Research Division Report #9

Performing Artsand Museums:A Critical Review

National Endowmentfo[ the Arts ~

November 1978

A Study by Paul DiMaggio, Michael Useem and Paula Brown, Center for the Study of Public Policy, November 1977

_pREFAdE_~ " ._

This Research Report had its origins in aseries of meetings organized in 1975 byPolly Buck, the National Endowment forthe Arts’ progra~L director for the per-forming arts. The directors of theprograms of theatre, dance, and musicparticipated in these meetings with theResearch Division staff to outlineresearch needs. The study of the per-forming arts audience was an area ofspecial concern. The performing artsprograms were receiving applications forsupport of audience studies from insti-tutions in their fields. The programdirectors recognized that many audiencestudies had been completed but wereunevaluated and could not be used for thegreatest advantage. A critical reviewwas thought to be needed before under-taking new audience studies. Subsequentl~at the direction of the Arts Endowment’sChairman, Nancy Hanks, the concept for acritical review of audience studies wasexpanded to include museums as well asthe performing arts.

Ih 19760 a competitive program solici-tation was released requesting proposalsfrom individuals or organizations wishingto undertake a critical review of ¯audience studies. The proposals wereevaluated and an award made to the Centerfor the Study of Public Policy inCambridge, Massachusetts. The project,.under the leadership of Michael Useem andPaul DiMaggio, with the assistance ofPaula Brown, included a diligent searchfor audience studies. Initially, it washoped that i00 to 150 studies would befound. A number of this magnitude wouldprovide an excellent basis for a criticalreview. However,, to everyone’s surprise,270 completed audience studies were foundthat were made available to the investi-gators. All but five have now been madeavailable to additional users, asexplained in the last paragraph of thePreface.

The critical review examined two kinds ofquestions. The first is about what pastaudience studies show when analyzed as aset. The second group of questions wasconcerned with the methodology ofaudience studies and the bringing togeth-er of the experience so that caveats andguidance for future audience studiesmight be developed. Both of these sidesof the project are presented in thisreport.

An important caution to the reader: thecoverage of institutions is not a.resultof a structured sampling procedure aimedat providing data on all American artsaudiences. Some categories of audiencesare poorly represented among the audiencestudies. We believe that this is anindication that few, if any, audiencestudies have been made by the institutionsserving these audience categories. Also,the study restricted itself to the liveperforming arts and to museums: it made noattempt to collect audience studies in themedia fields. Research Division Report#4, Arts and Cultural Programs on Radioand Television, contains useful infor-mation about the audience measurementprocedures utilized in the media fieldsof radio and TV. (Copies of this reportare available on request.)

During the study, the investigators met,corresponded, or talked with about~ 600individuals who had been involved with oneor more audience study projects. In avery real sense, this study report is adistillation of the combined experience ofthese people. Their contribution to theproject was vital and is greatly appre-ciated by the Arts Endowment. It is hopedthat this report will become a benchmarkthat will allow their experiences to beused effectively by future audienceinvestigators and thereby contribute toboth the improvement of the art of study-ing the audience and to the capability ofarts and cultural institutions to servetheir audiences.

The investigators collected 270 audiencestudies. Of these, five were retained asconfidential with respect to further exam-ination beyond the project. The remaining265 audience studies have been brought tothe Arts Endowment’s offices and areorganized as a study collection. Thesestudies are identified in Appendix II.Visitors to the Arts Endowment may makearrangements to examine these studies bycontacting Mrs. Chris Morrison0 Librarian,National Endowment for the Arts, 2401 EStreet N.W., Washington, D.C. 20506,telephone (202) 634-7640.

Research DivisionNational Endowment for the ArtsOctober 1978

SUMMARY

INTRODUCTION

CHAPTER i: THE NATURE OF THE ARTS PUBLIC

The StudiesBasic Demographics

GenderAgeEducationOccupationIncomeRace and EthnicitySummary of Demographics

Additional Issues in Audience ResearchThe Arts Audience Over Time~dience StructureEconomic and Political Impact

CHAPTER 2: QUALITY AND IMPACT OF ARTS AUDIENCE STUDIES

The Arts Audience SurveyPredicting Quality in Arts Audience StudiesFactors Predicting Research QualityThe Correlates of QualityPredicting the Utility of Arts Audience Studies

CHAPTER 3: ORGANIZATIONAL FACTORS AFFECTING RESEARCH UTILITY

The Purposes of Audience ResearchPolitical factorsOpportunityGeneral concerns

The Impact of Audience StudiesInstrumental applicationsPolitical applications

The Role of Audience Research Findings in Arts ManagementFactors Promoting Research Utility

Study attributesPersonal commitmentExternal factors

Factors Preventing Research utilityStaff turnover and lack of resourcesHostility and lack of interestResearch planningcommunication and follow-throughReport contentStudy executionTechnical quality

Conclusion and Recommendations

CHAPTER 4: AN AGENDA FOR ARTS AUDIENCE RESEARCH

APPENDIX I: BIBLIOGRAPHY

APPENDIX II: LIST OF STUDIES

6

8

i013

34

42

424344465O

56

56

58

6063

66

70

72

75

79

ii

Table 1

Table 2

Table 3

Table 4

Table 5

Table 6

Table 7

Table 8

Table 9

Table l0

Table ii

Table 12

Men in Audiences by Art Form

Age of Audiences by Art Form

Percentage of Audiences in Five Educational Categoriesby Art Form

Occupational Distribution of Audiences

Occupational Distribution of Audiences by Art Form

Median Income of Audiences by Art Form

Time Trends in the Gender, Age, Education, Occupation,and Income Compostion of Performing Arts Audiences

Deviation from Average Study Quality byInvestigator Background, Resources, andInsitutional Setting

Simple Correlations and Regression Coefficients ofAudience Study Quality with Investigator Background,Resources and Institutional Setting

Deviation from Average Audience Study Internal andExternal Utility by Investigator Background, Resources,Institutional Setting and Quality

Frequency of Instrumental and Political Applicationsof Audience Research Results

Frequency of Factors Cited Affecting Use of AudienceResearch Results

14

17

20

22

24

30

35

47

49

52

59

64

iii

Audience research for-the performing artsand museums has increased rapidly inrecent years to the point where it is nowco~anonplace. We have assembled andanalyzed a number of these studies withtwo purposes in mind. First, we areinterested in the social profiles of thesegroups of attenders. Who is the culture-consuming public? Our answers to thisquestion should be of use to arts adminis-trators, government agencies, otherresearchers, and the concerned public.Second, we are interested in the qualityof audience research. How well donetechnically is the work? Further, howuseful are the results of these surveysand questionnaires to the organizationswho undertake them? Ours is the firstsystematic evaluation of the quality ofaudience research, and we hope ourfindings will be helpful not only toother researchers but to those managersand directors who contract for audiencestudies and who review the results. Inthe course of both discussions--onaudience composition and on researchquality--we review and summarize per-tinent literature in the field.

There is a need to know who the culturalaudience is--and is not--before policydecisions are made on how to expand anorganization’s activities or on how to bemore responsive to a wider range of thepublic. Our data on the composition ofperforming arts audiences and museumvisitors are based on 270 studies, bothpublished and unpublished, which we listat the end of the report. We gatheredthis material after a search of libraries,indexes, and bibliographies. A mailedinquiry to over 1,200 cultural institu-tions and organizations brought more than600 responses.

Nearly all of the studies we received weredone after 1970, although originally wehad hoped to include earlier ones as well.These were done for institutions whichvaried widely in size, function, andlocation. They do not, however, representa precise cross section of the culturalactivities involved. In the field of theperforming arts, the studies were ofaudiences for theatre, dance, ballet,classical music, and opera. For museumsthey were of visitors to art, history,science, and other museums. We did notreceive audience studies for jazz, folk/ethnic music, popular art, and did notseek studies on any of the media arts ofradio, television, and film. The studies

came from forty-one states and theDistrict of Columbia. They were notweighted according to audience size, butare what we were able to collect.

For categorical variables (gender, edu-cation, occupation, and race) we tabulatedpercentages of respondents in thosecategories used in the greatest number ofstudies. For continuous variables (ageand income) we calculated median figuresfor each audience studied.

Gender. The stereotype of the arts as apredominately feminine activity did nothold true. Women only slightly outnum-bered men in relation to their percentageof the population as a whole. The men inthe audiences were a median 43 percent forthe performing arts, 46 percent formuseums, and 49 percent for the populationAudiences for ballet and dance were themost heavily female (60 percent) andvisitors to science and history museumswere the most heavily male (53 percent),but there were large variations fromaudience to audience both within andamong cultural types. For example, therewere more men than women in one-quarterof the studies of the performing arts andin two-fifths of the studies of museums.The median percentage of men ranged from31 to 58 in the performing arts and from30 to 71 in museums. In other words, theoverall figures on gender would be a poorpredictor of the make-up of a particularaudience. The day of the week and, perhaps,the content of the event had some effecton the gender of the audience.

A__q~. The median age of visitors tomuseums was thirty-one and of audiencesfor the performing arts was thirty-five.This age profile is similar to that of thegeneral population since the figures liebetween the median age of the entire U.S.population (twenty-eight) and the medianage of the population sixteen and over(forty). The data from the studies wascomplicated by a variety of restrictionson the age of respondents.

Again, variability was great within andamong the types of perfornting arts andmuseums. Audiences for opera and class-ical music were older than for theat~-e andballet/dance. There was less variabilityamong museums but science museums hadaudiences younger by a median two yearsthan art museums. The seasons of the yearand the time of performance in some casesaffected audience age. Audiences were

younger in the sunder: the median age grew -muSeum-visitors as°�.0mpar~-t~. 59_percentolder from weekend evenings to weekday ¯ ~professionals among¯per-forming .artsevenings to matinees (the oldest)~ The~e ¯ audiences. Art museums differed fromis some evidence that content of th~ent ~-¯other types-of museums with a higher ratioand that geographic region affect audience of professionals. In fact, they had theage.

Education. Educational attainment is thesingle most important variable in thesocial profile of attenders. Perhaps thisis because schooling provides formaltraining in the arts, a social milieuencouraging participation, opportunitiesto attend, and a family habit which ispassed down to children. Althoughaudiences varied considerably, medianeducational attainment was consistentlyhigh. Among attenders, 30 percent hadgraduate training, 54 percent had four-year college degrees (as compared to 14percent of U.S. adults), 22 percent hadno schooling beyond high school (74percent of U.S. adults), and 5 percent hadnot completed high school (38 percent ofU.S. adults).

It is likely that children were under-represented in audience samples althoughthe extent cannot be determined fromanalysis of the studies.

Median education was higher for performingarts audiences than for museum visitors,higher for ballet and dance than fortheatre, and higher for art museums thanfor science and history museums. Althoughfor education, as for other variables,museums served a somewhat broader publicthan the performing arts, culturalaudiences were closer to each other thanto the general public.

Occupation. The complexity of classifyingoccupations into manageable categories,especially given the vast array of schemesin the studies we analyzed, means againthat tendencies are more important thanexact figures here. Even with this qual-ification, however, one of the moststriking findings was the very highpercentage of professionals amongattenders and the very low percentage ofblue-collar workers. Professionals madeup a median 56 percent of employed personsin the audiences but were only 15 percentof the employed work force: blue-collarworkers were 4 percent of employed personsin the audiences but 34 percent of theemployed work force. (At that, the per-centage of blue-collar may be slightlyoverstated due to classificationambiguities.)

same percentage (59) as did the performingarts, which differed little among thevarious types.

Teachers (including college and university)were especially numerous among profes-sionals with 21.0 percent of the attendersoverall but 4.1 percent of the work force.In comparison to their share of theprofessional work force, they exceededtheir share of professionals ammngattenders by a third.

If professionals were high in all meas-ures, blue-collar workers were low. Onlyamong "other museums", which showed 17percent blue-collar, was there any changein the pattern. Blue-collar workers were2.8 percent of performing arts audiencesand 3.1 percent of art museum visitors.And excluding museums other than art,thirty-four of the fifty-two studiesshowed blue-collar workers in numbersamounting to less than one-tenth of theirshare of the work force as a whole.

Among other occupations, managers hadhigher percentages in the performing artsaudiences than in the work force, thoughnot nearly so high as professionals. Somestudies merged professional and managerinto one category. When we did so todiscover a rough index of high statusoccupations, we found that 69.5 percent ofattenders fell into this combined categorywhich made up 25.5 percent of the work

force.

Clerical/sales and homemakers had percent-ages below their share of the work forcebut the latter were highly variable fromaudience to audience (5 to 52 percent).Students were attenders to a very highdegree, though again the range was great(0 to 63 percent). Retired and unemployedwere consistently low in relation to theirshare of the population.

Again, museum visitors were somewhat m~rerepresentative of the public than theperforming arts audiences. For example,there were 42 percent professionals among

Income. High income is associated withcultural participation but is not thecause, at least not nearly to the degreethat education and occupation are. Whenall three factors are controlled, incomedoes not seem to predict attendance whereeducation and occupation do.

Since it is sensitive, private information,income figures are liable to distortion(there were nonresponse rates up to 29percent). All figures from the studieshave been converted into mid-1976 dollarsfor comparability. In this regard, a

median figure for the U.S. as a whole" change ~n tim~-ov~.~he~l~st fifteen years

would be a family income of $14,000.for the.performing;art-S~ ~the only studies

..... -~-for which we had-sufficient data. Changes

Throughout the studies~familyinc~m~;~as /~ ~imight .be g.oing .on within or among audi-consistently above this baseline. For the ences for individual art forms but theperforming arts, the-median was $19,000and for museums $17,000. When audiencesfor outdoor dramas were excluded, themedian income for the performing arts washigher, $20,250. Also, once this group ofstudies was disregarded there was a gooddeal of consistency from type to type inthe performing arts.

AS they did with regard to education andoccupation, museums attracted a somewhatmore representative cross section of thepublic in terms of income. But still,only one museum study reported a medianfamily income below the general population.

Race. Very few studies reported attend-ance by race or ethnic background. Basedon this scarce data, we found thatminorities--blacks and persons of Hispanicand Oriental background--were 7 percent ofthe audience but 20 percent of the popu-lation. Museums other than art were moreinclusive with Ii percent minorityattendance. Blacks made up 3.0 percentof the attenders but 12.3 percent of theurban population. Selected comparisonsof individual studies with urban areastatistics also showed this pattern ofunderrepresentation.

The low ratios of minorities are probablydue to the fact that these groups on theaverage are younger, have less educationand lower incomes, and are less likelyto work in professional occupations.

A review of cross-sectional surveys inwhich people reported on their ownattendance habits showed wide variationfrom place to place and time to time, butin .general the figures for minorityattendance are much higher than in theaudience surveys.

In sum, the studies show that the culture-consuming public is more educated, hashigher incomes, and has higher status jobsthan the general public. Museum visitorswere somewhat more representative of thepublic than performing arts audiences.The difference may be attributed in partto the lower median age of the museumpopulation. Theatre audiences hadslightly lower scale demographic profiles~han the other performing arts, and artmuseums were higher scale than the othertypes of museums.

We could find no evidence that audienceswere becoming more democratic. None ofthe variables showed any significant

necessary aggregates of our.data wouldconceal these.

In our review of audience studies we alsolooked at the frequency of attendance,economic impact, and public attitudestoward the arts. We discovered thatfrequent attenders are more educated andprobably have higher incomes than infre-quent attenders. Also, frequent attendersof one of the performing arts are likelyto go to performances of the other arts(frequent theatregoers are the exception).The few studies which examined economicimpact showed that cultural institutionsseem to draw visitors to a city and theresultant spending benefits certainsegments of the economy substantially.And finally, studies of attitudes towardthe arts showed widespread public supportwith majorities or near majorities infavor of government subsidy. Local moneytends to be preferred to federal, fundsfor institutions are preferred to fundsfor artists, and museum support is pre-ferred over the performing arts. Ingeneral, public attitudes seem to begrowing more supportive along with theincrease in government funding.

Good audience research is scarce. Thebest single study of cultural audiences todate has been William J. Baumol andWilliam G. Bowen’s work in 1966, Perform-ing Arts--The Economic Dilemma. (Theirdata showed a more elite audience thanours, probably because they analyzedprofessional performing arts only.) Apartfrom this landmark work, quality variestremendously. One of our main interestshas been to measure the factors affectingquality and to relate these to the use towhich audience research findings are put.We hope our analysis will provide guide-lines--though certainly not hard and fastrules--for those organizations consideringaudience research.

Detailed questionnaires were sent to ll2directors of studies done since 1970. Ofthese, 86 responded with the informationrequested on such factors as the profession,education, and experience of the investiga~

tor: organization conducting the research,the project’s budget, the research methodsemployed: and the applications made of theresults. These questionnaires along withthe study reports themselves were ratedaccording to a checklist of specificresearch procedures. We measured bothinternal validity (whether the explanatiodoffered is the true cause) and external

validity (whether the results can be gen- cent of ~he’use~ ~[[~i~et±ng 120eralized) and developed a single quality ~.percent), further.researcn (12 percent),scale. We used multiple regression~.~.-.~~. L. programming (6 percent), and politicalanalysis to determine the eff~cts"~f@a~h reasons (22perbent internal and 12study characteristic, holding all others.constant.

The result was that budget and investi-gator’s profession were of greatestimportance in producing quality. Forexample, a budget of at least $i,000 anda study director with a background insocial science, or in a miscellaneouscategory of other research-relatedprofessions, did very well on the qualityscale. Budget and profession explained63 percent of the tota! variation inquality.

percent external).

Research findings entered into policydecisions in marginal and indirect ways.It was not only a matter of results beingapplied when they supported existingattitudes and ignored when they did not--though this was characteristic--but it wasalso that results might fly in the face ofother priorities: "Data step on toes," asone study director put it. Further, mostcultural organizations have severelylimited means to follow up on the directimplications of audience research.

We also developed utility scales tomeasure the application of research topolicy, based on the reports of the studydirectors. These looked at internal use(e.g., ticket pricing and exhibit content)and external use (e.g., public relations,funding, and audience expansion policy).We found no significant correlationbetween use, technical quality, and thecorrelates of quality--budget, invest-igator profession, and institutionalsetting. The only factor which had anysort of importance was the interaction oftwo variables: experienced in-houseinvestigators produced studies that werefound to be more useful than those doneby either inexperienced in-house oroutside investigators. And these twovariables produced a small and inconsis-tent relationship which explained lessthan i0 percent of the variation in use.

The extent to which the studies wereapplied varied sharply and neitherquality, its variables, nor common senseexplained why. In order to discover theanswer, we looked intensively at twenty-five audience studies which were selectedon the basis of type, region, andcurrency. Also, we conducted forty-twosemi-structured interviews with studydirectors and with those responsible forstudy use. As additional background,unstructured interviews were conductedwith twenty-five more directors and users.

Through our interviews we learned that,contrary to conventional wisdom, researchwas not undertaken to solve specificproblems and findings were applied in avariety of ways. The chief motives forundertaking audience research were forpolitical leverage or because the oppor-tunity was offered gratis or out of avague sense of concern for more infor-mation of some sort. The results--seventy-seven applications were mentioned--were used for physical planning (29 per-

In addition to being applied when theyconfirmed the suspicions of administrators,results were also likely to be used whenthey were championed by an influentialperson, when the authority of outsidersgave them legitimacy, and when researcherswere involved in staff deliberations.Along with the limits on use imposed bylack of funds, results were not used whenthere was high staff turnover (a commonproblem in cultural organizations) andwhen they were confusingly reported, notfollowed up on, or seemed trivial orinconclusive.

It is clear that research findings con-tributed to policy in highly indirectways as reinforcing, suggestive,expressive, and symbolic gestures thatdepend little on the precise technicalmethods employed. The lack of concernover technical quality in audienceresearch is a rational response to theenvironment in which these organizationsoperate.

But that environment is changing, in partbecause of budget pressures and a generalshift in attitudes toward researchplanning. It is likely that better usewill be made of better research, and werecommend the following: support forsystematic planning in the arts with someconsensus as to the role of audienceresearch: the creation of an informationclearinghouse to publicize and disseminatearts research: the establishment of localconsortiums for cooperative arts researchto aid institutions that cannot affordtheir own work: and workshops on socialscience methods for managers and admin-istrators of cultural institut~.ons.

There are a number of gaps in our know-ledge about audiences which requireseveral approaches. First of all, we needon a national basis routine gathering ofdescriptive statistics over time. These

should be from a sample stratified ~cc-oTd- It-seems to .us that the :more local studiesing to institutional type, region, degree -~ which-are-publ-ish~d[i-~.it~--9~etter for allof urbanization, programming policy,

concerned and if a-clearlnghouse were

professional status, and ticket prices." ~ --available there would be a pool of infor-

....... ~ ~...~.%~-~~i~. - omation all_ could usefully draw upon.

Individual organizations need to stand-ardize their survey data in order to makeresults more useful to themselves and toothers. In conducting a survey, theyshould base their demographic categorieson census schemes: any other or specialcategories should be highly differentiated.A technique which can be used easily andto good effect by local organizations iscross-tabulation- And information can beincreased with little added effort byusing census frequencies for metropolitanresidents and by asking how often respon-dents attend during a given period. Also,quasi-experimental design--controlledstudies before and after limited policychanges--is a useful methodology thatcould be employed more often.

Nonattenders, who are of great interest toarts managers, pose a problem for audienceresearch and may require special attentionthrough in-depth interviews.

A major question of audience research iswhether there is one audience or many, andwhether any distinctions can be madeaccording to audience types and theirresponsiveness to such questions as priceand program content. Other largequestions have to do with the process ofsocialization of arts audiences and thepublic for the arts in forms other thanlive performance.

Certainly cultural organizations canimprove the quality and use of audienceresearch by shifting their priorities.But the systematic use of audienceresearch on a wide scale after the fashionof governmental agencies and privateindustry may be prohibitively expensive-

To the general play-goer, it is presumed that the mostinteresting part of a theatre is behind the scenes. Toactors and actresses, naturally enough, the chief inter-est lies with the audience - Before the Footlights .....I never tired of studying the many-headed animal - theAudience. I love to take it up in its different elements,and ponder it - looking out from a cozy corner in a stage-box, myself unobserved. (Logan, 1871: 291)

Although research on the arts audiencedates back to the museum visitor studiesin the 1920s (Robinson, 1930) and FederalTheater Project performances in the 1930s,such research appears to have been under-taken on a large scale only in the lastdecade or two. Beginning in the 1950s--with the museum studies of de BorhegyiHanson, and their colleagues (1968) andAbbey and Cameron (1959, 1960)--andcontinuing in the 1960’s--with the perform-ing arts surveys of Baumol and Bowen(1966)--gathering information aboutaudiences in museums and performing artsinstitutions has grown to the point whereit is nearly con~onplace. Of more than600 arts organizations responding to onerecent survey (Johnson and Prieve, 1976),23 percent had conducted audience surveyswithin the previous five years. We founda similar situation: out of 612 artsorganizations, 27 percent had undertakensuch studies in recent memory, and manyothers were preparing to do so. Further-more, the generous cooperation wereceived from overworked and question-naire-weary individuals in theatres,museums, orchestras, and other artsinstitutions was itself an expression ofa keen interest in the subject ofaudience surveys. In fact, a surprisingnumber of arts managers sought our adviceon specific aspects of the design orexecution of audience studies.

In relation to the growing study of artsaudiences, our report has two aims.First, we have gathered research on thecomposition, attitudes, and preferencesof arts audiences and have put togethera description of important features ofthe American arts public. In doing so,we drew upon reports, questionnaires, andother materials from more than 250research projects-

The second aim has been to assess thequality and utility of the arts audienceresearch. This report represents thefirst evaluation of research in this

area: and it is one of the first to studyexplicitly both how well research has beencarried out by social-scientific standardsan__d how useful it has been to the organi-zations on whose behalf it was undertaken.(For a brief but illuminating study ofmarketing research by symphony orchestras,see Wainwright (1973)).

We began to gather our information in thefollowing manner. First, an exhaustivelibrary search was conducted for publishedaudience studies and an inquiry form wasmailed to over 1,200 museums, performingarts organizations, arts councils, andother organizations concerned with thearts. The form requested information on,and copies of, any audience research withwhich the recipient had been involved orwas acquainted. This search eventuallyyielded materials on 270 studies.

Second, a longer survey form was sent tothe directors of each of more than 100studies that we had obtained by January l,1977. The survey, based on a review ofrelevant methodological materials and onmore than two dozen unstructured inter-views with arts administrators andresearchers, requested information on thestudy director, conducting organization,research budget and funding, researchmethodology, and policy applications.Eighty-six directors responded within theallotted time of approximately threemonths.

Finally, structured interviews with forty-two directors and users of twenty-fiveaudience studies were conducted in orderto better understand the purposes ofaudience research and the reasons why somestudies yield more useful findings thanothers. The research project selectedfor case study represented a cross sectionof art forms and study types.

Our findings and methodology are reportedin three chapters. Chapter One presentsa synthesis of data on audience

composition reported by the studies in our~ _and t~pist~ all ~t~-e~ally great skill,possession. We have analyzed information ¯

good humorl an~ fo£~l~~H~e~ A~ainst not- ....

on gender, age, educational attainment, -inconsiderableodds0 she succeeded in

occupation, income, and race of-.~~.!_" /-maintaining order and organization in aattenders for various art forms. Inaddition, Chapter One presents inforrmation on changes in audience compositionover time, differences between frequentand infrequent attenders, and the findingsof studies of the economic impact of thearts and of public attitudes towardgovernment financing of the arts. ChapterTwo, based on the survey of study direc-tors described above, analyzes thedeterminants of the technical quality andthe effects of quality on the use ofaudience studies in policy decisions.Chapter Three draws on the case studyinterviews and describes the reasonsaudience studies are undertaken, the usesthey serve, the ways they enter thedecisionmaking process, and the factorsinfluencing their use. Finally, ChapterFour presents an agenda for furtherresearch.

It should be noted that references areprovided in two ways. References toaudience studies are indicated in thetext by the study number (e.g., #17) andreported in the List of Studies given inthe back. Other references are cited byauthor and year of publication and arereported in the Bibliography.

This report may be useful to arts managersand policy makers in several respects.

~ singuiar[y centripetal environment.

Thanks, also, to John Case, Nancy,Lyons,Larry Baker~ Lee Aitken, Barbara Beelar,and the other residents of the Center forthe Study of Public Policy, who gave us acomfortable and friendly home and providedgood advice and moral support on numerousoccasions.

We are also thankful to Joseph Farrell,Bradley Morison, and Joseph Zeigler, whowent beyond the call of courtesy to makeavailable to us both their research andtheir extensive experience in the study ofarts audiences. We are indebted, as well,to Donald Newgren, who donated to us alarge collection of museum visitor studiesthat he obtained during a study of museumresearch: and to Lorraine Brown of theResearch Center for the Federal TheatreProject at George Mason College for makingcopies of several Federal Theater Projectsurveys available to us.

The following people also provided valu-able information at various stages of thisproject in numerous ways: Duncan Cameron:Irving Cheskin: James Copeland: StephenCouch: David Cwi: Kevin R. Diels: EmilyFarnham: Carol Ann Feldman: Arnold Foster:Janet Gracey: Rena Hall: Walt Haney: PhilHyde: Christopher Jencks: Samuel Kaplan:

First, it presents a comprehensive overview Jerome Karabel: Jeff Klopek: Robertof audience composition and makes clearthe limits of the information now avail-able.

Second, it compares the results ofaudience research carried out undervarying circumstances and illuminateswhy much of that research is less thansatisfactory. While there is no easyrecipe for ensuring that audience researchcan be both useful and of the highestquality, the material in these chaptersdoes indicate the complexity of the pro-cesses leading to good research and toresearch use. These chapters alsoprovide insights into those aspects ofresearch management about which somethingcan be done. However, it should be notedthat this report presents no guidelinesfor conducting audience studies. Fordetails on how to go about surveying anaudience or set of visitors, see Baumoland Bowen (1966: Appendix IV-l): Camer¢~nand Abbey (1960a, 1960b, 1961): Mann(1966): and Newgren (1972).

Lakota: Richard Light: Margaret Parsons:Laurie Perman: Richard Peterson: E. ArthurPrieve: Richard Sheldon: Harris Shettel:Ann smith: Mark Sumner: Mary Tassick:Ellen Thurston: Harrison White: ThomasWolf: and Peter Zeisler.

The opinions expressed and interpreta-tions offered are those of the authors anddo not necessarily reflect the views ofany of the individuals or organizationswhose assistance has been acknowledged.Any errors or misinterpretations are oursalone.

Paul DiMaggioMichael UseemPaula Brown

Our deepest thanks go to Jane Gallop, whoserved as this project’s administrator,librarian, accountant, editor, secretary,

-THE’tNATURE~OFTHE ARTS PUBLIC

The nature of the public for the arts inthe United States has been a source ofcontroversy and speculation for much ofthis country’s history. Alexis deTocqueville, the liberal French aristocratwho studied American democracy during the1830s, noted then that America’s Puritansimplicity and unbounded resources pro-vided more fertile soil for commerce thanfor art. Nonetheless, he suggested, asthe frontier closed and the Puritanlegacy was diluted, the natural tendenciesof democracy might eventuate in unprece-dented public involvement in the arts."Not only will the number of those whocan take an interest in the production ofmind be greater," he wrote, "but the tastefor intellectual enjoyment will descend,step by step, even to those who, inaristocratic societies, seem to haveneither time nor ability to indulge inthem" (Tocqueville0 1956: 162).

Tocqueville predicted the democratizationof both the production and appreciationof art as the United States became moremature. A half century later ThorsteinVeblen, the iconoclastic economist,presented a more pessimistic view.Having witnessed the rise of great for-tunes that Tocqueville had not forseen,Veblen feared that the arts (as well asmost aspects of culture, learning, andmanner) had become the playthings of therich--baubles and badges of social stand-ing less respected for their beauty orintrinsic merit than for their rarity andexpense. High culture, thought Veblen,would remain the preserve of the wealthybecause only they had the leisure toattend to it and the power to definewhat, in fact, would be considered "art"(Veblen0 1899).

The opposing perspectives of Tocquevilleand Veblen had been echoed in debatesthroughout this century. Recently, forexample, some writers have discerned acultural "boom," asserting that the arts,while previously the monopoly of an elite,have become central to the lives ofmillions of Americans. Alvin Toffler,perhaps the most optimistic spokesman forthis position, cites the rise of a massivemiddle-class constituency for the arts.In earlier years, he says, the artsaudience was composed of European-oriented,rich, alienated intellectuals andaspiring artists: but more recently

"millions of Americans have been attractedto the arts, changing the composition ofthe audience profoundly." While not allAmericans are part of the culture boom,"a major step toward democratization has,indeed, been taken." As a result, the"rise of a mass public for the arts can,in its way, be compared with the rise ofmass literacy in the eighteenth century inEngland" (Toffler, 1965: 34,51).

Other writers have taken a less sanguineview. Sociologist Herbert Gans arguesthat high culture remains the preserve ofa small circle of aficionados and adiverse "user-oriented" public thatincludes art patrons, collectors, highlyeducated professionals, and businessexecutives. But the masses are stillnot reached, for in the view of Gans highculture continues to serve "a small publicthat prides itself on exclusiveness"(Gans, 1974: 77).

Sophisticated critics and analysts havefailed to agree on whether the art publicis mass or elite. The reason, in part, isa matter of definition. Should the term"art" be restricted to paintings hangingin major museums, serious theatre, musicplayed by symphony orchestras, and tra-ditional or experimental opera and ballet?Or should we also include commercial andcommunity theatre, jazz, crafts, foreignfilms, and "pops" orchestras? By elite,do we mean the rich and the top executives,or does the elite also encompass theupper-middle classes and the college edu-cated? And does the arts public consist,say, of anyone who makes an annual visitto a local art museum, or should the termbe restricted to serious consumers of atleast one of the traditional art forms?Much of the disagreement about the com-position of the arts audience can beattributed to imprecise language. Yeta significant part of the controversy isalso due to the state of the research, forhowever the terms are defined, goodresearch on the public for the arts hasbeen--and to a great extent, still is--relatively scarce and inaccessible,difficult to compare, and often equivocalin its findings.

An exception is William Baumol and WilliamBowen’s careful and extensive study of theaudience for the professional performingarts. It remains a landmark work since

its publication in 1966. Their assess ....--~ that were-between~5-~and~~0 percent blue

ment indicated that Veblen’s insights" collar (#!6).~ And_.one. e~rlier-study of

were more accurate than Tocqueville’s--~ ~.-.the Milwaukee Public Museum revealed that

On the much-touted cult-ura~ boom-.~f~i~h~~-.~ -~ _visitors were nearly representative of

1960s0 they wrote: "evidence of a modestthe American public: of the employed

expansion in performing arts activity .... " visitors, only a tenth were professionalsthough by no means negligible, is farfrom universal and can hardly be called acultural explosion" (1966: 36). Compar-ing the performing arts audience to theurban population as a whole, they notedthat its members were somewhat younger,far more educated, of higher occupationalstatus, and more affluent. Over 55 per-cent of the men surveyed had done graduatework (as compared to 5 percent of theadult urban population as a whole), whileonly 2 to 3 percent of employed maleswere blue-collar workers (as opposed to60 percent of the urban population).Frequent attenders were of an even higherstatus than infrequent visitors. Baumoland Bowen conclude that even "if therehas been a significant rise in the sizeof audiences in recent years, it hascertainly not yet encompassed the generalpublic... Attempts to reach a wider andmore representative audience, to interestthe less educated or the less affluent,have so far had limited effects" (1966:96).

There is no work comparable to Baumol andBowen’s which deals with museums, althoughthere exists a fifty-year-old traditionof museum research in the United States.Most museum research before 1970 wasbehavioral, concerned not with whovisitors were but with how they respondedto and learned from exhibits. The fewstudies which were nonbehavioral gen-erally indicated that, except for thegreater proportion of children, museumvisitors were similar in most respectsto audiences for the performing arts.Economic and educational profiles looknearly identical. An early study of theBoston Museum of science, for instance,indicated a well-educated and prosperousclientele: a third of the adult visitorswere in professional or technical occu-pations and over half were collegeeducated (#246). More recently, a 1969yearlong survey of almost 5,000 visitorsto the smithsonian Institution found that48 percent of the adults were profes-sionals, 60 percent had family incomesexceeding $i0,000’, and 70 percent hadsome college education: only 14 percentwere in blue-collar or service occupations(#264). Nonetheless, the studies variedin their findings. While one study foundthat only 3 to 5 percent of the 1969visitors of three Manhattan museums wereblue-collar workers, museums in neigh-boring Brooklyn, Yonkers, and Newark werediscovered to attract visitor populations

and nearly half were laborers (#106).Overall, the early research suggested ahighly affluent visitor population, onewith greater diversity than that for theperforming arts but still closer toVeblen than to Tocqueville.

until recently, however, the paucity ofavailable studies made any generalizationssuspect. Only in the past few years hasthere been a large enough volume ofresearch so that an effort to develop ageneral portrait of the arts audience isnow feasible. Such an effort could answermany questions: Has the audience for theprofessional performing arts changed inthe decade since Baumol and Bowenexecuted their study? Who goes tomuseums? Who are the frequent attendersand how do they differ from individualswho go only once? Does arts attendanceresult in economic benefits for neigh-boring institutions?

Many of the studies we have used in exam-ining these questions are of low technicalquality. Often little care has beengiven to selecting a set of respondentstypical of~the audience about which theresearchers want to learn: questions arephrased in an imprecise manner: or impor-tant information affecting the audience’scomposition has been left out of the finalreport. But in an aggregate of more than250, the bulk of these studies helps usachieve a degree of certainty that wecould not expect from one or two alone.If a study of one museum’s audience, forexample, tells us that a disproportionatenumber of visitors are women (or men), wecan say nothing about the visitors ofother museums. If, however, twenty orthirty studies, with differing strengthsand faults, report the same findings, wecan begin to generalize with some confi-dence. In addressing these issues we are,of course, limited by the focus and natureof the studies assembled. And in thisreport’s concluding chapter we make somereco~endations about the sort of researchthat is needed to resolve a number ofimportant questions that currently avail-able studies cannot satisfactorily answer.

The issue of th~s chapter--audience com-position, attitudes, and behavior--is notsimply academic. Information on audiencesis of vital interest to individualsconcerned with managing the arts, to thosemaking general policy for the arts, andto the public which has an important stake

in these decisions and policies. The -e~timate-th~ �o~p6si~~he-audiencearts are increasingly dependent upon _ ; ~ ~or the ~rts in the United States. Wepublic and corporate benefactors-f~r.~_-~.-- . begin by looking at the basic demographictheir economic survival,. Such donors~may variables-~age, sex, education, income,want to know just whom their contributionsare serving. Particularly for publiclyfunded arts institutions, establishingthe nature and breadth of the clienteleto whom services are delivered may becritical to soliciting further support.

If, as many have suggested, exposure tothe arts is both personally rewarding anda social good~ it is important to knowhow widely the arts are being distributed.Before making efforts to expand the artsaudience or to develop art programs moreresponsive to public concerns and inter-ests, we should know what groups are notparticipating, why they do not do so, andwhat programs have successfully attractedthem.

occupation, and race--and we characterizethe arts audience in terms of each,~ withspecial attention to variations among artforms. Then we turn our attention to aset of more specific questions. Has theaudience composition changed over time?Is there one or are there many audiencesfor the arts? What has been the impactof the arts on local economies? And whatdo we know about attitudes toward thearts?

THE STUDIES

Understanding the audience for the artsis also crucial for a range of decisionsthat face managers and policy makers atevery level. Information on publicattitudes to the arts, the composition ofexisting audiences, and the spendinghabits of arts attenders can be used toestablish policies for public and privatesupport. For instance, information ondiffering habits and preferences for per-formance times and ticket prices can beused to set schedules and establishadmission prices. And information aboutwho attends and how they learn aboutexhibits and performances would be help-fu! in using scarce promotional resourcesmore efficiently.

While the tempo of audience research hasincreased, some arts managers continue tofeel that they know their public, thatthey have an intuitive grasp of theirclientele’s nature and needs, and thatresearch is superfluous. What data thereis on the question makes these claimsappear dubious. In the course of a studyof the public for the Royal OntarioMuseum, Abbey and Cameron (1961) askedthe museum staff to estimate the educationand income levels of their visitors. Thestaff’s estimates varied widely from thestudy’s findings: while the staff estim-ated that 20 percent of the adult visitorshad a college or university education, infact the percentage was 41: and while thestaff put the percentage of adult visitorswith incomes in the highest category at

~i0, the actual percentage was 39. It isour sense from conversations with indi-viduals in the arts that such discrep-ancies are not atypical.

In the remainder of this chapter we usefindings from the available studies to

Although audience surveys have been con-ducted for years, very little of theresearch has been published and many ofthe studies have been lost or buried inthe institutions that conducted them.The resulting lack of centralized infor-mation about the utility, design, orresults of audience research has proveda serious hindrance to every level ofarts organization from the local symphonyorchestra to the regional arts council.To help remedy this situation, we wishedto acquire as many reports of audiencestudies as were available. After aninitial review of published audiencesurveys, we identified three basic kindsof studies of audiences for museums andthe live performing arts. These threetypes of studies were: (i) attendersurveys, in which the audience of aspecific museum or performing arts organ-ization is surveyed, with questionsconcentrating on social of economiccharacteristics, motivations for attend-ance, and related issues: (2) cross-sectional surveys, in which a sample of alocal, regional, or national populationis surveyed, with questions focusing onfrequency of attendance at museums and/orperforming arts events, attitudes towardcultural organizations and issues, andthe social and economic characteristicsof attenders and nonattenders: (3) impactstudies, in which the impact of a museumexhibit, arts performance, or otherfeature of a cultural organization on anaudience is evaluated.

We tried in a variety of ways to obtainas complete a set of audience studies aspossible. First we conducted an extensivebibliographic search to create a list ofpublished studies conducted after 1950.

i0

Our review of thirty-five standard indexe~~. invoigedin .ahdie~ee~~e~arch- Finally,and bibliographic sources yielded approx- ¯ queries-were plac~di~ne~ghh ~rhs-relatedimately forty-five references. In " " ~-periodicals and newsletters (e.g.,addition, we also consulted-tweiv~’~~!- -~ American~ymphony Orchestra.Leagueinstitutional libraries, such as those ofthe Massachusetts Council for the Artsand the Center for Arts Information inNew York City.

However, most audience studies have neverbeen published, and in order to acquirethese we directly approached likelyorganizations. We compiled a list ofover 1200 arts organizations--museums,performing arts organizations, regional,state, and local arts councils, supportorganizations for specific art forms,and foundations involved in funding thearts. The museums and performing artsorganizations on our list were selectedfrom the Art Museum Directory and theNational Directory of Civic Centers andPerforminq Arts Orqanizations on thebasis of size, as we felt that the largerorganizations would be more likely tohave conducted an audience survey or toknow of other institutions which had.(Inquiries were mailed to all instru-mental music and theatrical organizationsreporting budgets of over $I00,000, allother performing arts organizations withbudgets of over $50,000 and all museumsreporting i00,000 or more visitorsannually.) To test this assumption, wedid, however, include i00 smaller museumsand performing arts organizations on ourlist.

Newsletter, Musical America New YorkTimes Book Review), requesting audiencesurveys. This effort yielded a numberof additional audience studies.

The response to this search for audiencestudies was greater than we had expectedwhen we set our initial goal to evaluateall published and unpublished audiencesurveys conducted since 1964. By the endof the third month of acquisition we hadobtained 160 studies and were stillreceiving new ones. Within nine monthsof the start of acquisition we had assem-bled materials on more than 250 audiencestudies.

In October, 1976, the director or managerof each organization was sent a letterdescribing our project and a brief formthat inquired whether the organizationhad ever conducted, con~nissioned, orparticipated in an audience survey. Ifthe organization had conducted a survey,we requested the name and address of thesurvey’s director and either a copy ofthe final report or information on how toobtain a copy. To those who wished it,complete confidentiality was Offered inregard to any materials that were sent tous. Respondents were also asked if theyknew of any other institutions that hadconducted audience surveys. The responserate to this inquiry u.ltimately rose toover 50 percent after a follow-up letterand second inquiry form were mailed toinstitutions that had not yet responded.%’hose organizations reported by ourrespondents to have done audience studieswere contacted by telephone or mail.

In addition to the bibliographic searchand mailed survey, our two major acqui-sitions efforts, an effort was made toacquire other unpublished audiencestudies by contacting individuals highly

Certain difficulties were encounteredduring the acquisition stage. Becauseremarkably few reports of audiencestudies have been published, the majorityof studies obtained through the librarysearch were museum studies, reflecting along tradition of visitor behavioralresearch that is unique to museums.(Such journals as Curator and Museum Newshave published reports of visitor studiessince the 1930s.) Other studies reportedin the published literature tended to belarge-scale, large-budget studies ofperforming arts audiences or populationcross sections.

Studies received in response to the mailedinquiry varied enormously in the amountof information reported. Some consistedof a questionnaire with hand-talliedresponses while others contained thoroughexplanations of methodology and extensivediscussions of results. Despite ourexpressed interest in studies conductedin earlier years, almost all the studiesreceived were conducted after 1970.Approximately 27 percent of the respond-ents stated that their organization hadplanned, conducted, or sponsored a studyand 20 percent reported familiarity withother audience research.

Efforts to follow up references obtainedthrough the mailed inquiry and biblio-graphic search met a substantial numberof obstacles. Often, people in aninstitution reported to have conducted anaudience study had no recollection ofhaving conducted it or, if they didremember, the survey report had longsince bee~, lost. This is due in largepart to the high turnover of employeesof arts institutions. Often whe~ theperson responsible for conducting orinitiating a study left the institution,so did the study. It was frequently

Ii

necessary to contact nearly every depart- ~ s-tatisti~s ~evi-ate ~_--[~q~-actual compo-ment within an institution before we .were .~ .sition of American audiences for the liveable to locate someone fanti!iarsurveys conducted as recentlymonths before. Also, despite an offer ofconfidentiality, fiveorganizationsrefused access to their surveys. In thisregard, it should be noted that we haveno way of estimating the number of surveysthat were never meant to come to publicattention. The number of explicitrefusals received obviously underrepre-sents the actual number of deliberatelyburied studies.

performing arts and for museums. Althoughmost of the studies eventually receivedwere from medium and small institutions,our inquiries were directed disproportion-ately at large and medium institutions.Thus, the larger institutions are over-represented in our data° at least incomparisc~ to the percentage theyrepresent of all arts institutions if notin comparison to the percentage of allannual visits and attendance for whichthey account.

Although the studies we collected wereof audiences for a wide range of insti-tutions, they do not represent a precisecross section. Surveys of attenders andnonattenders in forty-one states and theDistrict of Columbia are included, aswell as several national cross-sectionalstudies. By art form, the studiesinclude: 74 studies of theatre audiences:44 studies of art museum visitors: 33studies of population cross sections: 32studies of visitors to natural history,general, anthropology, and other relatedmuseums and exhibits: 19 studies ofscience museum or science exhibitvisitors: 16 studies of classical musicaudiences: 14 studies of those attendingseveral kinds of arts institutions: 12studies of visitors to history museums:ii studies of visitors to arts centers:7 studies of opera audiences: and 6studies of ballet and dance audiences.(Since calcualtions for specific variableswere based on subsets of these studiescontaining relevant data, and since manystudies provided data on more than oneaudience or set of audiences, distri-butions provided in specific tables inthe text of this report indicate theactual number of studies on which anygiven finding is based.)

In addition to a range of types, thesestudies include surveys of visitors andaudiences for institutions that cover thefull range in size. They do not includedata on audiences for jazz, fo--~/ethnicmusic, or the popular arts or data onaudiences for art as transmitted bybroadcasting or mechanical means. (Forinformation on audience research inradio and television, see Katzman andWirt0 1977.)

Since we attempted to acquire as manystudies as we could, and since nothir.~ isknown about the universe of all studiesconducted or about the representativenessof institutions that conduct audiencestudies, there is undoubtedly some biasin our data. We can only speculate asto the extent to which our summary

There is some reason to assume that thelarger institutions in the larger citiesdraw a more affluent and well-educatedpublic than smaller or conununity-basedinstitutions. On the other hand, sincethe quality of studies was so uneven,since response rates and total numbers ofrespondents varied so greatly, and sincenecessary data were not available, therewas neither a powerful rationale for northe possibility of weighting institutionsby total attendance in calculating overallfigures for audience composition. Theeffect of granting data from small insti-tutions equal weight with data from majorinstitutions would, we think, tend tocompensate for any inflation of highstatus categories caused by a dispro-portionate number of studies from majorinstitutions.

The audiences from which data have beendrawn may be unrepresentative in severalother ways. We do not know if audiencesthat are studied are systematicallydifferent from audiences that have notbeen studied. Out of the universe of allaudience studies that have been conducted,we could speculate that we gathered alarger percentage of published than ofunpublished studies, of recent than ofless recent studies, of studies for whichreports were written than of studiesyielding no formal reports, of major in-house or academic studies than ofproprietary studies, of studies oforganizations with relatively low staffturnover than of studies of organizationswith relatively greater staff turnover,and of demographic and opinion surveysthan of studies of exhibit evaluation orperforming arts impact. Given the numberand diversity of studies from which con-clusions are drawn, we do not think thatthese factors strongly bias findings oneway or the other. Nonetheless, thestatistics provided in this chapter mustbe seen as estimates rather than asscientifically rigorous descriptions ofthe public for museums and the liveperforming arts.

12

BASIC DEMOGRAPHICS

Gender

It is believed in many quarters that thepublic associates the arts with femininityand that this association inhibits menfrom attending the arts. The TheatreConununications Group, in a 1967 report onaudience development, suggests thattheatre-going "repudiates for many peoplethe all-American, red-blooded image ofwhat is supposed to be ’all-right’ for aman to do and still be considered ’all-man’" (Theatre Communications Group,1967: 31). Consequently, arts audiencesare dominated by women, according to thisbelief. Thus, an early study of asymphony audience concluded that the "sexdifference in @ymphony interest andattendance--more women than men--is borneout by statistic after statistic, studyafter study. The in-concert survey, thein-home interviews, and hundreds ofacademic studies irrefutably prove thepoint." The attendance difference can betraced to an underlying personalitydifference, according to this study, for"women have greater esthetic appreciationfor music, as they do for art and liter-ature, than men, who place greateremphasis on theoretical, economic,political, and practical-success values"(#64: 15). Arts policies are shaped bythis perception. Audience developmentstrategies to "de-feminize" the arts haveappeared, such as Bradley Morison’s (1968)effort to move news and publicity of theGuthrie Theatre from the woman’s page tothe sports section of Minneapolis news-papers, similarly, dance companies haveoccasionally emphasized the athleticprowess of dancers in promotional material.

unre~014~d:qu~st~bn.i-~n~doubt in partbecause-studies hag~sharpl~ V~ried inthe gender ratios reported. Resolutionof.the .issue requires a systematicassessment of gender ratios acrossstudies, and in this section we summarlzethe findings of seventy-two audiencestudies (which constitute all of thestudies in our possession reporting sexcomposition). In turning to thesestatistics, it is useful to keep twopoints in mind. First, there may be aresponse bias. Baumol and Bowen (1966),for instance, suggest that when surveyforms are distributed to couples attendingan arts performance, husbands will tendto assert the "male prerogative" andcomplete the questionnaires themselves,thereby inflating the male proportion inthe audience; but Book and Globerman(1975) have argued the opposite, suggest-ing that the male prerogative in thisinstance would actually be to delegatethe task to the wife, thereby inflatingthe female proportion in the audience-Such arguments aside, the true extent ofthe bias either way has not yet beenmeasured, although one study suggeststhat a slightly greater tendency for mento complete audience questionnairesincreases the observed male proportion by4 to 7 percent above the true percentage.In this study, groups entering a museumwere approached and asked to volunteerone person to respond to an interview.In one instance, 54 percent of the volun-teers were men, while only 50 percent ofthe groups were men: in another case, therespective percentages were 58 and 51(#121).

Other evidence, however, seemingly contra-dicts the belief that arts audiences areheavily female and that the publicconsiders attendance to be a feminineactivity. In a recent national survey ofattitudes towards the arts, respondentswere asked if "The arts are too effeminatefor most men to feel comfortable takingpart in them." While 18 percent of thepublic agreed with this view, an over-whelming majority--65 percent--rejected it(#7: 34). And Baumol and Bowen’s (1966)survey of the audiences of more than 150pwofessional arts organization perfor-mances revealed that men were in themajority, composing 52 percent of theaverage audience.

The true gender composition of the artsaudience remains a controversial and

The second point to keep in mind wheninterpreting the results of these studiesis the presence of sampling error. Thesamples considered have range from underi00 to over i0,000 respondents: the mediansize is approximately 500. Statistically,we are 95 percent confident that the truepercentage in a sample of 500 is somewherein between 4 points above and below theobserved percentage. With studies of thisscope, then, if 40 percent of the respond-ents are male we can be nearly certainthat males are indeed a minority of theaudience: but if 48 percent are male,such a conclusion cannot be drawn withgreat confidence.

Many of the seventy-two studies containinginformation on sex composition reportedresults for separate times and perfor-mances, and consequently data wereavailable on i12 distinct audiences (67 inthe performing arts and 45 for museums).The median percentage of men reported in ..the studies is displayed in Table 1.While the percentage of men in the U.S.

13

Table 1 Men in Audiences by Art Form

Number of Studies withinEach Percentage Range

Median percentage

Art Form percentage Range 28.1-42.0 42.1-57.0

All Museums 46.0 30-71 13 28

Art Museums 43.0 30-59 13 16

History Museums 48.5 44-53 - 4

Science Museums 52.0 43-71 - 8

All Performing Arts 42.5 31-58 30 35

Ballet and Dance 42.0 31-50 9 4

Theatre 43.5 32-58 15 17

Orchestra 44.5 33-54 5 7

Opera 46.1 41-58 1 7

57.1-72.0

8

7

2

TotalNumberofStudies

49

3O

4

15

67 ,’~

13 ~i

33

12

9

population is 49, the median percentage - (fewer than:hal-f ~f~[~r~Ig .age_._women are

of men observed in the museum studies .... - ~ ..employed). This rimefactor may accountwas 46 and in the performing arts was -_ for as much as i0 percent or more of the43. It is evident tha~ womenpar~±~i~a~ein arts audiences in proportions greater.than their share of the public as awhole, but the extent is very modest.Moreover, the gender ratio varied exten-sively from audience to audience: themale percentage ranged from 30 to 71percent in the case of museums and from31 to 58 percent for the performing arts.In fact, men outnumbered women-in aquarter of the performing arts studiesand two-fifths of the museum visitorsurveys. We have been unable to identifythe factors that account for the strikinggap between the average male percentagereported in the performing arts studiessurveyed here (43 percent) and the aver-age male percentage (52 percent) foundin the performing arts surveys conductedby Baumol and Bowen (1966).

The median figure for sex compositionvaried among the art fornm. Art museunlsdrew a higher proportion of women (57percent of their visitors on average),while history museums attracted equalrepresentation of both sexes and sciencemuseums were slightly favored by men(52 percent of the visitors). Withinthe performing arts, ballet and danceacquired the largest female audience (60percent on average): opera drew thelargest proportion of males, though menstill did not constitute a majority (46percent). Even within these art forms,the sex composition varied widely: operaaudiences, for instance, ranged fromthree-fifths men to three-fifths women,and art museum, visitors varied fromthree-fifths men to two-thirds women.

-variation-in-sex composition~ A study ofvisitors to New York’s Natural HistoryMuseum found that 52 percentof the-week-day visitors-were men, contrasting with59 percent on Saturdays (#203). Anotherinquiry revealed that while men and womenwere equally represented on Sundays amongmuseum visitors in the New York metro-politan region, the composition shiftedto 62 percent women on Thursday (#16).Similarly, studies of performing artsaudiences in the states of New York andWashington found that the proportion ofmen in the audience fluctuated by i0percent depending on the time of theperformance (#73: #63).

There is some evidence that the content ofthe performance or exhibit may attract menand women differently. For instance, theproportion of men in the weekend audienceof various productions of the JoffreyBallet ranged from 33 to 44 percent (#94).And a study of visitors to the ChicagoArt Institute discovered that i0 percentmore women attended during a week in whicha special Monet exhibit was on temporarydisplay than during three other weeks(#135). Factors associated with geo-graphic region may also influence the sexcomposition. Thus, 51 percent of the NewYork City performing arts audience arewomen, 53 percent of the New York stateaudience are women, and 62 percent of theWashington state attenders are women.However, the regional factors accountingfor this variation have not yet beenidentified (#73: #63).

Along with a slight overall tendencyfor women to outnumber men in arts audi-ences, but not among history and sciencemuseum visitors, it is clear that thesex ratio varied enormously. In otherwords, the median figures representstatistical tendency and are poor pre-dictors of the composition of an actualarts audience.

Although a fraction of the wide variationobserved in audience gender ratios isundoubtedly due to sampling fluctuationand to the use of nonprobability samplingtechniques (which can introduce systematicbias), a substantial part of the variationstems from other factors. Perhaps o~greatest significance is whether thevisiting or.performance time is duringa workday. Weekdays are obviouslyunattractive for most working people, andthe labor force participation rate of menis approximately twice that of women

The age composition of the audience forthe arts has interested arts adminis-trators for a number of reasons. Aprofile of the age of the audience, ofcourse, can help direct audience develop-ment efforts towards one age group oranother. Recently, for instance, therehas been a movement to make the arts moreaccessible to older Americans by offeringtransportation, ticket discounts, andspecial performance times (Johnson andPrieve, 1976). It is also believed thata young attender may grow up to be anold attender and,. while the link betweenattendance in one’s youth and in one’sprime has not yet been fully described,arts managers often view a young audiencewith an optimistic eye to the future. Theage composition of the audience alsoraises other interesting if more academic

15

questions. Is culture an acquired taste?~ a comparably smali~ercen~age 6f visitorsDoes the age composition of the audien-ce " ¯ under sixteen.differ from that of the generallation? On the latter question, mostobservers believe the.difference is small.Johnson and Prieve (1976), for instance,asked 605 arts administrators whether theythought that the age breakdown of theiraudience was roughly equivalent to that ofthe community: 60 percent said yes, while18 percent felt their audience was youngerand 16 percent felt it was older.

Two factors should be kept in mind whenexamining the age data. First, some ofthe studies in our possession restrictedtheir subject population to those indi-viduals over a certain age. Ten of fortymuseum studies surveyed only those oversixteen years of age and eight includedonly those who were over ten years of age.Likewise, nine of the performing artsstudies restricted their sample to thoseover sixteen and seven limited theirsample to audience members over ten. Toexamine whether this restriction made anysystematic difference in the results ofthe studies, we compared the median agesreported by the studies that did restricttheir samples with studies that did not.Surprisingly, there were no systematicdifferences.

It may be that many of the studies actual-ly limited their sample population but didnot state so in the report. Also, it ispossible that the study procedures werefrequently biased against the very youngbecause of the difficulties of obtainingreliable data from them. Another possi-bility is that the population undersixteen is indeed negligible, althoughavailable evidence suggests that this isthe case only for the perfroming arts andart museums. Studies of history andscience museum visitors that explicitlydid not restrict their sample oftenreported substantial numbers of youngchildren. The Nassau County HistoricalMuseum in New York, for instance, reportedthat 40 percent of their visitor popu-lation was under thirteen (#2) and theFranklin Institute in Philadelphia foundthat 39 percent of their visitors wereunder twelve and 4 percent were under fiveyears of age (#234). However, science andhistory museum studies generally reportfar greater numbers of children attendingthan do art museums. The MinneapolisInstitute of Art found that the proportionof visitors undez thirteen was under 3percent in 1970 and 1971 (#247), and theMuseum of Fine Arts in Boston reportedthat only one in fifty of their visitorswas under sixteen (#17). Studies of per-forming arts audiences show, on the whole,

Another factor that may affect studyresults is the presence of response.bias.P.erhaps youths defer to adults whenresponding to surveys, thus making theaudience appear older than it really is.The New York State Museum, for example,found few respondents under 14 years oldin one survey but noted that the actualproportion in attendance was approx-imately representative of the youngpopulation at large (#121). In thisstudy, groups entering a museum wereapproached and asked to volunteer oneperson to respond to an interview. Theinterviewer also collected data on thegroup composition. The age compositionof the group was inferred from the group’seducation levels.

Eighty-two of the studies in our posses-sion contained data on the age compositionof 145 distinct audiences. Most of thesestudies presented the data as the percentof the audience falling within various agecategories. Unfortunately, the agecategories varied widely. We havecomputed the median age for each audience(see Table 2) and to allow for comparisonbetween the age composition of each artform, we also have found the median of themedian ages. We refer to this number asthe median age of the art form.

The median age of 105 audiences of theperforming arts was thirty-five, while themedian age of 40 museum visitor popula-tions was thirty-one. This difference isconsistent with the results of two studiesof the arts audience conducted by theNational Research Center of the Arts(NRCA). They found that the median age ofthirty-seven for the performing arts inNew York State was five years older thanthe median age of the museum visitorpopulation (#73).

The summary statistics indicate that themedian age for the performing artsaudience was in the middle to latethirties while the median age for themuseum visitor population was in theearly thirties. These figures liebetween the median age of the entire U.S.population (twenty-eight) and the medianage for the population sixteen and over(forty). On the average, audiencesexhibited age profiles in a range similarto that of the entire population. However,specific audiences frequently divergedfrom this central tendency, and there wasa great difference in the average ageswithin and among the performing art andmuseum types.

16

Table 2 Age of Audiences by Art Form

Median Rangeof of Number of Studies within Each Age RangeMedian MedianAges Ages 19-22 22-36 27-30 31-34 35-38 39-42 43-46 47-51

All Museums 31 19-51

Art Museums 31 26-51 - 1 6 6 2 2 1 1 19

History Museums 33 28-42 - 2 1 1 1 - - ~5

science Museums 29 19-40 2 1 8 4 - 1 - - ]~6

21-49 5 7 14 23 22 21 8 5 105

TotalNumberofStudies

Ballet and Dance 33 30-38

Theatre 34 21-48 5 6 12 9 13 I0 3 2 60 .~

Orchestra 40 24-49 - 1 1 2 3 8 3 2 20 1

Opera 41 33-44 _ - - 1 3 3 2 1 i0,

Ballet and theater attracted the youngest "". attracte~a ~o~ngefi!a~di~dbe than otheraudiences of the performing arts while ~ plays (#3~)..opera and symphony drew the oldestaudiences. The NRCA studies-identical age rank ordering of the fourperforming art forms in their studies ofaudiences in New York State and WashingtonState. Baumol and Bowen (1966) identifiedalmost the same pattern except that theaverage age of the opera attender washigher than that of the symphony attender.

The median age for the science museumvisitor was two years lower than that ofthe art museum visitor, a difference thatwas not as great as among the variousperforming arts forms. NRCA also foundthat the museum visitor population wasolder in art museums than in sciencemuseums..

The age composition of the audience mayvary systematically with the seasons ofthe year, with the summer attractingyounger visitors. NRCA found that themedian age of performing arts audiencesin New York State was thirty-three in thesummer and thirty-eight in the fall thoughthe same did not appear to be true of themuseum visitor population. However, theBoston Museum of Fine Arts (#17) did finda slight seasonal variation: the medianage of the winter visitors was twenty-eight, while the average for the summervisitors was twenty-six. The results ofthe other rm/seum studies yield no consis-tent pattern. The Natural History Museumin New York (#203) found no variation, butthe Chicago Art Museum (#135) found thatvisitors were younger in June and Novemberthan in February and March.

The time of the performance seems toaffect audience age. NRCA found that themedian age for weekend evening perfor-mances was consistently lower than it wasfor matinees. In Washington State themedian age for weekend evenings was forty,for weekday evenings forty-two, and formatinees forty-nine. In New York State onweekend evenings it was thirty-five, onweekday evenings thirty-seven, and onmatinees forty-six. Audiences for theJoffrey Ballet (#94) showed this samepattern: the median age for the weekendevening audience was thirty-one while formatinee audiences it was thirty-three.

There is some evidence that differentprograms have greater attraction for cer-tain age groups than others. Thepreviously mentioned study of the JoffreyBallet reports that the median age of theaudience for a performance promoted as arock evening was younger than for otherperformances, and Moore’s study of Broad-way theatre audiences found that musicals

-NRCA reports differences in the age com-position of different regions. The medianage of the performing arts audience in oneregion (Southern Tier Central, FingerLakes) of New York State was thirty-three,while the median age was forty-four inanother region (New York City suburbs,mid-Hudson). They also report a highermedian age for both the performing artsaudience and the museum visitor populationin Washington State than in New York State.However, the reasons for this regionalvariation are unclear.

Education

Of all the characteristics of individualsthat studies frequently measure, aperson’s educational background appears tobe the best predicator of his or herattendance at museums and live performingarts events. The Ford Foundation notedfor example, that while frequency ofattendance at a variety of performingarts was related to both income andeducation, the latter factor was by farthe more important of the two. Individ-uals with much education but little moneywere more likely to attend the theatre,symphony, opera, and ballet than peoplewith high incomes but little education(115: 14-16). Similarly, analysis of anational cross-sectional study of resi-dents of cities and suburbs foundeducation to be a better determinant ofattendance at concerts, plays, museums,and fairs than income or occupationalstanding (Gruenberg0 1975).

There are reasons why individuals witheducation, particularly higher education,might be expected to attend more artsevents than their less educated peers.Schooling exposes students to formaltraining in the arts and, perhaps moreimportant, to a social milieu in whichthe arts are performed, exhibited, anddiscussed (The Arts, Education andAmericans Panel, 1977).

Then, too, arts attendance is a habitthat one develops over a period of time.A person may enjoy opera, but if perfor-mances are not locally available, orthere is no one to go with, he or she isunlikely to a~tend. By the same token,one may find modern painting incompre-hensible, but if one’s friends frequentgalleries and museums, sooner or lateroneis likely to give it a try.Education, particularly higher education,

18

provides both an environment in which th~~. -public ;at:la~ge-?~[~~ample° 30 percenqarts are relatively accessible and a group.- of the_typical aud~encehad sDme graduateof peers who attends with regularity~ .....

training: 54 percent had at-least~ .~, ~ .~%.~.~!. _-- ~ acquired .a.bachelor’s degree, compared toFinally, a disproportionate number of men 14 percent for the adult population in

and women who acquire a higher educationhave parents who are also well educated.Children of the well educated are morelikely than others to have been exposed tothe arts when they were young and mayalready frequent the arts by the time theyreach college (DiMaggio and Useem, 1978).

To learn about the educational attainmentof the American arts audience, we analyzedthe results of seventy-one studies report-ing findings of 108 audiences for theperforming arts and museums. In doingthis, we faced several methodologicalproblems. First of all, different studiesused different categories. Since medianeducation levels could not be calculatedfor every study, it was necessary todescribe audience educational compositionby reporting the percentages of anaudience that fell in five categories ofeducational attainment.

A second problem involved differences insampling designs used in the variousaudience studies. Of the 107 audiencesfor which findings were r.eported in atleast one of our five education categories,57 indicated a minimum age criterion hadbeen used to exclude audience members fromeither the sample or the analysis. Mini-mum ages, when reported, differedconsiderably. Three studies excludedaudience members younger than ten, threeused cutoff ages from thirteen to fifteen,and twenty cutoff ages from sixteen toeighteen. Studies of sixteen audiencesasked for the education of the householdhead only, one survey excluded "non-adults," another excluded "students," andone included only nonstudents eighteenyears and older. Twelve studies reportedthe educational attainment of only thoserespondents aged twenty-five or over.Also, it is likely that children wereunderrepresented in samples that did notexplicitly exclude them: completingquestionnaires is difficult and adultsmay tend to answer on behalf of children.

The extent of this underrepresentationcannot be determined. Differences in therespondent age criterion no doubt affectthe findings to some extent. Nonetheless,major differences in the findings ofstudies with different exclusionaryprinciples did not appear, s6 all studiesare pooled in the analysis here.

As expected, the educational attainment ofthe arts audience surveyed was substan-tially higher than that for the adult

general (see Table 3). Only 22 percent onaverage had not attended any college,compared tO 74 percent of the public as awhole, and only 5 percent were not highschool graduates, in contrast to 38 per-cent of the general adult public.

There was considerable variation amongstudies. Two of the most extreme figuresin the individual studies--6 percent ofthe audience had graduate training and 57percent had not completed high school--were reported in a study of the MilwaukeePublic Museum in 1962-63 (#35). Sincealmost half the respondents were agedseventeen or younger and more than three-quarters were less than twenty-four yearsof age, this accounts for much of theextremely low educational level. A studyof the same institution two years later,excluding children under thirteen, foundonly 25 percent of the visitors to be non-high school graduates (#108).

The educational attainment of live per-forming arts audiences was somewhat higherthan that of museum visitors. Some, butnot all, of the discrepancy is attribut-able to the greater representation ofyoung people still in school among museumvisitors.

As anticipated, those studies whichexcluded children under ages ranging fromi0 to 15 had a higher median percentage ofnon-high school graduates (24 percent)than those excluding visitors under theages of from 16 to 20 (where it was 7percent). The median percentage of non-high school graduates in studies with noexplicit exclusionary rule was 16 percent,probably reflecting unreported de factoexclusion of younger visitors. At theother end of the educational scale, studiesthat excluded only the very young reporteda median of 24 percent of visitors withcollege degrees, while studies that drewthe line higher recorded a median of 43percent. Studies that did not explicitlyexclude anyone reported an audience medianfor college graduates of 45 percent, againsuggesting that the young were under-sampled. However, even the set of museumsthat excluded their younger visitors fromthe survey reported audiences slightlyless well educa~ ed than the typical per-forming arts audience.

Among the performing arts, ballet anddance audiences included slightly aboveaverage proportions of well-educatedattenders: theatre audiences included

19

Table 3 Percentage of Audiences in Five EducationalCategories by Art Form

Art Form

Educational Level

At Least At Least High School Less Than High

Post-BA Training College Graduate Some College Graduate or Less School Graduate

Median Range (N)1 Median Range (N)

All Museums 17.5

Art Museums 22.0

Other Museums2 13.5

All Performing Arts3 32.0

Theatre3 32.7

Classical Music 37.5

Ballet and Danc~4 45.5

Median Range (N) Median Range (N) Median Range (N)

6-35 (13) 41.1 10-66 (23) 72.3 30-93 (18) 27.6 8-69 (18) 9.0 4-57 (23)

18-35 (5) 48.0 41-66 (9) 83.5 75-90 (6) 17.0 10-25 (6) 5.5 4-16~(8)

6-20 (8) 34.4 10-53 (14) 59.6 30-93 (12) 40.4 8-69 (12) 13.1 7-57,(14)

9-66 (42) 61.8 23-87 (53) 83.0 62~95 (44) 17.0 5-38 (45) 4.0 1-19.,~45)

20-50 (24) 58.0 23-80 (27) 82.7 56-93 (25) 17.1 8-44 (26) 4.0 1-15 !~.~21)

21-66 (8) 63.0 46-87 (9) 83.4 63-95 (8) 14.6 5-37 (8) 1.7 1-19 (8)

20-50 (5) 65.0 55-73 (i0) 87.1 77-92 (5) 12.9 8-23 (5) 3.0 1-5

Opera 37.3

All Museums andPerforming Arts5 30.0

U.S. Populationover 24 Years of NA6

Age, 1975

29-49 (5) 61.8 49-75 (7) 83.0 67-94 (6) 18.8 7-33 (6) 4.1 2-7

6-66 (73) 54.0 10-87 (97) 78.0 30-95 (83) 22.0 5-69 (84) 5.0

13.9 26.3 73.7 37.5

I(N) = Number of studies.

2Includes science, history, natural history,anthropology, and general museums.

3Excludes audiences of outdoor dramas.

:(6)

1-57 (72)

5Number of studies exceeds sum of other categories ..............due to inclusion of regional studies reportingattendance of all, undifferentiated art forms.

6NA = Not available

4Dance audience percentages apart from balletavailable only for two educational levels--at least college graduate and less than highschool graduate.

slightly below average proportions. An~~. stddi4s of hati~i~-~-locaI populationsamong museums, art museums attracted a . :. have consistently:~Ound h~gh~ ratesofmore highly educated public than did .... ’ attendance among professionals andhistory, science, and ~other~museums~~~. o managers, than any other group (#’s 73,though still not so well educated as theaudiences for the performing arts.

Overall, the proportion of college grad-uates reported for the arts audiencesand museum visitors exceeded the pro-portion of the adult population withcollege diplomas in all but one ofninety-seven audience studies: and thepercentage of individuals who had notcompleted high school was below thenational level in seventy-one of seventy-two audiences. Both exceptions are dueto presence of students still in highschool. And in seventy-eight of eighty-three audiences for which findings areavailable, the proportion of attenderswith at least some college training wastwice that for the general public.

The studies that we reviewed show audi-ences to be somewhat less educationallyexclusive than did the Baumol and Bowenstudy of the performing arts (1966).There were discrepancies. The relativelyhigh proportion of individuals with nocollege education reported in some of ourstudies of opera audiences is surprising,for example. But most differences can beattributed to Baumol and Bowen’s exclusionof respondents under the age of twenty-five and to the restriction of theiraudiences to the professional performingarts.

In sum, it is evident that visitors tomuseums and audiences for the live per-forming arts in these studies hadconsiderably more education than thepublic at large. Museums appeared toserve a somewhat broader public than didthe performing arts. Nonetheless, interms of educational attainment, themuseum visitors and performing artsaudiences surveyed were far more similarto one another than either group was tothe general public.

Occupation

115, 137, 142).

This tendency is not surprisingl For onething, those occupational groups that showthe highest rates of attendance are alsothose with the highest educational attain-ment. Blue-collar workers, who attendleast, also have the least education.Moreover, one’s job determines to a greatextent the social milieu in which onespends leisure time. The participationof a lawyer, teacher, or physician in thearts may be rewarded with respect byassociates and peers: among these groups,attendance at the theatre or symphony isan accepted way of spending a socialevening. By contrast, a carpenter or busdriver with a penchant for the artsperhaps receives less encouragement fromfriends and co-workers.

To better understand the occupationalcomposition of American arts attenders, weanalyzed the results of fifty-nine studiesof ninety-six audiences that askedrespondents to report their occupations.Our findings were consistent with theexpectation that art audiences aredominated by individuals in higher statusoccupations. Professionals, who consti-tuted 15 percent of the employed civilianlabor force in 1975, composed a median 56percent of employed persons in the artsaudiences surveyed (see Table 4). Con-versely, blue-collar workers typicallyconstituted a mere 4 percent of employedrespondents in the arts audience surveyedas compared to 34 percent of the employedcivilian labor force.

Although the summary statistics arestriking, .the reader should be cautionedthat the median figures are to be regardedas approximations. The classificationschemes used in the audience studies wereso varied that comparability was estab-lished only with great difficulty. Theoccupational categories we used here aredesigned to be compatible with as manystudy findings as possible and to becomparable to the classifications used bythe United States Census.

Next to education, occupation is perhapsthe demographic characteristic mostclosely related to involvement in thearts. Gruenberg found occupational statusa more significant predictor of attendanceat cultural events and institutions(concerts, plays, museums, fairs, andadult education classes) than income,second only to educational attainment(Gruenberg, 1975). And cross-sectional

Categories used to report occupation insome study reports were vague enough toencompass those employed in several moreconventional categories. For examp~.e,many studies used an occupational categorycalled "business," which may in some caseshave included business secretaries andclerks as well as executives while

excluding managers of public and nonprofitconcerns. Because most studies reported

21

Table 4 Occupational Distribution of Audiences

Median Percentage ofOccupations of 1 Percentage of U.S~ Employed RespondentsEmployed Persons Employment (1975) in Arts Audience

Professionals 15.0 55.9 65

Teachers 4.1 22.1 22

Artists, Writers, 1.0 8.2 8Entertainers

Managerial 10.5 14.9

Clerical/Sales 24.3 14.6

Service 13.7 3.7

Blue-Collar 33.0 3.7

Farmworkers 3.5 -

Major Activities of Percentage of U.S. Median Percentage of

Persons Unemployed or Population Aged All Respondents in

Not In the Labor Force 16 Years or Over Arts Audience

Homemakers 23.1 14.0

Students 5.5 18.0

Retired, Unemployed

Number of Audience StudiesReporting Informationfor this Category

51

41

13 ~i~; ~

71 iI ~,I ’~

_4

Number of Audience StudiesReporting Information forthis Category

78

80

11.2 4.5 65

IU.S. Census categories and audience categoriesare only approximately comparable due to varyingclassification schemes used in arts audiencestudies.

2Source: U.S. Bureau of the Census, StatisticalAbstracts, 1976 (Washington, D.C.: U.S.Government Printing office, 1976): U.S. Bureau ofLabor Statistics, Handbook of Labor Statistics,1976). Figures for U.S. population aged 16 orover exclude military personnel.

3Because these are medians and because not all occupa-tional categories were given in each study, antion of the median percentages across occupations willnot equal I00 percent. Thus, these figures offer onlyan approximate distribution of the arts audience byoccupation.

4None of the arts audience studies contained an occupa-~tional category for farmworkers. Thus, they wereeither distributed among the other categories or veryfew, if any, were found in the arts audiences.

occupation as a percentage of total -It s~uld be not~dLthat~~the U.S. Censusrespondents rather than as a percentage of.category ~f.profes-si.~n~which -we followemployed respondents, results from m~ny -’.---includes not only such individuals asstudies had to be recompu~ed.-cases, study categories were merged to fitinto the scheme used-here. When studyresults could not be reliably altered tofit our classification system, the find-ings were dropped. Our categories, then,represent a compromise. While thefindings can be used with confidence toassess general similiarities anddifferences among art forms, greatprecision should not be attributed to thefigures reported here.

An additional caveat: Even in those caseswhere audience studies used classifi-cations similar to census categories, onlythe most experienced analysts canunerringly place specific occupations intotheir appropriate general categories. Forexample, airplane pilots are consideredprofessionals, ship pilots are managers,and airplane stewardesses are serviceemployees; registered nurses are profes-sionals, while practical nurses areservice employees: an inspector is blue-collar unless he is a constructioninspector, in which case he is managerial.Few people on either end of an audiencesurvey--visitors responding to forced-choice occupation questions or codersclassifying open-ended ones--can beexpected to have mastered the precisetechnical census system, and a degree oferror is to be expected.

Finally, a word about "status." Thisconcept is based on the average educationand income within an occupation. In turn,the mean educational attainment and incomeclosely relate to the appraisals in cross-sectional surveys by respondents who areasked to rate the status or prestige ofsets of occupations. The several schemesused to assign status to occupation yieldhighly similar rankings (Haug, 1977).

Professionals. As noted, one of the moststriking consistencies in the occupationaldistribution of the arts audiencessurveyed was the very high representationof professionals. They were present innumbers proportionately greater than theirshare of the population in every one ofthe sixty-five arts audiences for whichappropriate data were reported. In allbut four of these audiences, the percent-age of professionals was at least twicet~,at of the work force as a whole, inforty-six audiences it was three times,and in more than a quarter of the audi-ences it was four times greater than inthe work force.

doctors, _!awyers, and architects but alsomembers of lower status professions suchas teachers, engineers, librarians, dieti-cians, social workers, and computerprogrammers. The number of respondentsfalling in this category may be under-stated since individuals in lower statustechnical professions seem to have beenincluded in the residual -white-collar"categories used by some studies. Forexample, in a 1976 study of the GuthrieTheatre audience (#122), in which onlyteachers, doctors, and lawyers were codedas professionals and a residual white-collar category was used, the professionalpercentage of the employed audience wasonly 40.4 percent, compared to 56.5percent in studies of the Guthrie audienceundertaken in 1963 and 1973 (#117, #126).The latter had precoded professional,technical, and clerical/sales categories.(For the few studies that includedseparate "technical" categories, "tech-nica!" respondents were included with"professionals" for this analysis.)

There was a significantly higher propor-tion of professionals in the audience forthe performing arts than for museums, 59percent compared to 42 percent (seeTable 5). The lower overall median formuseums was the result of the proportionsat museums other than art (which comprisedeleven of the seventeen museum studiesfound). These reported a median 42 per-cent professional representation. The sixart museum visitor studies had a median of59 percent of professionals, the samefigure as for the performing arts.

Except for this profile within a museumcategory, findings were remarkably uniformfor the various art forms. Among theperforming arts, median professionalpercentages ranged from 56 percent for thetheatre to 61 percent for classical musicaudiences. These figures are similar tobut slightly lower than Baumol and Bowen’sfindings (1966) on occupation. The dis-crepancy is probably attributable to thepresence of a greater proportion ofrelatively major institutions among whoseaudiences they sampled.

One group of professionals--teachers--appeared to play a special role in thearts audience. Overall, teachers (includ-ing college and university faculty)constituted 22 percent of the artsaudiences for which findings wereavailable. This figure was more than fivetimes their percentage of the employed ..

23

Table 5 Occupational Distribution of Audiences by Art Form

Art Form

All Museums

Art Museums

Other Museums3

All Performing Arts

Ballet and Dance

Theatre

Orchestra

Opera

OccupationsI of Employed Persons (Percents)

Professional &Managerial

Median Range (N)2

63.3 17-96 (32)

77.1 56-96 (16)

53.2 27-72 (16)

70.9 49-95 (42)

74.6 61-88 (9)

69.5 49-95 (23)

ProfessionalOnly

Median Range (N)

42.2 12-73 (17)

59.2 31-74 (6)

41.9 12-50 (ii)

59.1 24-80 (44)

59.6 55-73 (8)

56.3 24-70 (25)

ManagerialTeachers Only

Median Range (N) Median Range (N)

23.1 15-33 (6) 9.6 4-27 (14)

23.1 15-33 (5) 9.0 4-27 (,’6)

- - - 10.2 6-22 (’ 8)J

17.9 6-33 (16) 15.6 4-27 (!33)- - - 15.2 7-22 ci[~)

17.9 6-33 (7) 16.0 4-27 (20)

75.5 64-87 (5) 61.1 50-80 (6) ...... r ’

- - - 58.3 50-70 (5) ..... ~-

(Table 5 continued)

Occupationslof Employed Persons

ART FORM

Clerical & Sales Blue-Collar

Median Range (N)z Median Range (N)

Major Activities of Persons Unemployed orNot In the Labor Force

Retired &

Homemakers Students Unemployed

Median Range (N) Median Range (N) Median Range (N)

All Museum~

Art Museums 14.3 4-22 (14) 3.1 0-12 (16) 13.0 7-22 (9) 22.5 0-40 (I0) 8.0

Other Museums3 16.0 5-28 (9) 16.7 4-45 (19) 15.8 6-26 (15) 20.0 10-57 (15) 3.3 1-9 ~ (12)

All Performing Arts 18.0 8-33 (15) 2.8 0-27 (34) 14.0 5-52 (51) 17.1 5-63 (51) 3.9 0-16 (40) ,,,

Ballet and Dance _ - - 2.7 1-7 (i0) II.I 6-32 (i0) 15.0 9-34 (i0) 3.0 1-5 ’ (9)

Theatre 19.7 8-29 (I0) 2.9 0-27 (15) 14.0 5-52 (27) 18.9 5-63 (27) 4.2 0-26’ (24)¯

_ - - 19.0 5-26 (7) 18.0 7-31 (7) .... -i, ~Orchestra - - - ’, ",

Opera _ - - 2.8 1-13 (5) 16.2 8-40 (6) 10.7 7-23 (6) ..... ~i’~i,~

,~’,i’~ ¯

iThe "Professional & Managerial" and "ProfessionalOnly" categories include teachers. The percent-ages for "Homemakers", "Students", and "Retired& Unemployed" are based on all respondents: thepercentages for the other categories are based onemployed respondents only. Percentages are notreported when fewer than five studies areavailable.

2(N) = Number of studies

3includes science, history, natural history,pology0 and general museums.

4Excludes audiences of outdoor dramas.

anthro- ,

civilian work force (4.1 percent).Assuming that the twenty-two audiences-for which this category was reported-~are-.-not systematically different f~omaudiences, the median percentage ofteachers among professionals in theaudiences (37.7 percent) exceeds thepercentage of teachers among professionals in the employed work force as awhole (28.2 percent in 1970) by more thana third. Thus teachers seem to be heavyattenders among heavy attenders.

A second professional group participatingin arts audiences at rates well abovetheir share of the population was, notsurprisingly, individuals in the arts.Although artists, writers, and enter-tainers comprised only 1 percent of theemployed work force in 1970, in eightaudiences for which findings were reportedthey accounted for a median 8.2 percent.A fraction of the high ratio may stem fromdubious sampling procedures, a possibletendency for researchers coding hand-written occupation responses to reportartists as a separate category if theywere particularly numerous, and thetemptation for some students and amateursto report an avocation as an occupation.

Manaqerial. The managerial category inthe United States Census covers a rangeof managers and administrators, includingexecutives, government officials, salesmanagers, school and hospital adminis-trators, union officials, and smallbusinessmen. However, the audiencestudies we looked at included executives,managers, business, and proprietors.As noted earlier, the "business" categorymay have some clerical/sales employeesand leave out some public administrators.Similarly the "executive" category mayleave out some proprietors and low levelmanagers. Nonetheless, it is assumedthat the Census category which we use andthe audience studies category are roughlyequivalent.

Managers were found to participate in artsaudiences in greater proportions thantheir share of the population but to alesser extent than professionals. Theycomposed a median 15 percent of employedrespondents in fifty-one arts audiencesbut only ii percent of the employed workforce in 1975. Their median proportion ofperforming arts audiences (16 percent) washigher than their share of museum visitors(9 percent and i0 percent for art andother museums, respectively).

~rofessi0nal/manaqerial. A number ofstudies merged professionals and managersinto a single category. In order to usethis information, we joined the profes-sional and managerial categories in otherstudies and pooled them with studiesreporting only professional/managerialpercentages. Professional/managerialpercentages may be taken as a rough indexof the representation of individuals inhigh status occupations in the audiencessurveyed.

Among employed respondents, the medianpercentage of professional/managerialworkers in seventy-seven arts audiencesfor which data were available was 69.5percent, more than double this group’sfraction of the employed work force as awhole (25.5 percent). As with profes-sionals alone, there was some disparitybetween art museum and other museumvisitors. The managerial/professionalpercentage for art museums was 77.1 per-cent, even higher than for any of theperforming arts, while the percentage forother museums was 53.2 percent, lower thanfor any art form reporting.

Clerical/sales. The clerical/sales cate-gory includes, among others, officeworkers, secretaries, sales clerks,advertising, real estate, stock and bondsales workers, and telephone,operators.However, some schemes specified thatsecretaries were included but others didnot. Residual white-collar categorieswere excluded from this analysis except ina very few cases where white-collar unam-biguously included only clerical/salesemployees. Since a number of occupationslisted as clerical/sales by the Census--for instance, bill collectors, mailmen,and teacher’s aides--are difficult toclassify, there may have been slightleakage from this category into business,blue-collar, or service categories.

Clerical and sales personnel composed asomewhat smaller percentage of audiencesthan their share of the employed civilianwork force. Of forty-one audiences, theymade up a median of 15 percent, while theyconstitute 24 percent of the full employedcivilian work force. The median for theperforming arts (18 percent) was slightlyhigher than for museums (14 percent).

However, since the number of museumaudiences reporting this category issmall and the percentage range within themuseum and the performing arts studies

Blue-collar workers. Along with theextremely high proportions of professionalsreported, the most striking finding in thestudies reviewed was the consistently lowpercentages of blue-collar workers relativeto their share of the population. In the

26

seventy-one audiences for which data were-, perf~rmi-dg:ar~s.w~S!~.~e~y~i~gimilar to thatavailable, blue-collar workers comprised ....

_for all_museums..~.Them~diah p~centage

a median 4 percent of those emp~oy.@d.~ " .... for art museums was somewhat lower thanThat the median is even this- high~i~~%~/~ " .-~ for o~her- museums but the ranges werepartly due to the inclusion of 19 "other sintilar. The ballet/dance audiences hadmuseum" audiences, which reported a muchhigher median blue-collar participation(17 percent). The median blue-collarshare of performing arts audiences wasonly 2.8 percent and blue-collar repre-sentation among art museum visitors wasa median 3.1 percent. Excluding visitorsto museums other than art museums, theproportion of blue-collar workers inthirty-four of fifty-two arts audiencesfor which percentages were reported wasless than one-tenth of their represen-tation in the work force as a whole. Inonly nine audiences was it as high astwo-fifths. Among art forms, medianblue-collar percentages were remarkablyconsistent: 2.7 percent for ballet anddance: 2.8 percent for opera: 2.9 percentfor theatre: and 3.1 percent for artmuseums.

Remarkably, blue-collar attendance is, ifanything, probably overstated. Blue-collar workers include individuals in theskilled trades (carpenters, shoemakers,television repairmen), factory workers,laborers, and some transportation workers(including bus, taxi, and truck driversand parking attendants). Holders of anumber of other low-status jobs (chamber-maids, janitors, busboys, dishwashers,bootblacks, elevator operators) areclassified in a separate "service"category. However, information on thepercentage of service employees wasavailable for only eleven of theseventy-one audiences that reported ablue-collar percentage. (Since theservice category also contains a numberof relatively high-status workers likestewardesses, sheriffs, and detectives,blue-collar and service categories couldnot be merged.) It seems likely that, instudies where percentages of serviceworkers were not reported, individualsin the service category (1.7 to 20.0percent of audiences were reported, witha median of 3.7 percent) were dividedbetween "blue-collar" and residual white-collar categories, thus giving some upwardbias to the totals of each.

Homemakers. The median percentage ofhomemakers in seventy-eight audiences forwhich appropriate information was avail-able was 14. While homemakers were thusstatistically underrepresented--in 1975they comprised 23 percent of the civilianpopulation over sixteen--variation amongaudiences was great, ranging from 5 per-cent to 52 percent for the audiences as awhole. The median percentage for the

the lowest median percentage of homemakersand the classical music audiencesthehighest, but again ranges were similar.

Students. Students participated in thearts audlences to a high degree, com-posing 18 percent of the average ofeighty audiences for which data wereavailable but making up only 6 percent ofthe civilian population over sixteen.Most of the students were enrolled incolleges. The only surveys reportingappreciable numbers of respondents lessthan sixteen years old were from museumsother than art museums, and their medianis not much higher than that for theaudiences as a whole. It would be inter-esting to know to what degree the highpercentage of students is a measure ofthe success of c~itural organizations inattracting them by special discounts andother incentives.

As with homemakers, the proportion ofstudents varied widely from audience toaudience, with a range from 0 percent(found in one study of members only(#181)) to 63 percent (an audience of astudent theatrical production (#127)).

Retired and unemployed. The median per-centage of retired and unemployed personsin sixty-five audiences for the arts withappropriate data was 5 percent, ascompared to ii percent of the 1975civilian population over sixteen. Thisfigure would seem to reflect the relativeimmobility and often severe financialdeprivation of individuals in both groups,as well as their relatively low educa-tional attainment. Percentages did notdiffer greatly among art forms.

In most cases, audience studies presenteddata both on retired and on unemployedpersons: sometimes a single categoryincluding both. In our analysis, per-centages of retired persons alone wereoccasionally included since the repre-sentation of the unemployed, where listedseparately, was consistently minuscule.Downward bias in the totals may resultfrom a possible tendency for individualswho are unemployed, underemployed, retired,or semiretired to report their regularoccupations.

Summary of Occupations. Audiences formuseums and the live performing arts werefound to include substantially moreindividuals in high prestige occupationsthan the public at large. The two most

27

striking findings in the materialswere the extremely high pro- . ¯ ~±nternal ~dministratiVe considergtions,

analyzed of professionals,.above-all. .... ~ ~s~ch as estimating the price sensitivityportionsteachers, and the extraordinarily ~Q~~ of the presentaudience, the level of

percentages of blue-collar workers,contributions the audience is capable of

Variation among art forms was relativelyminor, with two exceptions. First,museums reported a less heavily profes-sional public than the live performingarts. Second, blue-collar workerscomposed a far higher percentage ofvisitors in museums other than art thanthey did in any other category. Severalother findings are also notable. Managerswere slightly overrepresented relative totheir share of the population in perform-ing arts audiences but not among museumvisitors. Clerical/sales personnel werestatistically somewhat underrepresentedin audiences for all the art forms, aswere homemakers. Students were greatlyoverrepresented relative to theirproportion of the public at large,although their participation variedconsiderably from audience to audience,and the retired and unemployed composedconsistently small percentages of audi-ences for all art forms.

Income

The notion that the audience for the artsis composed of an economic elite is afantiliar one. A study of the MinneapolisSymphony (#65) describes the popularstereotype of the symphony audience as oneof "extreme wealth, snobbery, ’our orches-tra,’ and long gowns and white ties andtails." While snobbery and long gownshave not yet been quantified, surveyshave repeatedly reported that museum andlive performing arts audiences haveconsiderably higher median incomes thanthe population at large. Baumol andBowen (#8) found that the median familyincome of the performing arts audiencewas roughly twice as high as the medianfor the total urban population, and theNRCA (#137) reported that people withhousehold incomes over $15,000 attendedthe arts more than twice as often asthose with incomes below this figure.

The relative affluence of the artsaudience has become an increasinglyimportant issue as arts organizations havesought government support. Some observershave warned that it is difficult tojustify public funding of the arts if theaudience is composed of a small, well-to-do segment of the populatiog. Where thisattitude prevails audience incomestatistics may not prove particularlyvaluable for soliciting public backing.

giving, and the participation of variousincome groups in the audience. At anyrate, it m~st be kept in mind thatalthough income may be associated witharts attendance, it is not necessarilythe cause of attendance. High income iscorrelated with having received a highereducation and holding professional ormanagerial occupations, and evidencesuggests that it is these latter factorsrather than income that determineattendance. When all three factors aretaken into account at the same time,education and occupation once controlledpredict attendance but income does not(#115: Gruenberg, 1975). Thus, theunderrepresentation of middle and lowincome groups is less the result of theirlower disposable income than of theirlower education and attainment and theirmembership in less prestigious occupa-tions.

Income distribution data were availableon eighty-eight audiences for museums andthe live performing arts. Two steps werenecessary to make the data comparable.First, virtually all studies reportedincome statistics by indicating theproportions.of the respondents fallingin various ~ncome ranges. For compar-ability, these range figures wereconverted to median incomes for eachaudience. Second, since the studiesanalyzed were conducted over a fifteen-year period, it was necessary to convertincome figures into constant incomelevels. Accordingly, the consumer priceindex was used to transform all mediansinto constant mid-1976 dollars.

Several additional problems should bekept in mind when interpreting theseincome figures. Personal income isgenerally regarded as sensitive infor-mation, and income data solicited throughquestionnaires or interviews is more proneto distortion than any other social char-acteristic considered here (thenonresponse rate for income questionsranged as high as 29 percent). Moreover,some studies requested family income,others sought household income, and stillothers failed to specify either (which insome instances was probably interpreted asa request for iqdividual income). Thismay introduce some downward bias. Whilestudies requesting household and familyincome yielded nearly identical medianincomes, surveys specifying neitherobtained median incomes which were on theaverage $2,591 below those eliciting

28

household income. No reliable procedure ~f twen~y-~eveh ~i~[~f-theatrewas available for adjusting these differ- ~..~audiencesreported m4dian incomes belowences. ~ " .... " ’~<~..i~ ~ ~ ~-

Finally, median real~family incomes forthe population as a whole increasedconsiderably in the 1960s and modestly inthe 1970s: median family income inconstant 1976 dollars was $10,778 in 1960,.$14,431 in 1970, and $14,476 in 1975. Anaudience with a median family income of$14,500 in 1976 dollars would be con-sidered relatively affluent were the studyconducted in 1960 but fairly representa-tive of the public were the surveycompleted in 1975. More than two thirdsof the studies reporting income data wereconducted during the 1970s, and thus afigure of approximately $14,000 for medianfamily income serves as a useful baselinefor comparison with study findings.

Median Family Income,U. S. Population

1960 $10,7781970 14,4311975 14,476

Consistent with the belief that the per-forming arts draw an upper-incomeaudience, the median income for seventyperforming arts audiences was $18,983,approximately $5,000 above that of theentire public (Table 6). Moreover,eighteen of the performing arts surveyswere conducted by the Institute ofOutdoor Drama. The outdoor dramas tendedto attract a more diverse audience, andthe median income for these studies was$15,249. By contrast, the median incomefor the performing arts studies withoutthe surveys of outdoor drama was $20,250.

The gap we found between the populationincome and performing arts audienceincome was somewhat less than thatobserved by Baumol and Bowen (1966),probably reflecting the greater diversityof audiences surveyed in the studiesreviewed here. Baumol and Bowen, forinstance, did not include as many uni-versity or outdoor performances in theirstudy, and the lowest median incomesconsistently are reported for these typesof audiences.

The performing arts studies reported arange of median audience incomes thatindicated considerable diversity in audi-ence composition from event to event.Nonetheless, nearly all of the assembledstudies found median incomes above thatof the general population. Excluding theeighteen outdoor drama surveys, only three

-that of the public at large, and all threeofthese ~ere 5f university theatre pro-ductions. No study of the otherperforming arts yielded median incomesbelow that of the general population. Theminimum median incomes reported for ballet,orchestral music, and opera were, respec-tively, about $2,000, $4,000, and $5,000higher than the population median. Ifoutdoor drama audiences are excluded, themajor performing art forms appear to drawmarkedly similar audiences: the theatremedian is $19,342 and the opera median is$21,024, with ballet and orchestral musicin between.

As has been previously observed in thecase of both education and occupation,museums attract a somewhat more repre-sentative cross section of the Americanpublic. The eighteen museum studiesreported incomes several thousand dollarsbelow the performing arts average thoughstill also several thousand dollars abovethe general population figure. (Only asingle museum study found a median incomebelow that of the general public.) Amongthe many factors that may account for thisdifference are the generally lower aclmis-sions charged by museums and the greaterappeal of museums for students and youngpeople. Though relatively few studiesare available on the separate museumtypes, as in the cases of occupation andeducation, art museums were found to drawa somewhat more affluent clientele thanscience or history museums.

Race and Ethnicity

The relative paucity of blacks and otherracial and ethnic minorities in artsaudiences has been commented on frequentlyand, indeed, has been a matter of someconcern to the arts community. In 1972,the American Association of Museums calledattention to the problem of making museumsrelevant and hospitable to intercity andminority people, noting that the movementsof the middle class to the suburbs and ofblacks, Mexican-Americans, and PuertoRicans to the core city "have left themuseum, an urban institution, to someextent a beached whale .... " (AmericanAssociation of Museums, 1972: 6).Museums have not been alone in recognizingthis dilemma. Recently, the KennedyCenter for the Performing Arts formed aspecial committee to find out why so fewof Washington’s many black residents wereattending the Center’s events.

29

Table 6 Median Income of Audiences by Art Form

O

Median of Range of Total NumberArt Form Median IncomesI Median Incomes of Studies

All museums$17,158 $13,394-$30,618 18

Art museums18,148 14,016- 50,618 i0

History museums 16,757 13,394- 29,055 3

Science museums17,269 14,765- 20,851 5

18,903 9,466- 28,027

20,082 16,452- 22,404 i0All performing arts

Ballet and dance

Theatre

Excluding

Including

Orchestra

Opera

fin constant

outdoor drama

outdoor drama

19,342 9,469- 25,784

16,819 9,466- 25,784

20,825 18,221- 28,027

21,024 19,017- 27,245

27

45

ii

5

mid-1976 dollars

Minorities--blacks and persons of Orient~ . metropQ~itan-pop~l.a~iO~2[_ ~n t~o studiesand Hispanic background--were underrep- . of attendance .at~.two~ of the. Smlthsonlanresented in most of the audiences for ..~ -..museums (#ii0 and #265), visi-tors werewhich data on race were acquired£~ii~T~.~_eY"

accounted for a median 7 percent ofthirty-five audiences as opposed to beingover 20 percent of the population as awhole. The median percentage of minor-ities for thirteen audiences of theperforming arts was 7, the same as it wasfor eleven audiences of art museumvisitors. Again, museums other than artwere more inclusive: for eleven sets ofsuch visitors the median percentage ofminorities was ii. As to specificminorities, while blacks constituted 12.3percent of the total urban population in1970, they represented a median 3.0percent of the fifteen arts audiences forwhich data were available. In a numberof studies outside the W~st Coast andSouthwest, individuals of Hispanic back-ground were not separated from the whitepopulation, thus depressing the minoritytotal. We surmise, however, from the fewstudies in these areas that did countHispanic people separately, that theygenerally made up a small percentage ofthe audience and that their exclusiondepresses the minority median by no morethan 1 percent.

Such overall figures should be interpretedcautiously because of the small number ofaudiences studied, variations in thedefinition of minority and, above all, thelarge differences in the populations ofminority groups in different locales. Theset of studies reviewed here, for example,contains data from the Washington, D.C.metropolitan area, where blacks composed24.6 percent of the population in 1970and from Washington State, where only onein fifty persons was black. Similarly,persons of Hispanic origin represented asubstantial portion of the populations ofLos Angeles and New York City (15.0 andii.i percent, respectively), but were amuch less significant presence in suchplaces as Boston or Montgomery, Alabama.For this reason, selected comparisonsare useful.

In all but one of fourteen audiences forwhich data on black attenders can becompared to census data, blacks wereunderrepresented relative to their numbersin the local population by ratios of upto eighteen-to-one. In five studies ofmuseums in the San Francisco area, whereblacks composed 10.6 percent of the metro-politan population in 1970, the highestblack proportion was only 3 percent (#111,#193, #194, #195, #265). In two New YorkCity audiences (#94 and #203), blacksrepresented 3 and 4 percent of attendersin contrast to over 16 percent of the

9 .and 5 perce~t black.

Data on audiences in the South differedlittle from other sections Of the Country.In JoffreyBallet audiences in threesouthern cities, blacks were underrep-resented in audiences by ratios of fromthree-to-one to thirteen-to-one (#38)relative to their share of local metropol-itan populations. "Nonwhites" (presumablyalmost all blacks) composed a rathersizeable 19 percent of visitors to aMontgomery, Alabama, art museum: but themetropolitan black population in that areais 34.4 percent. Only among sun, hervisitors of a Boston art museum (#17) wereblacks represented in proportion to theirnumber in the metropolitan population atlarge. Finally, nonwhites constituted arelatively high 16 percent of New YorkCity theatregoers in one study (#73).

It should be noted that for many insti-tutions a large portion of the visitorpopulation consists of tourists from out-side the relevant Standard MetropolitanStatistical Area (SMSA). Out-of-townvisitors have been found to compose between22 and 30 percent of visitors to theMetropolitan Museum in New York (#3, #16),between 12 and 55 percent of visitors toNew York’s Whitney Museum and the Museumof Modern Art0 and between 2 and i0 percentof visitors to museums in Newark andBrooklyn (#16). (These figures vary byday of week.) Percentages of out-of-SMSAvisitors to Baltimore museums and per-forming arts institutions range from 2to 14 percent (Cwi and Lyall, 1977).A strict comparison would have to takethese figures into account.

Individuals of Hispanic origin appearto have similarly low p~rticipationrates, although here the pattern isless clear. They ranged from 0.8 to3.2 percent of four sets of San Franciscomuseum visitors, while they constituted7.4 percent of the metropolitan popu-lation. Only 8 percent of the SanAntonio Joffrey audience (#138) and 5percent of American Museum of NaturalHistory visitors (#203) were found tobe Hispanic, but 37.5 and ii.i percentof San Antonio and New York City residents,respectively were of Hispanic origin in1970. The most anomalous findings onHispanic attendance appeared in a s~rveyin Washington ~State (#63), where Hispanicpeople composed from 5 percent of danceaudiences to 12 percent of history museumvisitors even though less than 2 per- _~cent of the state’s population is ofHispanic origin. If the findings are not

31

the result of unique methodological from 50 to~24 percent"foraft museums, and

aspects of the study, the high rate of- - .f~0m 44 to 23 percent for theatre (#137).

Hispanic arts attendance in Washing~-~,~..~. _o . _ .State is remarkable indeed and deserves Altho~gh most of the cross-sectional sur-

further study.

Information about minority attendancehabits can also be gleaned from sixcross-sectional studies undertaken bythe National Research Center of theArts. These surveys--two national, oneof New York State, one of California,one of Winston-Salem, North Carolina,and one of the New York Borough ofQueens--asked respondents if they hadattended each of several kinds of artsperformances and museums in the previoustwelve months. Relative responses ofwhites and nonwhites varied widely fromplace to place and time to time. InNew York State, virtually equal percen-tages of whites and nonwhites reportedattendance in every category except"concert or opera," where 36 percentof whites had attended as compared to23 percent of nonwhites. In Queens,slightly higher percentages of whitessaid they attended theatre and classicalmusic performances, but slightly morenonwhites attended dance (#190). InWinston-Salem a higher percentage ofwhites than nonwhites reported attendingall the performing arts (#201). InCalifornia white attendance was higherthan black and Hispanic attendance fortheatre, classical music, art museums,and science and natural history museums,but a substantially higher percentageof blacks reported attendance at danceevents. Hispanic respondents indicatedless attendance than blacks or whites atall the performing arts, but reportedattending museums more than blacks (#42).Consistent with the California results,a cross-sectiona! survey of the attitudesof Amarillo residents found blackrespondents relatively more enthusiasticabout classical music and Hispanicrespondents preferring the visual artsto theatre, classical music, or dance.

The two national surveys are ratherperplexing for although the second was areplication of the first and found rathersimilar rates of attendance among whites,attendance by nonwhites was sharply lowerin the second. The first survey, under-taken in 1973, showed roughly equalattendance at all the arts except fortheatre, where more whites reportedattendance, and dance, where greaterattendance was reported by nonwhites (#7).In the 1975 replication, however, whiteattendance substantially exceeds non-white in every category, with nonwhiteattendance dropping from 48 to 18 percentfor science and natural history museums,

veys do show relatively small disparitiesbetween the attendance behavior of whitesand minorities~ their findings must beinterpreted cautiously. Informationbased on people’s recollection isobviously considerably less reliable thaninformation obtained from people atactual arts events, and cross-sectionalstudy respondents may often defineattendance in idiosyncratic ways. Theresults of these and other differencescan be seen when the findings of a cross-sectional study of New York Stateresidents is compared with the results ofa statewide New York survey of individualsactually attending arts performances.Although nonwhites reported slightlyhigher attendance rates than whites fortheatre, ballet and dance, and museumsin the cross-sectional survey, nonwhiteswere consistently underrepresented in theactual audiences. This underrepresen-ration may reflect greater overreportingby nonwhite respondents: peculiaritiesof sampling: disproportionate attendanceby nonwhites at events excluded from theactual audience surveys: a tendency formany whites to attend frequently whilemany nonwhites attend only once ortwice a year: or some combination of theabove.

While the existing data does not permita definitive assessment--for example, nosurveys of museums or performing artscompanies appealing predominantly tominority group members were available--it seems likely that blacks and otherminorities are generally underrepresentedin performing arts audiences and amongmuseum visitors relative to their shareof the population. Since a higher per-centage of minorities than whites are veryyoung, poor, without college educactions0and/or employed in blue-collar or serviceoccupations--all categories with dispro-portionately low participation in artsaudiences--this is not in itself surpris-ing. In 1975, 34.4 percent of the blackpopulation, and only 26.1 percent of thewhite population, was under the age ofsixteen. The median income for whitefamilies in 1975 was $14,268 compared toa median of $8,779 for black families.Similarly, 63.2 percent of black civilianemployed persons were blue-collar orservice workers as compared to 44.3percent for whites. And the average blackperson twenty-five years of age or olderhad completed 10.9 years of schoolingcompared to a white median of 12.4.Although existing data do not permit anassessment, it is likely that poverty and

32

lack of education, rather than cultural - -found to vary-by-~~~time of perforfactors or racial exclusion, are respon- ~ . mance (day of wee~"~~me.ofd~y-), and thesible for the low level of minoritya~ts .... particular content of the performance orattendance. Only o~e audience 6~%~i-~itor _-~ -exhibit. - One final source of variationstudy (#193) reported educational attain- is that the composition of the audiencement by race. This study foun~ that thepercentage of black visitors who werecollege graduates was even higher (by afew percentage points) than the compar-able figure for white college graduatevisitors. Where data permits, furtheranalysis should be performed to assessattendance rates by whites, blacks, andHispanic persons of equal educationalattainment and comparable occupationaland income levels.

Summary of Demographics

The studies in our sample indicateconsistently that the audience for thearts is more highly educated, is of higheroccupational status, and has a higherincome than the population as a whole.Only one study out of ninety-seven foundthat the proportion of the audience witha college education was lower than thepopulation at large. Every one of thesixty-five studies which reported occu-pation found that the audience wascomposed of a substantially greaterproportion of professionals than thegeneral population, and only four ofseventy-six studies found that the medianincome of the audience was lower than themedian income of the population at large.

appears to differ slightly for differentart forms.

Museum visitor populations were somewhatmore representative of the American publicthan were the performing arts audiencessurveyed. The museum surveys foundsmaller proportions of professionals andthe well educated had lower median incomesthan did studies of performing artsaudiences. The differences found betweenthe museum visitor population and per-forming arts audiences may be attributablein part to the lower median age of themuseum visitor. There were differencesbetween the visitors to the various kindsof museums. The art museum visitorpopulation was better educated, wealthier,older, and composed of more professionalsthan visitors to history, science, orother museums. Among the performing arts,theatre audiences were somewhat lesseducated and less wealthy, and they werecomposed of a smaller proportion ofprofessionals than audiences for the otherperforming art forms.

ADDITIONAL ISSUES IN AUDIENC~ RESEARCH

Although women were slightly overrepre-sented in the arts audience, the genderratio varied extensively and one-quarterof the performing arts audiences in oursample and two-fifths of the museumvisitor populations were composed of moremen than women. The median age of thearts audience was close to the median ageof the population at large but variedwidely from audience to audience. Thefew studies which examined the racial orethnic composition of audiences indicatethat minorities were present in propor-tions smaller than their share of relevantmetropolitan populations.

All of the variables studied showedconsiderable change from audience toaudience. Some of this can be attributedto the differing methodologies, such asresponse categories, methods of sampling,and presentation of results. Some maystem from changes within an audience.Certain characteristics of audiences were

In this section we look at changes in thecomposition of arts audiences over timeto determine if the "reach" of museumsand the live performing arts has becomebroader, narrower, or remained the same.Also, we explore the differences betweenfrequent attenders and infrequentattenders and evaluate the evidence onaudience overlap among art forms: towhat extent does each art form have itsown devoted following and to what extentis it correct to speak of one artsaudience? Finally, we examine two impor-tant genres of audience research that donot deal with demographic composition.These are studies of the economic impactof spending by arts audiences on localeconomies and surveys of public attitudestowards government funding of the arts.

33

The Arts Audience Over Time

To examine whether the America~for live performing arts have been pro-gressively democratized over the pastseventeen years, we have evaluated trendsin five major i~dicators--gender, age,education, occupation, and income.During certain years, particularly somein the 1960s, we had few surveys to workfrom. Some of these have been groupedtogether in a span of several yearsto provide a more stable estimate ofaudience composition. At least six areincluded within each time period (withthe exception of one period for the dataon education). Furthermore, because ofthe relatively small number of museumstudies available for some of the periods,the analysis is limited to performingarts studies only. It should becautioned that the pre-1965 studiesinclude a number conducted by Baumol andBowen (between eight and thirteen,depending on the social characteristic).As we have already noted, these studiesyielded social profiles that were sig-nificantly more elite than those foundby most other audience surveys. Sincerelatively few other early studies areavailable, these surveys dominate theearly and mid-1960s audience compositionfigures, and this should be kept in mindin examining trends based on this period.

Gender. The proportion of men in the per-forming arts audience shows little changeover time, though there is a slight dropin recent years (Table 7). In mostperiods, the percentage of men varies fromthe low 30s to the low 50s, indicatingthat there is far more variation in gendercomposition from event to event thanbetween time periods.

A__q~. There is no indication of any trendtoward younger audiences.

Education. The proportion of the perform-ing arts audience with at least a collegeeducation evidences no decline over time.While the education level appears tofluctuate considerably between the firstthree time periods, much if not most ofthe change reflects special features ofthe studies conducted during theseperiods. Thirteen of fourteen pre-1967studies were executed by Baumol and Bowen,while seven of fifteen studies during the1967-1972 period were conducted onaudiences of university productions.(None of the post-1972 studies were ofcampus audiences.)

professional ~or~e~s-~[~lcollar.~workers--~e see littlechange over the.past..sevent~eny~ars-

Income. Income trends mirror thosereported for the other social indicators.The average income for 1960-1967 wasrecorded at a figure markedly higher (inconstant mid-1976 dollars), but again thisis based almost entirely on the Baumol andBowen surveys of prominent performing artsaudiences. It is notable that the medianincomes reported for audience studiesconducted within a time period vary farmore than do the averages between theperiods.

In short, our data do not reveal anystriking changes in the composition ofthe audience over the past one and one-half decades. However, we caution thatthe heterogeneity of the audience studiesevaluated here may have concealed subtlertrends. For example, if audiences forone art form were becoming increasinglymale while audiences for another morefemale, such a change would not bediscernable in our data. Similarly, iftheatre audiences in major cities werebecoming more diverse, while theatreaudiences in smaller cities and suburbswere becoming less so, no change wouldbe observed. Moreover, any changes inthe audience of particular organizationsor sectors would not be reflected in theaggregate figures we have considered.It is possible, for instance, that theaudience for professional dance compan-ies--or any other art form--is undergoinga significant broadening while theaudience for certain other arts forms isremaining stable or even narrowing.

Another way to examine time trends, andone which eliminates problems emanatingfrom the aggregation of studies ofdiverse institutions, is to comparestudies of the same arts organizationwhich have been conducted at differenttimes. In twenty-nine cases we havemultiple studies of an organization’saudience. However, the research method-ologies were usually so different and theidea of time-series data so absent amongthe studies that a meaningful comparisoncould be made in too few instances.

Occupation. Combining two indices of theoccupational composition of performingarts audiences--the percentages of

Audience Structure

In most audience studies attention israrely directed at one particularlycritical difference among audiencemembers: frequency of attendance. Somepersons are veterans of many performances

34

Table 7 ~.Time Trends in the Gender, Age, Education,Occupation, and Income Composition ofPerforming Arts Audiences

TotalMedian of Range of Number of

Social Character and Time Period Medians Medians Studies

Gender: Percent Men

1960-65 56 45-58 I0

1966-69 46 32-54 7

1970-71 42 36-51 ii

1972-73 45 33-54 ii

1974-75 37 35-43 9

1976 39 34-54 13

Age: Median (in years)

1960-67 37 33-45 9

1968-70 36 24-46 6

1971-72 41 34-42 8

1973 30 21-35 i].

1974 36 22-43 8

1975 38 29-48 12

1976 33 21-45 18

Education: Percent withCollege Degree or More

1960-66 72 61-86 14

1967-72 47 21-66 15

1973 63 55-65 6

1974 67 54-74 4

1975 57 48-65 7

1976 65 34-76 13

(Table 7 continues on following page)

(Table 7 continued)

Occupation:

Percent professional

1960-691970-741975-76

Percent blue-collar workers

1960-691970-741975-76

Income: 1976 Dollars

1960-671967-701971-731974-751976

65 48-80 ii57 50-63 i059 24-73 20

2.4 1-5 82.8 0-5 143.0 1-7 ii

23,40719,01719,68418,98320,004

19,342-28,02716,819-25,229

9,466-27,24515,292-23,20214,003-21,004

iii0127

ii

or visits, others rarely visit, and still----.Studi4s i~volvlng-~0~@i~£h~h ~ single typeothers are in the audience for the first .. ~ of arts organizati~~itY9i~ally reveal thattime. (Some are also there for the last. " frequent attenders of one type-of insti-time.) A national cross-sec~iOna~%~i#~ey -rut±on .also tend to be frequent attendersin 1975 reports that 47 percent of thepublic had attended at least one theatre,classical music, or dance performanceduring the previous twelve months: 52percent had visited a museum. Of the per-forming arts consumers, 62 percent hadmade one to five visits, while 38 percenthad gone even more often: of the museumvisitors, 58 percent went to the museumfive times or fewer° while 42 percent hadvisited more frequently (#137). Mostaudiences contain a mixture of regular andirregular arts consumers, and for somepurposes defining the relative proportionsis useful.

Growing total attendance can reflect anincrease in the number of individualsdrawn to the arts, an increase in thefrequency of visits, or both. One organ-ization experienced in audience researchdistinguishes between the "reach" and"frequency" of an audience. Reachdescribes the percentage of a communitywhich attends an arts institution atleast once during a one-year period,while frequency is the average numberof visits made by attenders during theyear (Morison and Fliehr, 1974). Theratio of reach to frequency can vary con-siderably from audience to audience. Forinstance, in one study of a park and atheatre in the park, it was found thatthe park’s reach was 6.0 percent (6 per-cent of the area residents had visitedthe park during the past year), whilethe theatre’s reach was only 2.5 percent(#118). On the other hand, the frequencyof the park visitor was 4.4 (of those everattending during the previous year, eachaveraged a little more than four visits),but the frequency for the theatre patronwas 5.4. In other words, the theatreattracted a smaller number of individualsthan the park, but it was a more committedclientele.

Reach is a good measure of an organiza-tion’s breadth of appeal, while frequencysignifies the extent to which the organ-ization has cultivated a regularconstituency. Though outreach programsare usually aimed at increasing theformer, some may actually be affectingthe latter. One art museum developed aspecial exhibit designed, in part, tobroaden the museum’s appeal. However, avisitor study revealed that althoughattendance did significantly increaseduring the exhibit, much of the expansionwas due to the return of regular visitorsrather than the appearance of newvisitors (#135).

of other institutions. An analysis ofcultural consumers in California revealsthat of infrequent museum visitorsloneto five visits during the previous year),47 percent had not attended a theatre,classical music, or dance performanceover the previous year and only 19percent had attended more than five times.By contrast, of frequent museum visitors(more than five times per year), only24 percent failed to attend one of theseperforming arts and 47 percent had goneto more than five performances over theyear (#42). There is even some evidencethat frequent arts attenders participatemore heavily in all leisure pursuits,such as sporting events, movies, thecircus, and creative activities (#’s 7,39, 42, 190, 203). The habits ofattenders of one art form differ fromthose in another. One study found, forinstance, that 63 percent of respondentswho had been to the theatre during atwelve-month period had attended no othertype of performing arts event. Bycontrast, only 36 percent of the audienceat the symphony, 25 percent at the opera,and 20 percent of the ballet had failedto attend at least one other type ofperforming arts event in the past year(#115). There are various ways ofmeasuring audience overlap, but howeverapproached, the results seem to indicatethat theatre audiences are the leastintegrated with those of the other per-forming arts (#’s 8, 115). Also, there issome evidence that slightly differenttypes of people frequent performing artsevents as compared to museums (#42).

Arts audiences distribute themselvesalong a continuum. Clustered at one endare those who frequently attend a varietyof arts events, and at the other end arethose who only occasionally sample asingle event. The habitual attendersgroup themselves in active social circles.Friendship and acquaintanceships areformed around a shared interest in thearts, cultural events are central topicsof informal discussion and exchange, andthere is the expectation that attendanceat, and knowledgeability of, the arts ishigh. Several studies report thatfrequent attenders are more likely thaninfrequent ones to hear about arts eventsthrough their social networks, to countcultural consumers among their friends0and to indicate that arts attendance isfashionable in their social milieu (#’s7~ 42, 64, 93).

37

This cluster of the arts audience is also r~ported ~ lower proporti~oh-~f men, anddistinguished from the occasional attend- " :’two studies found no difference~ Sim-ers by its social character.. Sixt.ee~.~--~i~ ~ ~ o~ilar!y, six studies concluded thataudience studies in our possession frequent a~tende£s were older thanexamined the relationship between fre-quency of attendance and education, andall sixteen found that regular visitorsare more highly educated than irregularvisitors for both museums and theperforming arts. A cross-sectional studyof Californians, for example, found thatof those who had not visited a museumduring the past year, 7 percent held acollege degree or more: of the infrequentmuseum visitors (one to five times), 18percent were college educated: and of thefrequent visitors (more than five times),31 percent held college degrees. Thecorresponding figures for the performingarts were 7, 18, and 43 percent, respec-tively (#42).

Those among the regular arts audiencealso tend to have higher incomes, thoughthe evidence here is less clear-cut thanfor education. Thirteen of seventeenstudies with relevant data report higherincomes for frequent attenders than forinfrequent attenders, but one studyrevealed no difference and three indicatedthe reverse. In all three of the lattercases, the audiences were for ballet ordance. For example, a study that includedballet audiences in New York State foundthat median income for frequent attenderswas $19,000, as compared to $19,400 forinfrequent attenders (#73).

infrequent visitors, three found theopposite, and two reported no age differ-ence.

Since frequent attenders are more likelyto be present in an audience for aspecific performance or to be museumvisitors on any given day, most audiencestudies are, strictly speaking, studiesof those present rather than of visitors.As we have seen, regular arts consumersare generally more highly educated andsomewhat wealthier than irregularconsumers. Thus social statistics basedon those present will tend to reveal asomewhat more affluent profile than ifthe statistics were based on all thosewho ever participate in arts audiences.

Economic and Political Impact

The precarious financial condition facedby many arts organizations and the growthof government interest in the arts haveled to an intensified search for ways ofjustifying public support. Increasingly,audience research has provided thefactual platform upon which rationalesfor public support of the arts have beenerected.

There is some evidence that income mayhave a stronger relationship to frequencyof attendance for the performing artsthan for museums. In one cross-sectionalstudy, for example, the income gap betweenfrequent and infrequent attenders is$2,900 for the performing arts but only$800 for museums (#42). Although museumadmission charges typically are cheaperthan performing arts tickets or arenonexistent, we suspect this explainslittle of the difference in attenderbackground. Studies of visitors tomuseums before and after the institutionof an admissions charge (Cameron andAbbey, 1962) or studies comparing "free"periods to times when admissions fees arecharged (#17) have found little variation.Then, too, professional sports and rockconcerts impose admissions fees comparableto those for the performing arts, yet suchevents, we suggest, often attract aconsiderably less "upscale" audience.

There was no decisive pattern for thegender and age composition of frequentversus infrequent visitors. Four studiesindicated that frequent attenders had ahigher proportion of men, six studies

Audience surveys may prove of practicalvalue for promoting public support inseveral ways. Social profiles can beused to demonstrate that a broad crosssection of the public is being reached byan arts organization and that, by impli-cation, the organization is performing avaluable quasi-public service. Anotherapplication of audience research to theacquisition of public backing is in theidentification of secondary economicbenefits of arts institutions for thelocal community. A third practical useis in demonstrating the educational valueof exhibits and performances for attenders,thereby showing that the arts serve publiceducation. Finally, attitude surveys ofcross sections of the public can be usedto document widespread support for thearts, so that spending on the arts byfunding agencies and legislative bodiesis made politically legitimate.

While demographic profiles have beenacquired in virtually all audience studies,few have examined the secondary economicimpact or the public appeal of the arts.Our-assessment of attitudinal and economicstudies, then, rests on a more tenuous

38

base than our assessment of the far moreextensively researched social profile ,~. important’’~ reasonf0r-the trip and 58

" . ~ _percent affirmed that the visit was "aquestions..- - ~~"i~¢~’~.!. ~-major" reason behind the trip. Comparable

Economic impact. Studies of the local .economic impact of the arts have not~solely relied on audience survey method-ologies. The direct and indirectconsequences of an arts organization’spayroll and purchases have been examined:efforts have been made to identify thelargely uncompensated contributions ofarts organizations to schools and otherlocal institutions: and the effects ofcultural resources on individual businessfirm decisions to locate in a communityhave been considered (see, for instance,#139: Arts Education and American Panel,1977: and Cwi and Lyall, 1977).

Audience research is particularly wellsuited to answering still other types ofeconomic impact questions: Are artinstitutions an important considerationin the decision of nonresidents to visita city? How large are the nonarts expend-itures during a visit to an arts insti-tution? What sectors typically benefitfrom the infusion of the associatedexpenditures?

Nine audience studies in our possession,all except one conducted in the mid-1970s,addressed one or more of these issues.One study was based on a survey of aBoston commerical theater audience (#4):a second was a survey of New York commer-cial theater audiences (#37): anotherinvolved a study of visitors to the NewYork Metropolitan Museum of Art (#3): afourth consisted of a survey of fourteenaudiences of nonprofit performing artsevents in Wisconsin (#29): a fifth andsixth were of performing arts and museumvisitors in New York State (#73) andWashington State (#63): the seventh wasbased on a survey of visitors to sevenmajor Chicago museums (#11): and two morewere surveys of audiences for a balletcompany (#94, #138).

There is no effective way of measuringhow cultural institutions draw visitorsto a con~nunity. As a result, thesestudies have relied on a technique whichyields suggestive but not definitiveinformation on this matter. Art organi-zation visitors are simply asked whetherthe presence of the institution was amajor factor in their decision to visitthe city. Thus, ~ng the nonresidentvisitors to the Metropolitan Museum ofArt (nonresidents comprised half of allvisitors), four-fifths reported thatthey had planned to see the museum priorto their arrival in New York City. Andof these, 24 percent indicated that their

intention t~-s~e ~e.~imus~um-_wa@_’’a fairly

levels of museum drawing power were foundin the Chicago study. Nonresidents. wereasked: "Was-a visit to the museum ormuseums an important reason for your tripto the city?" Nearly 50 percent indicatedit was the "main reason," and 85 percentattributed at least some importance tothe seven museums in stimulating theirtravel plans. The number of city visitorswho would not have come were the museumsunavailable cannot be fixed with anyprecision using these figures, but it isclear that a substantial proportion areattracted to the city largely as "culturaltourists." Since cultural consumers tendto be highly affluent, the arts may beparticularly effective in attractingthose who are most likely to make sub-stantial personal expenditures duringtheir visit to the metropolitan area.

The expenditures by visitors on nonartsgoods and services varied considerably.Patrons of the Boston theatre spent $6.40on the average: in several cities, between$5.00 and $14.00 were spent aside fromthe performance by persons attendingballet: New York State residents whoattended arts events spent an average$7.80 on nonarts activities while out-of-town patrons spent $14.30: Washingtonresidents spent $6.70 along with attend-ing a performing arts event: Wisconsinperforming arts audience spent $1.90 perperson in attending one of fourteensurveyed events but $15.80 in attendinganother: out-of-town visitors to Chicagomuseums spent $16.00 on the average: andout-of-town visitors to the New YorkMetropolitan Museum of Art typicallydisposed of $85.00 (a median figure).

If these amounts are used to estimatetotal annual expenditures, the directaggregate impact on the local economyis considerable. In Boston, visitors ofthe single theatre alone contributed$3.9 million to the local nonarts economyduring one season and $6.6 million duringanother season when attendance rates werehigher (nonresidents were not distin-guished from residents in this study, soonly a fraction of these totals repre-sents the infusion of outside capital).In Chicago, out-of-town visitors of theseven museums contributed $76.5 millionto the economy, and in New York nonresi-dents passing through a single museumwere responsible for approximately $187million in expenditures annually. Thesefigures represent direct outlays, andthere are additional indirect economicbenefits as the money changes hands

39

several times before entering savings or ~0 percent-endorsestate~ ~overnmenttax accounts. A multiplier of two is- i.:’...backing in this circumstance: and 63often used to estimate the-recycling-~" : " - ~percent back local government inter-effects, and so the combined-dir~c~ ~dd- vention (#42). " Comparable patterns areindirect economic impact may be as muchas double the above figures.

Not surprisingly, virtually all of thespending is concentrated in the usualtourist industries. For instance, inthe Chicago study (#ii) of the totalmuseum related expenditures 29 percentwent to restaurants, 27 percent retailstores, 21 percent lodging, and 9 percenttransportation. Thus, it is evident thatcertain sectors of the local economybenefit considerably from purchases bycultural tourists. It remains to bedemonstrated that the whole economy, themunicipal government, and the localpublic also benefit from this sectoraleconomic impact. It has not been shownthat the benefits outweigh any addition-al tax burden borne by local residentsresulting from government underwritingof art organization deficits. Also,it has not been shown that most of themoney spent on activities associatedwith attending arts events would not havebeen spent in the absence of such events.Still another important issue which hasyet to be addressed empirically is thelocal economic impact of public sponsor-ship of the arts relative to governmentinvestment in other areas or institutions.

recorded for Winston-Salem, NorthCarolina, (#201) and Anchorage, Alaska,(#93): the percentages supportingfederal, state, and local govvrnment

°financing are 49, 60, and 64 in theformer region and 47, 69, and 74 in thelatter. In Boston, more than half (57percent) of the city’s residents favoredexpansion of a city-sponsored culturalprogram from a summer season to a year-round basis (#62). And in Salt LakeCity, Utah, a majority of the public (58percent) would urge a greater allocationof the municipal budget to culturalevents (#166).

Political impact. Although the economicbenefits have not yet been decisivelydemonstrated, it appears that publicsupport for government subsidy is alreadywidespread. This conclusion emerges fromten studies we have assembled on publicattitudes toward government underwritingof the arts. Eight of the studies arecross-sectional surveys of the public(including two national studies), and theother two are of performing arts andmuseum visitors in two states. Nine ofthe studies have been conducted since1973, and the tenth was done in 1970.Seven of the inquiries were carried outby a single organization--the NationalResearch Center of the Arts. (#’s 7, 42,63, 73, 93, 137, and 201: the others are#’s 62, 66, and 187).

Within certain regions of America,majorities or near majorities endorse thegeneral principle that the governmentshould help finance cultural organi-zations that are running deficits, withlocal help clearly preferred over federal.Among California residents, for instance,49 percent subscribe to the positionthat the federal government "should helparts and cultural organizations in thearea if they need financial support":

The apparently high levels of publicsupport in these regions may be due tothe question-sensitive nature of thisissue (though conceivably there could beregional pockets of high support forgovernment involvement). When a nationalsample of the American public was askedin 1973 whether "cultural organizations(should) have to pay their own way, orshould .... be able to receive directgovernment funds to help support them,"only 38 percent adopted the latterposition, while 34 percent indicated thatcultural organizations should rely ontheir own means and 28 percent reportedthat it depended on the circumstances orwere undecided (#76). Even greaterskepticism is evident whenthe issue isgovernment support for artists ratherthan cultural organizations. O~ly 31percent of the California public agreedthat "professional artists should receivehelp from (the) California state govern-ment if they need financial assistance tocontinue their artistic profession" (#42),and in 1975 only 29 percent of the Ameri-can public endorsed federal supportfor needy artists (31 percent endorsedsupport by state or local government)(#137).

The level of public support for inter-vention varies widely according to thespecific type of cultural organizationinvolved, with museums faring far betterthan specific kinds of performing artsorganizations. Thus, while 38 percentof the general public in a 1973 surveyagreed with the principle that "culturalorganizations such as museums and symphonyorchestras" should be eligible for govern-ment underwriting, far smaller proportion8urged such eligibility for specific kindsof performing arts organizations. Onlyllpercent of the public would like tosee opera receive public funds: thepercentages for commercial theatre,

4O

nonprofit theatre, ballet and dance, and -~ a-high p~io~ity-w~r~-;�Onfronted ~ithsymphony orchestras stood at only 5,-12, .~.�oncretep01itical Choices- Some evidenceii0 and 16, respectively.. BY .c°ntr~st.~ ~ -: indicates that a Substantial part of the

government subsidies for museums~£e~ ~ar publicis prepaJed to have the government

greater support. The.percentagesendorsing government support for art,science, and history museums were 41, 55,and 57, respectively (#7).

There is some indication that the levelof support has grown in recent years asgovernment spending on behalf of culturalorganizations has itself expanded. In a1975 survey of the general American pub-lic, the percentages accepting the ideaof local government support for operahad increased to 33 percent, for theatreto 38 percent, for ballet and dance to 33percent, and for symphony orchestra to37 percent, similarly, subsidies forart, science, and history museums werenow supported by 46, 64, and 64 percentof the public, respectively (#137).

The rank order of the level of publicsupport for the various art forms closelyparallels the degree to which the formsattract a socially elite audience. Themore representative an art audience isof the general public, the more widespreadis public support for government financ-ing of the art form. This is hardlysurprising, for one would expect interestin government support for the arts tocorrespond to the benefits perceived.Among those attending performing artsevents and museums in the states ofWashington and New York, over 80 percentfelt that government assistance should beprovided performing arts organizationsand over 90 percent felt that it shouldgo to museums (#63:.#73). Similarly, incross-sectional surveys two of the bestpredictors of individual willingness toendorse government involvement are theindividual’s educational level (alreadyshown to be one of the best indicatorsof arts attendance) and whether the indi-vidual is an active arts consumer. Inthe 1973 national survey, 22 percent ofthose with an eighth-grade educationagreed that the government should supportcultural organizations, while 50 percentof the college educated took this position.Twenty percent of the nonattenders but 64percent of the frequent attenders (thosein the top decile of the attendance rate)shared the view that government subsidiesfor the arts were desirable (#7).

While large segments of the public agreein principle that government support forthe arts is appropriate, it is less clearthat these segments would give the arts

intervene in at least a very modestfashion. In Anchorage, for instance, 71percent of the residents assert that theywould be willing to pay an additional fivedollars in local taxes to support communitycultural activities (#201): 54 percent areso inclined in California (#42), and 58percent of the 1975 national populationwould be willing to undertake thisnominal payment (#137). A fivefoldincrease in the tax burden, however,results in many fewer supporters: 20percent of the California respondents and41 percent nationally would support atwenty-five dollar increase in their taxesto underwrite the arts (#42: #137). Again,willingness to undertake this burden ishighly correlated with whether the indi-vidual is a cultural consumer. However,it is also clear that the arts still rankfar below other priorities for most ofthe public. When a national sample wasasked in 1975 to evaluate the importanceof various community services, the artsrated below health0 transportation,education, law enforcement, housing, andrecreation facilities. Similarly, whenasked whether federal spending should beincreased in a number of areas, respon-dents ranked the arts far below education,health, public transportation, and housing,with only defense and welfare spendingrated significantly less preferable thanthat of the arts (#137).

It is evident from available audienceresearch, then, that strong minorities ofthe public (and in some cases majorities)are in agreement with the genera! prin-ciple that the government should beinvolved in funding cultural organizations,though there is less support for directfunding of artists themselves. Support isstrongest among those segments who standto benefit most directly from increasedgovernment backing. However, while theseresults are suggestive they cannot be usedto determine whether this public supportfor the arts is--or could be--mobilizedin the political process. We do not know,for example, whether the arts lobby hasa more willing public to mobilize onbehalf of art spending than do otherinterest groups on behalf of other, com-peting priorities. Nor do we knowwhether public attitudes toward govern-ment arts poli~ies become translatedinto voter preferences during electioncampaigns.

41

OF AUD CE

Arts institutions and organizationsconcerned with the arts have alreadyundertaken a great many studies ofaudiences, and the tempo of such researchappears to be increasing. Arts managersand policy makers have studied audiencesin order to assess public attitudes, todetermine the composition of the publicthat particular institutions serve, tohelp decide on prices and hours, toprovide baseline data for market develop-ment programs, and to estimate the impactof arts activities on local and stateeconomies.

Such research has been greeted with acombination of skepticism and enthusiasm.An increasing segment of the arts commun-ity seems to feel that institutions "inneed of practical advice miss a goldmine of wisdom by neglecting to surveytheir audiences" (Wainwright, 1973).Others assert that the research is oftrivial importance, an expensive way offinding out what is already known.

Has audience research been of value tothe arts? To answer this question wemust ask two more in turn. First, hasthe technical quality of audience studiesbeen sufficiently high to provide infor-mation that, if acted upon, will permitmanagers and policy makers to accuratelypredict the impact of their decisions?Second, has the research been planned insuch a way that the individuals respon-sible will be willing and able to use itsresults? Research can be of the highesttechnical quality, but if it does not leadto recommendations that decision makershave power to implement, it will not beuseful. Similarly, if research providesdata directly relevant to pressingdecisions but the research is shoddilyexecuted, policy based on that researchis more likely to have unfortunate con-sequences.

The purpose of this chapter is to discoverthose factors most closely related totechnical quality and policy utility ofarts audience research. Our strategy hasbeen to rate the quality and utility ofeighty-six studies of arts audiences andto ascertain the relationship betweencertain characteristics of the studiesand their scores on the quality andutility scales. Organizations thatconsider sponsoring or undertaking

audience research may use these findingsas guidelines against which to measuretheir own assumptions about such issuesas what kind of research to do, whetherto do research in-house or contract out,what kind of researcher to hire, and howmuch to spend.

THE ARTS AUDIENCE SURVEY

Our discussion is based upon an intensiveexamination of eighty-six studies of artsaudiences and on completed surveys fromthe directors of these studies. Inaddition to reports on museum visitorsand performing arts audiences, weexamined cross-sectional surveys of localor national populations designed toacquire information on exposure to and/orattitudes toward the arts. Most of theeighty-six studies employ traditionalsurvey techniques, although some studiesuse quasi-experimental designs (Campbelland Stanley, 1966). They were undertakento provide information for a variety ofpurposes, ranging from fund raising,audience expansion, and marketing toplanning facilitie~ setting ticket prices,and lobbying legislatures.

We described our acquisition procedures inchapter one. Within three months, we had127 audience studies which had been under-taken since 1970. Studies conductedbefore 1970 were excluded on the groundsthat study directors would find itdifficult to recall essential proceduraldetails of their research. We estimatethat at least 400 audience studies havebeen conducted since then and so the 127located for this inquiry can be assumedto be reasonably representative. Somebias towards more recent studies andtoward studies of above average qualityand utility may have resulted from ourprocedures.

Two types of information were compiled.First, each study report was coded by tworaters on a variety of quality dimensions.Second, a twelve-page survey form wassent to directors of i12 studies. (Fif-teen study directors could not be located

42

or were deceased.) After a second mail- the eigh~y-~iX ~udi~6~?~-t~i-es--- Theing and several telephone contacts, ~ - ..~.first set Was used wish the questionnaireusable forms were received frgm-e~hty~.-~-_-, which had been completed by the directorssix of the directors, for a response of thestudies:°the second set wasrate of 77 percent. The study audienceswere distributed among the various artforms as follows:

Art Museums 21History Museums 14Science Museums 7Ballet 12Dance 6Jazz 7Folk and Ethnic Music 4Chamber Music i0Orchestras 17Commercial Theatre 7Nonprofit Theatre 32Opera iiCross-sectional Studies 13

The total exceeds eighty-six because manystudies surveyed audiences of more thanone art form.

PREDICTING QUALITY IN ARTS AUDIENCESTUDIES

By technical quality we refer to theextent to which a study is properlyconceptualized and executed in accord-ance with the norms of scientific invest-igation. Previous efforts to assess thetechnical quality of research havegenerally relied on generalized assessmentsby peers or specially trained reviewers(e.g., Persell, 1971: Gordon and Moris,1975: Yin et al.0 1976) or on itemizedassessments in which raters, counting thenumber of specific procedures, generate ascore on a quality index (e.g., Gephart,1965: Bernstein and Freeman, 1975: Yinet al., 1976: McTavish et al.0 1977).While there is merit in both methods,because of resource limitations only thelatter is used here. Drawing on a numberof standard discussions of preferredtechnical procedures in social research(e.g., Kerlinger, 1973: Bernstein, 1976:Campbell and Stanley, 1966:Lin0 1976),on a~ exhaustive list of seventy-fivedesirable technical research featuresdeveloped by McTavish et al. (1977), andon observations of factors specificallyrelevant to arts audience research (Mann,1972: O’Hare, 1974: Cameron and Abbey,1960b), we established two sets ofcriteria for evaluating the quality of

employed by two raters who evaluated thereports available on each audiencestudy-1

We divided the quality criteria into twodomains. Following a distinction elab-orated by Campbell and Stanley (1966)and by others (e.g., Bracht and Glass,1968: Bernstein, 1976), these domainscan be referred to as internal validityand external validity. Internal validityrefers to the extent to which an investi-gator can eliminate alternativeexplanations as causes. Externalvalidity refers to the extent to whichthe researcher can generalize from theindividuals studied to a larger population.

Internal validity of each survey isassessed using nine items on the investi-gator’s questionnaire and ten items fromthe research report assessment. Theseitems include whether the survey waspretested, trained personnel were usedin the administration of the study,multivariate statistical techniques wereemployed, and a valid linkage was madebetween the survey’s data and the con-clusions drawn. External validity isassessed with ten items on the investi-gator’s questionnaire and eight itemsin the report assessment dealing withsuch issues as sample selection, samplesize, testing for response bias, and useof tests of statistical inference.2 Eachitem was dichotomized into high- and low-quality categories. Quality scales wereformed by summing the number of times anaudience survey fell into the high-qualitycategory.

While some of these items may appearesoteric, each can have a significantimpact and offers the potential of dis-torting research findings. For example,imagine a situation in which poorlysupervised theatre ushers are responsiblefor inserting survey forms in programsand placing them on every other seat:the usher responsible for the front ofthe house places the programs in thecorrect manner: the usher for the middlerows inserts the surveys properly butforgets to collect them: and the usherresponsible for the rear falls ill at thelast minute and is replaced by someoneunfamiliar with the survey procedure whofails to distribute any questionnaires.The audience members in the front rowseats dutifully fill out and return theirforms and, when the program has finished,the researcher has a total response rateof about 30 percent. The researcher doesnot bother to check the representativeness

43

of the seats from which completed forms ~ standard-deviation-higher in-.technical

were gathered. When the results are ..... " ~. quality than the average audience survey.calculated, he or she is .surp.ri.se~d%.to~~.~.-," .-~ In the study .a .questionnaire was dis-find that the crowd is older and m~re tributed to randomly chosen visitorswell-to-do than expected. If the theatremanagers do not choose to ignore thesurvey findings, there is a danger here.They might launch an expensive campaignto recruit younger and less affluentpeople to their performances, withoutrealizing that the findings simplyreflected the fact that audience memberswho purchase more expensive tickets aregenerally older and more affluent thanthose in the less expensive seats, whowere unrepresented among the returnedquestionnaires (Baumol and Bowen, 1966).Because the respons@ was biased, andbecause the investigator failed to takethis into account, the audience surveycould mislead its sponsors.

While this hypothetical case is extreme(though perhaps not so unusual as onemight hope), it indicates the problemsthat can result from poor researchtechniques. Failure to pretest question-naires may result in answers that areuseless or misleading. Failure to usemultivariate statistical techniques maylead readers to infer that one factor isresponsible for a second when, in fact,they are both caused by a third. Failureto sample properly may result in general-izations about an entire visitorpopulation on the basis of responses froman unrepresentative group. Thus, theinternal and external quality scales areimportant elements for measuring validresearch.

We discovered that the internal andexternal quality scales were stronglyassociated: studies high on one scaleare likely to be high on the other. Theinterscale correlations are .566 for theinvestigator questionnaire items and .733for the report assessment data. Accord-ingly, the internal and external validitydimensions for each data source werecombined into a general quality measure.3

Similarly, using this single qualitymeasure, we found that ratings from theinvestigator questionnaire items and thereport assessment data are also highlycorrelated (.579). Thus, these too werecombined to form a single overall qualityscale that serves as our technicalquality measure.4

The variation in research qualitymeasured by this scale can be illustratedby comparing studies that fall high andlow on the index. An example of a highquality study is a social profile s%rveyof the visitors to a major metropolitanart museum: this study is a full

during four time periods selected torepresent the seasons of the year. ~Thosedistributingthe forms were trained andclosely supervised. Nearly 5,000 visitorswere approached, more than 95 percentprovided usable responses, and bothpopulation variability and the width ofpreferred confidence intervals wereconsidered in selecting this large asample. The analysis was facilitated bya computer. Although neither scalingnor multivariate techniques were employed,the results were weighted to adjust forthe sample frame and tests of signifi-cance and confidence intervals wereestablished. The study report includeda discussion of the research design(though previous audience research wasignored), valid linkages were drawnbetween the data and conclusions, andthere was a discussion of the policyimplications accompanied by concreterecommendations. The report lacked asynopsis of its basic findings as well asa statement of the study’s limitations.

For purposes of comparison, we haveselected a low quality study of theaudience for a single performance of anonprofit theatre: its quality is a fullstandard deviation below that of thetypical audience study. The survey formwas not pretested and those who adminis-tered the survey were not carefullysupervised, but a probability samplingprocedure was employed. The sample size,however, was not based on considerationsof statistical i~ference, a response rateof approximately 50 percent was obtained,and no effort was made to adjust forpossible response bias or for the sampledesign itself. The analysis was under-taken without the aid of a computer,simple bivariate statistics were the mostcomplex data analyses performed, and thereport presented little more than thedistributions of respondents among thevarious response categories. The researchdesign, policy issues, policy impli-cations, and study limitations were no-where discussed.

FACTORS PREDICTING RESEARCH QUALITY

The quality of a research study is, wethink, a function of the resources thatan investigator can mobilize. Such

44

resources include the investigator’s " distinguished..The~f~-i~.the profes-personal capacities and background and a A-~-~-sion of the investigator, since differentvariety of external factors,_such~S:~"~S" -~_professio~s hold varying definitions ofor her colleagues, audiences, careerincentives, time, and financial support.For instance, if the intended audiencefor a report is not well-equipped tojudge its methodological rigor, theinvestigator is less constrained to main-tain orthodox methodological standards.similarly, a shortage of funds can forcethe investigator, whatever his or herpersonal standards, to employ lessacceptable but more economical techniques.

In our hypothesis, then, predicting thequality of a study is partly a matter ofidentifying the investigator’s researchcapacities and the necessary environmentalsupport. To this end, we analyzed theinvestigator’s experience and background,the organizational setting of the studyand the financial resources available.

acceptable research procedure. Bernsteinand Freeman found that variations inprofessional norms between social-sciencedisciplines ~id have consequences forresearch quality (1975: 118). Evensharper differences may be expectedbetween investigators affiliated with thesocial sciences and those identified withthe marketing or arts management profes-sions.

The personal capacity of the investigatorto conduct high quality research dependson his or her level of training andresearch experience. It is true that inan analysis of 236 major federal eval-uation studies initiated in 1970,Bernstein and Freeman found that the ¯researcher’s level of formal traininghad little bearing on study quality (1975:115). Yet the absence of an effect offormal training may not be universal. Itwill be examined here through the variableinvestiqator deqree, the highest formaldegree obtained by the study director.5

Investiqator experience, our measure ofrelevant research experlence, will beassessed by the sum of the number ofsurveys the investigator had conductedprior to the audience study in question.6

Financial resources that affect thequality of a research product include thesize, quality, and organization of theresearch staff, library and computerfacilities, and disposable funds for thepurchase of ancillary research materials.A convenient, albeit approximate,aggregate measure of project financialresources is the total study budget.Although Bernstein and Freeman found nosignificant impact of budget on qualityfor their evaluation studies, theyexcluded studies with total expendituresunder $i0,000. M~st of the eighty-sixarts audience studies considered herewere conducted with more modest resources.Only ten of the directors report costs of$i0,000 or more ($150,000 was the mostexpensive), and the median cost was amere $471.

Three sets of institutional factors thatmay affect research quality can be

A second potentially significantinstitutional factor is the nature of theorganization in which the investigatorworks. The scientific method is perhapsbest established in academic institutions,less so in nonacademic research organi-zations, and least so in artsorganizations. Studies of research inother fields have yielded conflictingconclusions about the relative qualityof academic and nonacademic research. Inan analysis of 140 studies of technolog-ical innovations in local services, Yinet al. (1976) found no relationshipbetween the kind of organization conduct-ing the study and the quality of theresearch. Yin and Yates’ assessment ofcase studies of urban decentralizationand participation (1975), however,indicated that higher quality studieswere conducted in academic institutions.Bernstein and Freeman (1975) report asimilar finding.

The third institutional factor is therelationship of the organization conduct-ing the study to the institution that isthe subject of the inquiry. An in-houseresearcher may have a stake in producingresults acceptable to his or her organ-ization, whereas an autonomous outsideresearcher may find it easier to maintainan independent, objective stance. On theother hand, in-house investigators maybe more sensitive to the research settingand, as a result, may develop moreappropriate research designs. Thecounterbalancing of these two factorscould explain the apparent inconsistencyof previous research on this issue. Yinet al. (1976) found that outsideresearchers did higher quality studiesthan insiders. Yin and Yates, however,found no relationship between thesefactors, and Bernstein and Freeman foundthat in-house investigators did somewhatbetter than their unaffiliated counter-parts.

We analyzed the institutional setting ofart audience studies in terms of thefollowing variables. Investiqator’s~rofession: thirty-one of the study

45

directors were primarily arts managers:The second, index of~inveStiqat~r back-

fifteen were in marketing: fifteen were . ~ground--the investigator’s highest

identified with a socialsciencedls~~!.~ " degree-does predict study quality.

cipline: and the remaining twenty-fivewere associated with a variety of otherresearch-related fields. Orqanizationtype in which the study director worked:twenty-seven were arts institutions:twenty-three were independent researchfirms (nonprofit and profit): andnineteen were academic institutions.Orqanization experience: the number ofsurveys of any kind that the organizationhad sponswred before the study inquestion." Finally, orqanization affili-ation of the study director: thirty-seven of the eighty-six studies wereconducted by internal researchers:forty-nine were not.

Researchers who hold Ph.D.’s and compar-able credentials conduct studies whichare, on the average, 4 to nearly 7 pointsabove average. Those with only B.A.’s orM.A.’s typically produce research that is4 to 5 points below average. (The F-testis significant at the .001 level.)

THE CORRELATES OF QUALITY

What is the actual relationship betweenthe technical quality of the audiencestudies and the various study character-istics expected to affect study quality?In answering this, our first step wasto examine the empirical relationship oftechnical quality with each studycharacteristic. Next, since these studycharacteristics are themselves empiri-cally interrelated, it was important bothto isolate the unique impact of eachcharacteristic controlling for theinfluence of the others, and to obtain anestimate of their joint, overall impacton quality.

We calculated the average quality of thestudies within each category of thepredictor variables. Then we subtractedthe average quality for all categoriescombined (15.40, with a standarddeviation of 8.45).8 Table 8 shows theresulting deviations from the overallmean.

First, it is evident that the investi-gator’s prior survey research experiencehas virtually no bearing on the qualityof his or her study. The average qualityof the studies conducted by highlyexperienced investigators (more than nineprevious studies) and by those withoutprior survey research experience is lessthan one point above average, whileinvestigators with moderate experience(one to nine studies) performed slightlybelow average research (-1.49). An F-testfor intergroup differences fails to meeteven the .05 level of statistical signif-icance.9

Study budget is also strongly correlatedwithquality. Audience researchconducted with less than $350 is morethan 5 points below standard, whileresearch performed with budgets of morethan $1,650 is 6 points above the mean(F-test significant at .001).

The factors related to the institutionalsetting predict variations in the qualitymeasure as well. Indeed, in this samplethe best predictor of all the variablesis investigator profession: studiesconducted by social scientists scorenearly a full standard deviation aboveaverage (7.01), while research carriedout by arts management personnel isthree-quarters of a standard deviation(6.40) below average. The nature of theorganization also makes a difference, butan organization’s prior experience withsurvey research does not. Investigatorsaffiliated with academic institutionsand private research firms generatestudies 2 and 4 points above average,respectively, while those in arts organi-zations produce research 5 points belowaverage. The quality of inquiriesconducted by organizations with extensiveexperience, however, is a statisticallyinsignificant 3 points above the qualityof research by moderately experiencedorganizations and only a single pointabove the studies of organizations withno prior experience. Finally, outsideresearch is clearly of higher qualitythan in-house studies. The mean qualityof the former is more than ~0Pointsgreater than of the latter.

In summary, then, by technical standardsthe best research in this sample is.produced by individuals with Ph.D.’s orcomparable degrees who are socialscientists affiliated with privateresearch firms or academic institutions.

Since the predictor factors are highlycorrelated among themselves, it isnecessary to examine their simultaneousimpact on quality if we are to isolatethe importance of each. For instance,both budget and type of organizationstrongly predict research quality butthese variables are also highly relatedto one another. The median budget ofstudies conducted in private firms,

46

Table 8 Deviation from Average Study Qualityby Investigator Background, Resources,and Institutional Setting

Study Characteristic

Investigator Research Background

Investigator experience

DeviationI (N)

More than 9 studies1-9 studies0 studies

Resources

Budget2

More than $1649$350-1649Less than $350

0.64 (23)-1.49 (23)

0.75 (36)

Study Characteristic

Investigator degree2

Deviation (N)

6.29 (23)-0.02 (21)-5.58 (26)

Institutional Setting

Investigator profession2

Social scienceOther research relatedMarketingArts

Organization type2

Private firmAcademicArts

7.01 (15)4.13 (25)

-0.66 (15)-6.40 (31)

4.13 (25)1.99 (26)

-5.02 (32)

Other advancedPh.D.MBAMABA

Organization experience

More than 12 studies1-12 studies0 studies

6.854 .ii0.74

-5.00-3.76

1.22-1.72

0.25

Organization affiliation2

External researchInternal research

(9)(27)(7)(19)(22)

(37)(37)

iDeviation from the overall mean 2F-test for intergroup differences issignificant at the .001 level.

academic institutions, and arts organi-zations are $6~260, $750, and $253respectively. We cannot tell fron~-h~efigures reported in Table 8 whetherbudget, type of conducting organization,or some combination of both accounts forthe variation in ,~ality.

To answer this question, we apply thestatistical technique of multipleregression analysis, which enables us toinspect the relationship between researchquality and any single predictor variable,while holding all other predictorvariables constant.

The predictor variables are entered intoa regression equation with quality as thedependent variable. Investigator degreeis entered in a dichotomized form, withthose holding a Ph.D. or related degreejoined in one category, and those withoutsuch degrees grouped in the other. Thelogarith~$c transformation of the budgetis used,±~ and investigator profession andorganization experience are entered assets of dummy (dichotomous) variables.Since investigator and organization exper-ience exhibited insignificant, zero-orderassociations with quality, they areexcluded from the analysis. Because ofthe high correlation between two othervariables--organizational affiliation andtype of organization--the less powerfulpredictor of the two, organizationalaffiliation, is also deleted.

The correlations and the regressioncoefficients of predictor variables withstudy quality are displayed in Table 9.The correlations are consistent with thepatterns seen in Table 8, but the stand-ardized regression (beta) coefficientsreveal that several of the predictorvariables have little impact on qualityonce other variables are controlled. Forexample, the substantial simple correl-ation of .48 for investigator degree isreduced to a beta value of -.02 once theconfounding effects of other variablesare removed. This means that whether aninvestigator holds a Ph.D. or comparabledegree has no direct independent impacton study quality. Rather, the highcorrelation resulted from the fact thatstudy directors with Ph.D.’s frequentlywere in the social sciences or otherresearch-related professions and had highbudgets with which to work.

The association between budget and qualityremains very high even after controllingfor the other bariables. The beta value of

~. The beta-coefficie~ts~f0r the £hree dummy"variables of investigator profession are

" olall statistically significant and rangefrom .19 for those in marketing to .28for social scientists and .39 for thosein other research-related disciplines.These beta coefficients signify that,other factors held constant, investigatorswho were not arts professionals generatedtechnically better research. Finally,although the simple correlations oforganization types are substantial, themore important beta coefficients are not.The beta value is -.08 for private firmsand .14 for academic institutions, neitherof which approaches statistical signif-icance.

Thus, although a number of factors areempirically associated with higher qualitystudies, it is evident that only two werefound to have a substantial directindependent effect: budget and theprofession of the study director. More-over, with only a little assistance fromthe other variables these two explain63 percent of the variance in studyquality. (Variance explained is derivedby squaring the multiple correlationcoefficient.) This means that we wereable to predict audience study qualityin this sample with considerable precision.

With certain caveats we will note in amoment, the unstandardized regressioncoefficients can be used to predict thelikely quality of a proposed audiencestudy. If the study were allocatedvirtually no budget and placed in thehands of an investigator primarily identi-fied with the arts, a quality index ofapproximately 6.2 could be expected: thisis more than a full standard deviation(8.3 points) below the average qualitylevel for all the studies. An investi-gator with a Ph.D. or related degreewould not improve quality, but increasingthe budget wculd have a dramatic impact.By this model, expansion of the budgetfrom $0 to $i,000 would add 5.8 points tothe score. (It would require an add-itional $i0,000 to bolster the scoreanother 5.8 points.) Employing a market-ing analyst as primary investigator meansan additional 4.0 points; a socialscientist adds 5.8 points: and a memberof a research-related profession (thesewere urban planning, architecture,engineering and applied mathematics, andpublic opinion polling) increases thescore 7.1 points. Whether the study isassigned to an investigator located inan arts organization, private firm, or

.63 exceeds that for any other variable and academic institution makes very little

indicates that one can best predict thedifference, though 2.4 points might be

quality of an arts audience study if oneadded if the academic setting is selected.

knows what funds were available to itsdirector.

48

Table 9 Simple Correlations and Regression Coefficientsof Audience Study Quality with InvestigatorBackground, Resources, and Institutional Setting

Study Characteristic rI beta B F163 P

Investigator Background

Investigator degree:Ph.D. or related .497 -.016 -0.26 0.02 n.s.

Resources

Log of budget .699 .627 5.83 41.92 4.001

Institutional Setting

Investigator profession

social science .284 .267 5.81 6.12 4.05

other related .399 .390 7.08 12.50 4.001

marketing -.061 .191 4.01 4.60 4.05

Organization type

private firm .315 -.082 -1.61 0.61 n.s.

academic institution .230 .138 2.38 1.82 n.s.

Constant

Multiple CorrelationCoefficient (R)

R-Squared

6.21

0.794 17.912 4.001

iKey: r=simple correlation; beta=standardized regression coefficent;F=F-test value (i and 63 degrees offreedom); p=statistical probabilitylevel.

0.631

(7O)

2F-value with 6 and 63 degrees offreedom.

Thus, if the studies reviewed here are PREDICTI~G.~HE ~T~L~[T~.I:~5~~-~T-S AUDIENCEtypical, expanding the budget from $0 to . ~STUDIES$I,000, selecting a marketing analys~ . -rather than an arts professi6nal,quartering the study in an academicinstitution rather than an arts organi-zation would increase expected quality byover 12 points to a total of 18.4. Onthe basis of these studies, one wouldpredict that if a social scientist werechosen in place of the marketing analyst,the score would rise to 20.2 and if amember of a related research professionchosen instead, the increment would bestill more, over 15 points for a total of21.5.

Of course these figures, a result ofmanipulating the data of the eighty-sixaudience studies, represent tendencies,not hard and fast laws. For example,some arts organizations have producedtechnically better studies than someacademically based researchers. Also,these figures rest on the assumptionsthat the eighty-six studies are repre-sentative of arts audience studies ingeneral and that the associations foundare genuine and do not reflect some otherset of underlying factors that influenceboth the predictor variables and researchquality. We believe both of theseassumptions are reasonable, but we areunable to prove either with our data.Finally, even if the relationships foundhave existed in the past, they will notautomatically continue to exist in thefuture. For example, if research userswere to become much more sophisticatedand demanding about research methodology,the technical quality of studies mightbecome less dependent upon the professionof the study director or the nature ofthe conducting organization. We do notsuggest that our findings be appliedsystematically to every research-planningdecision, rather they are a description ofthe factors affecting the quality ofresearch that has been done in the pastsix years. They should be seen assuggestive guidelines only.

Clearly, however, arts audience researchvaries enormously in its technical qual-ity, and the evidence presented heresuggests that much of this variation isa direct consequence of two elements ofthe research process-ithe resourcesavailable for the study’s execution andthe professional identity of the principalinvestigator. The other elements in ourhypothesis seem to have little immediateimpact on the quality of the finalresearch product.

Through an assessment of the availableliterature and through informal discus-sions with thirty individuals involvedin audience research and arts management,we identified ten areas in which theresults of audience research are oftenapplied. These ten areas were aggregatedinto two subgroups, the internal oper-ations of arts organizations and theirexternal relationships. Internal policyquestions included such matters as theevaluation of selection of exhibits orworks to be performed, the developmentof educational programs, and the estab-lishment of ticket prices and of hours orperformance times. External policyissues had to do with planning publicrelations campaigns, designing strategiesfor approaching funding sources, anddeveloping or evaluating audience expan-sion programs. The respondent was askedto rate the actual usefulness of his orher study in each of the ten policy areas.An internal utility scale was created bysumming the ratings of seven internalitems, and an external utility scale wascreated from the sum of the ratings forthree external items-12

A high or low rating on these scales canbe illustrated by referring to the twostudies used earlier to exhibit the mean-ing of the quality index. The survey ofvisitors to an art museum rated nearly onestandard deviation above average in over-all utility (assessed by combining the twoutility measures). This survey proved ofhigh value to the museum for its publicrelations efforts, development ofstrategies for recruiting new visitors,the assessment of an arts developmentplan, the evaluation of the drawing powerof a particular exhibit, and the develop-ment of educational materials related tothe museum. In contrast, the survey of anonprofit theatre audience rated one-halfstandard deviation below average inutility. The only area of high’appli-cation was in the theatre’s audiencedevelopment plans.

We considered the possibility that ratingswould be biased by the respondent’srelationship to the research and to itsapplication. In half the cases (54percen~ the respondent reported that heor she was the person "primarily concernedwith managerial or policy applications of

50

the study’s findings," and half (55 - " is ~ide~pre-a~-a~6~-~tC~:fe~. e~a~ policypercent) reported that they were "prin-. ..... - makers-over thereliability of appliedcipally involved in making the.decii~on ..... social research(Caplan, 1976) suggeststo finance or fund the audiencei~t~dyl~ that these-users, at least, are highlyA comparison of the average internal andexternal utility ratings of these groupsindicates that their assessments do notsubstantially differ. Directors involvedin applying results are slightly morelikely to note utility than are otherinvestigators (1.30 and 0.29 pointdifferences for internal and externaldimensions, respectively), but, contraryto expectations, funders are slightlyless likely to provide a high rating thannonfunders (-0.66 and -0.69 point differ-ences). Since none of the observeddiscrepancies approaches statisticalsignificance, we assume that these factorsdo not substantially bias the utilityratings.

Students of social research have ident-ified a number of factors that affectwhether study results are applied,although few of their hypotheses havebeen subjected to empirical test. Ingeneral, these factors have to do withsuch general concerns as: the charact-eristics of the study and investigator,such as study quality and substantiveconclusions, investigator reputation, andproject resources: the characteristicsassociated with the potential user, suchas the user’s attitude toward andexperience with social research and thepolitical environment into which theresearch is received: and the features ofthe investigator-user interaction,including the study’s timeliness, thedegree of cooperation in the design andexecution of the study, and the means bywhich study results are co~m!nicated(Caro, 1971: Rossi and Williams, 1972:Weiss, 1972, 1977: Caplan et al., 1975:Cohen and Garet0 1975: van de Vall et al.,1976: Rein and White, 1977).

sensitive to the issue of researchquality. A study by Weiss-and Bacuvalas(1977), in.which 155 federal, state, andlocal mental health officials were askedto rate brief descriptions of actualresearch studies, found that of fivestudy characteristics evaluated researchquality was the best predictor of will-ingness to consider the findings inmaking relevant decisions. On the otherhand, Patton et al. (1977), in intensivecase studies of twenty evaluations ofhealth programs, concluded that method-ological rigor played a very minor rolein determining the extent to whichevaluation results were utilized.

In isolating the impact of quality,however, it is important to separate thedirect impact of quality itself from thejoint effect of some underlying factor onboth quality and utility. One correlateof quality that may also affect utilityis the nature of the organization con-ducting a study. Although outsideinvestigators may produce research thatis higher quality than that conducted bytheir in-house counterparts, van de Valland his colleagues have argued thatresearch done by insiders is more likelyto be used (van de Vall, 1975: van deVall et al.0 1976). Consistent with thisthesis is Caplan’s (1976) finding thattop federal officials make extremelydisproportionate use of research conductedwithin their own agencies. While artsaudience studies differ from the kind ofapplied social research that has been thesubject of these previous studies, wefelt it was important to look at therelationship between utility and qualityand, also, between utility and thecorrelates of quality.

We are primarily concerned in thischapter with only one of these factors--technical quality of the research--andwe expected that high quality researchshould be more useful than research oflesser merit. The quality of evaluationresearch, for instance, has been shown toinfluence whether the program underevaluation is concluded to be a successor failure. Reliance on faulty studiesmay lead to fundamentally misdirectedpolicy decisions (Mann, 1972: Yin andYates, 1975: Gordon and Morse, 1975: Yinet al., 1976).

Relatively little research, however, hastested the assumption that high qualityresearch is applied more widely than poorresearch. Evidence that skepticism

The average ratings for usefulness ofaudience studies are displayed in Tablei0. The most notable finding is thatnothing we have measured, neither qualitynor its correlates--investigator back-ground, resources, and institutionalsetting--has any substantial impact onresearch utility, at least as perceivedby study directors. Although somedifferences are apparent for organizationexperience, organization type, andresearch quality, none of these approacheseJen a minimum level of statisticalsignificance. C~ntrary to expectations,the relationship between utility andquality is small and inconsistent. Highquality research has an internal utility --rating 0.13 below average, medium-qualityresearch 0.66 above average, and

51

Table 10

Study Characteristic

Deviation from Average Audience Study Internaland External Utility by Investigator Background,Resources, Institutional Setting and Quality

DeviationIStudy Characteristic

Deviation1

Internal (N) External (N)

Investigator Research Background

Investigator experience

More than 9 studies -0.31 (21) -0.49 (21)

1-9 studies 0.24 (19) 0.77 (19)

0 studies 0.i0 (20) -0.03 (20)

Investigator degree

Other advancedPh.D.MBAMABA

0.46 (8) 0.99 (8)0.99 (20) -0.19 (21)0.17 (6) 0.41 (6)

-0.66 (16) 0.ii (17)-0.66 (20) -0.41 (20)

Resources

Budget

More than $1649$350-1649

-0.42 (20) -0.00-0.44 (16) -0.13

(20)(17)

Institutional Setting

Investigator profession

Social science 0.75 (8) 0.50 (9)

Other related -0.25 (24) -0.39 (24)

Marketing -0.44 (ii) 0.04 (ii)

Arts 0.17 (28) 0.17 (29)

Organization type

Private firm 1.16 (19) 0.44 (19)

Academic -1.45 (23) -0.55 (23)0.55 (27) 0.20 (27)

Quality

Quality index (points)

High (20-37) -0.13 (21) 0.14

Medium (11-19) 0.66 (25) 0.04-0.51 (26) -0.15

Previous organization experience

More than 12 studies -1.991-12 studies -0.03

No previous studies 1.01

Organization affiliation

Internal research 0.27External research -0.26

(21)(25)(27)

iDeviation from the overall mean: internal andexternal utility.

2(N) = Number of studies.

(13)(12)(26)

(31)(32)

-0.910.720.13

(13)(12)(26)

(31)’

low-quality work 0.51 below average. Ourdata indicate that there is extensive useof audience research in decision-rdak~n~that it varies from study to study butthat none of the factors considered hereinfluences the extent to which researchis applied.

Although the bivariate relationshipsbetween the utility measures andpredictor variables are largely insub-stantial, it is possible that three-variable or higher order interactioneffects may be present. Among the mostlikely candidates is an interactionbetween investigator experience andorganizational affiliation. It can beargued that the effect of investigatorexperience on utility will be more pro-nounced if the research is internallybased than if it is conducted outside thearts organization. When the research isinternally based, an investigator withprior survey experience is likely todesign a study that is more responsive tothe specific policy conditions andproblems of the arts organization. Whenthe research is externally based, however,the prior experience of an investigatoris less likely to result in such specialsensitivity.

Though the relatively small number ofcases on which the statistics are basedrenders any conclusions highly tentative,the patterns are consistent with expecta-tions. Among studies housed within artsorganizations, investigators with atleast some prior survey experience producestudies which are on average 1.87 pointshigher in internal utility and 1.32points higher in external utility thanthose studies carried out by inexperiencedanalysts: the corresponding correlationsare .320 and .329 (F-test significant atthe .05 level in both cases). By contrastinvestigator experience actually has amodest negative effect on utility when theresearch is housed outside the artsorganization. The difference between theresearch of experienced and inexperiencedinvestigators is -1.48 points for internalutility and -1.05 points for external

Utility: the correli~ion-s are; r~pec-¯ tively, -.169 and ~.248 (F-test not~-significant.).. The differences are notlarge, but they do suggest that priorresearch experience only makes for betteruse of the results when the researcher ison the staff of the arts organization.

But we are still left with a puzzle. Ourindependent variables enable us to predictthe technical quality of arts audienceresearch with an unusually high degree ofaccuracy. But neither research quality,the common-sense explanation, nor any ofthe underlying variables that predictedtechnical quality so well, seem to havea major effect on whether researchfindings are applied. In contrast toexplaining 63 percent of variance inquality, we can predict only 9 percent ofthe variance in internal utility and 6percent in external utility.

To some extent, the absence of anassociation between quality and utilitymay be a product of the nature of artsaudience research and arts policy.Research-based policy in such areas aseducation and health has a long traditionand is often carried out at the federallevel. Policy makers are in a position todraw on the best and to disregard theworst. By contrast, arts policy is youngand largely decentralized. Most ofresearch that we studied was performed bylocal institutions, with few resourcesand little cumulative experience.13 Itmay be that many of those who would useaudience studies are not sufficientlyaware of research standards to use themcritically and selectively.

Even so, this explanation is not initself satisfactory. The extent towhich studies are applied varies sharplyfrom case to case and something must becausing this variation. In the absence ofclear answers related to quality or itscorrelates, we conducted open-endedinterviews with individuals who had eitherdirected arts audience studies or beenresponsible for applying their results.The next chapter reports our findings.

53

i. Quality measures based on the invest-igator questionnaire information could besomewhat inflated, since there may be atendency for investigators to reportgreater conformity to the canons ofscientific inquiry than occurred inpractice. By contrast, quality measuresbased on our own report assessment maysomewhat underestimate quality, since thefailure of the report to mention apreferable methodological feature iscoded as if absent from the study.

2. The internal and external validityitems were the following (a study wasscored as high quality on an item if itincluded the procedure described):

Investiqator’s questionnaire internalvalidity: survey pretested: trainedfield staff: survey administrationdirectly supervised: survey measuresbased on measures used in previousstudies: bivariate statistics used:tables with more than two variablesused: multiple regression and relatedtechniques employed: other multivariatetechniques utilized: computer-basedanalysis.

Report assessment internal validity:procedures or instrument pretested:trained research staff: conventionalmeasurement techniques employed: previousresearch discussed or used: scalingtechniques employed: visitors distin-guished from visits: bivariate analysis:table analysis: multivariate analysis:valid linkage between data and conclu-sions.

Investiqator’s questionnaire externalvalidity: some sampling procedure u~ed:sample size of at least 500: responserate of at least 60 percent: width ofconfidence intervals a consideration inestablishing sample size: populationheterogeneity a consideration in estab-lishing sample size: response biasassessed: weighting used for responsebias, sample frame, or both: tests ofstatistical inference used: confidenceintervals established: analysis ofvariance employed.

Report assessment external validity:sample and/or population clearly defined:sample definition appropriate: randomsample principles employed: sample biaschecked: respondent representativenesschecked: tests of statistical inferenceused: weighting used as a result of

sample design:, generalizability of find-ings described.

3. Six additional items were added to theten internal and eight external validityitems in forming the quality scale basedon the report assessment data. Theseitems were: research and policy issuesconceptualized: research design described;implications of study results discussed;specific policy reco~endations offered:nontechnical sunm~ry of results included:results compared with those of othersurveys.

4. The score of the audience studies onthe overall quality scale ranges from 0to 37, with a median between 15 and 16.The mean is 15.40 and the standarddeviation is 8.45.

5. The highest earned degree is coded asfollows: (i) high school diploma;(2) college B.A. or B.S.; (3) M.A., Ed.M.:(4) M.B.A., D.B.A. (professional businessdegrees): (5) Ph.D, Ed.D.

6. The investigators were asked in thesurvey: "At the time of the study, howmany previous audience studies or othersurveys had the director participated inor directed?"

7. The investigators were asked: "Atthe time of the study . . . how muchprior experience had the conductingorganization had with (previous audiencestudies or other surveys)?"

8. For example, if studies conducted bypeople with brown eyes had an averagequality of 20.00 and those conducted bypeople with blue eyes had an averagequality of i0.00, the value of brown eyeswould be 20.00-15.40 or +4.60, and thevalue of blue eyes would be 10.00-15.40,or -5.40.

9. An F-test indicates how likely it isthat an observed intergroup differencecould occur by chance alone rather thanas a result of a social process. If anF-test is significant at the .I0 level,for instance, there is a i0 percentlikelihood that the differences observedin the quality of two g~oups of studiesreflects a chance occurrence and it doesnot indicate that the two groups actuallydiffer in their quality. A researcher,then, would generally argue that theobserved difference was not substantialenough to signify a tr~e difference. On

54

the other hand, if the F-test is signifi- i2~ .Each itemwas ra~ed*Onthree~pointcant at the .01 level, there is only a-one . =scale (l=not useful., 2=somewhat useful,percent chance that the d.iffer~nce..be~.~n- _~h±ghly.~s~ful).. The question was asthe groups is the product of a chance follows (the mean and standard deviationoutcome, and the researcher is more con-fident that the difference reflects areal social process.

10. The importance of the internal-external distinction in research locationis further corroborated by a separateanalysis of the externally conductedresearch alone. Studies vary in thedegree of cooperation between the invest-igator and the arts institution whoseaudience was the subject of the study.If external housing of research isimportant for producing high quality, itcan be reasoned that the highest qualityexternal studies should be those con-ductedby investigators with greatestindependence from the subject institution.This possibility can be examined bydividing the externally conducted studiesinto three categories: (i) no cooperation(respondents characterized their study asone with "no consultation in the designand analysis of the study, all decisionsmade by conducting organization"):(2) moderate cooperation ("subject insti-tution formally reviewed study design andanalysis, but most study decisions madeby conducting organization"): (3) strongcooperation ("subject institution hadapproximately equal voice in studydesignand analysis" or "determined most of thestudy design and analysis"). As antici-pated, the mean quality of the nocooperation studies (n=14) is 1.88 pointsabove the average external study quality(which itself is 2.76 points above theoverall average): the moderate cooperationstudies (n=17) have an average qualityidentical to that of all external studies:and the strong cooperation (n=6) studiesare 4.39 points below the externalaverage. Thus, the critical advantage ofexternal research housing for qualityappears to be that the investigator isfreed of nonscientific constraints fromthe institution that is the subject ofthe study.

ii. The logarithm of the project budgetis used on the assumption that the mar-ginal utility of each additional dollardeclines as the total budget rises.

for the rating of each item appears inparentheses): "To what extent were the(audience) study’s findings actuallyutilized? Please rate the . . utilityof the study for each of th~ followingareas:

Internal Utility

(1) select exhibits or works to be per-formed (1.62: 0.90)(2) evaluate exhibits, performances,programs (1.77: 0.94)(3) develop educational or informativematerials (1.63: 0.83)(4) decide on hours and/or performancetimes (1.48: 0.83)(5) decide on admission or ticket prices(1.52: 0.91)(6) decide on organization management or

personnel (1.28: 0.74)(7) initiate or evaluate arts development

plan (1.76: 0.92)

External Utility

(I) promote public relations (1.96: 0.84)(2) gain or maintain support from fundingsources (1.59: 0.77)(3) develop or evaluate audience expansionstrategies (2.20: 0.89)

The mean and standard deviation of theinternal utility scale are 13.63 and3.81: for the external utility scalethese values are 5.72 and 1.96. The twoscales exhibit relatively high internalconsistency in that there is a markedtendency for a high rating on one of thescale items to be associated with a highrating on the other scale items. The 21item-to-item correlations among theinternal utility scale items range from.22 to .77 and they average .44: therange for the 3 external utility item-to-item correlations is .35 to .51, withan average of .43.

13. In fact, research on museum visitors,which is part of a tradition dating backto the work of Robinson in the 1920s, wasfound to be significantly more highlyutilized than were studies of performingarts audiences.

55

CHAPTER 3 - ¯ ~ --

ORGANIZATIONAL FACTORS AFFECTING ~SEARCH UTILITY

TO better understand the subtle insti-tutional processes by which audienceresearch is put to use, we haveintensively examined twenty-five audiencestudies. These studies included all ofthose in our possession that had beenconducted between 1974 and 1977 in theNew England and Middle Atlantic regions.Among the studies were surveys concernedwith economic impact, general planning,specific planning, exhibit effectiveness,and members or subscribers. Elevenmuseum studies were included (six artmuseums, two history museums, one sciencemuseum, and two other museums), as wereten performing arts organization studies(five theatres, two classical musicorganizations, one opera, one ballet, andone other). There were also two cross-sectional studies and two surveys of thoseattending a number of different artsevents. In each instance we attempted tointerview both the study director and theperson most likely to have been in aposition to use the research results.However, in eight instances either thestudy director was the key user or inter-views with only one of the two individualscould be obtained. Forty-two semi-struc-tured interviews were completed: theyaveraged forty minutes in length andranged from twenty to ninety minutes-I

As additional background material,unstructured interviews were conductedwith twenty-five other individuals whohad commissioned, directed° or attemptedto use the results of audience research.

Our interviews with the directors andusers of a wide range of studies revealedthat the conventional view of the deci-sion-making process provides a poor guideas to what really happens when arts organ-izations sponsor audience studies. Onemight expect that research is undertakento help solve a specific problem.

enters the decision-making process. And,finally, we discuss the factors that arecritical in the use of audience-studyresults.

THE PURPOSES OF AUDIENCE RESEARCH

Not one of the twenty-five studies forwhich directors and/or users were inter-viewed was undertaken primarily togather information necessary for aspecific managerial decision. Instead,they were begun because of factors suchas the need for political leverage, theappearance of an unexpected opportunityto have a free study conducted, and avariety of other concerns only indirectlyrelated to specific organization decisions.While most managers exhibited a lively

¯ curiosity that influenced the content ofthe survey questions, the need for datafor specific decisions was never astudy’s raison d’etre.

Political factors

The most frequently cited reason forundertaking an audience study was poli-tics, prominently mentioned for ten ofthe twenty-five studies. Politicalpurposes included acquiring evidenceuseful in seeking funding, gainingleverage in internal policy debates, andappeasing members of the organization’sboard of directors or other influentials.

The initiation of research for the sake

However, the researchers and arts managers of seeking outside financing is illus-who shared their experiences with us trated in the case of one study undertakenportrayed a different process. Theiraccounts explained the perplexing lack ofconnection between research technicalquality and utility. Thev also suggestlessons for those who would undertakeaudience research themselves. In thischapter, we describe the purposes forwhich the research was initiated. Then,we illustrate the varied ways in whichaudience research was applied. Third,we explain the ways in which research

to document public support for a newperforming arts facility. Said the studydirector: "A committee (of bankers andbusinessmen) set about to raise money toget (the local government) to take overthe theatre for the county once it wasrenovated. The study was a spin-off ofthat effort .... It was done to provethat there was a market and to gainadditional support to get the county toapprove and accept a building." In

56

~-_ ~ -- ;- : ...... ~.< ~i~-_-:. -another instance, an economic impact study . dane in response to pgessure f±0m-influ-was done to illustrate the importance of " ential membership com~ztees or membersa beleaguered theatre district, to ar[~i;}:~i~-. ~. o-f~boards of. directors. One inquiry wasurban economy. The city was ready to actand "the research had clout because itdocumented the obvious." In yet anothercase, a cross-sectional survey was com-missioned by a municipal government todocument an existing arts council’sfailure to meet local art needs. Thesurvey results contributed to the re-signation of the old council and thecreation of a new one. Finally, onearts counci! conducted a study essen-tially for the purpose of announcing itspresence and increasing its scope ofoperations.

Other research was commissioned for usein internal debate. Individuals neededadditional ammunition for their positionsand were confident that a research studywould support their cause. Though thestudy instigator may have been open topersuasion, the primary motive was tocompile data for a position rather thanto resolve an issue. One theatre manager,for instance, in explaining his reasonsfor surveying the audience of a summerdrama festival immediately after becomingmanager, stated: "In the summer, (thetheatre) did seven shows in rolling rep,which I think is insane itself, and (thetheatre was) doing about 50_percentbusiness .... I had the feeling that (thetheatre) should be delivering a morepopular product, and the survey helpeddocument this. The next year we providedmore popular plays and got 90 percentbusiness." Being new to the particularjob, this experienced arts manager neededto suggest the value of an alternativepolicy before instituting a controversialchange, and he (correctly) anticipatedthat a survey would support his ownpreference for more popular fare. Sim-ilarly, a new director for a rathertraditional museum saw in a wide-rangingmembership study a fulcrum for change:"I had been at the (museum) a little overa year as director and felt it wasimportant to see how we appeared to ourmajor constituency, the membership. Wehad been in business for a long whileand certain things continued to be donebecause they had always been done thatway, without our knowing what our memberswanted." Still another museum visitorsurvey was initiated for evidence tocombat pressure for an admission fee.The converse purpose motivated one studyof another arts facility: a survey wasundertaken to justify the institution ofan admission fee to a skeptical statefunding agency.

Finally, audience studies are occasionally

undertaken of a performing arts institu-tion because of a membership committee’sconcern with what it thought was anoverly "elite" audience. The study’sfindings, however, were lazgely ignoredby management. This was also the outcomeof study initiated at the behest of achairperson of a museum’s membershipcommittee. The administration of themuseum regarded the survey questionnaireas "silly" and the disappointed studydirector concluded that her study "wasjust an exercise." She observed: "Igot a lot of experience and a lot offrustration. I didn’t know who to tellthe results to or who would listen tome. "

Opportunity

The second most common general motivationfor undertaking audience research was theappearance of an unexpected and relativelyfree opportunity to undertake a study.This was a principal consideration ineight of the twenty-five cases weexamined. Arts managers took advantageof occasions for inexpensive research tosatisfy a kind of free-floating curiosity.Volunteer labor, the availability ofoutside funding, or both were usually thecatalyst. In one instance, museum admin-istrators were in the process of preparinga grant application for federal funds.It was a near certainty that the museumwould receive the grant, and at the lastminute an affiliated researcher revisedthe proposal to include a visitor survey.At another museum, when questioned aboutthe timing of a visitor study, thedirector said: "Simple, funds becameavailable .... (A federal agency) madefunding available for the purpose so (themuseum) used the occasion to do a study."Volunteer outside labor was the motivatingfactor in other instances. In one case,a county-wide attenders/nonattendersstudy was included in a larger audiencedevelopment program only after a univer-sity professor stepped forward, suggestedthe study, and prqmised to design thequestionnaire and provide student labor.A theatre study was undertaken when abusiness school student with an outsidegrant took it on as a su!mner job. Theinitiative for such studies often restedwith a single individual prepared to takeadvantage of an opportune situation. Oneresearcher, hired as a consultant foroverall planning, defined his role toinclude carrying out a visitor study.

57

The museum "didn’t so much want the study decisions as physical"planning; marketing,done as they, kicking and screaming, ~ - programming, or further research.-

grudgingly allowed me to do~ it._" . ~.%:~b:~i; ~ tWentyrsix (~34 percent) were basicallypolitical, related either to internal

General concerns

The third major reason for undertakingaudience research, cited as most impor-tant in six of the twenty-five studies,was a vague sense of concern, a feelingon the part of managers that they areworking in a vacuum and that certainkinds of background information, usuallynot clearly specified, would be good tohave. In several cases, for instance,museums were about to undertake long-range physical planning and felt thatthey needed "some input" from visitorsor wanted "to get some idea about theaudience." One outside researcher com-plained that a museum representativeapproached him with "vague, vacuousquestions." Another said of his clientson an arts council, "they vaguely sug-gested doing a survey of general goals."A museum gallery director spoke of thedifficulty he had in fixing goals for astudy of his visitors, and an in-houseresearch director for a performing artsinstitution described his study as a"first feeble attempt at research ....Some of it was stabbing in the dark."The studies were inspired by a genuinedesire to learn more and a sense that solittle was known that any increment inknowledge would be worthwhile.

politics or to external lobbying and fundraising. Instrumental usage was made oftwenty of t ~e twenty-five studies, whilepolitical application was made of eighteenof the studies. Instrumental applicationscan be further divided into physicalplanning, marketing, research, and pro-gramming: political usage can be dividedinto internal and external politics.

Instrumental application

The most frequently mentioned use ofaudience research was for the instrumentalarea of physical planning, cited forfourteen of the studies and representing29 percent of all applications (Table ii).In nearly half of these cases, researchfindings were useful for decisions aboutthe orientation of museum visitors (e.g.,signs, information desks, guide training,brochures) or about visitor conveniences(e.g., special bus services, restaurantfacilities, roadway markings, cleanerwashrooms). Audience research was alsocited as influencing decisions aboutticket and admission prices, performancetimes and museum hours, exhibit labelingand design, exhibit acquisitions policies,and performance sites. More generally,studies were said to have an indirectinfluence on architectural planning andto increase staff concern with visitororientation.

THE IMPACT OF AUDIENCE STUDIES

Despite the many reasons for which thesestudies were undertaken and the wide rangeof their quality, once they were completed,arts managers did use the results exten-sively. The reasons for this apparentparadox--widespread application ofresearch undertaken for diffuse or non-instrumental reasons--will be explainedin the sections that follow. Here wedescribe the range and extent of appli-cations reported.

Participants in all but two of the twenty-five studies mentioned at least oneexample of study impact, and multipleuses were cited in many cases. Ofseventy-seven uses described, fifty-one(or 66 percent) were broadly instrumental°related to such specific organizational

In marketing, the second most importantarea of instrumental application, audienceresearch helped with decisions to changethe target of marketing efforts and tochange the themes of promotional materials.More generally, studies were also givencredit for stimulating institutionalthinking about audience composition,marketing, and audience development.

Surprisingly, in 12 percent of theapplications the directors and usersreported that they used the results inconnection with more research. Sixstudies were used to encourage researchbeyond the institution sponsoring thestudy: three studies aroused enthusiasmfor further research within the sameorganization. Finallh, study results hada direct effect on programming choices oron the thinking of administrators aboutprogramming in 6 percent of the appli-catiOns.

58

Table 11 Frequency of Instrumental and PoliticalApplications of Audience Research Results

ApplicationNumber of StudiesCiting Applicationl

Instrumental--total 20

Physical planning 14

Marketing 12

Research 8

Programming 5

Political--total 18

Internal politics 14

External politics 9

Number ofApplications

Percent of AllApplications

51

22

15

9

5

66.2

28.6

19.5

11.7

6.5

26

17

9

33.8

22.1

11.7

All applications 23 77 i00.0

iMore than one citation was found in some studies.

Political application

Internalpolitical consequences we~ ........cited for fourteen of the twenty-fivestudies, representing 22 percent of alluses mentioned. Such political usesincluded increasing trustee interest,selling administrators on the value ofmarketing, aiding the reorganization ofa local arts council, providing leveragewith parental or affiliated organizations,sparking the withdrawal of some membersto form new institutions, and makingcurators more secure in their positions.Of the application areas described here,internal political uses were the mostoften unexpected by those who had otherpurposes in mind at the time the studywas conceived.

Use in external political areas wasmentioned in nine of the studies andrepresented 12 percent of all instancescited. Audience research results wereused to seek funding from municipal andstate governments and from privateindividuals and concerns. No intervieweesexplicitly indicated the results wereuseful in approaching the federal govern-ment.

It is evident, then, that audienceresearch, whatever the reason it isundertaken, has payoffs for arts organ-izations in a wide range of substantiveareas. Even research that .is poor byorthodox standards of social scienceinquiry played a useful role in the delib-erations of art managers. The conven-tional view of research holds that it ismost powerful when it is mostsophisticated, that good research,designed to address specific problems,is used to make specific decisions aboutthese problems. While this ideal modelmay characterize a few of the studies,for the most part audience research ishighly variable in quality, is rarelydesigned with specific decisions in mind,yet is reported as being highly useful.This could reflect a lack of researchand managerial sophistication among artsadministrators, but we think not. Rather,just as research is not undertaken forthe purposes commonly supposed, researchfindings do not play the role in rationaldecision-making that has been attributedto them. To understand how audienceresearch becomes applied, let us look moreclosely at the ways in which study find-ings have affected arts management.

¯ . THE ROLE OF AUDIENCE¯ RESEARCH FI.NDINGS INARTS MANAGEMENT

The most notable feature of the impact ofresearch findings on arts management isthat it is invariably a marginal one.This is true in several senses. First,arts managers usually have at least someadministrative experience, are oftenaware of the limitations of research, andrely on their own experience and judgmentto assess research conclusions. Researchfindings are used selectively in thecontext of a complex background of pre-viously acquired knowledge and beliefs.For instance, a performing arts managercited an audience study--the technicallimitations of which he was fully aware--as influencing his decision to changepromotional strategies for a series ofpublic performances: "It helped us refocusour promotional efforts in the (outdoordrama series). I’m not totally trustfulof the results, but they did show a largenumber of people heard about the concertsin the community newspapers, which wehadn’t expected, and even if it’s onlyhalf as large as the survey indicated, itis very economical advertising- We’reputting more money into the neighborhoodpress." Studies frequently serve toreinforce preferences already held ordecisions already favored. The resultsof one study, said a theatre manager,"followed exactly what my gut was saying.I just wanted to be sure I was right."

Conversely, when research results contra-dict strongly held positions or views,they are likely to be ignored despitehigh technical quality and clear-cutpolicy implications. Thus, one well-executed museum visitor study hadvirtually no impact even though itcontained implications for museum designand visitor orientation- As one personacquainted with the study recounted:"(The museum staffers) were skeptical,first because they could not believe that(the research director) knew more aboutthe public than they did, and secondbecause they did not feel that knowingabout the public had anything to do withhow the galleries should be handled ....The major criticism of (the researchdirector) was that he was an outsiderwho lacked a depth of knowledge basedon years of experience. He was not

6O

criticized on any specifically method- . ~arginal role in managerial p~ii~-ological grounds: his critics didn’t ..... m~king, its impact is highly indirect asknow what methodology was~" .In some%~i~.~-~." .~ell... Research contributes in-circuitous,instances, studies provided material for often unexpected, ways to the policythose on all sides of a~debate. One some- " process.what cynical research veteran observed ofanother museum study: "I’m a bit jaun-diced against this study, I have to say.People have pulled out of it what theywanted. They picked and chose what theyneeded to support their position. It’sa predictable use."

A second sense in which audience studiesare marginal is that decisions into whichthey enter usually involve competingpriorities. Even when participants takethe accuracy of findings for granted andagree on the implications, differences invalues strongly affect their willingnessto implement the findings. In one typicalinstance, the audience at a performancestrongly preferred an earlier curtaintime, but action on this finding wasthwarted by the need for a tight rehearsalschedule. Similarly, many museum direc-tors and curators balanced their findingson visitor needs against their commitmentto other museum functions. One museumdirector put it this way: "My chiefpurpose is to preserve the collection: mysecondary purpose is to offer programsand services which will maintain publicsupport." Indeed, our interviewees citedmany instances of administrative orcuratorial resistance to research impli-cations that were perceived as implicitlypopulist. Thus, research frequently con-fronts vested interests, making directapplication problematic. As one directorof a performing arts audience study putit, "In general, data step on toes."

Finally, audience research findings aremarginal because they often addressproblems which in organizations withlimited funds and staff, are givenmarginal attention. One performing artsmanager favored a marketing strategysuggested by a study (to arrange a dinnerpackage with a neighboring hotel), butnoted that the "hand-to-mouth" existenceof his organization precluded arrangingfor even such a minimal innovation.Similarly, several individuals in artscouncils felt that other demands on theirtime had prevented them from fullydisseminating the results of audiencestudies they had undertaken. And onefestival director attributed an inabilityto use research results to the precariouseconomic existence of his organization:"One of the restraints on the implemen-tation of new policy was that the festivalis just so poor."

If research results play a largely

In many cases, the studies are used lessto suggest solutions to problems than tocatalyze action on a burning issue or tosymbolize a point of view: "I think thatthe survey results basically gave us adata base to support many of the thingswe had an inclination about already ....But there was nothing cataclysmicallydifferent from what we had expected. Itsimply gave us a statistical base fromwhich to work." In one ~useumwhere alabeling study was undertaken, the spe-cific findings have been largely ignored,but staff people arguing for more labelmaterial often cite the study to bolstertheir position.

In other cases, directors or usersmentioned that study findings foundapplication but were at a loss to assessthe findings’ relative weight in thedecision-making process, again suggestingthat the effects were largely intangible.One sponsor of an internally managed pub-lic opinion poll, the results of whichwere used in a successful lobbying effort,said of the study, at one point: "It wasdefinitely effective in our case and atour level of government." Severalmoments later, however, he thought that:"It is hard to attribute anything directlyto the report. The biggest thing wasimpact--much of what was found was veryobvious, but they never (had done) any-ring about it .... They needed some kindof incentive." The effect was morecatalytic than decisive. Another in-houseresearch director noted that she usedsurvey results mainly to legitimizedecisions already reached.

In several other instances, staff membersof arts organizations assumed the role ofchampions of a survey, using it repeatedlyin arguments over issues involving thepublic. In these cases, data was broughtto bear in the decision-making process,but its use was largely symbolic,representing more generalized commitmentsto such principles as service, bettervisitor orientation, or the value ofmarketing. In these cases the researchwas simply part of a much broader processof discourse and contention over organi-zational values and aims.2

The research process itself is at timesas influential as the study findings.An audience study can serve to focusattention on certain aspects of an artorganization’s management or environment.

61

One researcher felt that a report of his

study of museum labeling wascompletely ignored, but noted-that ~"number of the staff had never thoughtabout the issues I was raising, and mycomments seemed to open their eyes." Inanother museum study, both the museumdirector and the researcher felt thestudy had heightened sensitivity tovisitor concerns. The researcher obser-ved: "I think it has made a generaldifference in how people see things.There is not yet a radical enougheffect .... But the idea of the question-naire has been accepted. That goes on alot now, whenever there is any contro-versy or question to be solved, peoplecirculate questionnaires to get visitoropinions. The idea of feedback fromvisitors has become more important. Evengoing out on the floor and observing andtalking to people has become more impor-tant. The basic change is the idea thatyou can’t sit behind closed doors andpredict visitor reaction, you have to goand find it out."

One of the most important applicationsof audience studies was not in solvingproblems but in finding them. Rationaldecision-making theory would suggestthat organizations monitor their environ-ments, note problems as they arise, andmake decisions accordingly. Research isgenerally seen as a part of the decisionprocess, undertaken to fill gaps ininformation Ineeded to make rationaljudgments on existing or future programs.More often, however, research appeared tohelp organizations scan their environment,to define problem areas where at mostonly vague concerns existed-3

Museum visitor studies were particularlyuseful in this respect. Studies ofvisitors to several museums led tonumerous, easily accomplished changes.Floors were renumbered, new signs posted,information desks installed. Severalperfornting arts institutions found thataudiences preferred different performancetimes and curtain times were changed.Surveys revealing audience socialcompostion sometimes led to greaterpublicity among overrepresented groups,at other times to publicity among under-represented groups. In some cases,statistical findings were less influentialthan longhand comments elicited at theend of survey questionnaires. Criticalassessments of the physical plant weredescribed as particularly useful, sinceorganizations could readily respond tomany of the recommendations. The impor-tance of the problem-signaling functionof audience studies provides a clue tothe lack of relationship between technical

quality and utility~~<In~ormati6h need_ not be precise to place an item-on an¯ organization’ s agenda.

Audience research, then, enters the policyprocess in a number of often unexpected

.and usually indirect ways. Its use isgenerally one of six types:

Problem-solving function. In a few cases,especially in the area of marketing,research is used to guide decisions onspecific issues. Spending for promotionand pricing decisions are typicalexamples.

Problem-findinq function. Frequently,research is used to monitor an organi-zation’s activities and environment.Identifying causes of visitor discontentis a common application.

Reinforcement function. Frequently, studyfindings are used to back up or legitimatepreferences or decisions of arts managers.Reinforcing a decision to alter program-ming would be characteristic.

Attention-focusinq function. Sometimes,even when the results are ignored, theresearch process itself focuses staffattention on some previously slightedissue. The importance of doing researchat all may be established only by thecompletion of an initial research project.

Expressive function. Occasionally, audi-ence studies are used to representsymbolically a commitment to suchprinciples as the importance of marketingor an organization’s responsibility tothe public.

Lobbyinq function. In many cases,research findings are used in efforts topersuade government agencies or otherinstitutions to provide financial assist-ance or otherwise support an arts insti-tution.

Poor research can of course, lead tounsound policies, especially if it isapplied automatically, in textbookfashion, to a decision. But we did notfind this to be the case. Rather thanhelping managers make specific decisions,the audience studies we looked at usuallyserved to reinforce opinions, persuadeoutsiders, or focus attention on somegeneral problem area or set of goals.Tne contribution of research to themanagement appeared to be suggestive orsymbolic rather than definitive, andresearch carried out poorly was as usefulas research that was well-designed andexecuted by orthodox standards. Even inthose few cases where research was

62

brought to bear on relatively specific . outside researCher (cite~ in four studies).questions, managers often had so littleinformation that they felt-anyhowever crude, reduced ambiguity andclarified alternatives.. If poor policy.resulted from poor quality research, itwas not so noted by the management.

FACTORS PROMOTING RESEARCH UTILITY

The research directors and users weinterviewed named a long list of factorsthat helped or hindered in applyingresearch results. In general, studieshad powerful effects in these situations:when their findings confirmed thesuspicions of arts managers: when aninfluential person actively sought touse them; when the authority of outsideresearchers lent legitimacy to theirfindings: and when researchers wereinvolved on a sustained basis in staffdeliberations. Studies failed to make animpact in these cases: when there washigh staff turnover; when organizationslacked the resources to use the findings;when influential individuals were hostileor indifferent to the research: whenresults were reported in a confusingmanner: andwhen report contents wereperceived as trivial or inconclusive.

Those factors which promoted use of theresearch can be grouped as three types:attributes of the study, features of thearts organization applying the results,and the political environment.

"~igh authority was derived from affili-ationwith ~ prestigious university orreputable marketing or public-opinionfirm. In a few instances individuals¯also benefitted from considerablereputations of their own. Authoritativedirectorship of the research ensured thattechnical challenges to the researchfindings would not be raised and ingeneral provided an air of legitimacy tothe research. Thus, one study aimed atlocal public officials gained credibilityfrom the firm’s long-standing trackrecord: "There was no skepticism overthe methods of the study. Most politi-cians were savvy about survey research,since they use it in polling all the time.And the people involved in the study,including myself, were already wellknown .... We were already highly visiblepeople when we came in to do the study."In another instance, a museum adminis-trator turned to a well-establishedmarketing firm for a visitor study aftera previous study had floundered from lackof credibility: "You have to have aprofessional prepare the study, bothbecause only a professional, an outsider,can prepare unbiased questions, and onlya professional knows the techniques fordoing these kinds of studies. Peopleworking in museums will prepare biasedquestions and don’t know how to conduct thestudy." Experienced outside researchersbring not only the needed technical skillsbut also the capacity to effectivelyinterpret the results based on statisticalprocedures. One performing arts managerin a university town turned to the busi-ness school for assistance because "theyhave much more expertise in designingsurvey instruments (and) they could explainto me what a cross-tabulation is, how tounderstand a chi-square."

Study attributes

The most frequently mentioned of the threewas study attributes. It was cited ascontributing to research use in twelve ofthe twenty-five studies (Table 12). Thesingle most important aspect here waswhether the research findings fit withthe preconceptions of the organizationmanagers (mentioned in eight studies).Use was high when the research served toreinforce attitudes. One study directorreported that the trustees of a performingarts festival were initially skepticalabout his study because of the relativelysmall sample, but nonetheless acceptedthe findings because they were expected.

Another attribute of a study whichenhanced its use was the authority of the

The third attribute was the presence ofunexpected results. Surprise findings,while neither confirming nor refutingstrong preconceptions, were importantin a few instances because they drewattention to new problems (cited in twostudies). One study designed to provideammunition for a struggle over admissioncharges found that the museum had apreponderance of first-time visitors anddrew from a broader public than had beenbelieved. The surprising nature of theseincidental findings led the museum toalter its scheduling. In a study ofnonvisitors done by another museum, thesurprise was that nonattenders wereindifferent rather than hostile tomuseums. The unexpected lack of publicantagonism had the effect of increasingmanagerial optimism about the value ofbroader marketing.

63

Table 12Frequency of Factors Cited Affecting Use ofAudience Research Results

~actor Affecting Use Number of Studiesfor Which FactorWas Cited

Factor Affecting UseNumber of Studiesfor Which FactorWas cited

Facilitating Factors

Study attributes--total12

Fit preconceptionsAuthority of outside researchers 2Surprising results

Organizational factors--totalii

Support of influential individuals9

Researcher involved in staff deliberations 4Small institution provided flexibility

for inovation3

Autonomy of department1

External political factors--total4

Politicians needed position legitimization 2Interest groups needed results for 2

lobbying

Inhibiting Factors

Organizational factors--total13

staff turnover broke momentumii

Lack of resources for implementation7

other problems preempted attention3

Lack of interest or hostilityi0

Low priority, uninterested64

Researcher viewed as outsider 4Hostility to public input

Planning

Lack of goalsNo intention to use results

i0

53

Inhibiting Factors (Continued)

Communication and dissemination--total

Results delivered without follow-through

Report confusing, too longResearchers unavailable for follow-

through as time passedReport recipients lacking technical

competenceReport did not reach right peopleconflict between researchers and admini-

stratorsReport never delivered

Report content--total

Findings obvious or trivialNo study of nonattendersOrganization interests changed during

time of studyFindings outdatedToo few questions addressedLack of negative feedback

Study execution--total

Inadequate fundsInadequate timeLack of opportunity for managerial input

Technical features of study--total

Low response rateSmall sampleLack of in-house expertise

10

64

2

8

-. - -Although in only two studies were unex-.. .c~ntained ihthis area.-we do thepected results explicitly cited as a :~--. research and then we disseminate thereason for the study’s ut±lity, other ~~ .... information ~o the areas that would beevidence suggests that theelement ofsurprise n%~y increase the likelihood thata study will be applied. In thirteenstudie~, the results were unexpected bythe researchers and managers: in elevenstudies they were not. All of the formerstudies had impact on some policy area,while four of the latter were deemed tohave virtually no impact.

Personal commitment

The second major set of factors contribut-ing to study utility had to do withcertain features of the organization(identified as important in elevenstudies). The most critical organiza-tional aspect was the commitment of anadministrator in the arts organizationto the research (cited in nine studies).Without such a commitment, research wasoften ignored.4 One administrator, who

served as advocate for an in-house reporttold us: "The only way for these studiesto get used is if someone is personallyinvolved and comnuitted to the data. Youhave to care enough to really push some-thing or it just won’t get used. This istrue of just about everything in themuseum world."

involved in the relevant (nonmarketing)decisions." More typically, however,supportive managers faced considerableresistance. In several cases sympathetichead aclministrators disassociated them-selves from research in order to avoidfurther polarizing divided institutions.

Since studies are rarely designed toprovide immediate information for specificdecisions, their use depends on familiar-ity and a cumulative process of acceptanceand learning. This is most likely tooccur if an in-house researcher isinvolved in staff deliberations on a day-to-day basis. One study director, forinstance, repeatedly discussed data atstaff meetings. For many months no finalreport was written: "I purposely didn’twant to write a final report or have afinal report floating around because thatwould have created closure on the project.I wanted people to feel that there was adata bank there to be used and possiblyadded to if there were more questionsthat needed answers." In another museum,the key administrator placed the officeof the research director next to that ofthe director of education, to ensure theywould frequently encounter one another inthe halls.

Another museum administrator explainedhis role in promoting application of avisitor study: "There is a mandate toimplement the report at all levels. (Thestudy director) has the license to roamaround the place and complain whenevershe sees something being done that goesagainst the findings of the study. Shetries persuasion and happens to be verypersuasive, and I stand behind her witha big stick."

Administrative backing of research usewas especially critical in small insti-tutions. One manager of a theatricalorganization, asked if he faced difficultyin implementing the findings of an in-house study, put the matter succinctly:"No. By virtue of the fact that I wasmanager of the companies, I could dowhatever I wanted to do."

In large institutions, even when keyachninistrators favor use, bureaucraticconflicts and resistance can hamper imple-mentation. In one case the relativeautonomy of a research-oriented departmentwas an aid to application. The marketingdirector explained: "The way the market-ing department works, it’s pretty self-

External factors

The third set of considerations contrib-uting to the use of audience studiesinvolved external political factors(cited as important in four of thetwenty-five cases). A receptive polit-ical climate significantly helped the useof study results. In two cases, localgovernment officials, wanted some furtherrationale for decisions they were alreadyprepared to make. In one, for example,an economic impact study of performingarts institutions in a city was done aspart of a public relations campaign tojustify improved lighting and policeprotection in the theatre district. Cityleaders were sympathetic--an importantcity official had, in fact, been muggedin one institution’s lobby--and welcomeda study with entirely predictable find-ings .~hich bolstered their position. Intwo other cases, lobbying groups quicklycapitalized on results useful to theircampaigns. One study director describedthe use of his study: "The communicationwas largely personal. We talked to keypeople, particularly on the (lobbying)committee and they talked to the

65

legislature.talk to the politicians one.byresearch was never formally prese~te~-~

The report was very limited in distribu-tion, never presented as a main support,only drawn on when it was useful."

FACTORS PREVENTING RESEARCH UTILITY

While many studies were applied exten-sively, others were not. The list ofinhibiting factors was a long one. Itcan be divided into the following areas:organizational factors: lack of interest,planning, communication, and follow-through: report content: study execution:and technical features of study (Table12).

Staff turnover and lack of resources

The financial people would .~simply lacked resources to i~pfe~ent_ recommendations- One museum s£udy was

opposed by that-institution’s educationdepartment because, in the study direc-tor’s words: "It was the attitude that weknow what is right and good to do but wecan’t do it anyway because resources arescarce, so why spend m~ney on this kindof research?" Several arts counciladministrators felt that studies they hadsponsored were inadequately publicizeddue to lack of staff time. Less directly,low salary levels contributed to thedepartures of some staff members who mighthave been instrumental in using studyresults. But perhaps the most criticalscarcity was that of funds to try newprograms. A performing arts institutionadministrator explained: "One of therestraints on the implementation of newpolicy was that (the institution) is justso poor. It was clear that a broad adver-tising campaign should be developed toattract tourists, but (the institution)didn’t have money or staff to do this.Our hands were tied." In several cases,management or financial crises intervenedto the extent that research results werelost in the shuffle.

The problem most frequently cited aspreventing use had less to do with thestudies themselves than with the organi-zations that commissioned them (identi-fied as important for thirteen studies).Of all the organizational factorshampering implementation, staff turnover,endemic to arts organizations, was theprime culprit. The use of research, aswe have seen, involves building and main-taining commitment, and arts institutions,perhaps because they are understaffed,seem to rely more on memory and less onmemoranda than other organizations, staffturnover poses serious problems forresearch use. In the case of studies oftwo performing arts organizations andone museum, administrators most involvedwith research projects left theirinstitutions and, while the findings wereuseful to them in their new positions,the studies had no impact on the institu-tions for which they were designed. Inthe case of two other museum studies, theadministrators who commissioned theresearch took jobs elsewhere, leavingstudy directors to face an indifferent orantagonistic staff. One museum wentthrough several directors within threeyears of a study’s conception. In twoinstances, the reluctance of caretakerstaff to make major decisions duringextensive search periods for new directorsmeant a reluctance to use the visitorstudies.

Another problem was that organizations

Hostility and lack of interest

A second set of factors involved indif-ference or hostility towards research onthe part of staff and management (citedfor ten studies). In some cases theresearch director was distrusted as aninexperienced outsider. One director ofa visitor study was perceived, accordingto a sympathetic governing board member,as "an outsider brought in by thetrustees. If the staff had their way,all outsiders would be dropped, even theoutside auditors: they think they knowall they need to know." similarly, amuseum director who had attempted todisseminate the findings of a study ofhis institution reported: "There aresome senior people in the museum worldwho literally won’t read the report, evenin a very short version. I’m friendlywith some of these people and they havefrankly told me that it is useless andthey won’t look at it. If you want toremain on friendly terms you just have tolaugh it off." Hostility to social sci-ence research also exists. One museumadministrator told us: "I think audiencestudies are absolutely hopeless--they area waste of time and the work force. Wetried here to use the questionnaire-typefor three different seasons.’ We wouldsit somebody down like a stooge to askthem questions, and we used observation,

66

and it was ridiculous. They are no good people, ve{baf di~cus~n~-~i~h articles,for anything at all. I’m just predisposed changes in priorities. Cultural institu-against questionnaires, they’re silly~.-~;-I.~ tions are becoming more important inget ten a week across my desk. The~~ ~f4 ....... people’s iivAs,th~re is more concernlike macaroni and cheese, you can get it ¯ with people’s rights, maybe leisure isanywhere, and the only question is whoseis better." Distrust of outsiders and ofsocial science methods in general is notexclusive to museums. It was cited bypersons involved in theatre and symphonyaudience research as well.

In some cases, particularly art museums,the staff doubted the relevance of publicopinion. Administrators contended that amuseum’s responsibilities to the publichave to be balanced against its duties inthe area of scholarship. This positionwas a source of complaint by one research-er: "There are people in establishedpositions who feel that it is entirelytheir prerogative to run the museum onthe basis of connoisseurship and that thepublic’s desires couldn’t be less rele-vant. They are very sensitive to art-historical standards; connoisseurship isthe religion of curators. They have hada lot of experience with people wantingcircuses for the hoi polloi and they seethat as very threatening to their posi-tions. Even a few who are sympatheticare afraid." One museum official notedthat some curators even refused to allowchairs or benches in their galleriesafter a visitor survey indicated a demandfor seating to combat museum fatiquebecause "they felt that tacky modern fur-niture would distract the visitors fromthe beauty of their .... masterpieces."

The presence of such attitudes did notrender all art museum studies useless byany means. For one thing, resistance topublic input is not universal. Mostinstitutions studied had several staffmembers or administrators sympathetic toresearch and the balance of opinion variedwidely from place to place. A number ofrespondents reported that financial hard-ships were making museum administratorsincreasingly responsive to public desires.As an administrator explained, interestin planning is increasing as a result oftwo pressures: "The first is financialand all the rest can be tied back to this.Financial pressures are facing all cul-tural organizations. Donors and support-ers are demanding a more businesslikeapproach: you are getting greater sophis-tication from everyone from trustees tostaff .... Also the public is becomingmore aware that the museum is a publicinstitution. Pressure comes from thepublic to make services more readilyavailable and indirect pressures are per-ceived by the trustees and others .... Itcomes in the form of pressures from

more important. It is not like the six-ties when blac~ groups applied pressureto museums by direct actio[,: that is notgoing on now. But it’s more like agroundswell--the impetus is internal, itcomes from the trustees and management,but that is just a reflection of thepresent-day world."

Researchers and sponsors managed to createsome enthusiasm for research--or at leasttolerance of it--by avoiding questionsthreatening to particular staff members,by presenting findings without recommen-dations, and by including museum staff inresearch design through soliciting ques-tions and feedback on study plans.

Perhaps more distressing to study direc-tors than hostility was the frequentindifference to their work. One research-er who carried out a visitor study in amuseum (after the director who hired himhad left) complained: "Working in (themuseum) was like working in a vacuum.Nobody cared. There were no obstacles,everybody was friendly and nominallycooperative, but they were very worriedabout the new exhibits and this wastaking up their time and energy .... Ihave no way of knowing if any of theresults were surprising, since the reportwas not read." A director of an in-housemuseum study complained, "If I hadn’tfollowed through, the results would havebeen buried immediately. I had to workhard to get people to even read theirreport." While such lack of interestseemed particularly characteristic ofmuseum adminstrations, it was by no meansrestricted to them. The director of aperforming arts audience study said: "Idon’t know exactly what use was made ofthe research .... The report was sent tothe (membership group) but I never gotany feedback from the board. I also gaveit to (administrators and board members)and said I would like to talk to themabout it, and that was the last I everheard from them. I don’t even know ifthey have ever read the whole report." Aperforming arts organization staff membercommitted to audience research resignedwhen he called a meeting to present thefindings of a study he had commissionedand only one person came. The directorof another performing arts study wasactually unable to find someone in thearts organization, which had undergoneextensive staff turnover, willing to re-ceive the report.

67

Research Planning~-bo~h their~authofs ~n~i~i~±ents .(in o~ly¯ one case was a report no~ prepared). One

A third set of factors detracZing fromI¯-~- university-affiliated researchersaid:

study use is related to planning(ci~’~¯~~ "The ~analyses were done by a graduate

for ten studies). Several researchersstudent working under me .... The student

and users complained about the absence ofclear research goals. A university-baseddirector of a performing arts audiencestudy commented that one "factor inexplaining its lack of utility is that(the study) was not aimed at any specificproblem." similarly, a museum official,discussing a visitor study in effect"donated" to his institution, said, "Therewas a problem in fixing the objectives ofthe study. (The study director) wantedus to state our objectives, but we foundthis difficult to do. The questions hefinally worked out seemed trivial to us."An academic investigator who directed across-sectional study for a local artscouncil noted: "There was a fair amountof interest in doing a survey. The prob-lem was a lack of understanding of what asurvey could do, a lack of proper expec-tations--and this was probably our fault,because it’s important in market researchto establish this first. People didn’treally know what to expect--they thoughtit was a good idea to do a survey andfind out something about the audience,but they have no clear idea about what touse the results for." In a few cases,studies were planned for internal poli-tical reasons rather than for the use ofresults. As mentioned previously, somestudies were performed to placate member-ship committees, and one study wasreportedly undertaken because of a per-sonal friendship between an administratorand a member of the research firm invol-ved. Finally, two studies suffered frombad timing, unavoidable because of theavailability of funds or personnel. Aperforming arts institution was surveyed(as part of a larger effort) just beforemoving into a permanent facility, render-ing some of the data irrelevant. Thepresence of major construction and itsattendant problems complicated the admin-istration of one museum study, pushingstaff energies to the limit and, thoseinvolved speculated, inflating the numberof respondents who expressed disappoint-ment in their visits.

Communication and follow-throuqh

Difficulties in communicating and dissem-inating the study information made up afourth set of factors which diminishedresearch use (reported to be significantin ten studies). Several study reportswere considered too long or confusing by

wrote a long report that was really notthat well written, and then he and acouple of people at (the arts councilsponsoring the research) sent out apamphlet .... For market research to bereally effective, it has to be presentedto small groups who have the opportunityto ask questions and really go over thething. I sort of have the feeling thatthat never happened in this case." Thedirector of another arts council that hadcommissioned an audience study felt thatthe findings would have been more power-ful if the report had been condensed, withfewer statistics. A museum administratorwho received one rather technical reportof a visitor study confessed that, althoughhe was interested in research and carriedthe results around for a while, he foundthe report so boring that he never readit. Two study directors complained aboutla~k of sophistication in their readers."It was apparent that most people (in themuseum) didn’t have any appreciation forsocial science research, of the most basicelements of experimental procedures,"observed one researcher. In general,however, researchers with specializedtraining appeared willing and able towrite their reports for an audience ofintelligent laypersons.

The key communication problems had to dowith an absence of follow-through oncethe final report was delivered. In eachof the three cases in which an artscouncil or umbrella group sponsoredresearch on a local cross section or setof audiences, inadequate communicationwith member arts organizations wasidentified as a critical defect of theresearch policy process. One in-housestudy director recommended that suchstudies be seen as two-stage endeavors,the first involving research, the secondconsisting of workshops and informalcommunications with specific arts insti-tutions. Another felt that, while aone-day workshop helped to make memberinstitutions more conscious of research,further efforts would have been valuable.In a third case, the director of a per-forming arts organization whose audiencewas surveyed as part of a larger effortcomplained: "Quite frankly, I have yetto have (the study) on my desk. I lookedat it briefly in (the study director’s)office, but it was such a cumbersomething .... We are absolutely not influ-enced by it because we have no knowledgeat all of what the data did show. That’san important point: make sure that the

68

institutions get to see the versely, inflated hop~s~’~-[ ......cooperatingresults. That seems simple." This was .......not a case of malicious neglect~ .In~i.~_t~ ~. -I~ Several other cases the report’s con-the study director, who headed a localumbrella arts group, urged us to speak tothis arts administrator as someone whohad used the report’s findings to goodeffect. Yet the arts administrator hadnot been given a copy of the study--"I’veasked for the results about four or fivetimes and I’m not going to ask anymore, Ihave other things to do .... and was quiteindignant.5

Lack of follow-through was also cited byone study director and one research useras a danger inherent in the use of studentlabor. A performing arts manager said ofa study undertaken with the help of abusiness school student: "I have a strongsense that there was other data we had notdragged out, that there was more therethan we were able to make use of. Thehazard of using a student is that onceher second year got underway, like us,she got busier and busier and less ableto work with us--that was a liability.If we do it again and cannot afford tohire a professional group who will do itin an elaborate fashion, if we do usestudents again, I am pretty sure that wewill assign it to someone and make itpart of a course load for a full year,not simply a means of summer support."Our interviews, as well as the experienceof many studies not considered here, indi-cate that graduate students and, in somecases undergraduates, represent an impor-tant resource to organizations that cannotafford to hire professionals. But whenstudent labor is used it is essential tomake sure that students have sufficientexpertise, that they receive adequatesupervision, that they will hold them-selves accountable for high-quality work,and that they will be available to par-ticipate in follow-through research,interpretation, or dissemination,u

Report content

A fifth difficulty had to do with thecontent of the reports themselves (iden-tified as important for nine studies).In two cases findings were perceived asoutdated due to changes in the audience.In two other instances, research usersstated that results were inconclusive orobvious, in another case that resultswere unexpected but of trivial importance,and in yet another that findings were "notdramatic" enough to make a difference. Tosome extent, these responses reflectinitial hostility to research or, con-

tent was unsatisfactory for relativelyspecific reasons. In the case of onemulti-institutional analysis, by the timethe report was delivered the sponsoringart council changed its interest tostudies tailored to the needs of specificmember institutions. (The study hadalready served its primary purpose ofpublicizing the council before its find-ings were available.) In another case, amuseum abandoned a major planning effort,and the visitor study conducted in con-junction with that effort was not irmned-iately usable. Users of three studies,one of severa! performing arts events andtwo of museums, regretted that the studiesthey had sponsored were not of widerscope. They were interested in thecharacteristics and attitudes of non-attenders, who had been left out of thestudy, as wel! as of those who~sed theirinstitutions. To some extent, this mayhave reflected the fact that as managersbecome involved in the research process,their questions grow more sophisticatedand become better defined. Finally,in-house directors of two museum studiesregretted the relative paucity of negativeevaluation from visitors, since specificcritism was considered particularly use-ful to management.

Study execution

Study execution (cited for eight studies)was also a factor in study utility.Directors or users of four studies saidthat study funding was inadequate. Oneart council staff member felt a strongerstudy could have been conducted hadmoney been available to survey nonatten-ders. An in-house museum researcherreported that his survey had not beenfully used in part due to the absence offunds for computerized data analysis.Another study director, who had volun-teered his services, acknowledged that hiscon~nitment to the project was underminedby the lack of compensation: "I was toobusy to pay much attention to (the dataanalysis) and I was involved in a numberof other projects. Frankly, if I hadbeen paid it would have been different."Two study directors regarded the level ofexpenditure on audience studies as animportant index of an institution’s com-mitment to the research process, whichaffected the inclination of the institu-tion to use the results. An in-housestudy director for a performing artsfestival observed that "if the project had

69

been given more money (by the board) the As one s~udQ"d~rect~~h~-:~: -"ol wasstudy would have had more impact because ~..anxious ~obe challengedon the methodo-the trustees would have-exp~cted.~.m~.-~i~ - logical quality of the work. Wheneverfrom it." similarly, an outside director - someone saidsomething slighting in a

of a study of a major performing artsinstitution, with much experience inmarket research, reported: "The studywas viewed as cheap by (the institution),which had the effect of lowering theconm%itment as well. When organizationsare not paying for the product, they areless committed to using it." In general,however, few of the directors or managersinterviewed felt that more money wouldhave noticeably improved their studiesor made their use more frequent. As weshowed in the previous chapter, fundingwas the major determinant of audiencestudy technical quality, but had noimpact on utility.

Two of the arts managers stated that re-search on their audiences had littleimpact because they lacked an opportunityto affect the study’s design. In mostcases, however, both in-house and outsidestudy directors reported soliciting userinvolvement in the design of the study.Usually, outside researchers consultedclosely with key administrators, and in-house research advocates tried to drawas many staff and administrators aspossible into survey planning. As oneresearcher put it, such consultation wasnecessary "to establish a political envi-ronment in which I could proceed."

Technical quality

In contrast to the extensive attentiongiven such matters as organizationalproblems, administrative and staff atti-tudes, communication and dissemination,and planning, references to technicalquality were almost entirely missing fromour interviews. Low quality was mentionedas a problem in only three of the twenty-five studies What is more, referencesto such factors as sample size, lowresponse rate, and lack of in-house re-search expertise in these three instancesinvolved casual, offhand observations.Although the studies varied widely inquality and many were poor by convention-al social research standards, in n__qo casewas low technical quality cited as amajor reason for disuse. In fact, allthree ~ho mentioned defects in technicalquality were among those who found theirstudies useful for specific managerialdecisions. Furthermore, in cases ofthose reporting the greatest amount ofhostility or indifference to research, noobjections were based on methodology.

staff meeting, I would call him on it,but they invariably withdrewi"

If hostility towards research findingswas never expressed on methodologicalgrounds, neither were such deficienciesmuch cause for distress when they werenoted. An in-house researcher referredto a first study as "stabbing in thedark," "a first feeble attempt at re-search," and stated that he was currentlyworking with university researchers todevelop a more sophisticated program.Nonetheless, he used the earlier study tosuggest marketing changes that were imple-mented. He believed only those resultsthat were very strongly reported and thathe himself found plausible. In anothercase, aboard of trustees overlooked thesmall sample surveyed in a performingarts audience study because "the findingspretty much were expected."

CONCLUSIONANDRECOMMENDATIONS

The willingness of arts managers to acceptfindings of research that does not meetregular technical standards is in largepart a rational response to the environ-ment in which their organizations function.Most art organizations have little time,money, or experience and could not striveto undertake high quality research evenif they wanted to. Also, most have hadvirtually no systematic information aboutthe composition, attitudes, or habits oftheir audiences. Any increment in know-ledge can be valuable. And finally, lackof concern with technical quality reflectsa recognition of the way in which researchfindings actually enter into the decisionprocess in arts organizations--as marginal,indirect, reinforcing, suggestive, ex-pressive, or symbolic inputs that dependlittle on the precise technical methodsemployed.

Seen in this way, the absence of a corre-lation between study technical qualityand study utility is neither mysteriousnor cause for great dismay. Nonetheless,if poor research is useful, good researchcan be even more useful, especially ifit is applied in a rational manner tospecific policy problems. We think suchuse of good research can become typical,and we offer the following recommendations.

7O

First, arts managers have to be committedto using high quality audience research -as a regular part of policy de!~ber~%i~..~_and planning. This depends, in turn, ontheir receiving the resources necessaryfor systematic planning and it meansreaching some concensus as to the roleaudience information should play in artsplanning. Second, an information infra-structure must be created in which bothbasic and applied research is conductedand widely disseminated throughout theart world. Until arts managers can easilydraw on a pool of information and cumu-lative knowledge about the nature andhabit of American arts consumers, eachwill continue to reinvent the wheel.Because of high staff turnover a profes-sional research memory must serve in placeof many transient individual ones. Third,an arrangement must be made that willpermit those arts organizations unable toafford their own high quality research toget the information they need. Local artsresearch consortia, much like cooperativefund-raising drives for the arts, shouldbe established and their limits and pos-sibilities tested. Finally, as part ofthis effort, managers must increase theiracquaintance with social research methods

through sho~rt tutorials~orother means.¯ The.services of individuals literate inresearch met_h0ds s.hould be made availableto institutions that are without accessto them. .

Certainly, there is some cause for opti-mism. Our formal interviews and informalconversations convince us that budgetpressures along with a general change inattitude mean that planning and researchin arts management is increasingly impor-tant. The research activities of theNational Endowment for the Arts and otheragencies may in time provide the infra-structure needed to minimize redundantresearch and reduce the cloud of uncer-tainty under which arts managers operate.And the development of programs in artsadministration and the appeal of artsmanagement to individuals with otherkinds of business and social sciencetraining promises to raise the level oftechnical knowledge upon which arts organ-izations can profitably draw. There is,then, reason to believe that if a studysimilar to this one is carried out in tenyears, its findings with regard to re-search quality and use will be different.

NOTES

1. Only post-1973 studies were includedto ensure that respondent recollectionswere relatively fresh: the regionalrestriction was imposed to minimize datacollection costs. However, the timerestriction resulted in the exclusion ofall examples of several major types ofstudies, and the geographic restrictionwas therefore relaxed to include fivemidwestern studies so that all types ofaudience studies were represented amongour interviews.

2. The idea of research as discourse isdeveloped by Cohen and Garet (1975) inan essay on social science research andfederal educational policy.

3. For a discussion of the contributionof research to problem-setting at thefederal level, see Rein and White (1977).

4. On the importance of leadership inthe utility of federal health programevaluations, see Patton et al. (1977).

5. Although cooperative audienceresearch efforts hold the promise ofrigourous and comparable studies fororganizations that lack the time orexpertise to undertake them alone, theirpotential has yet to be realized. Onereason is that such studies are usuallycarried out for broadly politicalpurposes. Also, local councils or otherconsortium organizations lack the staffand resources to provide an adequateaccount of research findings and to helpmember organizations make use of them.

6. On the positive value of researchalliances with university faculty andstudents, see Wainwright (1973).

71

CHAPTER 4

AN AGENDA FOR ARTS AUDIENCE RESEARCH

Perhaps the first priority for audienceresearch should be the routine gatheringof descriptive statistics about theaudience over time. Such statisticscould be gathered through a regularnational survey of audiences for astratified random sample of arts insti-tutions. Thus far, the population ofarts institutions has not been fullyspecified. However, improvements maymake systematic sampling possible in thenear future. The Census of Business,which had already included data onperforming arts institutions, in 1977added data on museums for the first time:and an economic data series is underconsideration by the National Endowmentfor the Arts. Institutions included inthe survey should be stratified by suchvariables as art type, region, degree ofurbanization, programming policy, amateurversus professional status, and ticketprice, community-based and predominantlyminority institutions, as well as freeand outdoor events, should be included.

Studies performed by individual institu-tions must be designed locally to addressthe specific needs of the organizationssponsoring them. But questions can bewritten so as to make a survey comparableto previous research. The benefits aremutual: others will be able to use theresults and those who undertake the sur-vey will be able to contrast their ownaudiences with existing baseline data.In general, demographic categories canbe patterned after census categorizationschemes, with additional categories addedas needed. When conventional categori-zation schemes are not used, then theuse of many categories for such variablesas education and occupation is preferablesince it is often possible to mergeresponse groups for purposes of compar-ison.

In sampling audience members, it isimportant to stress that forms should becompleted by those who actually receivethem, and not other members of theirparty or family. Such a practice, forexample, would minimize any biasingeffects of tendencies for older men (orwomen) to take on the task of completingquestionnaires for othe9 family members.Questions on educational attainmentshould differentiate between high schooland non-high school graduates, between

individuals with some college, graduatesof two-year colleges, graduates of four-year colleges, and those with graduatetraining or degrees. Occupationalcategories are difficult at best. Theuse of standard census categories inpre-coded questionnaires, or requestsfor precise occupational descriptions tobe coded by investigators with referenceto census listings, would minimizeconfusion in this area. Researchers mayalso reduce response error by specifyingthat the respondent be currently employedat least one-half time in the occupationreported. Where income information isrequested, family income should bespecified. Where racial or ethnicinformation is requested, categoriesshould be made specific and clear. Thecategory "nonwhite," for example, mayinvite ambiguous responses and missimportant differences.

Local organizations can also greatlyincrease the information gained fromsurveys by making use of cross-tabulations--that is, joint frequencydistributions in which audience membersare placed in cells formed by cross-classifying two or more variables. Cross-tabulations require little statisticaltraining, only marginal added effort, andanswer a wide range of managerial andother questions. For example, if onewants to see if those audience membersreporting lower educational attainmentare primarily young people continuingtheir education, one can do a cross-tabulation of age and education. If onewants to assess an audience’s occupationallevel independent of gender, one cancross-tabulate gender andoccupation. Amanager who wishes to predict the effecton audience composition of an across theboard increase in ticket prices can gainsome insight by cross-tabulating ticketprice and education (or occupation orincome) and comparing those in the mostexpensive seats to those in the leastexpensive. A marketing specialist aimingat a particular income group can cross-tabulate income and source of informationto see if different kinds of advertisingreaches different kinds of visitors orattenders. ¯

In addition to using cross-tabulations,researchers can inexpensively increasethe information yielded by surveys in two

72

other ways: demographic frequencies canbe compared to census frequencies formetropolitan residents as a-visits and visitors can be distinguishedby asking respondents to note how manytimes they have attended an institutionover a suitable time period (such as theprevious t~.elve months).

Local organizations should be urged topublicize their own research findingsand to make them available to other artsorganizations. In general, arts institu-tions do not appear to be competing forthe same dollars. Individuals whofrequently attend one art form orinstitution seem frequently to attendothers as well. Promotional energiesmay more profitably go towards expandingthe total arts market for an area thantowards dividing up the existing public.At any rate, audience studies rarelycontain enough surprising, embarrassing,or definitive information to give aninstitution a competitive edge.Although we offered to maintain audiencestudies collected for this study on aconfidential basis, we had few requeststo do so, and such requests were almostalways related to a specific and unusualinstitutional consideration. In mostcases, then, those who undertake studiesseem willing to disseminate their results.What is needed is a clearinghouse forsuch research in which organizations canshare information to their mutual benefit.

In addition to the need for comparabledescriptive data on audiences over time,other critical questions about artsaudiences require more focused studies.Many of the arts managers we spoke withwanted information about nonattenders,the people whom direct audience surveyscan never reach (though cross-sectionalstudies, of course, do so). Do individ-uals fail to attend museums and theperforming arts because of lack ofinterest, antipathy, inconvenience, orprices? Such information is critical toattempts to enlarge the market for thearts and to meet the public’s desire forgreater accessibility to the arts. Tounderstand nonattendance, in-depth inter-views may be necessary to get beyondinitial responses to questionnaires andto reach deeper motivations. Dependingupon the targets of a market developmentplan0 such studies could be focused onindividuals demographically similar toattenders (for example, their next-doorneighbors) or on individuals fromsocioeconomic groups with low attendancerates.

Research on the relationship of attend-ance at one art form to attendance at

others, indicates that;.~i-th-the p~ssible- ~xception of theatregoers, aficionados of

one art form a!so.attend others. Suchresearch, however, is at a rudimentarystage. Is there one arts audience ormany? Do conditions vary from city tOcity with, say major arts centers likeNew York having multiple publics andsmaller cities having a single culturalpublic? Furthermore, what is the respon-siveness of arts attendance not only toprice but also to content? If the operaraises its ticket prices or alters itsprogramming, will audiences go to thetheatre? Will they stay home and watchtelevision? If an art museum changes itsexhibits policies or raises its admissionfee, will visitors go to science museumsor to the aquarium or to a football gameinstead? It has been observed that tele-vision viewers watch television ratherthan tuning in specific programs. Is thesituation similar in the arts? To whatextent can institutions use programchanges to draw larger or new audiences,or to experiment with new offerings with-out fear of losing the existing audience?We know little about the answers to thesequestions.

Many people in the arts have stressed aneed to expand audiences to include thosenot already reached. The audiencesanalyzed in this study tended to sharesuch characteristics as high educationalattainment, high incomes, large percent-ages of professionals, and smal!percentages of blue-collar workers andminorities. Yet there were some strikingexceptions. Intensive analysis ofinstitutions that draw on unusually wideaudiences may provide insights that otherorganizations can use.

An often useful but neglected methodologyis the quasi-experimental design. If aninstitution is contemplating some changein admission price, time of performance,or other program policy and wants toassess its effect on audience composition,controlled studies of audiences before andafter limited changes can be of greatvalue. In such research it is importantto consider alternative explanations forany change found. If this is done, pre-test/post-test studies can be a powerfulmanagement tool.

Another issue about which little is knownand much curiosity exists is the processof socialization into arts attendance:how early does it begin, how important isthe family, and how important is theschool? One easy way to begin to assessthe importance of family background is toask respondents questions about theirparents. We know nothing about the

73

relationship between the educationalultimateiy ~im~tter[~’~-~cided on the

attainment or occupation of the parents ~..basis of values and priorities, For

and a person’s attendance.. If th@~.~t~-example, while research has usually

is acquired early in life, family back-ground may be almost as important asone’s own education or occupation.

A more thorough examination of sociali-zation into the arts must go beyondsurveys to more focused interviews andstudies of children and teenagers. Weknow that a person’s educational attain-ment is the best predictor of his orher attendance at museums and performingarts events. But why is this so? Is itbecause people who stay in school along time come from families where thearts are cultivated from an early age?Is it because formal training in the artsin high schools and colleges creates anappetite? Is it because colleges providestudents with a culturally oriented peergroup and large quantities of free time?Or is it some combination of these andother factors?

If there are many serious gaps in ourknowledge about the public for museumsand for the live performing arts, weknow even less about the public for thearts in other forms. How many peopleenjoy theatre, dance, opera, andclassical music on television and radio?Are these the same people who attendlive performances or is it an entirelydifferent group? Do media presentationsserve as a substitute for live perfor-mances and exhibits, or do they onlywhet consumers’ appetites? (The inter-ested reader should consult Arts andCultural Proqrams on Radio and Televlslonby Natan Katzman and Kenneth Wirt (1977).)What about art books and phonographrecords? Are such mechanical reproduc-tions a supplement to or substitute forvisits to art museums and nights at theopera? until we learn more about thosewho consume the arts in their mediaforms, we can only speculate about thesize and breadth of the arts audience asa whole.

revealed that the arts attenders arewealthier, better educated, and employedin more prestigious occupations than thepublic at large, audience research cannotindicat~ whether this situation is good,bad, or indifferent. Some institutionsare committed to broadening the socialcomposition of the audience, and it seemsclear that such efforts can bear fruit.Among the studies we assembled were a fewof audiences containing quite diversecross-sections of the American public.And in the midst of the Depression,audiences for the Federal Theater Projectincluded many employed blue-collar workers.Other institutions have found it easier,and financially critical, to developfurther those segments of the publicalready attending. Different prioritiesfor expansion dictate differing researchdesigns. Such priorities must be madeexplicit in order to make the best use ofresearch.

The kind of research to be conducted andthe extent of research carried out is

Especially at the local level, researchis part of a process of planning andadministration: and planning is some-thing relatively new to the arts, aboutwhich there is some disagreement.Planning and research both cost money.The best development and use of artsaudience research will require money fora research infrastructure, money forstaff time to execute and follow throughthe implications of research, and moneyto permit institutions now living fromcrisis to crisis to become involved inlong-range planning. Arts institutionshave some capacity to improve researchby shifting their own priorities. But,ultimately, systematic use of research ona wide scale, after the fashion of manygovernment agencies and private industry,may be prohibitively expensive. The levelof resources allocated to the arts fromamong competing national priorities is,of course, a product of the politicalprocess, and the constraints of thisprocess will, indirectly critically shapethe role that such research can play.

74

APPENDIX_

BIBLIOGRAPHY

Abbey, D. S. and Duncan F. Cameron

1959 The museum visitor: I--surveydesign. Toronto: RoyalOntarioMuseum.

1960 The museum visitor: II--surveyresults. Toronto: Royal OntarioMuseum.

1961 The museum visitor: III--supple-mentary studies. Toronto: RoyalOntario Museum.

American Association of Museums

1972 Museums: their new audience.Washington, D.C.

The Arts, Education and Americans Panel

1977 coming to our senses: the signi-ficance of the arts for Americaneducation. New York: McGraw-Hill.

Baumol, William and William Bowen

1966 The performing arts: the economicdilemma. Cambridge: The M.I.T.Press.

Bernstein, Ilene N., editor

1976 validity issues in evaluativeresearch. Beverly Hills: Sage.

Bernstein, Ilene N. and Howard E. Freeman

1975 Academic and entrepreneurial re-search: the consequences of diver-sity in federal evaluation studies.New York: Russell Sage.

Book, S. H. and S. Globerman

1975 The audience for the performingarts: a study of attendance pat-terns in Ontario. Toronto:Ontario Arts Council.

Bracht, Glenn H. and Gene V. Glass

1968 The external validity of experi-ments. American EducationalResearch Journal 5 (November):437-474.

Cameron, Duncan F. and D. S. Abbey

1960a Visits versus visitors: an analy-sis. Museum News 29 (3): 34-35.

1960b Investigating a museum’s audience.The Museoloqist 77 (December): 2-7.

1961 Museum audience research. MuseumNews 39: 34-38.

1962 Museum audience research: theeffect of an admission fee. MuseumNews 41 (November): 25-28.

Campbell, Donald T. and Julian C. Stanley

1966 Experimental and quasi-experimentaldesigns for research. Chicago:Rand McNa!ly.

caplan, Nathan

1976 Social research and national policy:what gets used, by whom, for whatpurposes, and with what effects?International Social Science Jour-nal 28: 187-194.

75

Caplan, Nathan, Andrea Morrison, and. ~ - Gordon,- Gerald,i~d ~EdwardV’- MorseRussell J. Stambaugh ~ - _ ~%~.~~:i~~.-- - 1975~ Evaluation research. ¯ In Alex

1975 The use of social science knowledgeInkeles, ed., Annual review of

in policy decisions at the nationalsocioloqy. Palo Alto: Annual

level. Ann Arbor: Institute for Reviews Inc., pP. 339-361.

socia! Research.

Caro, Francis G., editor

1971 Readinqs in evaluation research.New York: Russell Sage.

Gruenberg, Barry

1975 How free is free time? Analysis ofsome determinants of leisure acti-vity patterns. Unpublished manu-script. Middletown: Department ofsociology, Wesleyan university.

Cohen, David K. and Michael S. Garet

1975 Reforming educational policy withapplied social research. HarvardEducational Review 45 (February):17-41.

cwi, David and Katharine Lyall

1977 Economic impacts of arts andcultural institutions: a model forassessment and a case study inBaltimore. washington, D.C.:National Endowment for the Arts.

Haug, Marie

1977 Measurement in social stratifica-tion. In Alex Inkeles, Jamescoleman, and Neil Smelser, eds.,Anuual review of sociology. PaloAlto: Annual Reviews Inc.,pp. 51-78.

Johnson, Alton C. and E. Arthur Prieve

1976 older Americans: the unrealizedaudience for the arts. Madison:Center for Arts Administration.

de Borhegyi, Stephan F. and Irene A.Hanson, editors

1968 The museum visitor: selectedessays and surveys of visitorreaction to exhibits in the ~ilwau-kee Public Museum. Milwaukee:Milwaukee public Museum.

Katzman, Natan and Kenneth wirt

1977 Arts and cultural proqrams on radioand television, washington, D.C.:National Endowment for the Arts.

DiMaggio, Paul and Michael Useem

1978social class and arts consumption.Theory and Society. 5 (spring).

Kerlinger, Fred N.

1973 Foundations of behavioral research.2nd edition. New York: Holt,Rinehart and Winston.

Gans, Herbert

1974 popular culture and hiqh culture:an analysis and evaluation of taste.New York: Basic Books.

Gephart, William J.

1965 Development of an instrument forevaluating reports for educational~esearch. Unpublished Ph.D. dis-sertati~n.

Lin, Nan

1976 The foundations of social research.New York: McGraw-Hill-

Logan, Olive

1871 The mimic world and public exhibi-.-tions: their historyr their morals.,and effects. Philadelphia: Newworld publishing.

76

Mann, J. Persell, Ca;01ine Hodges~ -~ ......

1972 The outcome of evaluative research~-~. ~ 197.1 The quality of research on educa-In Carol H. Weiss, ed.~ E~aluatt~i~d~ ...... tion: an empirical study of re-action programs. Boston: Allyn searchers and their work. Newand Bacon, pp. 267~282. York: Bureau of Applied Social

Research.

Mann, P. H.

1966 Surveying a theatre audience: meth-odological problems. British Jour-nal of Sociology 17: 380-387.

Rein, Martin and Sheldon White

1977 Policy research: belief and doubt.Policy Analysis 3 (Spring): 239-271.

McTavish, Donald G. et al.

1977 Assessing research methodology:the structure of professionalassessments of methodology. Socio-logical methods and research ~(forthcoming).

Robinson, Edward S.

1930 Psychological problems of thescience museum. Museum News 8 (5):9-11.

Morison, Bradley.gnd Kay Fliehr

1968 In search of an audience: how anaudience was found for the TyroneGuthrie Theatre. New York: PitmanPublishing.

1974 Reach: a new method of measuringan audience. In Alvin H. Reiss,Arts Management Handbook. New York:Law-Arts Publishers, pp. 126-129.

Newgren, Donald

1972 A standardized museum survey: amethodology for museums to gatherdecision-oriented information.Ph.D. dissertation, Syracuse Uni-versity. Ann Arbor: UniversityMicrofilms.

O’Hare, Michael

1974 The audience of the Museum of FineArts. Curator 17 (June): 126-158.

Patton, Michael Q. et al.

1977 In search of impact: an analysisof the utilization of federal healthevaluation research. In Carol H.Weiss, ed., Using social researchin public policy making. Lexington:Heath, pp. 141-163.

Rosenberg, Harold

1966 The anxious object: art todaZ andits audience. New York: CollierBooks.

Rossi, Peter H. and Walter Williams,editors

1972 Evaluating social programs: the-ory~ practice~ and politics. NewYork: Seminar Press.

Simon, Herbert A.

1965 Administrative behavior. New York:Free Press, 1965.

Theatre Communications Group

1967 Toward a new audience: a reporton a continuing workshop in audi-ence development. New York

Tocqueville, Alexis de

1956 Democracy in America.New American Library.

New York:

77

" weiss, -Ca~01 ~. ~".-~d~~-~- -Toffler, Alvin ~...

1965 The culture consumers. Baltimore: -1972 Evaluation research: -methods of

................. ~ssessinq proqram effectiveness.Penguin Books. Englewood cliffs: Prentice-Hall.

van de vall, Mark

1975 utilization and methodology ofapplied social research: four com-plementary models. Journal ofApplied Behavioral science Ii (i):14-38.

Weiss, carol H. and Michael J. Bucuvalas

1977 The challenge of social researchto decision making. In carol H.weiss, ed., usinq social researchin public policy makinq.. Lexing-ton: Heath, pP. 213-233.

van de Vall, Mark, Cheryl Bolas, and TaiS. King

1976 Applied social research in indus-trial organizations: an evaluationof functions, theory, and methods.Journal of Applied BehavioralScience 12 (April-May-June): 158-

177.

veblen, Thorstein

1899 The theory of the leisure class:an economic study in the evaluationof institutions. New York: TheMacmillan company.

Yin, Robert K., Eveleen Bingham, andKaren A. Heald

1976 The difference that quality makes:the case of literature reviews.socioloqical Methods & Research 5(November): 139-156.

Yin, Robert K. and D. Yates

1975 Street-level qovernments: assess-inq decentralization and urbanserviceg... Lexington: Heath.

wainwright, June M.

1973 Svmphony orchestras and marketinqresearch proqram{, vienna, Va.:American Symphony Orchestra League.

78

APPENDIX iI

LIST OF STUDIES

Kimmel, Peter S. and Mark J. Maves."Public Reaction to Museum Interi-ors," Museum News 51(1): 17-19(Sept. 1972).

Twelve museums and galleries inthe San Francisco-Oakland area.

Smits, Edward J. "A Suburban MuseumLooks at Its visitors," Museum News42: 30-34 (May 1964).

See also #’s 200, 204

Nassau County Museum on Long Is-land.

Yankelovich, Skelly and White, Inc.A Study of Out-of-Town Visitors tothe Metropolitan Museum of Art. NewYork: Metropolitan Museum of Art,Sept. 1975, 38p.

See also #’s 230, 250

Arts Administration Research Insti-tute. "The Economic Impact ofBoston’s Legitimate Theatres." (Pre-liminary Report.) cambridge: ArtsAdministration Research Institute,Dec. 1974, 9p.

Haacke, Hans. "Hans Haacke’s GalleryVisitors’ Profile," Artforum ii:44-45 (June 1973).

Boice, Bruce. Comments on Haacke’ssurvey, Artforum ii: 46 (June 1973).

John Weber Gallery in New York City.

Gardner, Howard and Lynn Goldsmith.Final Report to the National Sndow-ment for the Arts Regarding thevisiting Specialist Grant. Minne-apolis: The Minneapolis Instituteof Arts. Sept. 1975, 75p.

See also # 247

8

National Research Center of the Arts,Inc. Americans and the Arts: ASurvey of Public Opinion. New York:Associated Councils of the Arts,1975, 162p.

National Research Center of the Arts,Inc. Americans and the Arts: High-lights From a Survey of PublicOpinion. New York: Publishing Cen-ter for Cultural Resources (avai-able from ACA Publications, P.O.Box 4764, Tulsa, OK 74104), 1974,36p.

See also #137

Nationwide public opinion surveyfor Associated Councils of theArts.

Baumol, William J. and William G.Bowen. "Chapter IV: The Audience"and "Appendix IV-l: Audience SurveyMethods." In William J. Baumol andWilliam G. Bowen, Performing Arts--The Economic Dilemma. Cambridge:M.I.T. Press, 1966. pp. 71-97, 443-467.

Boston Symphony Orchestra, "AudienceSurvey - 1964 (print-outs), Baumol &BOWen (20th Century Fund Study onthe Economi~s of the PerformingArts)." Separate print-outs forFeb. 16, Mar. 9, Mar. 19, Mar. 24,Mar. 28, Apr. 5, and Apr. i0, 1964.Boston: Boston Symphony Orchestra,1964, 7p.

National multi-art form survey.

Kaali-Nagy, Christina and Lee G.Garrison. "Profiles of Users andNonusers of the Los Angeles MusicCenter," California Management Review15 (2): 133-143 (Winter 1972).

79

I0

ii

Raymond, Thomas C., Stephen A.Greyser, and Douglas Schwa!bei.~"Broome County veterans-Men~fia~~-Auditorium and Perfroming ArtsTheatre (B)." In Thomas C. Raymondet al., Cases in Arts Administration.cambridge: Institute of Arts Administration0 July 1971. pp. 1-18.

Tauber0 Edward M. and P. weissenberg."Audience Development for CommunityTheater: opinion Leaders and Wordof Mouth Communication," ~Arts Review 2(3): 525-534 (1971).

See also #71

New York--survey of all countyarts events for 1970 season.

(Leo Burnett Company, Inc.) "TheImpact of Chicago’s Museum visitorson the City’s Economy." Preparedfor council of Museum Directors.(Chicago: Leo Burnett Company,Inc.), Jan. 1977, 18p.

See also #’s 135, 136, 179-182, 236,242, ~43.

12 Washburne, Randel G. and J. AlanWagar. ,,Evaluating visitor Responseto Exhibit Content " Curator 15(3):248-254 (Sept. 1972).

See also #’s 232, 233

Study in four park visitor centersin the Pacific Northwest.

13 Langley, Stephen. "Chapter Thirteen:Community Relations-" In StephenLangley, Theatre Management in Am-erica: Principle and Practice. NewYork: Drama Book SPecialistsTPubli-shers, 1974. pP. 275-288.

Conducted by Theatre Division ofBrooklyn College on six perform-ance nights of a major production-

14 Cerami, Joseph M. "Chapter V: Au-dience Profiles for Selected Music,Drama, and Opera Performances atLong Island Colleges 1969-1970."

In Alian.Easton~-~eommuni~y Suppor%of the PerforminqArts--Selected

.Problems of Local and National Inter-es___t. Hofstra University Yearbook ofBusiness, series 7, V. 5. New York:Hofstra University, 1970. pp. 275-284.

15 Shettel, Harris H. and Pamela C.Reilly. An Evaluation of ExistingCriteria for Judqinq the Quality ofScience Exhibits. Pittsburgh: American Institutes for Research inthe Behavioral Sciences, Nov. 1965,81 p. (Clearinghouse for FederalScientific and Technical Information,TiD-22703/General Miscellaneous, andProgress Reports (TID-4500).)

Shettel, Harris H. "Exhibits: ArtForm or Educational Medium?" MuseumNew____~s 52(1): 32-41 (Sept. 1973).

Responses to effectiveness of U.S.Atomic Energy Commission’s Atomsin Action exhibit (designed inAmerican Museum of Atomic Energy,Oak Ridge, Tennessee)-

16 Johnson~ David A. "Museum Attendancein the New York Metropolitan Region,"Curator 12(3): 201-230 (Sept. 1969).

Interviews of visitors to sixmuse~ ¯

17 O’Hare, Michael. "The Audience ofthe Museum of Fine Arts," Curator17(2): 126-159 (June 1974).

"The Public’s Use of Art--~isitor Behavior in an Art Museum,Curator 17(4): 309-320 (Dec. 1974).

"Why Do People Go toMuseums: The Effect of Prices andHours on Museum Utilization-" Work-ing Paper. Cambridge: Laboratoryof Architecture and Planning, Schoolof Architecture and Planning, MIT,Aug. 1974, 22p.

See also #224

Boston Museum of Fine Arts.

8O

18

19

20

21

22

23

Kaiser, Michael. Letter to C. 24 -Ellis, John W. "American ShakespeareRichardson, Skylight Theatre, Mil- " ...... iThea[re Audience Questionnaire, 1975waukee, Wisconsin, dated Jan..-l~Z~ib<~.~~.." _- - Season: -Response Analysis,-Intro-1975, 3p. ductory Material." Stratford, Conn.:

Includes copy of profile comparingSkylight Theatre audience withaggregate data for other regionaland major opera companies.

Permaloff, Anne et al. "A Survey ofPatrons of the Montgomery Museum ofFine Arts." Paper submitted to Mont-gomery Museum of Fine Arts, April 15,1976. Montgomery: Auburn Universi-ty, 1976, 29p.

American Shakespeare Theatre, 1975,9p. (No data.)

. "The Economic Impact ofthe American Shakespeare Theatre onStratford, Connecticut." Stratford,Conn.: American Shakespeare Theatre,1975, 7p.

25 Fife, Camille. "Indiana RepertoryTheatre Audience Survey, 1975-76Season." Report. Indianapolis:Indiana Repertory Theatre, 1976, 24p.

Ballet West. "Statistics of BalletWest Aspen Poll. " Computer print-out. Salt Lake city: Ballet West,n.d.

Trumpet in the Land. "An Analysisof Audience Questionnaires From the1975 Season of Trumpet in the Land."Mimeographed. Dover, Ohio: Trumpetin the Land, (1975), 3p.

See also #221

Dourdounas, George (John R. Baylessand William E. Crocken, project di-rectors). "Initial Pilot Survey ofRural Central Pennsylvania." Uni-versity Park, Pa.: PennsylvaniaState University, 1975, 10p.

"Survey Results." Univer-sity P~k, Pa.: Pennsylvania StateUniversity, 1975, 10p.

See also #’s 74,150

Mail survey for Pennsylvania StateFestival Theatre.

26

27

28

New Jersey Shakespeare Festival."New Jersey Shakespeare Festival -Audience Questionnaire Responses,July 5th through August 19, 1972."Madison, N.J.: New Jersey Shake-speare Festival, 1972, 4p.

See also #27

Rodgers, F. L. Memorandum from F. L.Rodgers to Executive Committee Re."Telephone Survey," dated Jan. 22,1974 and "Summary of Telephone Inter-views." Madison, N.J.: New JerseyShakespeare Festival, 1974, 5p. + 3p.

See also #26

Interviews with renewed and non-renewed subscribers and individualplay attenders.

(Foran and Greer, Inc.) "BlossomSurvey, 1974." Results. CuyahogaFalls, Ohio: Blossom Music Center,1974, 2p.

Cleveland Symphony Orchestra.

Billings Symphony Orchestra. Copyof questionnaire and summary ofresults of audience survey takenApr. 27, 1975. Billings, Mont.:Billings Symphony Orchestra, 1975,3p.

29 Byrne, June F. Letter to RichardThomas of the Milwaukee Symphony,dated Aug. i0, 1976. Attachments area summary table and questionnaire.Madison, Wisc.: Graduate School ofBusiness, Center for Arts Administra-tion, University of Wisconsin, 1976,3p.

81

Byrne, June F. et al. Hiqhliqhtinqthe Economics of Not-for-Profit.~ArtsOrganizations. Madison,wisconsin Arts council, Inc., 1976,8p.

Respondents were 140 nonprofitarts organizations in Wisconsin(includes survey form used forfourteen audiences to documenteconomic impact of the arts).

Parsons, Lee"~-’iSYs~emi£ic Testingof Display Techniques for an Anthro-pological. Exhibit." In Stephan F.deBorhegyi and Irene A. Hanson, eds.,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee PublicMuseum, 1968. pp. 81-98.

See also #’s 47, 106-109

30

31

DeLuca, Barbara and Carol Bullard.Arts and Arts-Related Orqanizationsin the capital District: The Struc-ture and Function of the Orqaniza-tions Themselves; The Nature andNeeds of the Audiences They Serve;The Pattern~ Impact~ and Potentialof Their Funding. Albany: AlbanyLeague of Arts, 1976, 90p.

Rochester Museum and Science Center.1973-74 visitors Survey. Report.Rochester, N.Y.: Rochester Museumand Science Center, 1974, 26p.

32 CONFIDENTIAL

33

34

35

Shettel, Harris H. "Exhibits: ArtForm or Educational Medium?" MuseumNew__s 52(i): 32-41 (Sept. 1973).

Effectiveness study of vision ofMan traveling exhibit.

Prague, Rochelle H. "The universityMuseum visitor Survey Project,"curator 17(3): 207-212 (Sept. 1974).

interviews, behavioral observation,and questionnaires of visitorsto university Museum in Philadel-phia.

deBorhegyi, Stephan F. "Testing ofAudience Reaction to Museum Exhi-bits," Curator 8(1): 86-93 (Dec.1965). Also in Stephan F. deBorhegyiand Irene A. Hanson, eds. The Museumvisitor: Selected Essays and Surveysof visitor Reaction to Exhibits inthe Milwaukee Public Museum. Milwau-kee: Milwaukee Public Museum, 1968.pp..75-80.

36 Nielsen, Richard. Questionnaire forKrannert Center for the PerformingArts Audience Survey. Urbana, Ill.:university of Illinois, n.d., 19p.

N(ie)isen, Richard P. "Types ofTheatre Information and Marketing-Audience Development Among the Ali-enated," Performing Arts Review14(1-2): 44-57 (1973).

Nielsen, Richard P. and charlesMcQueen. "Attendance Types ofPerforming Arts Events and Explana-tions for Attendance and Non-Atten-dance at the Krannert Center for thePerforming Arts." Paper. Urbana,Ill.: university of Illinois, n.d.,18p.

Nielsen, Richard P. and charlesMcQueen. "Audience Segments of theKrannert Center for the PerformingArts." Paper. Urbana, Ill.: uni-versity of Illinois, n.d., 14p.

,’Performing Arts Consumero

Behavior: An Exploratory Study."Paper. Urbana, Ill.: university ofIllinois, n.d.0 19p.

Nielsen, Richard P., charles McQueen,and Angela B. Nielsen. "PublicPolicy and Attitudes on Tax Supportfor Live Artistic communicationsMedia," American Journal of Econo-mics and sociology, April 1976,pp. 149-160.

Nielsen, Richard P., Angela B.Nielsen, and charles McQueen. "At-tendance Types of Performing ArtsEvents and Explanations for Atten-dance and Non-attendance," Perform-ing Arts Review 6(1): 43-69 (1975).

,.Explanations for Public’sWilling~ess to Pay Additional Taxesfor Support of the Arts," Performin~Arts Review 5(3-4): 199-209 (1974).-

"Performing Arts AudienceSegments: A case Study," Performing

82

37

38

39

Arts Review 6(2): 301-312 (1975-76) -40

Champaign and Urbana, Ill. home.~survey. " - - " ’~~ ~

Moore, Thomas Gale. "V: The Audi-ence." In Thomas Gale Moore, TheEconomics of the American Theater.Durham, N.C.: Duke University Press,1968. pp. 69-89.

See also #98

Description of 1960 Playbill-sponsored survey of Broadwaytheatergoers.

41

Moore, Thomas Gale. "V: The Audi-ence." In Thomas Gale Moore, TheEconomics of the American TP.eater.Durham, N.C.: Duke University Press,1968. pp. 69-89.

¯ "New Audience Survey Iden-tifies ~laygoers." In Alvin H.Reiss, Arts Manaqement Handbook.New York: Law-Arts Publishers, 1974.pp. 142-144.

"Non-Newspaper AdvertisingBrings Few Theatregoers to BroadwayPlays." In Alvin H. Reiss, ArtsManaqement ~Handbook. New York: Law-Arts Publishers, 1974. pp. 146-149.

1962 survey of audiences in sevenBroadway theaters.

42

43

Farrell, Joseph. Letter to JanetGracey and Richard Kirshner, datedApr. 17, 1973. In Theatre Develop-ment Fund, An American Dance Theatre.New York: Theatre Development Fund,n.d. 7p.

National Research Center of the Arts,Inc. Arts and the People: A Surveyof Public Attitudes and Participationin the Arts and Culture in New YorkState. New York: National ResearchCenter of the Arts, Inc., 1973, 225p.

Reiss, Alvin H. "Key Study ShowsPublic Receptive to Arts." In AlvinH. Rei~s, Arts Manaqement Handbook.New York: Law-Arts Publishers, 1974.pp. 122-126.

44

(Pri~ve%ESA~t~[~-~O~estionnairefor Wisc0nsin UniOn Theater Survey.Madison, WisC." (Center for Arts

--Administration, 1976), 5p.

Study of university theater audi-ence to be completed in 1977.

Guthrie Theatre. "Guthrie Theatre1974 Audience Survey" questionnaire.Minneapolis: Guthrie Theatre, 1974,2p.

Namakkal/Eringer Marketing, Inc. Re-port on composition of the audienceof the Guthrie Theatre for the 1974-75 season. Minneapolis: Namakkal/Eringer Marketing, Inc., n.d., 33p.

See also #’s 104, 117, 122, 126, 199

National Research Center of the Arts,Inc. Californians and the Arts: ASurvey of Public Attitudes Toward andPariticipation in the Arts and Cul-ture in the State of california. NewYork: National Research Center ofthe Arts, Inc., 1975 (?), 230p.

National Research Center of the Arts.Californians and the Arts: SummaryReport. Sacramento: California ArtsCommission, 1975, 13p.

Richardson, Raymond J. and John F.Maxwell. "Lake Placid Center forMusic, Drama and Art: Its EconomicImpact; A Study for the New YorkState Council on the Arts." Platts-burgh, N.Y.: Technical AssistanceCenter, Dec. 1973, 30p.

Report of resident and touristsurvey.

Arts Development Services, Inc. ADSMarket Study 1975: A Summary Report.Buffalo: Arts Development Services,Inc., (1975), 16p.

(Gottstein, James F.) PotentialMarkets for the Arts in Western NewYork. Buffalo: Arts Development Ser-vices, Inc., (1975), 72p.

Mail survey of Buffalo area.

83

45

46

47

48

J. R. Taft Corporation. An Evalu-ation of the Impact and Effective- ~

ness of Affiliate Artistsr Inc..Washington, D.C.: J. R.-Taft~C°r .....

~

poration, Comlab, Inc., Oct. 1971, 49

61p.

Evaluation of school arts program.

Nash, George. "Art Museums as Per-ceived by the Public," Curator18(1): 55-67 (1975).

Art Museums as Perceivedby the ~ublic. New York: New YorkState council on the Arts, 1974,31p.

See also #’s 238, 239

Survey of individuals in downtownManhattan and Rochester, visitorsto the Whitney Museum of Art, andstudy of outreach programs of theMemorial Art Gallery at the uni-versity of Rochester. 51

Screven, C. G. "The Application ofProgramed Learning and Teaching Sys-tems Procedures for Instruction in aMuseum Environment." In Stephan F.deBorhegyi and Irene A. Hanson, eds.,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee PublicMuseum, 1968. pp. 167-174.

The Application of Proqram-ed Learnlng and Teachinq SystemsProcedures for Instruction in a Mu-seum Environment: Final Report..Washington, D.C.: U.S. office ofEducation, Bureau of Research (ED048 745), Dec. 1967, 96p.

The Measurement and Facili-tation ~f Learning in the MuseumEnvironment: An Experimental Analy-sis. Washington, D.C.: SmithsonianI--~titution Press, 1974, 91p.

See also #’s 35, 106-109

52

53

54

50

Metropolitan Opera. "Me~ sampler--Survey Findings." New York: Metro-politan Opera, 1976, 5p.

See also #’s 50, 51

Telephone survey of participantsin special Sampler program.

Metropolitan Opera. introduction tothe Met in the Parks Survey. NewYork: Metropolitan Opera, n.d., 40p.

See also #’s 49, 51

interviews with attenders of out-door concert programs; includestables and graphs of responses.

(veitch, Patrick.) "cancelled Sub-scribers--Survey Findings." NewYork: Metropolitan Opera, (1976),5p.

See also #’s 49, 50

Mail survey by Metropolitan Opera’sMarketing Department-

(Rothman, Steve.) .Stage/West Audi-ence Survey - 1975: Survey Questionsand Sample Response; Schedule C."springfield, Mass.: Stage/West,1975, 3p.

New Hampshire Music Festival. "1975New Hampshire Music Festival AudienceQuestionnaire." Tally sheet andcopy of survey. Center Harbor, N.H.:New Hampshire Music Festival, 1975,ip.

CONFIDENTIAL

Study of use of self-teachingmachines at Milwaukee PublicMuseum.

Barter Theatre. "Audience Survey:Barter Theatre, 1975." Questionnairewith results. Abingdon, va.: Barter

55 Homans, Richard. "saint Louis Sym-phony Orchestra Survey." St. Louis:University of Missouri, MarketingDepartment, June 1976, 7p.

Telephone survey of subscribersand their non-subscribing neighbors.

84

56

57

58

59

60

61

62

(Carver, James.) "Results of the ¯ ~.63Questionnaire." Copy of question~ .....naire and tally sheets. ~_Kalamaz60~i:~." _- o.civic Players, n.d., 4p.

Oregon Shakespearean Festival Asso-ciation. "Survey for the Instituteof Outdoor Drama: King John, Aug.27, 1969," and "Survey for the Insti-tute of Outdoor Drama: King John,August 31, 1969." Ashland, Oregon:Oregon Shakespearean Festival Asso-ciation, 1969, 12p. (Tabulationsheets only.)

See also #’s 208-221

64

Holth, Henry. Memorandum to Membersof the Executive Committee, datedJuly i, 1976. Attachments: "HoustonBallet Audience Questionnaire" withtabulations and "Some Notes on theAudience Survey Results." Houston:Houston Ballet, 1976, 6p.

See also #59 65

Houston Ballet. "Houston BalletTelephone Survey." Questionnairewith tabulations. Houston: HoustonBallet, 1974, 3p.

See also #58

Brookgreen Gardens. "BrookgreenGardens: 1970 Survey." Brief reporton results. Murrells Inlet, S.C.:Brookgreen Gardens, (1970), 4p.

See also #61

66

Brookgreen Gardens. "BrookgreenGardens: 1973 Survey." Brief reporton results. Murrells Inlet, S.C.:Brookgreen Gardens, (1972), 3p.

See also #60

Becker Research Corporation. Atti-tudes of the Boston Public Toward"Summerthinq." BOston: Becker Re-search Corporation, Oct. 1970, 50p.

Telephone survey sampling BOstonresidents.

67

National- Reseafch<Cente@ o£ the Arts,Inc. A Study of Washingtohians’ At-tendance, at. Performing Arts Eventsand Museums. New York: NationalResearch Center of the Arts, Inc.,1976, l19p.

For Washington State Arts Commis-sion.

Beldo, Les. A Report of Three Sur-veys: A State-Wide Survey Conductedby The Minnesota Pollr MinneapolisStar and Tribune; An In-Concert Sur-vey and A Personal Interview SurveyBoth Conducted by Mid-Continent Sur-veys, Minneapolis, Minnesota. (Min-neapolis): Campbell-Mithun, Inc.,Mar. 12, 1956, 89p.

See also #65

For Minnesota Symphony OrchestraAssociation.

Earl craig Associates, Inc. TheConstituencies (Actual and Potential)of the Minnesota Orchestra: A Reportto the Minnesota Orchestra! Associ-ation, n.p.: Earl craig Associates,Inc., 1975, 43p.

See also #64

Audience, community leader, and so-cial institution representativeswere interviewed.

Geltner, Frank J. and Tim Wason.Articulture and the Eugene-SpringfieldArea: Surveys of Public Opinion andArts Organizations. Report. (Salem,Oregon): Oregon Arts Commission,Sept. I., 1976, 50p.

Oregon Arts Commission. "Summary ofResults: 1976 Eugene-SpringfieldArts Impact Study" and summary ofdata tabulations. Salem, Oregon:Oregon Arts Commission, 1976, 5p. +3p.

Telephone interviews.

Phoenix Symphony. "Phoenix Symphony1971 Questionnaire" with tabulationsand summary sheet. Phoenix: PhoenixSymphony, (1972), 5p.

85

68

69

Free Public Theatre Foundation. . ~ ~Questionnaires with response .tabulartions for Romeo and ~Jullet, ~Aug~-~i~8"~ " -~31, 1975 and Shakespeare and HisPeo_q_o~!~, Aug. 25-31, 1975. Los Ange-les: Free Public Theatre Foundation, 74

1975, 2p.

Music Theatre of Wichita. "Enter-tainment Survey" questionnaire,"Entertainment Survey Results," and"Telephone Survey Results." Wichita:Music Theatre of Wichita, (1973), 3p.

Statewide surgeY~Of~ audien~c~ forperforming arts and museums.

cober, Rodney L. A PsychoqraphiCLife Style Analysis of Interqenera~tional Continuity in the Developmen~of the Rural Theatre Audience. Thesis,Business Administration. universityPark, Pa.: Pennsylvania State uni-versity, Graduate School, Departmentof Marketing, May 1977, 220p.

See also #’s 22, 150

7O American Ballet Theatre. "AmericanBallet Theatre Audience Survey" ques-tionnaire and questionnaire withdata tabulations. New York: Ameri-can Ballet Theatre, n.d., 4p. + 6p.

75 Becker Research Corporation. TheBoston symphony Orchestra: An Apprai-sal by Subscribersr Former Subscribers,and Potential Subscribers. Boston:Becker Research corporation, June1972, 53p.

71

72

73

(weissenberg, Peter, Mary FrancesGordon and Joseph A. Yacura.) Reporton audience characteristics for theperforming arts in Binghamton, NewYork. (Endicott, N.Y.: Joseph A.Yacura, 1976), 32p.

See also #I0

Mittler, Elliott and Walter Wallner.A Membership Study of the Los AnqelesCounty Museum of Art, June 1967. LosAngeles: LOS Angeles County Museumof Art, 1967, 35p.

Report of telephone survey of mem-bers and survey of membership prac-tices of 27 other other U.S.museums.

National Research Center of the Arts,Inc. The New York Cultural Consumer.New York: National Research Centerof the Arts, Inc., May 1974, 264p.

National Research Center of the Arts,Inc. The New York Cultural Consumer.New York: New York Foundation forthe Arts, 1976, 264p.

National Research Center of the Arts,Inc. Questionnaire for study Th___~eNew York Cultural Consumer. NewYork: National Research Center ofthe Arts, Inc., n.d., 8p. (Sent byMunson-Williams-Proctor Institute,utica, N.Y.)

76

77

78

79

Arts Development Associates, Inc.A Report on Dialoque on the Arts: APublic Planninq Project of the Nebras-ka Arts council. Minneapolis: ArtsDevelopment Associates, Inc., Aug.1976, 73p.

statewide dialogue program includedcitizen questionnaire distributedat meetings and printed in news-papers.

Gourd, William. information Proces-sinq in the Theatre: Sex Differencesin Response to ’The Homecominq’ and’Private Lives.’ Washington, D.C.:U.S. Department of Health, Education,and welfare, National Institute ofEducation (ED 098 622), Apr. 1974,17p.

Baltimore Museum of Art. "Instruc-tions to Survey Volunteers." Balti-more: Baltimore Museum of Art, 1974,2p.

Albany Institute of History and Art."A Comprehensive Membership Surveyby the Albany Institute of Historyand Art" questionnaire and "Resultsof a comprehensive Membership Surveyby the A.I.H.A." summary. Albany:Albany Institute of History and Art,n.d. brochure + 5p.

86

8O Smithsonian Institution. Feasibility 87Study of an Aerospace Museum in. the .-~.~Western United States. PreP~r@~f.0r ~the committee on Scienc~ an~A~trD~ -

~

nautics, u.s. House of Representa-tives, 92nd Congress, 2nd Session,and the National Aeronautics andSpace Administration by the Smith-sonian Institution, Nationa! Air andSpace Museum. Washington, D.C.: 88Government Printing Office, 1972.pp. II, III, 160-237, 362-363.

Potential audience survey of publicin area near Ames Research Center.

Hunhi~gto~ ~ib£~~-: ~mpressions ofthe-Vis~tors~" ~includes data tables.San Marino, calif.: HuntingtonLibrary~ J~ne-July 1972~ 5p.

See also #88

Huntington Library. "A Survey ofVisitors." Brief report and datatables. San Marino, calif.: Hunting-ton Library, 1973, 3p.

See also #87

81 CONFIDENTIAL 89 CONFIDENTIAL

82 walker Art Center. Questionnaire.Minneapolis: Walker Art Center,(1976), ip.

In current use.

9O Denver Symphony Orchestra. "DenverSymphony Questionnaire - January,1972: Results - June, 1972." Denver:Denver Symphony Orchestra, 1972, 8p.

83

84

85

86

(Barton, Genevieve.) Questionnairefor Roberson Center for the Arts &Sciences audience survey. Bingham-ton, N.Y.: (Genevieve Barton, 1976),brochure.

Results in process of compilation.

Oklahoma Theater Center. "Resultsof Informal Survey of Audience TakenDuring Performances of U.S.A. DuringJanuary 1976" and "The OklahomaTheater Center Audience Profile 1976"questionnaire with tabulations. Okla-homa city: Oklahoma Theater Center,1976, 2p. + 4p.

Garden Theatre Festival. "Analysisand Evaluation of the Autumn GardenTheatre Festival 73 (project #four)."Los Angeles: Garden Theatre Festi-val, n.d., 13p.

See also #174

Theatre Tulsa. Membership question-naire. Tulsa: Theatre Tulsa, Mar.29, 1976, 4p.

91

92

93

Arts Development Associates, Inc.20 "Flagship" Communities of NewYork State: A Report on ResearchConducted on Needs~ Potential Spon-sors~ and Facilities for a PossibleTourinq Program by the New York StateCouncil on the Arts. Minneapolis:Arts Development Associates, Inc.,May 1974, 55p.

See also #92

Foster-Pegg, F. Peregrine and JosephWesley Zeigler. The 1975 New YorkState Medallion Tour: A Report onthe Backqround and Findings of theNew York State Council on the ArtsDemonstration Tour of Theatre. Min-neapolis: Arts Development Associ-ates, Inc., June 1975, 75p.

See also #91

National Research Center of the Arts.Anchorage~ Alaska: Public Perspec-tive on the Arts and Culture; Reportof a Survey Conducted for AnchorageArts Council. New York: NationalResearch Center of the Arts, Jan.1975, 221p.

87

94

95

96

97

National Research Center of the Arts,: -Inc. The Joffrey Ballet Audience~ ~ _A Survey of the sprinq 1976 Se~S-~h~~ ¯at the city Center Theater. NewYork: National Research uenter ofthe Arts, Inc., June 1976, 133p.

i00

See also #138

Taylor, James B. et al. Science onDisplay: A Study of the united~tates Science Exhibit~ Seattleworld’s Fair, 1962. Seattle: Insti-tute for sociolo~cal Research, Mar.1963, 184p.

interviews of visitors.

Milton, Paula Mae. Florida Stateuniversity; A Descriptive Studv o~Audience Attitudes: A Survev o~Selected Audiences in professionalLEducational and ’CO~unity Theatre~in Dader Broward~ and Palm Beachcounties in the State of Florida.~issertation, Florida state univer-sity, Aug. 1967, 73p.

Molette, carlton W., II. concepts.About Theatre: A survey of Som~colleqe Students in the Florida coun~ties of Broward, Dade, and Palm Beac~h~omparinq Those Who Have Had a The-atre and .Drama Apprec.iation course~-ith Those Who Have Not. Disserta-t--ion, Florida state university, Aug.1968, ll5p.

Includes surve~~--~£-fOrming artsat tender s-.

Newgren,.Donald A. A StandardizedMuseum Survey: A Methodoloqy for.Museums to Gather Decision-Oriente~~nformation Dissertation, Syracuse-"?- .. ",~-- ~-- Arbor" Unl-versity Microfilms. 1973, 282p.

Includes surveys of museums inSyracuse and Grand Rapids.

101 Arkansas Arts Center. Arkansas ArtsCenter: Public Opinion studyr Jannuary 27, 1971, Project #PS 10-121R~.Little Rock: Arkansas Arts Center,1971, 53p.

Telephone survey.

102 community Service Bureau, Inc. Re__~port and Recommendations: ArkansasArts Center~ Little Rock~ Arkansa~-Dallas: COmmunity Service Bureau,Sept. I, 1971, 48p.

Leading citizen interviews.

103 McKee, David T. Profiles and Pre-ferences: An Audience survey of Sub-scribers~ occasional Patrons and the~eneral Public for Reqional Theatre--~n Seattle. Dissertation, universityo~f washington, 1972. Ann Arbor: Unl-versity Microfilms, 1972, 184p.

98

99

Romano, Albert. October 1971 In-Theatre SurV~. New York: Metro-~edia, Inc., 1971, 103p.

See also #37

Playbill questionnaire survey ofaudiences of sixteen shows in Newyork city.

cherry, Kathleen Ann, Patricia AnnMitchell, and Bradley C. Morison.Brid~es: A Report on an Exploration~f the possibilities in a Herita@e --and Cultural Brid@e concept for F.,’rgo/~oorhead. Minneapolis: Arts Devel-~-pment ~ssociates, Inc., Oct. 1976,138p.

104 Morison, Bradley G. and Kay Fliehr.In Search of an Audience: HOW anAudience Was Found for the Tyron~~uthrie Theatr~e. includes surve~~estionnaire and report. New York:Pitman Publishing corp., 1968, 230p.

See also #’s 41, 117, 122, 126, 199

105 Arts Development ASSociates, Inc.An Evaluation Report on the Illinois~rts council Bicent~.~nial Theater~our and a comparative Analysis o~ffFour Major Theater Tour Pro]ects.~inneapolis: Arts Development ~sso-ciates, Inc., Sept. 15, 1976, 122p.

88

106

107

108

Niehoff, Arthur. "Audience Reaction-~.in the Milwaukee Public Museum: The ~winter Visitors." In StephandeBorhegyi and Irene A.- Ha~s~~eds~,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee Public iiiMuseum, 1968. pp. 21-31.

See also #’s 35, 47, 107-i09

Niehoff, Arthur. "Characteristicsof Audience Reaction in the Milwau-kee Public Museum." In Stephan F.deBorhegyi and Irene A. Hanson, eds.,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee PublicMuseum, 1968. pp. 9-16.

See also #’s 35, 47, 106, 108, 109

112

Cooley, William and Terrence Piper."Study of the West African Art Exhi-bit of the Milwaukee Public Museumand Its visitors." In Stephan F.deBorhegyi and Irene A. Hanson, eds.,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee PublicMuseum, 1968. pp. 143-165.

See also #’s 35, 47, 106, 107, 109

113

~ter.views~io~i~itors to theNational Museu~ of Natural Historyand the National Museum of Historyand Technology.

Colvin, Claire. A Membership Studyof the Fine Arts Museums of SanFrancisco and the Asian Art Museum~July 1976. San Francisco: Fine ArtsMuseums of San Francisco, 1976, 94p.

See also #’s 192-195

Arts Development Associates, Inc."...For a Town of Any Size!" A Planfor Developing and Enriching theCultural Life of the Sioux city Area.Minneapolis: Arts Development Asso-ciates, Inc., Aug. 1975, 155p.

Interviews with community leadersand audience survey questionnaireused at 29 different culturalevents.

Arts Development Associates, Inc.A Report and Recommendations on theQuad cities Cultural Survey. Minne-apolis: Arts Development Associates,Inc., (May 1975), 99p.

Audience survey of 19 culturalevents in Davenport, Iowa-Moline,Illinois area.

109

Ii0

Abler, Thomas. "Traffic Pattern andExhibit Design: A Study of Learningin the Museum." In Stephan F.deBorhegyi and Irene A. Hanson, eds.,The Museum visitor: Selected Essaysand Surveys of visitor Reaction toExhibits in the Milwaukee PublicMuseum. Milwaukee: Milwaukee PublicMuseum, 1968. pp. 103-141.

See also #’s 35, 47, 106-108

Krasnegor, Rebecca. SmithsonianAudience Survey: Summary of BSSRPretest Experience and Recommenda-tions for the Conduct of ~n AudienceSurvey. Washington, D.C.: Bureauof Social Science Research, Inc.,Oct. 1967, 50p.

See also #’s 128, 264

114 Zeigler, Joseph Wesley. A Report onthe Marketability of a Center StaqeTour in the Middle Atlantic States.Minneapolis: Arts Development Asso-ciates, June 1974, 188p.

Zeigler, Joseph Wesley. Steerinq theCenter Stage Study: A Report onMethodoloqy. Minneapolis: ArtsDevelopment Associates, Inc., Aug.1974, 35p.

See also #’s 198, 202

Communities studied were Dover,Del.; Frederick, Rockville, andSalisbury, Md.; Long Branch andVineland, N.J.; Allentown-Bethlehem,Hershey, and Scranton-Wilkes Barre-Hazelton, Pa.; Charlottesville,Norfolk, and Newport News-Hampton,Va.: and Clarksburg, W. Va.

89

115 Ford Foundation. The Finances of.the Performinq Arts. Volume I: ¯ASurvey of 166 ProfessionalNQnPr~fitResident Theaters, Operas,SY~Pnies, Ballets, and Modern Dance com-panies. Volume Ii: A Survey of thecharacteristics and Attitudes ofAudiences for Theater, Opera, Sym-phony, and Ballet in 12 U.S. cities.New York: Ford Foundation, 1974.

Reiss, Alvin H. "Lowry DiscussesNew Ford Foundation Survey." InAlvin H. Reiss, Arts ManaqementHandbook. New York: Law-Arts Pub-lishers, 1974. pp. 34-38.

120

Dept ~ of - Hdalh~.i!Ed~at ion, _andwelfare,OffiCe of-Education, Bureauof Research (contract NO. 0ECI-6-

¯050245-i015), May 1969, l14p.

Children’s Museum of Boston.

Morison/Fliehr Associates. ProjectFuture: A Ten Year Plan for the Dev-elopme%t of Audience~ Funds andFacilities for the Trinity SquareRepertory comDany, Providence~ RhodeIsland. Minneapolis: Morison/FliehrAssociates, Mar. 19, 1968, l18p.

116 American conservatory Theatre. Re-port of 1976 audience survey. SanFrancisco: American ConservatoryTheatre, 1976, 73p.

Subscriber survey.

ll7 Arts Development Associates, Inc.A Decade Later: A Report and Anal-ysis of the Guthrie Theatre Audienceof 1973 and How It compares With theFirst Season. Minneapolis: ArtsDevelopment Associates, Inc., Apr.1974, 57p.

Arts Development ASSociates, Inc.Ten Years Later: A Report and Anal-ysis of the Guthrie Theatre Audienceof 1973 and HOW It Compares With theFirst Season. Draft I. Minneapolis:Arts Development Associates, Inc.,Apr. 1974, 57p.

See also #’s 41, 104, 122, 126, 199

118 zeigler, Joseph Wesley. The ArtparkAudience:¯ A Report on Research Doneby Our Company in Season 1976. In-cludes materials from "Artpark -Evaluation II: 1976 Season, Nov. 9& I0, 1976, Rensselaerville, Newyork" meeting. Minneapolis: ArtsDevelopment Associates, Inc., Oct.1976, 54p. + meeting materials.

Audience at Artpark and public inwestern New York and Niagara Fron-tier were studied.

121 New York State Education Departmentand Janus Museums consultants, Ltd.The 1966 Audience of the New YorkState Museum: An Evaluation of theMuseum’s visitors Proqram.- Albany:university of the State of New York,State Education Department, Divisionof Evaluation (U.S. Dept. of Health,Education, and welfare, ED 044 621),Jan. 1968, 60p.

122 Guthrie Theatre. "Interim Results:1976 Guthrie Audience Survey." Min-neapolis: Guthrie Theatre, 1976, 5p.

See also #’s 41, 104, 117, 126, 199

123 walker Research, Inc. children’sMuseum Imaqe Study. Prepared forthe Children’s Museum of indianapolis.indianapolis: Walker Research, Inc.,1975, 77p.

Telephone survey of adult residentsof metropolitan indianapolis.

124 weiss, Robert S. and SergeBoutourline, Jr. "The communicationValue of Exhibits," Museum NEWS,NOV. 1963, pp. 23-27.

See also #246

Observation of and interviews withvisitors to the Boston Museum ofscience.

119 Nicol, Elizabeth H. The Developmentof validated Museum Exhibits. FinalReport." washington, D.C.: U.S.

9O

125 Klinzing, Dennis. Determininq Audi-ence Profile and Effectiveness ofPublicity. Washington, D.C~.~U~S.. -~_-Dept. of Health, Educati0n,ih~:W~ifare, National Institute of Education(ED 083 649), Au~. 1973, 13p.

Two audience surveys done at theUniversity of Delaware Theatre.

126 Batten, Barton, Durstine and Osborn,Inc. Preliminary Report: An Analy-sis of Those Attendinq the GuthrieTheatre Durinq 1962-1963. Minnea-polis: Twin cities Marketing &Research Dept., Batten, Barton,Durstine and Osborn, Inc., 1963, 18p.

127

"Adult EduCah~i6n~P&r£ic~ant Study"Questionnaire. Green B~y: Green

.Bay.Adul~ Education Council, n.d.,4p.

128

Reiss, Alvin H. "Survey Shows WhenAudiences for Different Arts FormsOverlap." In Alvin H. Reiss, ArtsManagement Handbook. New York: Law-Arts Publishers, 1974. pp. 133-135.

See also #’s 41, 104, 117, 122, 199

132 Actors Theatre of Louisville."Actors Theatre of Louisville Audi-ence Survey" questionnaire with sub-scriber and single ticket buyerresponse rates. Louisville, Ky.:Actors Theatre of Louisville, n.d.,2p.

133 Research and Educational PlanningCenter. Status of the Arts and Cre-ative Activities in the State ofNevada: A Statewide Survey. Reno:Research and Educational PlanningCenter, College of Education, Uni-versity of Nevada, (1976), 37p.

Stack, Christopher D. An Examinationof Lawrence University Audiences.n.p., n.d., 30p.

Danquist, Gerald A. et al. A Market-ing Study of the Smithsonlan NationalAssociates. Cambridge: Harvard Uni-versity, Graduate School of BusinessAdministration, Apr. 28, 1971, 109p.

134 Long Wharf Theatre. Data tabulationsof Long Wharf Theatre subscriptionaudience survey. New Haven, Conn.:Long Warf Theatre, (1976), llp.

See also #’s ii0, 264

129 CONFIDENTIAL

130

135 Leo Burnett U.S.A. The Art InstituteSur__~. Prepared for the Art Insti-tute of Chicago. Chicago: ResearchDepartment, Leo Burnett U.S.A., Nov.5, 1975, 62p.

See also #’s ii, 136, 179-182

Questionnaire survey conductedover one year for one week periodseach season.

Hopkins Center, Dartmouth College.Questionnaire and "Summer 1974 Sta-tistics" compiled by Ellen Feldman.Hanover, N.H.: Hopkins Center,Dartmouth College, 1974. 2p. + 6p.

See also # 151

136 Leo Burnett U.S.A. The Art InstituteGeneral visitor Survey and FocusGroup Research. Prepared for theBoard of Trustees of the Art Insti-tute of Chicago. Chicago: ResearchDepartment, Leo Burnett U.S.A., Oct.18, 1976, 17p.

See also #’s ii, 135, 179-182

131 Dowling, William D. Characteristicsof Adult Education Participants~Green Bay, Wisconsin. Green Bay:Green Bay Center, University of Wis-consin, (Mar. 1962), 48p.

Green Bay Adult Education Council.

137 National Research Cent&r of the Arts.Americans and the Arts: A Survey o~the Attitudes Toward and Participa-tion in the Arts and Culture of th~

United States Public. New York:Associated Councils of the Arts, 1976,60p.

91

138

139

140

141

See also #7

National Researchcenter of the Arts,Inc. The Joffre7 Bal~et Audience onTour. New York: Natlonal ResearchCent------er of the Arts, Inc., July 1976,60p.

See also #94

Report of surveys done on touraudiences in San Antonio, Houston,and New Orleans.

Joint committee on cultural Resources.In Search of a Reqional Policy foxthe Arts: Phase II. Baltimore:Johns Hopkins university Center forMetropolitan Planning and Researchand Regional Planning council,(1975?), 52p.

See also #202

145

zeigler, Joseph Wesley. A Report o~nthe Pittsburqh Audience for Theatr9y~s. New York: Arts DevelopmentA--~ociates, May 1975, 37p.

Leader interviews and mail question-naire of Pittsburgh area residents.

Century Research corporation. Th___~eArts in Arlinqton: 1974 Survey ofthe publiq. Prepared for the CountyBoard, Arlington county. Arlington,Va.: Century Research corporation,July 1974, 61p.

interviews of county residents.

Lucarelli, Anth6hy! A, Memorandum toDr. Albert S.~Miles, re "Final Re-

¯ sukts:- Student Preference Survey-March i, 1974," dated Mar. 16, 1974.Tabulation sheets attached. River-side, calif.: Performing Arts Pre-sentations, university of california,1974, 3p.

Lucarelli, Anthony A. Questionnairefor proposed 1977 city resident sur-vey. Riverside, calif.: PerformingArts Presentations, university ofcalifornia, (1976), 4p.

zeigler, Joseph Wesley. "An Analysisof the Albany symphony OrchestraQuestionnaires - spring, 1976."(Analysis, questionnaires, and datacharts included.) New york: ArtsDevelopment Associates, Inc., 1976,llp.

Comparison of audiences at symphonyperformances in Albany, schenectady,and Troy, New York, and generalpublic on Albany League of Artsmailing list.

146 country Music Hall of Fame and Museum."Survey 1976." Results of SurveyAug. 6-23, 1976. (Nashville: countryMusic Hall of Fame and Museum, 1976),5p.

See also #’s 147, 148

147 country Music Hall of Fame and Museum."Survey." Results of survey NOV. 12-14, 1976. (Nashville: country MusicHall of Fame and Museum, 1976), 6p.

See also #’s 146, 148

142 Arts Development Associates, Inc.Buildinq an Audience for OPERA/OMAHA:.A Report and Recommendations for the1976-77 Season. Minneapolis: ArtsDevelopment Associates, Inc., June1976, 24p.

Audience questionnaire used at twoperformances.

143 Shreveport symphony. ..shreveportsymphony Audience Survey" question-naire, shreveport, La.: ShreveportSymphony, n.d., ip.

148 country Music Hall of Fame and Museum."survey comparison 1973-1975." (Nash-ville: country Music Hall of Fameand Museum), n.d-, 7p.

See also #’s 146, 147

149 (George Eastman House.) "visitorSurvey" questionnaire- (Rochester,N.Y.: George Eastman House), n.d.,2p.

Ongoing-survey of visitor residences.

92

150 Heitman, G. and W. E. Crocken. "SomeObservations on Theatre AudienceComposition, Preferences and.2~.cep-.tions." Paper for Califofni~M~a~e-Iment Review. University Park, Pa.:Pennsylvania State University, n.d.,17p.

Heitman, George and W. E. Crocken."Theatre Audience Composition, Pre-ferences, and Perceptions," Cali-fornia Management Review, 19(2):85-90 (Winter 1976).

See also #’s 22, 74

Audience surveys done at Pennsyl-vania State University (both theFestival Theatre and UniversityTheatre) in 1973.

151 Hopkins Center. Questionnaire.Hanover, N.H.: Hopkins Center,Dartmouth College, June i, 1976, 10p.

See also #130

Questionnaire for Dartmouth seniors.

152 (California Museum of Science &Industry.) "Patronage." Results ofattendance survey. (Los Angeles:California Museum of Science & Indus-try), n.d., ip.

See also #’s 251, 252

153 (Bureau of Government Affairs.)Results of survey for North DakotaCouncil on the Arts & Humanities.Grand Forks, N.D.: Bureau of GOvern-ment Affairs, (1974), 3p.

Aud ! eric e-Membe~ <’ in~e~ vi-ewf .-.. (Min-neapolis) : Arts Development Asso-ciate.s, .Ju.ly 26, 1970, !2p.

See also #154

Telephone interviews.

156 Zeigler, Joseph. "The Future of Jazzat St. Peter’s." Excerpts fromJoseph Zeigler, "The Common: AnExtraordinary Place." New York: Arts-Development Associates, Apr. 23, 1976.pp. 1-4, 21-22.

157 Theatre Development Fund. "Surveyof Times Square Theatre Centre."Includes "Comments on the PreliminaryTKTS Survey" by William J. Baumol.New York: Theatre Development Fund,Sept. 1973, 20p.

See also #’s 158, 159

Surveyed patrons of day-of-performance half-price ticketprogram.

158 Baumol, William and Hilda Baumol.Last Minute Discounts on UnsoldTickets: A Study of TKTS. Report i.New York: Theatre Development Fund,1974, 53p.

Theatre Development Fund. "Figuresfrom the Baumol’s published survey..."In Theatre Development Fund, TheatreDevelopment Fund: A Proqress Report~1974-75. New York: Theatre Develop-ment Fund, n.d.0 p. 15.

See also #’s 157, 159

154 Morison/Fliehr Associates. "Prelimi-nary Report--l: Center Opera CompanyAudience Research." (Minneapolis):Morison/Fliehr Associates, Dec. 30,1968, 40p.

See also #155

Interview survey of Twin Cities’women judged representative ofpotential audience.

159 Baumol, William. "Survey of Usersof the Lower Manhattan Theatre Centre."New York: Theatre Development Fund,Feb. i, 1975, 14p.

See also #’s 157, 158

160 CONFIDENTIAL

155 Colburn, D. "Center Opera Company:Summary of Attendance Review and

93

161 Bureau of Business Research, WestTexas State university. An Awareness L~

and Attitude Study to~Determ~n~t~ "Characteristics Leading to Attendanceand Participation in Selected Fine~Arts in Amarillo. Prepared for FineArts council cf the Amarillo chamberof commerce. Amarillo: Bureau ofBusiness Research, West Texas Stateuniversity, Feb., 1971, 37p.

162 Monmouth county Arts Council. "Audi-ence Survey" questionnaire and"Results of Showcase II Question-naires." Red Bank, N.J.: MonmouthCounty Arts council, n.d., 5p.

168

Audience Research Study."_ In ThomasC. Raymond et al., Cases in ArtsAdmihistration. cambridge: Insti-tute of Arts Administration, July1971. pp. 1-21, 1-28.

Lambertville, New Jersey, musicaltent theatre, 1959.

Rankin, Senath. The wilmington AreaArtist series: A Study of the 1974-75 Season. washington C.H., Ohio:Senath Rankin, June 1977, 33p.

cincinnati Symphony Orchestra.

163 Green Mountain Guild. Two question-naires; one with tabulations. (WhiteRiver junction, Vt.): Green MountainGuild, n.d., 2p.

169 virginia Museum. "Members’ Surveys,"(questionnaire only) virginia MuseumBulletin 33(9): i0-ii (May 1973).

See also #’s 170, 253

164 New Hampshire Performing ArtS Center."Audience Survey for the New Hamp-shire Performing Arts Center" ques-tionnaire and letters from E.P.Jancewicz, Sept. 23, 1976, andTimothy G. Jones, Nov. 3, 1976, ana-lyzing data. Manchester, N.H.: NewHampshire Performing Arts Center,1976, 2p + ip. + 2p.

165 Rivas, Frank W. An Assessment ofAttitudes Toward Music. Preparedfor National Assessment of Educa-tional Progress, a Project of theEducation Commission of the States.Washington, D.C.: U.S. GovernmentPrinting office, Sept. 1974, 31p.

On-going national student surveydesigned to evaluate music edu-cation in the U.S.

166 Grossman, carol. "Report on Feasi-bility of Music Voucher." New York:Theatre Development Fund, June 22,1976, llp.

Includes interviews with theater-voucher users.

170 (virginia Museum.) "council Hostess.Information Sheet on visitors toMuseum" questionnaire. (Richmond:virginia Museum of Fine Arts, 1975),ip.

See also #’s 169, 253

171 Lob, Diana Friedel et al. MarketResearch Survey: How People ConsumeChica~o Theater at the organic The-ater, Second city Theater~ and St.~Nicholas Theater. Chicago: univer-sity of Chicago, Graduate School ofBusiness, June i, 1976, 120p.

See also #’s 172, 175-177

172 Alsberg, Eric et al. The Developmentof a Subscription Plan for the organ-ic Theater. Chicago: i~niversity ofchicago, Graduate School of Business,June 1976, 54p.

See also #’s 171, 175-177

Focus group interviews-

167 Raymond, Thomas C., Stephen A.Greyser, and Douglas Schwalbe."St. John Terrell’s Music circus"

94

173 Connecticut Theatre Foundation. ASurve7 of Audience Opinion: Monda--y,19 Auqust Throuqh Saturdayr 24 Auqust, 1974. Compiled by Rita~.Ma~etBarrows. Westport, Conn. : Connecticut Theatre Foundation, Inc., 37p.

Report of audience survey at West-port Country Playhouse.

179 Chilson, Barby et al. "I Think IWent .... Backwards": A MarketinqResearch Project for the Art Insti-tute of Chicaqo and Marketinq 353.(Chicago: University of Chicago,

Graduate School of Business, 1975),15p.

174 Chernack, Peter A. Report on theGarden Theatre Festival (4th AnnualL.A. Performinq Arts Festival)~Barnsdale Park, July 2-25, 1976.Los Angeles: Garden Theatre Festi-val, Nov. 1976, 179p.

Garden Theatre Festival. Copies ofcompleted survey forms. (Los Ange-les: Garden Theatre Festival, 1976),87p.

See also #85

180

See also #’s ll, 135, 136, 180-182

Chesterfield, Jim et al. FocusGroup Interview Study of Member~and NOn-Members of the Art Instituteof Chicaqo. (Chicago: Universityof Chicago, Marketing Class), i0June 1975, 51p.

See also #’s ii, 135, 136, 179, 181,182

175 Goonasekera, A. et al. "A Study ofthe Sources of Information of theOrganic Theater Audience." Paperprepared for the Marketing Communi-cation Program. Chicago: Universityof Chicago, Graduate School of Busi-ness, Spring 1976, 33p.

See also #’s 171, 172, 176, 177

176 Wasso, Louise et al. "Near NorthSide Film Audiences as a Market forthe Organic Theatre." Paper preparedfor Business 353, Advertising Manage-ment. (Chicago: University of Chi-cago, Graduate School of Business),June i0, 1976, 27p.

See also #’s 171, 172, 175, 177

177 Edison, Marcia et al. "The OrganicTheater Marketing Study." Paperprepared for Business 353. (Chicago:University of Chicago, GraduateSchool of Business), Spring 1976,33p.

See also #’s 171, 172, 175, 176

181

182

Louer, Robert, Brian Copp, and JayRodrian. The Art Institute Member-ship: Profile, Attitudes, Tastes-Group VI - Report - The Mailed Survey.(Chicago: University of Chicago,Graduate School of Business, 1975),24p.

See also #’s ii, 135, 136, 179, 180,182

Art Institute of Chicago.

Art Institute of Chicago. "Prelim-inary Review of the Findings of theUniversity of Chicago Marketing ClassSurvey of the Art Institute - June1975." Chicago: Art Institute ofChicago, July 25, 1975, 2p.

(David, Harry.) Art Institute Mem-bership: General Profile; 4-ClusterSolution. (Chicago: University ofChicago), Dec. 15, 1975, 17p.

See also #,s ii, 135, 136, 179-181

Reports further analyses of datafrom studies #181 and #135

178 Davis, Harry L. The Chicaqo Sym-phony and Its Audience: A SUmmR~yRe_~. Chicago: University ofChicago, Graduate School of Business,

183 Balling, Robert. "The SpringfieldArt Center: Its Relationship toOur Community." Springfield, Ohio:(Springfield Art Center), 1973, 9p.

95

184

185

186

187

Residence distribution of membersand students.

Temple, Robert E. Excerpts from"Report to the Board, Oct. 18, 1976,"profiles from 1976 season parkinglot survey, and audience question-naire. (Maryville, Tenn.): SmokyMountain Passion Play Association-1976, 5p.

-off ~h:e- Boro~gh~:o£~een~ _~n~ Th~l~ -~iews on and Participatzon zn uneArts and CUlture. New York: Nation-al Research Center of the Arts, Inc.,June 19750 194p.

191 (Andreasen0 Alan R.) Questionnairesfor in-progress survey of symphonyand theatre attendance. Urbana0 Ill.:university of Illinois, 1977, 37p.

Consumer demand for the arts inthe South.

Story, Janet. Questlonnaire summary,July-October, 1976, " ,,QuestionnaireResponses Mentioned More Than Once, "and ,,Questionnaire- " (Portsmouth,N.H.: Strawbery Banke0 Inc., 1976),5p.

192 Cavala0 John. "Media Enhancement ofa Musetun Gallery." (San Francisco:Fine Arts Museums of San Francisco(M.H. de Young Museum)), n.d., 16p.

See also #’s iii, 193-195

Music Theatre of Wichita, Inc.Questionnaire with tabulations.Wichita: Music Theatre of Wichita,Inc., 1975, 3p.

Interviews of M.H. de Young Museumattenders.

Season ticket holders.

Gisler0 John F. The People and theArts. Prepared b~ the University--of---~tah, Bureau of Economic andBusiness Research and Rocky MountainArts Research Group. Salt Lake City:Salt Lake council for the Arts, Oct.1976, 171p.

Interviews of Salt Lake area resi-dents.

188 Sidney Hollander Associates. Geo____lgraphic Origins of Attendance at th~Baltimore Museum of Art~ Mid-JanuaryThrough Mid-June, 1971. Baltimore:Sidney Hollander Assouiates, 1971,5p.

189 (Oklahoma Arts and Humanities Coun-cil.) "Section G: Organizations,Spare Time" questionnaire and com-puter print-out of tabulations.(Oklahoma City: Oklahoma Arts andHumanities council, 1976), 26p.

190 National Research Center of the Arts,Inc. The Arts in the Borough ofQueens: A Survey of the resident@

193 McElroy, Guy and Cleveland Bellow.Museum audience survey, 1974-75.San Francisco: M.H. de Young Memo-rial Museum, 1975, 25p.

See also #’slll, 192, 194, 195

194 M.H. de Young Memorial Museum.-Acoustiguide Survey Final Tally."(Questionnaires with tabulations)San Francisco: M.H. de Young Memo-rial Museum, n.d., 9p.

See also #’s iii0 192, 1930 195

195 M.H. de Young Memorial Museum."American Art: An Exhibition fromthe Collection of Mr. & Mrs. John D.Rockefeller 3rd." Questionnaireswith tabulations and summary. SanFrancisco: M.H. de Young MemorialMuseum, n.d., 14p.

See also #’s iii, 192-194

196 Miller, William F. "Staying Away:Cost, Not crime, Spoils Attractionof Night Life Here for Suburbanites,"newsclip from The Cleveland Plain ~Dealer, (fall 1976?).

96

Playhouse Square Foundation. "Play- . :.202 .Cwi,-David. "~cent@r Stage" residenceHouse Square Free Theater Study~ ........ data from PhaSe II, Regionil Planning(Cleveland, Ohio: Playhouse.~r~.~_- .~Council-Johns Hopkins audience survey.Foundation), n.d., 3p. Baltimore: Johns Hopkins University,¯

Center for Metropolitan Planning andSurvey of audience for free theatre. Research., (1976?), 24p.

See also #’s 114, 139, 198

197 "Fresno Philharmonic Audience Survey,"American Symphony Leaque Newsletter13(2): ii (Jan., Feb. 1962).

Part of larger study done for Re-gional Planning Council (#139).

Questionnaires distributed to con-cert audiences.

198 Arts Development Associates, Inc.Excerpts on Center Stage from The1973 Reqional Tours of the Cent-~Staqe of Baltimore, MD and theGuthrie Theatre of Minneaplis, MN:A Documentary Evaluation. Minneapo-lis: Arts Development Associates,Inc., July 31, 1973, 105p.

See also #’s i14, 202

199 Arts Development Associates, Inc.Excerpts on Guthrie Theatre fromThe 1973 Reqional Tours of the Cen-ter Staqe of Baltimore, MD and theGuthrie Theatre of Minneaplis, MN.Minneapolis: Arts Development Asso-ciates, Inc., July 31, 1973, 59p.

See also #’s 41, 104, 117, 122, 126

200 (Harbour, John E.) "Chapter VI: Sum-mary." In John E. Harbour, "Who Arethe Visitors? An Analysis of theVisitors to th~ Nassau County Musettm."Thesis, Cooperstown Graduate Pro-grams, State University College atOneonta, N.Y., 1975. pp. 92-96.

See also #’s 2,204

203 American Museum of Natural History.A Profile of Consumer Use and Evalu-ation: The American Museum of Nat-ural History: Based on a Survey ofAttendance July 1974-July 1975. NewYork: American Museum of NaturalHistory, Jan. 1977, 303pp.

204 Culp, Marion et al. "Nassau CountyHistorical Museum Visitation Survey1967." Paper prepared for SocialScience ReSearch Technician Course,Hofstra University, Hempstead, N.Y.1966-1967, 26p.

See also #’s 2, 200

205 Williams, Monte G. "Quincy SymphonyOrchestra General Opinion Survey:Spring, 1976." Quincy, Ill.: QuincySymphony Orchestra, 1976, 10p.

Report on questionnaire and inter-view study.

206 Cleveland Area Survey. "ClevelandArea Survey -- 1977" questionnaire.Cleveland: Cleveland Area Survey,Case Western Reserve University,1977, 47p.

Survey of visitors to four museumsof the Nassau County, Long Island,Museum.

201 National Research Center of the Arts,Inc. Winston-Salem/Forsyth County:Public nerspective on the Arts andCulture. New York: National Re-search Center of the Arts, Inc.,July 1974, 189p.

Cross-sectional study of NorthCarolina county.

207 Wadsworth Atheneum. Results of avisitors’ survey at the WadsworthAtheneum: Preliminary Report. Hart-ford: Wadsworth Atheneum, Nov. 1976,9p.

Wadsworth Atheneum. "Visitor Survey"and "Visitor Survey: Part II"questionnaires. Hartford: WadsworthAtheneum, 1976, 6p.

97

208~. ~215

Institute of Outdoor Dr~na. "Na-tional Audience Survey 1976:-~.McIntosh Trail (Peach~ree~Ci~’~Ga.)." Chapel~Hill: Institute ofOutdoor Drama, 1977, 4p.

See also #’s 57, 209-222

Instituteof~tdo0r Errata-. "Na-tional Audience Survey 1976: Loui-siana Cavalier (Natchitoches, La.)."Chapel Hill: Institute of OutdoorDrama, 1977, 4p.

See also #’s 57, 208-2140 216-222

209 Institute of Outdoor Drama. "Na-tional Audience survey 1976: Crossand Sword." Chapel Hill: Instituteof Outdoor Drama, Jan. 1977, 4p.

Se~ also #’s 57, 208, 210-222

216 Institute of Outdoor Drama. "Na-tional Audience Survey 1976: TheCow, non Glory." Chapel Hill: Insti-tute of Outdoor Drama, 1977, 4p.

See also #’s 57, 208-215, 217-222

210

211

212

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: TheLost Colony (Manteo, N.C.)." ChapelHill: Institute of Outdoor Drama,1977, 4p.

See also #’s 57, 208, 209, 211-222

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: TrailsWest." Chapel Hill: Institute ofOutdoor Drama, 1977, 4p.

See also #’s 57, 208-210, 212-222

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: StephenFoster (Bardstown, Ky.)." ChapelHill: Institute of Outdoor Drama,1977, 4p.

See also #’s 57, 208-2110 213-222

213 Institute of Outdoor Drama. "Na-tional Audience Survey 1976: FromThis Day Forward." Chapel Hill:Institute of Outdoor Drama, 1977, 4p.

See also #’s 57, 208-212, 214-222

214 Institute of Outdoor Drama. "Na-tional Audience Survey 1976: Wilder-ness Road (Berea, Ky.)." ChapelHill: Institute of Outdoor Drama,1977, 4p.

See also #’s 57, 208-213, 215-222

217

218

219

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: Beyondthe Sundown." Chapel Hill: Insti-tute of Outdoor Drama, 1977, 4p.

See also #’s 57, 208-216, 218-222

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: strikeat the Wind (Pembroke, N.C.)."Chapel Hill: Institute of OutdoorDrama, 1977, 4p.

See also #’s 57, 208-2170 219-222

Institute of Outdoor Drama. "Na-tional Audience Survey 1976: TheSword of Peace." Chapel Hill: In-stitute of Outdoor Drama, 1977, 4p.

See also #’s 57, 208-2180 220-222

220 Institute of Outdoor Drama. "Na-tional Audience Survey 1976: Con-quistador (Hot Springs National Park,Ark.)." Chapel Hill: Institute ofOutdoor Drama, 1977, 4p.

See also #’s 57, 208-219, 221-222

221 Institute of outdoor Drama. "Na-tional Audience Survey 1976: Trumpetin the Land (New Philadelphia,,~Oh-)-"Chapel Hill: Institute of OutdoorDrama, 1977, 4p.

See also #’s 21, 57° 208-220, 222

98

222

223

Institute of Outdoor Drama.tional Audience Survey 1976:(Canyon, Tex.)." Chapel Hill: In~ " ~ "stitute of Outdoor Dr.ama, 1977.~-~p~-~ -.

See also #’s 57, 208-221, 223

"Na- -~225 Co£e,--Luhy-Ruggles~:~i-~i-~-Artists Infor-Texas ...... mati6n ~Ure~u---~~f~-Quiz, May

1976~"Report on survey results andcopy of questionnaire. New Orleans:

Artists InfOrmation Bureau, 1976,8p. + 3p. -.

Texas Panhandle Heritage Foundation,Inc. "Texas" 1965-76 materials.Includes "Cumulative Statistics on’Texas’ Musical Drama, 1966-1976":questionnaire forms used during 1966-1967, 1968-1970, 1972-1976, and 1976.Canyon, Tex.: Texas Panhandle Heri-tage Foundation, Inc.

Texas Panhandle Heritage Foundation,Inc. "Survey Reports: ’Texas’ Musi-cal Drama 1976," "Report on the Busesto ’Texas’ 1976," and data sheets onnumbers of cars parked in 1976 sea-son. Canyon, Tex.: Panhandle Heri-tage Foundation, Inc., 8p.

Texas Panhandle Heritage Foundation,Inc. "Report of Audience SurveysTaken in 1975." Canyon, Tex.: TexasPanhandle Heritage Foundation, Inc.,lp.

Texas Panhandle Heritage Foundation,Inc. "Texas" survey data for 1970.Canyon, Tex.: Texas Panhandle Heri-tage Foundation, Inc., lp.

Texas Panhandle Heritage Foundation,Inc. "Audience Survey Taken Friday,August 8, 1969." Canyon, Tex.:Texas Panhandle Heritage Foundation,Inc., 2p.

Texas Panhandle Heritage Foundation,Inc. "’Texas’ Statistics for 1968"and "Occupations Represented at the’Texas’ Show in Two Surveys Taken in1968." Canyon, Tex.: Texas Pan-handle Heritage Foundation, Inc., 3p.

General opinion study of public"monumental sculpture."

226 Wedge, Eleanor F. Nefertiti Graf-fiti: Cor~nents on an Exhibition.Brooklyn: The Brooklyn Museum, 1976,161p.

Analysis of comments written byvisitors to Egyptian art exhibit.

227 Conrad, Jack. Profile of the Play-goer: Study of a Theatre Audience.Memphis: Southwestern at Memphis,1969, 137 p.

Front Street Theatre, Memphis

228 Shakespeare Festival of Dallas."Shakespeare Festival of DallasAudience Questionnaire" and "Shake-spearean Festival of Dallas July,1976 Questionnaire Results." Dallas:Shakespeare Festival of Dallas, 1976,4p.

229 Screven., C.G. "The Effectiveness ofGuidance Devices on Visitor Learning,"Curator 18(3): 219-243 (1975).

Glass exhibit at Renwick Gallery,Smithsonian Institution.

See also #222

Materials listed above consist ofbrief narratives and tables ofdata.

230 Metropolitan Museum of Art. "TheMetropolitan Museum and Its Public:Highlights of the Yankelovich Survey."In Metropolitan Museum of Art, AnnualReport 1973-1974. New York: Metro-politan Museum of Art, 1974. pp. 15-19.

224 Geer, Abigail. "A Study of the Roleof the Art Museum as Collector orEducator." n.p., 1976, 78F.

See also #17

Study of visitor behavior at BostonMuseum of Fine Arts.

See also #’s 3, 250

231 Wagar, J. Alan, Gregory Lovelady,and Harlan Falkin. Evaluation Tech-niques for Interpretation: Study

~-Results From an Exhibition on Energy:USDA Forest Service Research Paper

99

PNW-211. Portland, Ore.: U.S. De-partment of Agriculture, ForestSer- ~ -.vice, Pacific Northwest Forest_andRange Experiment Station,1976~°13P

Pacific Science Center, Seattle.

tor Attitude -ada~A~f@h@ss-at anAqua~ium~" curato~ 20(1): ~8-52(March 1977).

See also #ii

John G. Shedd Aquarium, Chicago

232 Dick, Ronald E.0 Erik Myklestad, andJ. Alan Wagar. Audience Attentionas a Basis for Evaluating Interpre-tive Presentations. USDA ForestService Research Paper PNW-198.Portland, Ore.: U.S. Department ofAgriculture, Forest Service, PacificNorthwest Forest and Range Experi-ment Station, 1975, 7p.

See also #’s 12, 233

Audio-visual presentations atvarious Pacific Northwest park,visitor centers.

237

238

Merrimack Valley Textile Museum."Survey Questionnaire-Merrimack Val-ley Textile Museum." North Andover,Mass.: Merrimack Valley TextileMuseum, (1967), ip.

Memorial Art Gallery of the Univer-sity of Rochester. "Adult WorkshopQuestionnaire." Rochester, N.Y.:Memorial Art Gallery of the Univer-sity of Rochester, (1971), 2p.

See also #’s 46, 239

233 Wagar, J. Alan. Cassette Tapes forInterpretation. USDA Forest ServiceResearch Paper PNW-207. Portland,Ore.: U.S. Department of Agricul-ture, Forest Service, Pacific North-west Forest and Range ExperimentStation, 1975, 20p.

See also #’s 12, 232

Pacific Northwes.t park visitorstudy, q

234 Borun, Minda. Measurinq the Immeas-urable: A Pilot Study of MuseumEffectiveness. Washington, D.C.:Association of Science-TechnologyCenters, Apr. 1977, l14p.

See also #248

Study of visitors to the FranklinInstitute Science Museum and Plane-tarium in Philadelphia.

239

240

241

Clay, Langdon F. Student question-naire. Rochester, N.Y.: MemorialArt Gallery of the University ofRochester, n.d., ip.

See also #’s 46, 238

Foster, Arnold and Thomas Juravich."Art Activities in the Tri-CitiesArea." Paper presented at the Con-ference on Community and Policy Re-search, Apr. 30, 1976, Albany, NewYork. Albany: State University ofNew York, 1976, 15p.

Cross-sectional survey of Albanyarea.

Exhibit Museum of the University ofMichigan. Questionnaire. Ann Arbor:Exhibit Museum of the University ofMichigan, (1956?), 2p.

235 Celender, Don. "Opinions of WorkingPeople Concerning the Arts." NewYork: O. K. Harris Gallery, 19750150p. ~

Survey of nonr~n~set of Minne-apolis residents.

236 Serrell, Beverly. "Survey of Visi-

242 Museum of Science and Industry. "At-tendance Survey." Results. Chicago:Museum of science and Industry, 1958,38p. -~ ~

See also #’s ll, 243

Geographical origin of visitors.

I00

243 Museum of Science and ~ndustry. 250 Me~qpo~itao M~mS~[ ~;t. "Ques-"Six-Day Survey Shows Museum visitors tion~aire ResponseiFe~. 17,-1971" byCome From All Over the World." In_ . ~ .. Frank Simon, questionnaire,- Jan. 1971,Museum of Science and Industry,.~-~acts. and "Results of Questionnaire: Jan-and Information. Chicag6:M~hm~- uary1971." New York: Metropolitanof Science and Industry, 1969, 13p. Museum of Art, 1971, ip. + ip. + 2p.

See also #’s ii, 242 See also~#’s 3, 230

244 Detroit Institute of Arts. "Ques-tionnaire," Detroit: Detroit Insti-tute of Arts, n.d. (2 completedquestionnaires)

See also #’s 245, 266-268

251 california Museum of Science andIndustry. "Questionnaire for MuseumVisitors." LOS Angeles: CaliforniaMuseum of Science and Industry, (1965-1968?), 12p.

See also #’s 152, 252

245 Detroit Institute of Arts. Question-naire. Detroit: Detroit Instituteof Arts, n.d., 2p.

See also #’s 244, 266-268

246 Museum of Science. "Lilly EndowmentSurvey: 1958." BOStOn: Museum ofScience, 1958, 13p.

See also #124

Visitor survey.

247 Minneapolis Institute of Arts. Ques-tionnaire and results from "Paintingin Italy in the 18th Century" and"Art Deco" exhibits. Minneapolis:Minneapolis Institute of Arts, (1971),8p.

See also #6

252

253

254

California Museum of Science andIndustry. Questionnaire. Los An-geles: california Museum of Scienceand Industry, 1971, ip.

See also #’s 152, 251

Virginia Museum of Fine Arts. Ques-tionnaire and results of survey fromMembers’ Bulletin, 1967. Richmond:Virginia Museum of Fine Arts, 1967,2p + Ip.

See also #’s 169, 170

Colonial Williamsburg. "visitorAnalysis-Summer 1964: Post-Image."(Questionnaire with tabulated respon-ses.) Williamsburg, va.: ColonialWilliamsburg, 1964, 3p.

See also #’s 255-259

248 American Institute for Research."An Evaluation of the EducationalProgram of the Franklin Institute:A Summary Report." Pittsburgh:American Institute for Research,Jan. 1960, 4p.

See also #234

249 Garvin, Alfred D. An Evaluation ofthe Cincinnati Science Center. cin-cinnati: University of Cincinnati,n.d., 49p.

Study of fifth graders.

255

256

Colonial Williamsburg. Questionnaire.Williamsburg, Va.: Colonial Williams-burg, Aug. 1968, ip.

See also #’s 254, 256-259

Colonial Williamsburg. Questionnaire.Williamsburg, Va.: Colonial Williams-burg, 1968, ip.

See also #’s 254, 255, 257-259

i01

257

258

Colonial Williamsburg. Question-naire. Will±amsburg, Va.: ColonialWilliamsburg, 1969, ip. " ~

See also #’s 254-256, 258, 259

Colonial Williamsburg. Question-naire. Williamsburg, Va.: ColonialWilliamsburg, 1970, ip.

See also #’s 254-257, 259265

See also #’s-Ii0, 128

Report o~ year-long visitor surveyat the National Museum of Historyand Technology and the NationalMuseum of Natural History.

Schwartz, Steve. Results of surveyof Oakland Museum visitors. Oakland:Oakland Museum, 1971, 19p.

259 Colonial Williamsburg. Question-naire with tabulated responses.Williamsburg, va.: Colonial Williams-burg, 1964, 3p.

See also #’s 254-258

260 Corning Glass Center. "visitor Sur-vey--Hall of Science and Industry"questionnaire. Corning, N.Y.: Cor-ning Glass Center, 1971, ip.

See also #’s 261-263

261 Hares, Doris. "Summary of the Tea-cher Evaluation of the Corning GlassCenter Tour Program for the Year1968-1969." Corning, N.Y.: ComingGlass Center, July i0, 1969, 12p.

See also #’s 260, 262, 263

262 Corning Glass Center. "The GlassCenter visitor." Corning, N.Y.:Corning Glass Center, n.d., 7p.

See also #’s 260, 261, 263

263 Wake, Robert B. Memorandum to J.P.Fox, Jr., re. "Results of VisitorSurvey July 3,4, and 5, 1971," datedJuly 6, 1971 and Memorandum to J.P.Fox, Jr., re. "Results of VisitorSurvey July 6 and 7, 1971," datedJuly 8, 1971. Corning, N.Y.: Cor-ning Glass Center, 1971, 5p. + 4p.

See also #’s 260-262

264 Wells, Carolyn H. Smithsonian visi-to_~r. Washington, D.C.: Smithsonian

266 Detroit Institute of Arts. "HighSchool Survey--Spring, 1969" ques-tionnaire with tabulated results andletter from virginia R. Harriman toteachers. Detroit: Detroit Insti-tute of Arts, 1969, 2p. + ip.

See also #’s 244, 245, 267, 268

267 Downs, Linda. Memorandum to Missvirginia Harriman, re. "Results fromQuestionnaire Sent to Detroit GradeSchools Regarding the Interest inthe Tour Program," dated Apr. 22,1970. Detroit: Detroit Instituteof Arts, 1970, 3p.

See also #’s 244, 245, 266, 268

268 Detroit Institute of Arts~ Question-naire for teachers on "Ancient World"tour and "Art to the Schools Evalu-ation Form." Detroit: Detroit In-stitute of Arts, n.d., ip. + Ip.

See also #’s 244, 245, 266, 267

269 (National Association of the DenverMuseum of Natural History.) Reportto the Board of Trustees~.DenverMuseum of Natural History on Surveyof visitors to the Denver Museum ofNatural History, 197~. Introduction,Summary, and questionnaire form.(Denver: National Association ofthe Denver Museum of Natural History,1974), pp. 1-2, 17-21.

270 Ewell, Maryo and Peter Ewell. "Plan-ning for Grassroots Arts Development:A Research Study of Nine Communitiesin Transition," Arts in Society .12(1): 92-109 (1975).

Rural communites in wisconsin

102