Patterns of Cultural Consumption Behavior

PATTERNS OF CULTURAL CONSUMPTION BEHAVIOR

Donald E. Sexton, Columbia University

ABSTRACT -

Five data sets concerning ballet, opera, symphony, theater, and museums were analyzed to identify segments whose members showed different patterns of cultural consumption. The results for the various data sets were quite similar with respect to the dimensions most informative in defining segments and with respect to segmentation structure. These findings should be of use to administrators of cultural institutions as they formulate marketing strategies and gather information fir marketing planning.

INTRODUCTION

Many studies regarding attendance at cultural institutions have been reported (DiMdggio, et. al., 1978). Most such studies have focused on cultural consumption of but one art form, comparing, for example, those individuals attending frequently and infrequently. The work reported here has a broader focus - the individual's pattern of consumption of several types of culture.

Frequencies of attendance by an individual at a number of cultural events were used as a basis for market segmentation. Similar analyses were performed on a variety of data sets. Two main issues were addressed: (1) Are there specific cultural events that appear to be the key identifiers of segments whose members display differing patterns of cultural consumption? (2) Is there a segmentation structure general to the overall market for cultural events?

DATA BASE

The larger of the data bases consisted of a sample of 41,099 individuals. In 1974, the National Research Center of the Arts, Inc. (NRCA), under a grant from the New York State Council on the Arts, distributed self-administered questionnaires to individuals while they were attending performing arts events and museums. More than 30,000 completed questionnaires were received from the performing arts audiences and over 10,000 from the museum visitors. Among other information, the questionnaires included responses regarding frequencies of attendance at various cultural events over the past twelve months.

Not all the 41,099 observations were analyzed for the results reported here. Due to computer size limitations, various randomly chosen subsamples were used as data sets. Specifically, some of the findings were based on 8,000 of the performing arts respondents and all 10,888 of the museum respondents. Other findings were obtained by analyzing 950 of the respondents who were attending ballet, 950 of those at the opera, 999 of those at the symphony, and 950 of those at museums.

A second data base was collected by the author in 1978 to estimate the size of the market for theater in Pittsburgh. A random sample of families in the Pittsburgh area skewed toward those with higher incomes and education levels, was selected. Nine hundred and sixty telephone interviews were completed. Frequencies of attendance at several cultural events during the preceding year were obtained for each respondent.

SEGMENTATIONDIMENSIONS

To identify those events to be used as segment dimensions, various portions of the NRCA data set were examined. Twelve-month attendance frequencies for the ten cultural events listed in Table 1 were factor analyzed (with orthogonal rotation).

The analysis was repeated for all the museum respondents (10,888), a random sample of the performing arts respondents (8,000), and a random sample of the respondents who were attending ballet when they completed the questionnaire (950). The results were quite similar for these three data sets.

In each analysis, three factors emerged to account for nearly all of the variation. Moreover, the factor loadings were similar in pattern and suggested that factor I be labeled the Classical Event factor, factor II the Museum factor, and factor III the Dance factor. Frequencies of attendance at opera, history museums, and ballet appeared to be the key variables in each factor. From a practical viewpoint, attendance frequencies at those three events - at a minimum - should be asked in any study attempting to behaviorally segment the audience for a cultural event.

TABLE 1

FACTOR LOADINGS FOR EVENT ATTENDANCE FREQUENCIES

SEGMENTATION STRUCTURE

Four subsamples of approximately 1,000 respondents each from the NRCA data base and the Pittsburgh theater sample were each divided into segments by using the Howard-Harris cluster analysis. Dimensions employed included attendance frequencies at all events and attendance frequencies at only opera, ballet, art and history museums. Both sets of dimensions led to similar results.

Two- to ten-group cluster solutions were derived. Generally, for all five samples, beyond the four-group solution, the smaller clusters fragmented while the largest cluster remained unchanged. Based on the rules of thumb that the segments be interpretable and that subsequent splits not substantially alter the largest existing cluster, the four-group solution was selected as the most informative partitioning for each of the five samples.

Inspection of the average attendance frequencies for the various cultural events for each segment suggested labels. For example, the ballet sample divided into (1) the "Lights" (who rarely go out), (2) the "All-Rounders" (who go everywhere), (3) the "Museum Goers" and (4) the "Dance Specialists" (Table 2).

TABLE 2

MEANS OF ATTENDANCE FREQUENCIES IN PAST 12 MONTHS

As shown in Table 3, the Light and All-Rounder segments were found for each sample, as one would anticipate. For the four NRCA samples, the remaining segments consisted of specialists - those who were interested primarily in either dance, opera, art museums, or history museums and who attended all other cultural events relatively infrequently.

These cultural specialist segments were also likely present in the Pittsburgh sample, but in small numbers. Unlike the NRCA date which consisted of respondents questioned while attending a cultural activity, the Pittsburgh sample was a random telephone survey of the population at large and would therefore include fewer specialists. The third and fourth segments in the Pittsburgh data were those who attended movies and professional sports. There appeared to be several segments of Movie-Goers. As more clusters were identified, these Movie-Goer segments differed mainly on frequency of movie attendance.

What is interesting about these findings is the simplicity of the segmentation structure - Lights, All-Rounders, and Specialists. In none of the data sets did any segment of appreciable size appear that consisted of frequent attenders at, say, two, three, or four cultural events. Overall, the respondents either went everywhere infrequently, everywhere often, or one type of place often. Such a segmentation structure -if it is generalizable - suggests cultural organizations may team with other, different cultural organizations to attract the All-Rounders or the Lights but must use their own resources (or those of organizations in the same cultural area) to pursue the rest of their market - the Specialists.

TABLE 3

SEGMENTATION STRUCTURES FOR FIVE SAMPLES

CONCLUSIONS

For various data sets, similar segmentation dimensions and similar segmentation structures were identified. While these results were in basic agreement, the analyses must be replicated with data collected in other geographical areas and with other methodologies before they can be generalized. Still, these findings suggest there may be key cultural events distinguishing differing patterns of cultural consumption and there may be relatively few basic patterns of cultural consumption.

REFERENCE

DiMaggio, Paul, Useem, Michael, and Brown, Paula (1978), Audience Studies of the Performing Arts and Museums, Washington, D.C.: National Endowment for the Arts.