or the issue of the focus of representation.It is easily conceiv-able that being responsiveto a district's will-the wants of its people-mayinvolve being unresponsiveto a district's interest-the needs of its people.With regard o the focus of representation,being responsiveto the electoraldistrict may produce unresponsivebehavior in the largerunit of which thedistrict s a partand,of course,vice versa.12In fact, a closer look at the question of representational ocus willrevealfurther the potentially multidimensional haracterof the phenomenonof responsiveness.The representativecan perceivehis constituency in amultitude of ways,13 thereby making the number of foci quite large. Onemight organizethese possible foci into three categories. The first categoryentails a geographic ocus; the representativemay perceivehis constituencyinterms of nation, region, state, district,orany other territorial evel of society.Thesecond categorywould includeparticular olidaryor functionalgroupingslike ethnic, religious, economic, and ideological groups, whether organizedor not. Finally, the representational elationshipmay have as foci individualpersons rangingfrom distinguishednotables to unknown clients in need ofhelp and to personalfriends.Representationalfocus, then, can differ a great deal in each ofthese three ways. The crucialpoint, however, is that the focus of representa-tion might vary with each of the four components of responsiveness.Whileone might find particularfoci, accordingto the three categories,for policyresponsiveness,one might find altogetherdifferent foci in regard o any ofthe other components of responsiveness.Any empirical combination ispossible within relevant logical constraints. Empiricalresearch has yet toaddressthe relationshipbetween modes of responsivenessand foci of repre-sentation, and untangle the web of complexity createdby the relationship.

248


18/23

The Puzzle of RepresentationResponsiveness ersusResponse

The generally confused and confusing use of responsiveness,especially when linked to notions of concurrence, s only symptomaticof a malaise that has come to characterize he scientific study of politics.The malaise is to substitute theory construction as a technique for sub-stantive theory or theorizing.A younger scholarin the field, Fiorina(1974,p. 24), after reviewingthe empirical researchon representationof recentvintage, has come to a similarconclusion. We quote him preciselybecausehe is not ignorant of or inimical to the new technological dispensationsofourtime:Too often it seems that the increasingavailabilityof electroniccomputingfacilities, data banks, and canned statisticalpackageshas encourageda concomitantdecline in the use of our own capabilities.Ratherthanhypothesizewe factoranalyze,regress,or causalmodel. Wespeakof empirical heoryas if it miraculously rowsout ofa cumulationof empirical indings,ratherthanas a logicalstructure ne mustcarefullyconstruct o explainthosefindings.

When Fiorina identifies data banks as one of the villains, hepresumablyimplies that the user of these facilities has grown increasinglyremote from his subjects of observation and lost touch with the humanityhe is supposed to understand.Indeed, there are today users of survey re-searchwho have never intervieweda single personin their lives. Not surpris-ingly, therefore, causal models are being reified as if they describedrealityratherthan being abstractions from reality. In the case of representationalresponsiveness, or instance, the causaldirection has been assumedto pointfrom the representedto the representative; he latter has been assumed tobe the object of stimuli to which he responds(or does not respond)in thefashion of Pavlov'sfamous dog. But such a model, even if one providesforinterveningattitudinal or perceptualprocesses,does not approximaterepre-sentational relationshipswhich are, above all, transactionsnot necessarilystructured n the ways of the S-O-Rparadigm.To appreciate the complexity of representational relationshipsas transactions, t is simply erroneous to assumethat responsiveness-what-ever component may be involved-is somehow the dependent variable n acausal structure. Responsiveness and response are not the same thing.On the contrary, a representativewhose behavior is purelyreactive-a con-dition that is hard to conceive on reflection but one that the concurrencemodel postulates-is the very opposite of a politically responsivepersonin Pitkin's sense. As that personhas been chosen, elected, or selected fromthe multitude or mass to be a representative,hat is, as he occupiesa superiorposition in the relationshipby virtue of his elevation, one should expecthim not merely to be reactive but to take the initiative. Whether he does

249


19/23

Heinz Eulau and Paul D. Karpsor not is, of course, an empiricalquestion; but the question cannot be an-swered by simply substituting an inappropriatemodel of causation forempirical observation and a viable theory of representation that wouldguideboth observationandanalysis.As alreadysuggested,the attractivenessof the notion of responsive-ness in the most recentperiodhas been due in partto the fusion of participa-tory and representationaldeas about democracy. But in the participatorytheory of democracy the leader-insofar as the model admits of leadershipat all-is largely a reactive agent guided by the collective wisdom of thegroup. He is at best the executor of the group'swill, indeed a human fac-simile of Pavlov'sdog. He reacts,presumably,but he is not responsive.Oneis in fact back to the instructed-delegate model in which thereis no roomfor discretionin the conduct of the representative.A causalmodel of repre-sentation that drawsits arrowsonly in recursive ashion from the representedto the representativecannot capture, therefore, the meaningof responsive-ness in Pitkin's sense. It excludes ab initio what is yet to be concluded.It is a grievouserror,againstwhich Fiorinawarned,to assumeandtoact as if the assumptionwere valid, that causalanalysis will automaticallyyield theory, or that by simple inversionof causalassumptions omethingmeaningful will come out of a causal analysis. Theorizinginvolves some-thing more than arbitrarily nvertingthe causal directions on the assumptionthat the resultant statistical structure will somehow reflect reality. It in-volves giving reasons and justifying the assumptions one brings into thecausal analysis. It involves going out on a limb, as it were, and sayingsomething substantive about the phenomenabeing investigated,ratherthanhiding behind the artifactual findings of a causal analysis that may beinappropriaten the firstplace.A next step in the study of representationas responsivenessmusttake off from the compositionalnature of the phenomenon.Thisstep cannotbe limited to simplisticmeasures ike congruenceor concurrence n connec-tion with one component of a complex set of transactionalrelationships.Any inferences one may make about the functions of any one componentof responsiveness n representativegovernment must be related to infer-ences one may make about the functions of other components. Otherwisethe puzzle of representation-having representativegovernment but notknowingwhat it is about-will continueto bewilderthe politicalimagination.

NOTESAn earlierversion of this paperwas presentedat the annualmeetingof theWesternPoliticalScienceAssociation,March31-April2, 1977.

250


20/23

The Puzzle of Representation 2511. Miller and Stokes (1963). A revisedversion is includedin Campbell,

Converse,Miller,and Stokes (1966, pp. 351-372). We shall be citingthe originalarticlebecause we are only interested here in the theoreticalaspectsof the analysiswhichremainedunaffectedby the revision.The particular nalysiswas partof a much largerstudy of representation onductedin connection with the 1958 congressionallections.2. In footnote 2 of their originalarticleMillerand Stokes refer to Eulau,Wahlke,Buchanan,and Ferguson(1959); HannaF. Pitkin'sthen unpublishedPh.D.dissertation 1961), which presumablyed to her later TheConceptof Representation(1967); de Grazia 1951); andFairlie 1940).3. The two most significant tudiesof the fifties in thisgenrewere:Turner(1951) and MacRae 1958).4. The operationaldefinition was expressedas follows: In each policydomain,crossing herankings f Congressmenndtheirconstituencies ivesanempiricalmeasureof the extent of policy agreement etween egislatorand district. The measureitself was expressedas follows: To summarize he degreeof congruencebetween egis-lators and voters, a measure of correlation s introduced Millerand Stokes, 1963,p. 49 and ft. 10).5. See, e.g., Stone (1976, p. 8), where one finds the bland statement:Representations conceived as congruenceor agreementbetweenthe behaviorof thelegislatorand the opinion of the constituency on comparablepolicy dimensions.Compare his also with Clausen 1973, p. 128): Given he principalorientationof thisbook, the policy orientation,representations furtherdefinedas the congruence f thepolicy requirements f the constituencywith the policydecisionsof the representative.6. Instead, to illustrate the constituency's sanctioning power throughelections, Miller and Stokes relied on data for a single Congressional istrict in a casewhich is both inappropriate nd deviant,involvingthe defeat of CongressmanBrookHays in the Fifth ArkansasDistrictwhere all votersin the sample(N=13) hadread orheard something about Hays and his write-inopponent. But, as Miller and Stokesadmit,the casewasinappropriate:he votersprobablyknew little aboutHays' egislativerecord in the previous Congressbut punishedhim for his non-legislative ole in theLittle Rock school crisis. The Hays case indicatedthe powerof an arousedelectoratein an unusualsituation;but even if theyknew the legislative ecordsof theirrepresenta-tives, electorates arerarelyso arousedoverany one of the manylegislative ssues withwhichrepresentativeseal.7. Wahlke(1971). The core ideas of this article were first presentedbyWahlke n a 1967 paperbeforethe SeventhWorldCongress f the International oliticalScienceAssociationn Brussels,Belgium.8. We could cite here, of course, as extensive institutional iteraturewhich has come to be neglected by behavioral tudents of representation.For aparticularlyuseful recent introductionthat paints a broadcanvas,see Birch (1971).9. The problem with causal analysesof phenomena ike influence or re-sponsivenesss that the direction of the relationships o which they presumably efercannotbe inferred rom the causalstructure f the statisticalmodel thatmaybe applied.The statisticalmodel assumesthe existence of a conceptualisomorphismbetween itsorderingof the variablesand their real-worldordering.The existence of a possibleisomorphismbetween the direction of a political relationshipand a causal relationbetween two variables n a statisticalmodel was broughtto the attentionof politicalscientists in a series of papersby HerbertA. Simon. Attemptingto define politicalpower,Simonfound that thedifficultyappeared o reside in a very specifictechnical


21/23

252 Heinz Eulau and Paul D. Karpspoint; influence,power, and authorityare all intended as asymmetrical elations. Itseemed to him that the mathematical ounterpartof this asymmetrical elationap-pearedto be the distinctionbetween independentand dependentvariables-the inde-pendentvariabledetermines he dependent,and not the converse. But,he pointedoutin a significantpassage hat causalanalystsseemat times to overlook, inalgebra, hedistinctionbetweenindependentanddependentvariable s purelyconventional-wecanalwaysrewrite our equationswithout altering heir content in such a way as to reversetheirroles. The problem,then, is one of giving operationalmeaning o the asymmetrythat is impliedin the definitionof influenceorpower: That s to say,for the assertion,'A has poweroverB,' we cansubstitute he assertion, A'sbehaviorcausesB'sbehavior.'If we can define the causalrelation,we can define influence,power, or authority,andvice versa. See Simon(1957, p. 5). The most significanterm n Simon'sexplicationofthe causalrelation s vice versa. It suggests hat the definitionof the causal elationand the definitionof the phenomenon o be causallytreated(hereinfluence)areinter-dependentevents. In otherwords: If we can defineinfluence,we candefinethe causalrelation. 10. UnfortunatelyFiorina then characterizeshe new-styleCongressmansan ombudsman. This attributionis inappropriatebecause an ombudsman, houghpresumably vailable or the settlementof grievances,s not the kind of hustler whomFiorinasees as comingon the stageof representation.Of course,both roles seemto beinvolved-thatof ombudsman nd thatof hustler.11. For example, Froman (1967, p. 36) found that in the 88th Congress(1963-1964) 15,299 bills and resolutionswere introduced n the House of Representa-tives, whereasonly 1,742, or a little over 11 percent, were reportedby committee.12. For the distinction between style and focus of representation, eeEulau,Wahlke,Buchanan, ndFerguson 1959).13. Fenno (1975) has also seen the need to decompose the concept ofconstituency.He suggeststhat congressmenperceiveseveraldistincttypes of constitu-enciesto whichthey respond n differentways.

REFERENCESBirch,A. H. 1971. Representation.London: St. Martin'sPress.Campbell,Angus,Converse,PhilipE., Miller,WarrenE., and Stokes, Donald E. 1966.Electionsandthe Political Order.NewYork:Wiley.Clapp,Charles.1963. TheCongressman:His Job as He Sees It. Washington: rookingsInstitution.Clausen,AageR. 1973. How Congressmen ecide.New York:St. Martin'sPress.Cnudde,CharlesF., andMcCrone,DonaldJ. 1966. TheLinkagebetweenConstituencyAttitudes and CongressionalVoting: A CausalModel, AmericanPoliticalScienceReview60 (March,1966): 66-72.de Grazia,Alfred. 1951.PublicandRepublic.New York:Knopf.Easton,David.1965.A SystemsAnalysisof PoliticalLife. NewYork:Wiley.Edelman,Murray.1964. The Symbolic Uses of Politics. Urbana:Universityof IllinoisPress.

1971. Politics as Symbolic Action. Chicago: Markham.Eulau,Heinz. 1967. ChangingViews of Representation, n Ithiel de SolaPool, ed.,ContemporaryPolitical Science: Toward Empirical Theory. New York:McGraw-Hill.


22/23

The Puzzleof Representation 253Eulau, Heinz, Wahlke,John C., Buchanan,William,and Ferguson,LeRoy C. 1959.' TheRole of the Representative: omeEmpiricalObservationsn the Theoryof Edmund Burke, American Political Science Review 53 (September,1959): 742-756.Eulau, Heinz, and Prewitt, Kenneth. 1973. Labyrinthsof Democracy: Adaptations,Linkages,Representation, nd Policies in UrbanPolitics.Indianapolis:Bobbs-Merrill.Fairlie, John S. 1940. 'The Nature of PoliticalRepresentation, AmericanPoliticalScienceReview40 (April-June, 940): 236-248,456-466.Fenno,RichardF. 1973. Congressmenn Committees.Boston:Little,Brown.

-----__. 1975. Congressmen in their Constituencies. Prepared for delivery at theannualmeetingof the AmericanPoliticalScienceAssociation.Fiorina, MorrisP. 1974. Representatives,Roll Calls,and Constituencies.Lexington,Mass.:D.C. Heath.

. 1977. The Case of the Vanishing Marginals: The Bureaucracy Did It,AmericanPoliticalScienceReview71 (March,1977): 177-181.Jackson, John E. 1974. Constituenciesand Leadersin Congress.Cambridge,Mass.:HarvardUniversityPress.Kingdon,John W. 1973. Congressmen'sVotingDecisions.New York:Harper ndRow.Lasswell,HaroldD., Merriam,CharlesE., and Smith, T. V. 1950. A Study of Power.Glencoe,Ill.: FreePress.Loewenberg,Gerhard.1972. Comparative egislativeResearch, n SamuelC. Patter-son and John C. Wahlke,eds., ComparativeLegislativeBehavior:FrontiersofResearch.NewYork:Wiley.MacRae,Duncan,Jr. 1958. Dimensionsof CongressionalVoting.Berkeley:Universityof CaliforniaPress.Matthews,Donald R., and Stimson,JamesA. 1975. YeasandNays: NormalDecision-Makingn the U.S. Houseof Representatives.New York:Wiley.Miller,WarrenE., and Stokes, DonaldE. 1963. Constituency nfluence n Congress,AmericanPoliticalScienceReview57 (March,1963): 45-56.Muller,Edward N. 1970. The Representationof Citizens by Political Authorities:Consequences for Regime Support, American Political Science Review64 (December,1970): 1149-1166.Patterson,SamuelC., Hedlund,RonaldD., and Boynton,G. R. 1975. Representativesand Represented:Bases of Public Supportfor the AmericanLegislatures.New York:Wiley.Peterson, Paul E. 1970. Formsof Representation:Participationof the Poor in theCommunityAction Program, AmericanPolitical Science Review 64 (June,1970): 491-507.Pitkin,HannaF. 1961. TheTheoryof Representation. h.D. dissertation,Universityof California,Berkeley.. 1967. The Concept of Representation.Berkeley:Universityof CaliforniaPress.Prewitt,Kenneth,and Eulau,Heinz. 1969. PoliticalMatrixand PoliticalRepresenta-tion: Prolegomenon o a New Departure rom an Old Problem, AmericanPoliticalScienceReview 63 (June,1969): 427-441.Simon,HerbertA. 1957. Modelsof Man.NewYork:Wiley.Stone, WalterJ. 1976. Representationn the United StatesHouseof Representatives.Ph.D. dissertation,Universityof Michigan.


23/23

254 Heinz Eulau and Paul D. KarpsTurner,Julius. 1951. Partyand Constituency:Pressures n Congress.Baltimore:Johns

HopkinsUniversityPress.Turner,Julius,and Schneier,EdwardV., Jr. 1970. Partyand Constituency.Baltimore:JohnsHopkinsUniversityPress.Verba,Sidney,andNie, NormanH. 1972.Participationn America:PoliticalDemocracyand SocialEquality.New York:Harper nd Row.Wahlke,John C. 1971. PolicyDemandsand SystemSupport:The Role of the Repre-sented, BritishJournalof PoliticalScience1 (July, 1971): 271-290.Zeigler,L. Harmon,and Jennings,M.Kent,withG. W.Peak. 1974. GoverningAmericanSchools.No. Scituate,Mass.:DuxburyPress.

Date post:	03-Jun-2018
Category:	Documents
Upload:	sourishjha
View:	219 times
Download:	0 times

Puzzle ofThe Puzzle of Representation: Specifying Components of Responsiveness Author(s): Heinz...

Documents