Richard R. Peterson, Pamela C. Hail and Roger H. Kern
*w
NATIONAL
ENDOWMENT
FOR THE ARTS
RGE RHD RUTS
PflflTICIPH
982-199/
Digitized by the Internet Archive
in 2012 with funding from
Boston Library Consortium Member Libraries
http://archive.org/details/ageartsparticipa2000pete
Research Diuision
Deport #42
AGE AND ARTS
PARTICIPA
582-1 99?
*»
NATIONAL
ENDOWMENT
FOR THE ARTS
Seven Locks Press
Santa Ana, California
Diehard A. Peterson
Pamela C. Hull
Roger H.Kern
Age and Arts Participation: 1982-1 997 is Report #42 in a series on matters of interest to the arts
community commissioned by the Research Division of the National Endowment for the Arts.
First printed in 2000
Library of Congress Cataloging-in-Publication Data
Peterson, Richard A., 1932-
Age and arts participation: 1982-1997 / Richard A. Peterson, Pamela C. Hull, Roger M. Kern
p.cm.
Includes bibliographical references.
ISBN 0-929765-86-9
1. Arts audiences-United States-Statistics. 2. Arts surveys— United States. 3. Baby boom
generation-Statistics. I. Hull, Pamela C, 1973- II. Kern, Roger M., 1967- III. Title.
NX220 .P48 2000
381'.457'0097309048-dc21
00-045625
Printed in the United States of America
Seven Locks Press
Santa Ana, California
800-354-5348
TABLE OF CONTENTS
Chapter 1.
Chapter 2.
Chapter 3.
Chapter 4.
Chapter 5.
Notes
Bibliography
List of Tables
Executive Summary
Introduction: The "Aging Arts Audience" Question
The Changing Age of the Arts Audience
The Contribution of Baby Boomers to
the Arts Audience
The Importance of Age as a Determinant
of Arts Participation
Correlates of Baby Boomer, Pre-Boomer,
and Post-Boomer Participation
vn
1
10
16
29
42
55
67
71
LIST OF TABLES
Table 1.1.
Table 1.2.
Table 2.1.
Table 2.2.
Table 2.3.
Table 2.4.
Table 2.5.
Table 2.6.
Table 2.7.
Table 2.8.
Table 2.9.
Table 2.10.
Table 2.11.
Table 2.12.
Table 2.13.
Table 2.14.
Median Age of Attendees, Overall Sample
and Benchmark Arts 13
Expected Age Group and Cohort Distributions 15
Age Group % Contribution to Total
Classical Music Attendances, by Year 18
Age Group % Contribution to
Total Opera Attendances, by Year 19
Age Group % Contribution to
Total Musical Attendances, by Year 20
Age Group % Contribution to
Total Jazz Attendances, by Year 21
Age Group % Contribution to
Total Theater Attendances, by Year 21
Age Group % Contribution to
Total Ballet Attendances, by Year 23
Age Group % Contribution to
Total Art Museum Attendances, by Year 23
Classical Music - Difference between % of
Total Times Attended and % of People Attending 25
Opera Music - Difference between % of
Total Times Attended and % of People Attending 25
Musical - Difference between % of
Total Times Attended and % of People Attending 26
Jazz - Difference between % of
Total Times Attended and % of People Attending 26
Theater - Difference between % of
Total Times Attended and % of People Attending 26
Ballet - Difference between % of
Total Times Attended and % of People Attending 27
Art Museum - Difference between % of
Total Times Attended and % of People Attending 27
VIII
Age and Arts Participation: 1982-11197
Table 2.15.
Table 3.1.
Table 3.2.
Table 3.3.
Table 3.4.
Table 3.5.
Table 3.6.
Table 3.7.
Table 3.8.
Table 3.9.
Table 3.10.
Table 3.11.
Table 3.12.
Table 3.13.
Table 3.14.
Table 4.1.
Sum of Seven Benchmark Arts with Attendees
Attending more Frequently than Expected
(Possible Range 0-7) 27
Cohort % Contribution to Total Classical Music
Attendances, by Year 31
Cohort % Contribution to Total Opera
Attendances, by Year 33
Cohort % Contribution to Total Musical
Attendances, by Year 33
Cohort % Contribution to Total Jazz
Attendances, by Year 34
Cohort % Contribution to Total Theater
Attendances, by Year 34
Cohort % Contribution to Total Ballet
Attendances, by Year 35
Cohort % Contribution to Total Art Museum
Attendances, by Year 35
Classical Music - Difference between % of
Total Times Attended and % of People Attending 38
Opera Music - Difference between % of
Total Times Attended and % of People Attending 39
Musical - Difference between % of
Total Times Attended and % of People Attending 39
Jazz - Difference between % of Total Times
Attended and % of People Attending 40
Theater - Difference between % of Total Times
Attended and % of People Attending 40
Ballet - Difference between % of Total Times
Attended and % of People Attending 40
Art Museum - Difference between % of Total Times
Attended and % of People Attending 41
Regression Results of Number of Attendances
on Age (standardized coefficients) 48
list of Tables
IX
Table 4.2.
Table 4.3.
Table 5.1.
Table 5.2.
Table 5.3.
Table 5.4.
Table 5.5.
Table 5.6.
Table 5.7.
Table 5.8.
Table 5.9.
Regression Results of Whether One Attends or
Not on Age (standardized coefficients) 49
Regression Results of Summary Measure of
Attendance on Age (standardized coefficients) 53
Regression Results of Classical Music Attendances
on Age by Cohort (standardized coefficients) 57
Regression Results of Opera Attendances on
Age by Cohort (standardized coefficients) 58
Regression Results of Musical Theater Attendances
on Age by Cohort (standardized coefficients) 58
Regression Results of Jazz Attendances on Age
by Cohort (standardized coefficients) 59
Regression Results of Theater Attendances on
Age by Cohort (standardized coefficients) 59
Regression Results of Ballet Attendances on Age
by Cohort (standardized coefficients) 60
Regression Results of Art Museum Attendances
on Age by Cohort (standardized coefficients) 60
Regression Results of Summary Arts Attendances
on Age by Cohort (standardized coefficients) 64
Regression Results of Summary Arts Attendances
(without Jazz) on Age by Cohort
(standardized coefficients) 64
EXECUTIVE SUMMARY
On February 12, 1996 an article titled "As Patrons Age, Future of Arts is
Uncertain" appeared in the New York Times (Miller 1996). It galvanised attention
on the question of the aging of arts audiences in the United States. The findings of
the National Endowment for the Arts Research Division Report #34, Age and Arts
Participation, released that same year, largely supported this assertion and helped
to energize the debate over the aging of arts audiences. While many interested in
arts policy echoed the fears of aging, the findings were hotly contested by some arts
presenters who said that they did not perceive their audiences as aging.
To bring further light to these issues, the National Endowment for the Arts com-
missioned Demographic Data Consultants of Nashville to revisit the issue of the
age of arts audiences with the newly available data from the 1997 Survey of Public
Participation in the Arts. Since Research Division Report #34 had made particular
note of the low rates of arts participation of the baby boom generation (those in
the United States born between 1946 and 1965), the Endowment asked
Demographic Data Consultants to pay particularly close attention to the arts par-
ticipation of baby boomers.
This executive summary highlights the prime findings of that study. It is divided
into five parts, each highlighting the key findings of the corresponding chapter in
the full report.
H g e and Arts Participation 1982-19117
Chapter 1
Chapter 1 sets the scene for the monograph by taking a first look at the age of
the audiences for the seven benchmark performing art forms in 1982, 1992, and
1997. Since the average age of the United States population has been increasing
over this span of years, the age of the arts audience for each of the seven bench-
mark art forms is compared with the age of the sample as a whole in each year.
The evidence for the three following conclusions can be found in Table 1.1.
• The audiences for all art forms, except opera, are aging faster than did the
entire sample.
• In 1982 only the opera audience was older than the entire sample. By 1997
the audiences for all art forms were older than the sample except for jazz,
and museum-goers have the same average age as the entire sample.
• The jazz audience is aging most rapidly. In 1982 it was eleven years
younger than that of the whole sample, by 1997 it was just two years
younger.
Chapter 2
In Chapter 2 we ask what distribution of young and older people we "see" if
we look out over the average audience for each benchmark art form in 1982, in
1992, and again in 1997? Is there indeed, a higher proportion of the audience with
graying hair, or are the larger numbers of young people born since World War II
taking the places of their elders in arts audiences over the span of years from 1982
to 1 997? The answer depends on which of the seven art forms one is talking about,
so each will be discussed separately. The data from which the following conclu-
sions are drawn are from Tables 2.1 through 2.7.
• The classical music audience is aging faster than the population as a whole.
In 1982 those under thirty years of age comprised 26.9 percent of the audi-
ence and by 1997 comprised just 13.2 percent of the audience. Over this
same span of years, those over sixty years of age rose from 15.6 percent to
30.3 percent of the classical music audience.
• By 1 997, a higher proportion of the classical music audience was over sixty
than was the audience for any other performing art form.
• In 1982 those under thirty years of age comprised just 17.8 percent of the
opera audience and by 1997 comprised only 13.3 percent of the audience
for opera. Over this same span of years, audience members over sixty rose
from 16.6 percent to 23.5 percent of the opera audience.
ExecutiuG Summary
While the opera audience was the oldest in 1982 and aged somewhat
through the years to 1997, it was the one art form whose audience aged
less rapidly than did the population as a whole.
The dynamics of the Broadway musical theater audience aging is similar to
that seen above but not as dramatically as for classical music. In 1982,
those under thirty years of age comprised 27. 1 percent of the audience and
by 1997 comprised just 16.2 percent of the audience. Over this same span
of years, those over sixty rose from 16.4 percent to 22.7 percent of the
musical theater audience.
In 1982 the jazz audience was unusually young in that just 5.0 percent of
the 1982 jazz audience was over sixty, while for all the other benchmark
arts, between 15 percent and 17 percent of the audiences was above sixty
years of age.
In 1982 fully 56.7 percent of the jazz audience was under thirty years of
age, but by 1997 these younger age groups had fallen dramatically to 23.2
percent of the jazz audience. Over the same span of years those over sixty
rose from 5.0 percent to 15.5 percent of the jazz audience.
In 1982 those under thirty years of age comprised 29.1 percent of the the-
ater audience, and by 1997 young people comprised just 16.7 percent of
the audience for theater. Over this same span of years, those over sixty rose
from 15.5 percent to 22.8 percent of the theater audience.
In 1982 34.1 percent of the ballet audience was younger than thirty years
of age; this was the second highest proportion after jazz. By 1 997 the pro-
portion of the ballet audience who were under thirty dropped to 16.1
percent, a level comparable to the other performing arts. Those sixty and
over comprised 15.4 percent of the ballet audience in 1982, and by 1997
had risen to 22.0 percent, a change comparable with the that taking place
for arts audiences generally.
In 1982 30.6 percent of arts museum attendances were by young people,
and by 1997 just 19.2 percent. This change is comparable with that of the
other art forms, but by 1997 art museums attracted the second highest pro-
portion of attendances by people under thirty (after jazz).
Unlike all of the other benchmark arts, the proportion of museum-goers
sixty and over has not increased appreciably over the fifteen years from
1982 to 1997. Later evidence introduced suggests the reason that not as
many older people frequent art museums is their impaired ability to walk
and stand for extended periods of time.
H g e and Arts Participation 1982-195?
Not all arts attendees attend with equal frequency. Observers have suggested
that older attendees are likely to buy season tickets while attendees under thirty
buy tickets event-by-event as time, money, and inclination dictate and thus gener-
ally attend less frequently. The information presented in Tables 2.8 through 2.14
show a somewhat more complex pattern:
• For opera, the predicted pattern held true in 1982 and 1997 with young
attendees attending somewhat less often than older attendees.
• For classical music, theater, and ballet, the pattern was the opposite in
1982 from what was expected. That is, young attendees attended more
often than older attendees. For all three forms, however, young attendees
attended less often than older attendees by 1997.
• In the case of jazz where season tickets are seldom offered, attendees in
their twenties attended more often than did both teenage attendees and
those sixty and above.
• Looking at all the art forms together (as seen in Table 2.15), we find that
arts attendees in their twenties get quite involved in the art forms of
their choice and attend often, while attendees in their thirties and forties
do not go as often, perhaps because of the competing demands of fam-
ily and work. Finally, in their later years, attendees come back to
attending more often.
Chapter 3
In Chapter 3 we follow respondents who are in the same birth cohort across the
survey years from 1982 to 1997. Two distinct predictions have been made about
the observed lower level of arts participation of baby boomers relative to their eld-
ers. One expectation is that they will "age into" arts participation as they embrace
midlife obligations and perspectives. The alternative prediction is that the lower
level of arts participation is a consequence of their early liberal experience and
will persist over the coming decades, while post-boomer cohorts, raised in a
more conservative atmosphere, will enjoy levels of arts participation compara-
ble to pre-boomers.
In 1982, baby boomer respondents were eighteen to thirty-six years of age, so
it was impossible to know what their mid-life experience would be. Now, fifteen
years later, is the appropriate moment to test these assertions. By 1997 post-
boomers comprised 21.5 percent, boomers 43.0 percent, and pre-boomers 35.5
percent of survey respondents.
fxecutiue Summary
The question asked in Chapter 3 is the same as the one asked in Chapter 2,
namely, what distribution of ages would be seen if you looked out at the typical
audience of each specific art form in 1982, in 1992, and again in 1997. The dif-
ference is that here we ask what cohorts, rather than age groups would
predominate. The following points highlight the findings shown in Tables 3.1
through 3.7. Starting with the youngest, cohorts are considered in turn.
• Post-boomer cohort 1976-1980: It is encouraging for future arts partici-
pation that this youngest group in 1997 attended five of the seven
benchmark arts somewhat more often than the sample as a whole, and
they were under-represented in the audience of only one art form, opera.
• Post-boomer cohort 1966-1975: This young cohort attended six of the
seven art forms less often than the sample as a whole, but their under-
representation was less pronounced in 1997 than in 1992 for three forms,
opera, musicals, and art museums.
• Late-baby boomers 1956-1965: Except for jazz, which more often appeals
to the young, late-boomers are clearly under-represented in arts audiences.
• Early-baby boomers 1946-1955: In marked contrast to late-boomers,
early boomers, contrary to all expectations, were over-represented in the
audiences of six of seven forms in 1982 and in 1997 as well. This starkly
contradicts the finding of Research Report #34. The reason for the differ-
ence is that in that analysis the authors took into account the participation
rates that the various cohorts should have attained given their education
and income. Given the boomer cohort's better education and higher
income relative to prior cohorts, their arts attendance was markedly less
than would be expected. Considering together the findings of these two
studies, demonstrates that early boomers attended the arts more often than
earlier cohorts but not nearly so often as would be expected given their
educational and financial advantages.
• Neither early- nor late-boomers are attending the arts more often as they
age, bringing into question the assumption that the boomer and later
cohorts will age into arts participation.
• War and Great Depression cohorts 1926-1945: Members of these cohorts
attend most art forms more often than the sample as a whole with the
exception of jazz.
• Roaring 20s cohort 1916-1925: Members of this cohort were at least
seventy-two in 1997, so it is not surprising that their arts attendance was
lower than for respondents as a whole in five of the seven art forms. The
6 Age and Arts Participation 1 9 B 2-1 9Q 7
only two forms in which they were somewhat over-represented in 1997
were classical music and opera.
• Pre-World War I cohort born before 1916: The youngest of this
group were sixty-seven in 1982 and eighty-two in 1997, so their under-
representation in audiences for all the art forms is not surprising.
Attendance does not go down gradually with advancing age, instead, it
plummets as one approaches seventy years of age.
In a fashion parallel to Chapter 2 where the focus was on age groups, the next
question is whether arts attendees of each of the cohorts attend more frequently
than would be expected by comparing their attendance rates with the average rate
of attendance of the entire sample. Because the findings suggest such groupings,
we will focus on three groups of adjacent cohorts: baby boomers, those born
before the baby boomers (pre-boomers), and those born since the baby boomers
(post-boomers). The conclusions drawn here are based on data in Tables 3.8
through 3.14.
• Among post-boomers, it was only for jazz that a few respondents account
for the cohorts' attendance. For all the rest of the art forms, a larger num-
ber of Generation X attendees attended fewer times on average. This
means a large number of people were sampling widely.
• For baby boomers generally, a large number of attendees attend infre-
quently and this trend grew more pronounced from 1982 to 1997. That
means that these boomers, like the post-boomers noted above, tend to
sample widely without showing a strong commitment to any arts form.
• In marked contrast, the attendance figures for pre-boomers are accounted
for largely by the frequent attendance of a relatively few people in these
cohorts.
Chapter 4
What is the importance of age relative to other factors in determining arts atten-
dance? This is the question addressed in Chapter 4. Ordinary Least Squares
regression analysis was used to look at the effects of age while controlling for the
effects of a number of other measured variables for each of the seven benchmark
arts. Also used, was a summary measure of the attendance at all the arts together.
It was important to control for the effects of the other factors because the
direct relationship between age and arts attendance was inconsistent and weak.
This was not unexpected because it is well known that arts participation tends
Executiue Summary
to rise gradually from the thirties through the sixties and then falls rapidly after
that age. However, with the controls in place, the results became clear and
strong as can be seen in Tables 4.1, 4.2 and 4.3. These results suggest that it is
not age per se but the many factors often associated with stages in the life-cycle
that influence arts participation.
• There is a significant positive relationship between age and arts attendance
for each of the art forms except jazz when the effects of the control factors
are taken into account. This result is even stronger when considering the
summary measure of arts attendance. This means that older persons attend
all the art forms except jazz more often than do younger people of the same
education, gender, marital status, income, etc.
• Education is, in every instance, the best predictor of participation in each
art form separately and also when they are combined together in the sum-
mary measure.
• Age is the second best predictor of arts attendance in four of the bench-
mark forms, and it is a significant predictor in every case but jazz.
• Age is the fourth most important predictor of the summary arts participa-
tion measure after education, income, and gender.
• The importance of the other control variables varies from art form to art
form in interesting ways. Since these findings are not directly relevant to
the contribution of age to arts participation, the reader is referred to the
text and tables to see their place in the mix of arts predictors.
Chapter 5
To what extent do the same factors determine arts participation for persons of
differing ages? A number of potential differences come to mind. For example,
respondents in their thirties are more likely to have their arts participation reduced
due to the presence of young children in the home. At the same time, chronic ill
health is likely to be a factor for more older people than for those who are young.
To assess the importance of influences during major phases of the life-cycle. As in
Chapter 4, the sample was split the into three parts: baby boomers, pre-boomers,
and post-boomers.
The same sort of analysis was performed as in Chapter 4 with one small varia-
tion, post-boomers under the age of 25 were asked whether they were full-time
students at the time they were surveyed. This variable is included for the post-
boomer regression analyses. The results of the OLS regression analyses for each of
the three age-groups on the seven benchmark arts are shown in Tables 5.1 through
8 Rge and Arts Participation 1982-1997
5.7. For a succinct summary of the findings for each art form please refer to the
text of Chapter 5. The most significant general findings are as follows:
• The set of variables taken together proved to be better predictors of pre-
boomers' arts participation than that of boomers or post-boomers. The set
of predictors of participation for those born during World War II or before,
including education, gender, and income, are not as important for younger
cohorts. The findings of Research Report #34 suggest that this is due to the
differences between the cohorts and is not simply a function of the age of
the respondents. Thus those interested in increasing the arts participation
of younger people will do well to look to factors other than those meas-
ured here.
• Even after dividing the sample into three parts based on age, being older is
a significant determinant of arts participation among pre-boomers or
boomers for five of the seven benchmark arts. The finding for pre-boomers
is surprising because, within that older group of respondents, age is nega-
tively correlated with arts participation. The finding of the regression
analyses mean that age is not, in itself, a deterrent to arts participation.
Rather, age is often associated with other causal factors such as health, edu-
cation, and income, which do correlate with arts participation.
• In twenty of twenty-one regression models, education is the best single pre-
dictor of participation in each of the arts. The single exception is ballet for
pre-boomers, where income, father's education, and being female are the
best predictors of attendance.
As in Chapter 4, a summary measure of arts participation was also created. The
respondent was given one point for each art form they had attended in the prior
year. Thus, since there are seven art forms, the variable ranges from 0 to 7. The
results for this summary measure of arts participation are shown in Table 5.8
• Even though the sample was divided into three parts on the basis of age,
age is still positively and significantly associated with arts attendance for
baby boomers and pre-boomers.
• The respondents' education was far and away the best predictor of arts
participation for all three age groups. The importance of education is fur-
ther underlined by the fact that the respondents father's education is
significantly associated with participation even for the pre-boomers who
fxecutiue Summary
were at least in their mid-fifties at the time of the survey. Further underlin-
ing the importance of education, the second most important predictor of
arts participation among post-boomers was being a full-time student.
Family income is the second most important predictor for boomers and
pre-boomers, and sixth most important for post-boomers.
Being female is the third most important predictor of the summary arts
measure for baby boomers and for pre-boomers, but gender hardly seems
salient for post-boomers.
Chronic health problems is the fourth best predictor for the pre-boomers
where we expected it to be most salient, but it is also the sixth best predic-
tor for boomers. As expected, health is not significantly related to arts
participation for post-boomers.
Finally, not being currently married, as expected, is significantly related to
arts participation but only for baby boomers and post-boomers.
CHAPTER 1 I HTflODUCTI OH: THE ICING ARTS AUDIENCE" QUESIION
Concern over the aging of performing and visual arts audiences in the United
States and elsewhere has been mounting over the past several decades. The ques-
tion was directly confronted in the 1996 National Endowment for the Arts,
Research Division Report #34 (hereafter referred to as Research Report #34). In
brief, the report found that in the decade from 1982 to 1992 there was clear evi-
dence of an aging audience in several of the performing arts disciplines, most
notably classical music and opera. This aging was due, the study found, largely to
the fact that "baby boomers" (people born between 1946 and 1965) did not attend
as often as would be expected from the participation rates of those born earlier.
The report evoked considerable comment and some criticism. Performing arts
professionals called for renewed efforts to attract younger people, and others pro-
moters questioned the findings, saying they did not see such a trend in their own
venues. At the same time, museum managers pointed with satisfaction to the find-
ing that museum audiences were not aging.
The recently completed National Endowment for the Arts 1997 Survey of
Public Participation in the Arts (hereafter SPPA) provides a good opportunity to
revisit the aging audience question and to take a fresh look at the question. For the
most part, therefore, we do not make the same kinds of analyses contained in
Research Report #37. There are three reasons for taking a different tack. First, the
1997 survey was fielded just five years after the prior one, so changes in the fun-
damental relationships reported in Research Report #37 cannot be great.
The second reason is to give those interested in arts audiences a more intuitively
clear way of seeing the aging question. Many practitioners in the arts world con-
sidered the complex multi-variant analyses reported in Part I, authored by Richard
Peterson and Darren Sherkat, to be incomprehensible, while they found Part II,
authored by Judy Balfe and Rolf Meyersohn too cluttered with tables relating age
of arts participants to other variables such as gender or education. Thus, though
many commentators accepted the conclusion that the performing arts audience is
aging, practitioners found it difficult to really understand the information they
needed and to apply the findings directly to their own situations.
The third reason why different methods are used here is that the 1997 arts par-
ticipation figures cannot be directly compared with those reported in the prior
surveys. This is because the methods of drawing the sample, choosing respondents,
and administering the questionnaire were all markedly different. For a detailed dis-
cussion of the differences see (NEA 1998a, 1998b). In consequence, the absolute
levels of arts participation in 1997 cannot be directly compared with those in
Chapter I 11
earlier years. The numerous comparisons we do make across years all compare the
proportion of the audience of a given age in one survey year with the proportion
in another survey year. Thus, to take one example from Table 1.1, which can be
found following the text of this chapter, the average age of the opera attendees was
43.5 years in 1 982 and 45 in 1 997, an increase of one and one half years. Over the
same span of years, the average age of all SPPA respondents — like that of the pop-
ulation of the United States — increased three years, thus, while the opera audience
aged, it did so only one half as fast as did the respondent sample as a whole.
Because the 1997 sample is not directly comparable with the other years, Chapters
4 and 5 focus entirely on the data from 1997.
In this and subsequent chapters we will use both the word "attendance" and the
word "attendee." Attendee refers to the individuals in the SPPA sample who report
participating in a particular arts activity at least once during the prior year.
Attendance refers to the aggregate of the participation of attendees in the particu-
lar arts activity. Thus, a person who reports going to four classical music concerts
in the prior year is counted as an attendee and contributes four units to the aggre-
gate attendance at classical music concerts for the year.
CHAPTER TOPICS
Chapter 2: The Changing Age of Arts Audiences. Chapter 2 shows the age dis-
tribution of the national audience for each of the benchmark art forms. Thus, for
the first time we will be able to see clearly the proportions of young, middle-aged,
and older people in the national audience for each art form in 1982, 1992 and in
1997. This kind of analysis makes any changes in the age of the audience for the
arts easier to see and to understand. In addition, focusing on the proportion of the
audience who are less than 20, 20-29, 30-39, etc. in each survey year circumvents
the problem, mentioned above, that the rates of reported participation vary widely
from survey year to year.
After first focusing on the age composition of arts attendance, the focus then
shifts to look at the age distribution of arts attendees. If young, mid-life, and older
attendees, on average, attend an art form the same number of times during a year,
the figures for "attendance" and "attendee" will be the same. If, however, as some
commentators assert, younger attendees sample a wide range of activities, attend-
ing any one art form less often than older participants who are more likely to buy
season tickets, then more young people and less older people will account for the
aggregate attendance.
Chapter 3: The Contribution of the Baby Boomers to Arts Audiences. One of
the prime concerns addressed in Research Report #34 was what seemed to be the
12 Age dud Arts Participation: 1QB2— 1Q97
lower arts participation of the baby boom birth cohort relative to their elders. A
birth cohort includes all those born in the same span of years, and the baby boom
cohort includes all those born between 1946 and 1965. Unlike an age group whose
members change every year, the members of a birth cohort always remain the same
as they age together year-by-year. Chapter 3 tracks the arts participation of the
baby boom, younger, and older cohorts as they age from 1982 to 1997. As in
Chapter 2 the national arts audiences for the seven benchmark arts is shown, but
here birth cohorts are focal, and the central question is, Do baby boomers form an
ever larger part of the arts audience as they mature? In a way parallel to Chapter
2, the focus is first on the cohort composition of the aggregate audience attendance
and then on the cohort composition of attendees.
Chapter 4: Age and Other Factors in Determining Arts Participation. As
Research Report #34 clearly showed, numerous other factors, such as gender, edu-
cation, income, health, combine with age in determining arts participation. To
more clearly distinguish the influence of age per se relative to the other factors,
regression analyses of participation are performed for each of the seven benchmark
arts. The focus is exclusively on the 1997 data to alleviate the incomparability
problem discussed above. The first section of Chapter 4 shows the influence of age
on participation relative to each of the other controlled factors. The final section
of Chapter 4 looks at the contribution of each of the control variables to predict-
ing arts participation.
Chapter 5: Factors Differentially Affecting Baby Boomers, Pre-Boomers, and
Post-Boomers Arts Participation. Different factors influence arts participation
over the life course. Divorce or death of a spouse is not often a problem for the
young, and children in the home is not often a problem for those past mid-life, for
example. To show the differential influence of the control factors for baby boomers
in comparison with pre- and post-boomers, regression analyses of participation are
performed showing the independent influence of age on participation, net of the
controlled factors. The focus here again is on the 1997 data to alleviate the incom-
parability problem discussed above.
FIRST LOOK AT ARTS AUDIENCES
The Median Age of Arts Attendees
While most of this monograph looks at age groups and birth cohorts, it is use-
ful to begin inspecting the data by looking at the median age1 of the audience for
each of the benchmark arts in the survey years 1982, 1992, and 1997. In effect, a
survey of the audiences at classical music concerts, ballet performances, opera, etc.
in 1982, 1992, and 1997 to see if they have aged over this period of time.
Chapter 1 13
The average age of the United States population has been increasing, so the key
question is whether the audience for each of the art forms is rising slower or
taster than is the population as a whole. The top line of Table 1.1 shows that
the median age of the survey respondent was 40 in 1982, 42 in 1992, and 43 in
1997.- These changes mean that, like the United States population as a whole,
the average age of survey respondents was rising one year in every five years
between 1982 and 1997.
Table 1.1.
Median Age of Attenders, Overall Sample and Benchmark Arts*
1982
1992
1997
Net Gain
Overall Sample
40
42
43
3
Opera
43.5
45
45
1.5
(+3.5)
(+3)
(+2)
Classical
40
45
46
6
(0)
(+3)
(+3)
Musical
39
43
44
5
(-D
(+1)
(+D
Theater
39
44
44
5
(-D
(+2)
(+D
Ballet
37
40
44
7
(-3)
(-2)
(+D
Museum
36
40
43
7
(-4)
(-2)
(0)
Jazz
29
37
41
12
(-11)
(-5)
(-2)
*Values in parentheses indicate years of difference from median age
These three median ages, 40 in 1982, 42 in 1992, and 43 in 1997, can be used
as the "expected" age of arts attendees for the three survey years. These ages would
be found if persons of all ages had attended the same amount. Any departure from
these expected figures signals that the audience for a benchmark art form is older
or younger than expected in the particular survey year. To be able to see these dif-
ferences at a glance, departures from the expected age are noted in parenthesis in
Table 1.1.
Taking a more detailed look at table 1.1, we can see the changing age of arts
audience in several ways.3 Looking at the column for 1982 we find that jazz, with
an audience averaging 29 years old, had by far the youngest of any of the bench-
14 Age and Arts Participation 1982-151?
mark arts. Opera at 43.5 years had the oldest audience, and the audiences of the
other five forms were at or below the average age of the sample as a whole.
The 1 997 column shows quite a different picture. While the sample as a whole
has aged three years, the audience for jazz has aged twelve years, nearly reaching
the age of the sample as a whole. What is more, the age of the audiences for all of
the other benchmark arts in 1997 is at, or above, the age of the sample as a whole.
Comparing the differences between the left-hand column for 1982 with the right-
hand column for 1997 shown in the "Net Gain" column at the right, we find that,
except for opera, the average age of audience members in all the arts has risen
faster than that for the sample as a whole, and that after jazz, this relative rise has
been greatest for the ballet, art museum, and classical music audiences.
Expected Age and Cohort Distributions
Just as it is possible to form an "expected" average age of arts participants, it is
possible to find the "expected" proportion of participants in each age group and
cohort during each of the three survey years. Thus, if each age group and cohort
contain the same proportion of attendees as all of the others in the survey year we
will find the proportions shown in Table 1.2. For example in 1982 we expect 20.6
percent of arts audiences to be in their thirties and in 1997 we expect 22.9 percent
of arts audiences in their thirties.
Linking these expected figures to those observed in the survey samples will be
focal in Chapters 2 and 3, but the expected distributions should be inspected before
moving to those chapters. Looking first at the upper half of Table 1.2, note the
clear decline of those between eighteen and twenty-nine, seen by comparing the
1997 column with that for 1982 and the corresponding rise of those in their for-
ties. The decline in the youth and increase in mid-lifers reflect the aging of the large
baby boom cohort over this fifteen-year period.4
Chapter 3 focuses on the changing contribution of Baby Boomers (respondents
born in the twenty years between 1946 and 1965), to the arts audiences over the
years from 1982 to 1997, and it compares these figures with the proportions that
would be expected if members of every cohort, on average, attended equally. These
expected frequencies are shown in the bottom half of Table 1.2. Over these fifteen
years boomers have comprised just over 40 percent of SPPA respondents. In 1982,
post-boomers were not old enough to be surveyed, but by 1997 they comprised
21.5 percent of the survey sample and pre-boomers ranks shrank proportionately
over the same time period.
Chapter I 15
Table 1.2.
Expected Age Group and Cohort Distributions'
Age Group Distribution of Samples (%)
1982 1992 1997
18-19
4.8
3.0
2.1
20-29
24.0
18.2
15.8
30-39
20.6
23.0
22.9
40-49
14.9
18.5
21.6
50-59
14.4
13.1
14.6
60-69
11.6
11.8
10.7
70 & over
9.6
12.4
12.4
Total 99.9 100.0 100.1
Cohort Distribution of Samples (%)
1982 1992 1997
1976-1980
4.6
1966-1975
15.2
17.4
1956-1965
22.1
22.5
23.5
1946-1955
21.8
20.3
20.7
1936-1945
16.0
14.3
12.9
1926-1935
14.5
12.0
10.6
1916-1925
12.9
10.0
7.7
Before 1916
12.7
5.6
2.6
Total 100.0 99.9 100.0
*Some column percentages do not total to 100.0 due to rounding.
CHAPTER 2: THE CHANGING AGE OF THE ARTS RUDI E HCE
In Chapter 1 it was observed that the average age of the audience for each of
the benchmark arts has increased between 1982 and 1997. Here in Chapter 2 this
observation is examined in greater detail by dividing the total sample into age
groups of ten years in length — twenty year-olds, thirty year-olds, etc. We focus on
the changes for each ten-year-long age group. In the first section the proportion of
the total audience represented by each of the age groups is discussed, that is to say
"attendance" as the word is defined in the 1997 Survey of Public Participation in
the Arts. In this analysis, the total number of attendances reported by survey
respondents is tallied, and the degree to which this representation is greater or less
than would be expected by chance is observed. In the second section of Chapter 2,
the focus is on attendees to see the degree to which a few respondents account for
the aggregate audience for the year by attending the arts many times and whether
the average number of attendees varies by age and art form. Only those relation-
ships that are statistically significant are discussed in the text.
PARTI ATTENDANCE
The concept of attendance used here represents what one is likely to see look-
ing out over the audience in each of the arts and observing what percent of the
audience is in their twenties, thirties, etc. This look is taken for the survey years
1982, 1992, and 1997. This seems such a simple way of asking about aging of the
audience; it is a wonder that it has not been focal in earlier survey reports. The rea-
son is that prior reports have asked other questions of the data.5 What is more,
though conceptually simple, answering this question involves a great deal of man-
ual calculation.10 To get this value we first ran cross-tab frequencies of age group
or birth cohort by the number of attendances, to learn how many respondents in
each age group or cohort attended each art form every number of times. For exam-
ple, in the 20-29 age group in 1997, 95 respondents attended jazz one time, 75
attended two times, 48 attended three times, etc., up to the maximum possible
value of 72 times in one year. Next, each value in the frequency cell was multiplied
by the corresponding number of attendances for each age group. (In the same
example, 95 times 1, 75 times 2, 48 times 3, etc.) Then these products were
summed for each age group or cohort to represent the total number of times that
respondents in this age group or cohort attended that benchmark art in the previ-
ous year. (For example, the 20-29 year old respondents attended jazz a total of
Chapter 2 I 17
1,084 rimes in 1997.) Finally, the sum for each age group or cohort was divided
by the summed total of every age group or cohort s attendances, that is, the total
number of attendances by all respondents in that year, to reflect the proportion of
attendances reported by each age group or cohort in relation to the others in each
sample year. (For example, the total for the 20-29 age group of 1,084 was divided
by the grand total of 5,123 attendances for the whole sample in 1997 to reflect the
21 percent share of total jazz attendances for this age group.)
Looking at the age group composition of arts audiences, recall that, as found in
Chapter 1, the survey population has aged somewhat from 1982 to 1997. To take
this aging of the population into account, the calculated percentage for each age
group with the expected percentages discussed in Chapter 1 and shown in Table
1.2 will be compared. In each of the tables displayed in Chapter 2, there is a num-
ber in parenthesis below each percentage. This represents the difference (plus or
minus) between the observed percentage and what would be expected if people of
all age groups attended the art venue with equal frequency.
Here and throughout the monograph, the art forms will be discussed in the fol-
lowing order: classical music, opera, musical theater, jazz, theater, ballet, and art
museum attendance. The tables are also presented in this same order.
The Classical Music Audience
Table 2.1 shows the age distribution of the audience attending classical music
concerts in 1982, 1992, and 1997. Looking across the second and third row of fig-
ures, the proportion of the audience in their twenties and thirties has gone down
dramatically over this fifteen year period, from 21.5 percent to 11.4 percent and
from 24.5 percent to 13.7 percent respectively.7 Over the same years the proportion
of respondents sixty years of age and older has nearly doubled, going from 15.6
percent to 30.3 percent.8
As noted above, the population as a whole has aged over this same period, but
an inspection of the Table 2.1 figures in parentheses shows that the audience for
classical music has aged faster than has the entire sample of respondents. The neg-
ative sign for those in their twenties shows that their attendance has always been
significantly below what is expected, and this difference has increased between
1982 and the two 1990s surveys. In addition, those in their thirties, whose atten-
dance in 1982 was 3.9 percent above what would be expected, had become 9.2
percent below what would be expected by 1997. The picture is reversed at the
other end of the age spectrum. In 1982, those over sixty were under represented,
but by 1997 the two oldest age groups attended classical music concerts more
than would be expected, given their proportions in the sample population. Taken
18 Age and Arts Participation 1982-191)7
Table 2.1.
Age Group % Contribution to Total
Classical Music Attendances, by Year
1982 1992 1997
18-19
5.4
1.2
1.8
(+0.6)**
(-1.8)**
(-0.3)
20-29
21.5
12.9
11.4
(-2.5)**
(-5.3)**
(-4.4)**
30-39
24.5
18.6
13.7
(+3.9)**
(-4.4)**
(-9.2)**
40-49
17.7
21.4
23.3
(+2.8)**
(+2.9)**
(+1.7)**
50-59
15.4
16.9
19.5
(+1.0)**
(+3.8)**
(+4.9)**
60-69
8.5
15.5
14.5
(-3.1)**
(+3.7)**
(+3.8)**
70 & over
7.1
13.4
15.8
(-2.5)**
(+1.0)*
(+3.4)**
Total 100.0 100.0 100.0
Values in parentheses indicate the difference of this observed percentage from the "expected"
percentage of a group in a total sample. These differences are statistically significant where indi-
cated (*p<.05; **p<.01). A statistically significant difference means that the probability of this
difference occuring merely by chance is less than 5% (for p<.05) or less than 1 % (for p<.01),
based on this sample. Therefore, we can reasonably conclude that this difference actually exists.
together, these findings show that the classical music audience is aging, and is aging
more rapidly than is the population as a whole.
The Opera Audience
A number of commentators have suggested that more young people have been
going to opera performances in the 1990s than were in the 1980s. The age distri-
bution of the audience attending opera performances in 1982, 1992, and 1997 are
shown in Table 2.2. Indeed, eighteen and nineteen year-olds composed one percent
of the audiences in 1982 and 1.6 percent in 1997. Though this increase is small,
the numbers in parenthesis are encouraging because in 1997 teens more nearly
approximate the expected contribution to the total audience for opera. Those in
their twenties show a different pattern. Their percent of the audience has fallen
from 16.8 to 11.7, but the differences from their expected attendance has nar-
rowed from -7.2 to -4.1.
Chapter 2 I 19
Table 2.2.
Age Group % Contribution to Total Opera Attendances, by Year
1982 1992 1997
18-19
1.0
1.2
1.6
(-3.8)"
(-1.8)"
(-0.5)
20-29
16.8
11.1
11.7
(-7.2)"
(-7.1)"
(-4.1)"
30-39
22.4
20.3
17.2
(+1.8)*
(-2.7)
(-5.7)"
40-49
22.9
23.1
27.5
(+8.0)"
(+4.6)"
(+5.9)"
50-59
20.2
18.7
18.4
(+5.8)"
(+5.6)"
(+3.8)"
60-69
5.9
17.1
11.7
(-5.7)"
(+5.3)"
(+1.0)
70 & over
10.7
8.6
11.8
(+1.1)
(-3.8)"
(-0.6)
Total 100.0 100.0 100.0
See footnote on Table 2.1 .
The proportion of those in their thirties has declined steadily, and their partici-
pation relative to the expected has moved from slightly positive (+1.8) to quite
negative (-5.7). Those in their forties have risen from 22.9 percent of the audience
to 27.5 percent, while the proportion in their fifties has remained nearly the same.
The proportion of the audience in their sixties and above has risen from 16.6 to
23.5 percent, and they have moved from being less represented in the audience
than would be expected (-4.4), to being a bit more than expected (+.4). Taken
together these figures suggest that the opera audience is older than that of the pop-
ulation, but it is not aging more rapidly than the sample as a whole.
The Audience Attending Musicals
The figures for the average age of the musical theater audience in Table 1.1
showed that, on average the audience for musicals is about the same as the total
sample and it has been aging at about the same rate as the total. The figures for the
musical theater audience by age group can be found in Table 2.3. They show that
both younger and older people are under represented. Unlike both classical music
and opera, the audience for musical theater is composed primarily of those from
30 to 59 years of age.
20 Age and Arts Participation 102-199?
Table 2.3.
Age Group % Contribution to Total Musical Attendances, by Year
1982 1992 1997
18-19
4.6
2.1
2.1
(-0.2)
(-0.9)**
(+0.0)
20-29
22.5
14.2
14.1
(-1.5)"
(-4.0)**
(-1.7)**
30-39
23.3
22.0
21.1
(+2.7)**
(-1.0)
(-1.8)**
40-49
16.6
21.8
22.8
(+1.7)**
(+3.3)**
(+1.2)*
50-59
16.6
15.7
17.1
(+2.2)**
(+2.6)**
(+2.5)**
60-69
10.3
13.8
12.1
(-1.3)**
(+2.0)**
(+1.4)**
70 & over
6.1
10.5
10.6
(-3.5)**
(-1.9)**
(-1.8)**
Total 100.0 100.0 100.0
See footnote on Table 2.1.
The Jazz Audience
The average age of the jazz audience is the youngest of all the benchmark arts,
as seen in Table 1.1. At the same time, jazz has experienced the greatest degree of
aging between 1982 and 1997. As Table 2.4 shows, in 1982 half of the audience
was in their twenties (49.6 percent), and 57.7 percent were under thirty years of
age. In 1997, however, barely one fifth (21.2 percent) were in their twenties and
less than one quarter (23.2 percent) were under thirty. All of the groups aged forty
and above have gained in their proportion of the jazz audience, but considering the
figures in parenthesis, all age groups fifty and older are still under represented in
the jazz audience, attesting to the fact that the audience for live jazz performances
is still younger than in the other art forms.
The Theater Audience
As seen in Table 1.1 the average age of those in the theater audience increased
five years, while the sample as a whole has aged three years between 1982 and
1997. This change, as seen in Table 2.5, is reflected across all the age groups. Those
age groups less than forty have a lower percentage representation in the audience
in 1997 than in 1982. Meanwhile, all those groups forty and older have increased
their representation in the audience, even those over seventy years of age, who in
1997 represented 11.5 of the audience, just less than one percent below the
expected percentage.
Chapter 2 21
Table 2.4.
Age Group % Contribution to Total Jazz Attendances, by Year
1982 1992 1997
18-19
7.1
1.7
2.0
(+2.3)"
(-1.3)"
(-0.1)
20-29
49.6
25.4
21.2
(+25.6)"
(+7.2)"
(+5.4)"
30-39
20.8
33.2
23.8
(+0.2)
(+10.2)"
(+0.9)
40-49
10.0
21.7
24.9
(-4.9)"
(+3.2)"
(+3.3)"
50-59
7.5
8.7
12.6
(-6.9)"
(-4.4)"
(-2.0)"
60-69
2.8
6.7
8.7
(-8.8)"
(-5.1)"
(-2.0)"
70 & over
2.2
2.6
6.8
(-7.4)"
(-9.8)"
(-5.6)"
Total 100.0 100.0 100.0
See footnote on Table 2.1.
Table 2.5.
Age Group % Contribution to Total Theater Attendances, by Year
1982 1992 1997
18-19
3.7
1.9
1.8
(-1-1)"
(-1.1)"
(-0.3)
20-29
25.4
13.7
14.9
(+1.4)"
(-4.5)"
(-0.9)
30-39
23.9
22.9
19.6
(+3.3)"
(-0.1)
(-3.3)"
40-49
17.9
21.7
23.8
(+3.0)"
(+3.2)"
(+2.2)*
50-59
13.7
16.6
17.2
(-0.7)
(+3.5)"
(+2.6)*
60-69
9.8
15.2
11.3
(-1.8)"
(+3.4)"
(+0.6)
70 & over
5.7
7.9
11.5
(-3.9)"
(-4.5)"
(-0.9)*
Total 100.0 100.0 100.0
See footnote on Table 2.1 .
22 Age and Arts Participation HB2-1QQ7
The Ballet Audience
As seen in Table 1.1 the average age of the audience for ballet was thirty seven
in 1982 and forty four in 1997, moving from three years younger than the entire
sample to one year older between 1982 and 1997. And just as we have seen in
looking at the theater audience, Table 2.6 shows that this aging has taken place
across all age groups. Those below forty represented 60.7 percent of the ballet
audience in 1982 and just 34.4 percent fifteen years later. At the same time, all age
groups forty and older have increased their proportion of the ballet audience. Only
those above seventy have not increased faster than would be expected from the
aging of the sample as a whole.
Art Museum Attendance
As seen in Table 1.1, the audience for art works has aged faster than for any
other art form except jazz. For the younger age groups particularly, as seen in Table
2.7, the pattern is like the audience for the other art forms. Respondents under
forty accounted for 56.1 percent of art museum goers in 1982 but just 39.7 per-
cent in 1997. Over the same period of years the proportion of respondents forty to
fifty nine have increased the most, increasing from 28.5 percent to 43.6 percent of
the museum audience. And, to a degree not seen for the other art forms, the
museum audience sixty and older has increased only slightly from 15.4 percent to
16.7 percent, far less than expected when compared to from the aging of the total
sample. Perhaps the physical activity involved in going through museum exhibits
deterred and continues to deter the participation of older people in this art form.
Chapter 2 23
Table 2.6.
Age Group % Contribution to Total Ballet Attendances, by Year
1982 1992 1997
18-19
2.4
3.8
2.0
(-2.4)"
(+0.8)
(-0.1)
20-29
31.7
16.1
14.1
(+7.7)"
(-2.1)
(-1.7)
30-39
26.6
26.1
19.3
(+6.0)"
(+3.U*
(-3.6)"
40-49
11.6
20.2
27.1
(-3.3)"
(+1.7)
(+5.5)"
50-59
12.4
10.4
15.5
(-2.0)"
(-2.7)"
(+0.9)
60-69
9.0
16.0
12.6
(-2.6)"
(+4.2)"
(+1.9)*
70 & over
6.4
7.5
9.4
(-3.2)"
(-4.9)"
(-3.0)"
Total 100.0 100.0 100.0
See footnote on Table 2.1.
Table 2.7.
Age Group % Contribution to Total Art Museum Attendances, by Year
1982 1992 1997
18-19
3.5
1.9
1.7
(-1.3)"
(-1.1)**
(-0.4)**
20-29
27.1
19.4
17.5
(+3.1)"
(+1.2)"
(+1.7)**
30-39
25.5
25.6
20.5
(+4.9)"
(+2.6)"
(-2.4)**
40-49
15.5
21.3
24.5
(+0.6)"
(+2.8)"
(+2.9)**
50-59
13.0
12.2
19.1
(-1.4)"
(-0.9)**
(+4.5)**
60-69
9.9
12.1
9.3
(-1.7)**
(+0.3)
(-1.4)**
70 & over
5.5
7.5
7.4
(-4.1)**
(-4.9)**
(-5.0)**
Total 100.0 100.0 100.0
See footnote on Table 2.1.
24 Hge and Arts Participation IQ82-1QQ7
Part 2 RTTEH DEES
Having looked at the age distribution of the arts audience as it has changed over
time, attention now turns from the question of "attendances" to the question of
"attendees." The question is whether attendees of all ages, on average, attend the
same number of times per year. This question is of importance because some com-
mentators have observed that the older people who do attend, attend more often
than young attendees do because, for example, they are more likely than younger
people to buy season tickets. Thus, to draw one hypothetical example, it may be
that fifty year old attendees are likely to attend five times each while those in their
twenties, on average, attend twice. If this were the case, two fifty-year-olds would
account for (2x5) =10 attendances, while it would take five twenty year olds to
account for 10 attendances (5x2) =10.
To quickly show the degree to which a few attendees may account for the lion's
share of the attendance, the proportion of attendees of a given age group is com-
pared with their contribution to the entire number of attendees. These comparisons
for the age groups in each benchmark art form are shown in Tables 2.8 to 2.14.
The figures in the tables are "difference scores" — that is the difference between an
age group's percentage of the total audience minus its percentage of all attendees.
The (+) sign in the upper-left-most cell of Table 2.8, for example, means that in
1982 relatively few teens accounted for the attendance of this age group at classi-
cal music concerts, because they represented a larger percentage of attendances
than of attendees. The (-) sign in the upper right-hand cell of the same table means
that in 1997 those teens who attended classical music concerts did so less often
than attendees in other age groups that year. The size of the numbers suggests
the degree to which a few or many people of an age group contributed to the
audience.9
The numerous no-zero differences scores shown in Tables 2.8 through 2.14
show that, indeed, attendees in different age groups do not all attend equally often.
The pattern of plus and minus scores, however, is more complex than the rela-
tionship suggested above. Table 2.8 shows that, in 1982, younger attendees went
to classical music concerts more often, while in 1997 older attendees went more
often. Tables 2.12 and 2.13 show that this pattern is roughly the same for theater
and ballet goers. Tables 2.9 and 2.10 show that opera and musicals attendees fit
the pattern described above with younger attendees going less often while other
attendees went more often all three survey years. The figures for jazz reported in
Table 2.11 show that attendees in their twenties consistently attended more often
than do those who are younger or older.
Chapter 2 25
In order to get an impressionistic summary measure of these difference scores,
the number of times attendees of a particular age attended more frequently than
would he expected across all seven benchmark arts in 1 982 and 1 997 was exam-
ined. The results of these tabulations are found in Table 2.15. The figures suggest
that attendees in their twenties attend often, while attendees in their thirties and
forties tend to attend less often. Finally, the figures show that those attendees who
are seventy years of age and older attend often as do those in their fifties and six-
ties in 1997.
Table 2.8.
Classical Music - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
+1.0*
-0.5
-0.3
20-29
+0.4
-1.4
-1.4
30-39
+0.7
-2.4*
-3.5
**
40-49
-0.2
-1.5
-0.8
50-59
-0.1
+0.5
+0.9
60-69
-1.9**
+2.7**
+ 1.9
*
70 & over
+0.3
+2.4*
+3.2
**
Total
100.0
100.0
100.0
These differences are statistically significant where indicated (*p<.05; **p<.01). A statistically sig-
nificant difference means that the probability of this difference occuring merely by chance is less
than 5% (for p<.05) or less than 1 % (for p<.01), based on this sample. Therefore, we can rea-
sonably conclude that this difference actually exists.
Table 2.9.
Opera Music - Difference between % of Total Times Attended
and % of People Attending
1982 1992 1997
Under 20
-2.2**
-0.9
-0.7
20-29
-1.5
-2.1
-3.2*
30-39
+3.3
-2.4
-2.1
40-49
+2.9
+ 1.8
+2.4
50-59
+1.7
+2.5
+0.6
60-69
-6.4**
+2.7
+ 1.3
70 & over
+2.1
-1.4
+ 1.6
Total 100.0 100.0 100.0
See footnote on Table 2.8.
26
Kge and Arts Participation: 1 9 B2-1 QQ 7
Table 2.10.
Musical - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
+0.4
-0.3
-0.4
20-29
-0.2
-1.6
+0.0
30-39
-0.8
-1.6
-0.8
40-49
-0.8
+0.1
-0.9
50-59
+0.5
+0.3
+ 1.0
60-69
+0.2
+0.8
+0.7
70 & over
+0.6
+2.3**
+0.3
See footnote on Table 2.8.
Table 2.11.
Jazz - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
-2.0**
-0.8
-0.2
20-29
+7.7**
+2.0
+3.1**
30-39
-1.6
+2.9
-0.3
40-49
-1.6*
+ 1.0
-1.4
50-59
-2.1**
-2.1*
-2.2*
60-69
-1.1*
-2.3*
+0.2
70 & over
+0.5
-0.7
+0.9
See footnote on Table 2.8.
Table 2.12.
Theater - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
-0.5
-0.7
-0.9*
20-29
+3.0**
-2.6*
+0.6
30-39
-0.1
-0.0
-1.5
40-49
+0.4
+0.2
+0.0
50-59
-2.3**
+ 1.0
+0.9
60-69
-0.2
+2.7**
+0.5
70 & over
-0.2
-0.6
+0.4
See footnote on Table 2.8.
Chap t g r 2
27
Table 2.13.
Ballet - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
-1.7*
+0.8
-0.4
20-29
+7.6"
-2.1
+ 1.3
30-39
-2.0
-1.4
-2.9
40-49
-4.4**
+ 1.2
+0.6
50-59
-0.1
-3.0
+0.3
60-69
-0.6
+4.3*
+ 1.5
70 & over
+ 1.3
+0.3
-0.3
See footnote on Table 2.8.
Table 2.14.
Art Museum - Difference between % of Total Times Attended
and % of People Attending
1982
1992
1997
Under 20
-1.0"
-0.7*
-0.8**
20-29
+0.2
-0.5
+ 1.1
30-39
+0.1
-0.8
-2.4"
40-49
-0.8
+0.4
+0.0
50-59
-0.7
-0.8
+3.4"
60-69
+ 1.2*
+ 1.7"
-0.4
70 & over
+0.9*
+0.8
-1.1*
See footnote on Table 2.8.
Table 2.15.
Sum of Seven Benchmark Arts with Attenders Attending more
Frequently than Expected (Possible Range 0-7)
1992
1997
Under 20
2
0
20-29
5
4
30-39
3
0
40-49
2
2
50-59
2
6
60-69
2
6
70 & over
6
5
28 Age and Arts Participation: 1 QB2-1 QQ7
Taken together, these patterns suggest that attendees in their twenties get quite
involved with one or more art forms, while attendees in their thirties and forties do
not go as often, perhaps because of the competing demands of family and job.
Furthermore, in their later years, attendees come back to attending more often.
There will be a better chance to understand why the frequency of attendance varies
by age in Chapter 5 where the predictors of arts attendance over the life-course are
explored.
CHAPTER 3 THE CONTRIBUTION Of BABY BOOMERS TO
TUT OUTS fl U D I E H C E
Are baby boomers (those Americans born between 1946 through 1965) less
likely than their elders to participate in the arts, as asserted in Research Report
#34? The 1997 data make it possible to see whether boomers are "aging into" arts
participation as some have predicted. In other words, if participation is due more
to the respondent's age than to their birth cohort, then the arts participation of
baby boomers will increase for the next twenty years.
This chapter focuses on the arts participation of boomers and the other birth
cohorts over the years from 1982 to 1997. Thus, while the changing age com-
position of each benchmark art's audience from 1982 to 1997 was central in
Chapter 2, here in Chapter 3 the focus is on the changing contribution of each
birth cohort to the arts audience.
In 1982, baby boomers were eighteen to thirty-six years of age, the youngest
respondents to that SPPA survey. By 1 997, baby boomers had become fully estab-
lished adults in the middle third of their lives, aged thirty-two to fifty one. In 1997,
boomers comprised 44 percent of the SPPA survey respondents, while the post-
boomers comprised 22 percent of respondents, and the pre-boomers comprised 34
percent of those surveyed.
As noted in Chapter 1, the term "birth cohort" or just "cohort" is the term used
by social scientists to refer to all those born in the same span of years. Unlike an
age group, say all 21 year olds, whose membership changes every year, the mem-
bers of a birth cohort always remain the same as they age together year-by-year,
decade-by-decade; thus once a baby boomer, always a boomer. It has been shown
that the experience of "late boomers" (those born between 1956 and 1965) has
been quite different from those of "early boomers" those born between 1946 and
1955 (Newman 1993). The early boomers grew up in the excitement of newfound
prosperity, the exuberance of the counterculture and the feeling that their lives
could make a difference in society. The late boomers grew up in the disillusioned
backwash of many of these ideas. While early boomers tended to easily find jobs
on the New Frontier or in the booming war plants, the late boomers entered a
much more competitive job market with most of the best jobs already taken by the
millions of early boomers. Since their formative experience was so different, we
analyze the arts participation of early and late boomers separately in this chapter.
To keep this same level of detail through out the age range, all birth cohorts are
defined as being those born in the ten years between the sixth year of a decade and
30 n g e and Arts Participation: 1 9B2-1 Q Q 7
the fifth year of the next decade. The only exceptions to this rule are for the cohorts
at the ends of the age range. Because of its rapid depletion, all those born in 1915
or before were lumped into one cohort. Those in this cohort were at least sixty-
seven years of age in 1982, and by 1997 they were at least eighty-two years old.
New cohorts have been added over the years, of course. In 1982 no cohort younger
than the boomers were old enough to be surveyed, but by 1997 there were two
new cohorts represented. Note however that this youngest 1997 cohort represents
only those born over a five-year period, this because only those born between 1976
and 1980 were at least eighteen years of age and thus eligible to be respondents.
Chapter 3 tracks the arts participation of all the cohorts as they age from 1982
to 1997. As in Chapter 2 the national arts audiences for the seven benchmark arts
is shown, but here birth cohorts are focal, and the central question is: do baby
boomers form a larger part of the arts audience relative to their proportion in the
whole sample as they mature? To get at this question the same approach is taken
here as that in Chapter 2 where the focus was on age groups. Our goal here is to
see what proportion of the arts audience comes from each of the birth cohorts.
Again, as in Chapter 2, the question is whether these proportions are greater or less
than would be expected if all cohorts had contributed equally to the audiences for
the arts. The focus is shifted briefly from the composition of the audience to ask to
what extent a few respondents tend to increase the cohort's contribution to the
audience by attending many times, and whether the average number of attendances
per attendee varies significantly by cohort.
Thus, the prime difference between Chapters 2 and 3 is that in the former we
focused on the changing age of audiences in each of the art forms. Here in Chapter
3 the focus is on the arts attendance of each of the birth cohorts over the years from
1982 to 1997.
Part 1 Cohort Attendance Rates
Birth cohorts vary greatly in size, so, for example, there were far more baby
boomers born in the United States than there were members of the two cohorts
that follow them. Indeed, that is why these Generation Xers, as they are now
called, have been referred to as the "baby bust" cohort. At the same time there
are relatively few people in the cohorts born early in the 20th century because
a considerable number of those born into them have already died. Given the
unequal size of cohorts, the arts audience of each cohort relative to their pro-
portion in the sample will be examined, so it is possible to see whether the
contribution a cohort would be expected to make if each cohort contributed
equally to the arts audience.10
Chapter 3 31
Tables 3.1 through 3.7 present the contributions of each of the cohorts to one
of the benchmark arts. In each table, the three upper left-hand cells are empty
because persons in these cohorts were too young at the time to be surveyed. The
figure in the upper right-hand cell of Table 3.1, for example, means that in 1997,
5.4 percent of the audience for classical music was from the youngest cohort sam-
pled, those born between 1976 and 1980." The numbers in parentheses in these
same seven tables show the contribution of each cohort to the audience relative to
the contribution it would make if all cohorts participated equally.12 The figure of
+0.8 in parenthesis in the upper right-hand cell of Table 3.1, for example, means
that the 1976-1980 cohort contributes eight tenths of a percent more to the 1997
audience for classical music audience than expected.11 Now each cohort will be
considered in turn.
Table 3.1.
Cohort % Contribution to Total Classical Music Attendances, by Year
1982 1992 1997
1976-1980
5.4
(+0.8)**
1966-1975
10.1
9.6
(-5.1)**
(-7.8)**
1956-1965
20.3
17.5
15.9
(-1.8)"
(-5.0)**
(-7.6)**
1946-1955
23.4
19.5
24.7
(+1.6)**
(-0.8)
(+4.0)**
1936-1945
21.0
19.2
16.4
(+5.0)**
(+4.9)**
(+3.5)**
1926-1935
15.1
15.4
15.3
(+0.6)*
(+3.4)**
(+4.7)**
1916-1925
10.9
14.8
10.2
(-2.0)**
(+4.8)**
(+2.5)**
Before 1916
9.3
3.4
2.5
(-3.4)**
(-2.2)**
(-0.1)
100.0 100.0 100.0
Values in parentheses indicate the difference of this observed percentage from the "expected" per-
centage of group in total sample. The differences are statistically significant where indicated (* p<.05;
**p<.01). A statistically significant difference means that the probability of this difference occurring
merely by chance is less than 5% (for p<.05) or less than 1% (for p<.01), based on this sample.
Therefore, we can reasonably conclude that this difference actually exists.
32 Age and Arts Participation 1 Q 8 2-1 99 7
Post-Boomer Cohort 1976-1 985 (-1980) Only in the 1997 survey were members
of the 1976-1985 cohort old enough to be surveyed for the SPPA, and only those
in the older half of the cohort years were at least eighteen years of age and thus eli-
gible to take part in the SPPA survey. Very little can be said about their arts
attendance. It is potentially encouraging for the future of the arts audience that
these youngsters attend five of the seven benchmark arts more often than would be
expected, and they under-attend only one art form, opera. It may be that their ele-
vated rate of attendance is due in part to their participation in conjunction with
school-related activities that give many of them easy access to arts performances.
Certainly, such a decline is in line with the attendance of younger baby boomers
whose attendance dropped from that of their youthful days in 1982 to 1997.
However, it will be possible to discuss this possibility further in Chapter 5 where
the focus is placed on the impact of school attendance on arts participation among
post-boomers.
Post-Boomer Cohort 1 966-1 975 The trend in the arts participation of these early
Generation Xers in the five years between 1992 and 1997 is very clear. They attend
six of the seven benchmark arts less than expected in 1992 and in 1997 as well.
What is more, their attendance, though still less than expected as shown by the
negative signs, improved for opera, musicals, and theater attendance. Jazz is, as we
have shown in Chapter 2, the one benchmark art that attracts more young people
than those who are older. And it is the only art form that this cohort attended more
than expected.
Later Baby boomers: 1 956-1 965 This cohort and all of the ones born earlier were
surveyed in 1982 as well as in 1997, so for these cohorts all comparisons of cohort
rates of participation will cover the full fifteen year time span.14
The arts participation of the later boomers, those born between 1956 and 1965
is well below their proportion in the sample as a whole. The only exception is jazz,
the form that distinctively appeals to young people, and even the cohort's jazz
attendance went down from 1982 to 1997. These late boomers were under repre-
sented in the audience of the other six art forms in 1982 and in 1997 as well.
What is more, their under-representation increased from 1982 to 1997. The
later boomers clearly show the low arts participation that has been widely
observed in earlier studies, and they do not seem to be attending more as they
mature into mid-life.
Chapter 3
33
Table 3.2.
Cohort % Contribution to Total Opera Attendances, by Year
1982
1992
1997
1976-1980
3.4
(-1.2)*
1966-1975
8.7
11.7
(-6.5)"
(-5.7)"
1956-1965
11.7
19.6
20.8
(-10.4)"
(-2.9)
(-2.7)*
1946-1955
19.3
18.4
25.7
(-2.5)"
(-1.9)
(+5.0)**
1936-1945
25.4
22.4
16.8
(+9.4)"
(+8.1)"
(+3.9)**
1926-1935
17.1
17.1
12.1
(+2.6)"
(+5.1)"
(+1.5)
1916-1925
14.4
10.9
8.0
(+1.5)*
(+0.9)
(+0.3)
Before 1916
12.2
2.9
1.5
(-0.5)
(-2.7)"
(-1.1)*
100.0
100.0
100.0
See footnote on Table 3. 1 .
Table 3.3.
Cohort % Contribution to Total Musical Attendances, by Year
1982
1992
1997
1976-1980
1966-1975
1956-1965
1946-1955
1936-1945
1926-1935
1916-1925
Before 1916
19.4
(-2.7)*
24.6
(+2.8)*
18.6
(+2.6)*
16.1
(+1.6)*
12.1
(-0.8)*
9.3
(-3.4)*
11.6
(-3.6)"
19.5
(-3.0)"
21.4
(+1.1)
19.1
(+4.8)"
14.3
(+2.3)"
11.7
(+1.7)"
2.5
(-3.1)"
4.6
(+0.0)
15.1
(-2.3)**
23.0
(-0.5)
22.3
(+1.6)"
14.7
(+1.8)"
12.1
(+1.5)"
6.7
(-1.0)**
1.6
(-1.0)"
100.0
100.0
100.0
See footnote on Table 3. 1 .
34 Hg e and Arts Participation: 1 9B2-1 997
Table 3.4.
Cohort % Contribution to Total Jazz Attendances, by Year
1982 1992 1997
1976-1980
5.2
(+0.6)*
1966-1975
18.2
23.0
(+3.0)**
(+5.6)**
1956-1965
44.1
32.2
25.8
(+22.0)**
(+9.7)**
(+2.3)**
1946-1955
29.6
26.4
21.8
(+7.8)**
(+6.1)**
(+1.1)
1936-1945
10.5
10.5
10.8
(-5.5)**
(-3.8)**
(-2.1)**
1926-1935
9.2
8.4
8.3
(-5.3)**
(-3.6)**
(-2.3)**
1916-1925
3.9
3.3
4.7
(-9.0)**
(-6.7)**
(-3.0)**
Before 1916
2.6
0.9
0.4
(-10.1)**
(-4.7)**
(-2.2)**
100.0 100.0 100.0
See footnote on Table 3.1 .
Table 3.5.
Cohort % Contribution to Total Theater Attendances, by Year
1982 1992 1997
1976-1980
5.0
(+0.4)
1966-1975
10.6
14.8
(-4.6)**
(-2.6)**
1956-1965
22.8
20.6
20.9
(+0.7)
(-1.9)**
(-2.6)**
1946-1955
22.3
22.1
23.7
(+0.5)
(+1.8)**
(+3.0)**
1936-1945
20.5
18.7
15.5
(+4.5)**
(+4.4)**
(+2.6)**
1926-1935
15.1
16.0
10.8
(+0.6)
(+4.0)**
(+0.2)
1916-1925
11.6
9.6
7.0
(-1.3)**
(-0.4)
(-0.7)
Before 1916
7.7
2.4
2.3
(-5.0)**
(-3.2)**
(-0.3)
100.0 100.0 100.0
See footnote on Table 3.1.
Chapter 3 35
Table 3.6.
Cohort % Contribution to Total Ballet Attendances, by Year
1982 1992 1997
1976-1980
5.6
(+1-0)
1966-1975
13.9
12.6
(-1.3)
(-4.8)**
1956-1965
25.5
25.7
22.3
(+3.4)"
(+3.2)*
(-1.2)
1946-1955
25.3
17.2
26.7
(+0.5)
(+1.8)**
(+3.0)**
1936-1945
18.8
16.8
14.4
(+4.5)**
(+4.4)**
(+2.6)**
1926-1935
11.8
11.0
12.0
(-2.7)**
(-1.0)
(+1.4)
1916-1925
9.6
13.2
4.9
(-3.3)**
(+3.2)**
(-2.8)**
Before 1916
9.0
2.3
1.4
(-3.7)**
(-3.3)**
(-1.2)**
100.0 100.0 100.0
See footnote on Table 3.1.
Table 3.7.
Cohort % Contribution to Total Art Museum Attendances, by Year
1982 1992 1997
1976-1980
5.2
(+0.6)**
1966-1975
15.1
17.2
(-0.1)
(-0.2)
1956-1965
22.8
23.8
21.8
(+0.7)**
(+1.3)**
(-1.7)**
1946-1955
25.1
22.8
24.5
(+3.3)**
(+2.5)**
(+3.8)**
1936-1945
20.3
15.4
16.3
(+4.3)**
(+1.1)**
(+3.4)**
1926-1935
12.3
13.4
9.4
(-2.2)**
(+1.4)**
(-1.2)**
1916-1925
12.0
7.8
4.5
(-0.9)**
(-2.2)**
(-3.2)**
Before 1916
7.4
1.6
1.1
(-5.3)**
(-4.0)**
(-1-5)**
100.0 100.0 100.0
See footnote on Table 3.1.
36 Age and Arts Participation: 1982-199?
Early Baby boomers: 1 946-1 955 The early boomers show a pattern of partici-
pation that is starkly in contrast with the picture of low arts participation of the
later boomers just discussed. Indeed their participation rate was higher than the
sample as a whole in six of the art forms in 1982, and that was true for all seven
forms by 1997. This finding jibes with the personal observations of many arts-
organization managers, but it goes against the conclusions drawn from NEA
Research Report #34. 15 The possibility that younger and older boomers are so dif-
ferent from each other will be taken up again in Chapter 4.
Late Depression and World War II Cohort: 1936-1945 Looking across Tables
3.1-3.7, the pattern of arts participation of the Great Depression and World War
II cohort is very clear. In 1982 and in 1997 as well, they were a larger part of the
audience than would be expected by their proportion of the sample in six out seven
of the benchmark art forms. It is also notable that over this span of years their over
representation in the arts audience became less pronounced. The only form for
which their participation was less than expected is jazz.
Early Depression Cohort: 1926-1935 Following the pattern just observed, the
arts participation of this cohort, who experienced the Great Depression in their
childhood and came to maturity during World War II, was high, being greater than
their proportion in the sample as a whole in four of the benchmark arts in 1982
and five in 1997. Their participation in jazz and museum attendance was less then
expected. The figures for jazz fit the pattern of a youthful jazz audience noted in
Chapter 2, but no explanation for their continuing low art museum attendance
comes readily to mind.
Roaring Twenties Cohort: 1916-1925 Members of this cohort, whose early
experience was the boom times following World War I, were between seventy-two
and eighty-one years-of-age in 1997, so it is not surprising, given their advanced
age, that their arts participation was below average for five of the seven benchmark
arts. Their participation was higher than expected only for classical music and
opera.
Pre-World War I Cohort: Born Before 1916 This "cohort" includes all those
born before 1916, thus they were at least 67 in 1982 and at least 82 in 1997. It is
not surprising, given their advanced age and lower education relative to baby
boomers, that they are not over represented in the audience of any of the bench-
mark arts. It is more surprising at first glance that their arts participation, while
Chapter 3 37
still low, more nearly approached, and in the case of classical music reached, the
level that would be expected in terms of their proportion of the sample as a whole.
The attendance rates of these older persons may be due to the fact that the factors
that make for more arts participation, such as higher education, greater wealth,
and an active mind, also make for better health so that by this time in life, surviv-
ing members of the cohort are more likely to have been inclined to arts
participation all along than is true of still young cohorts.
Part 2: Attendees' Hates of Attendance across Cohorts
Having looked at the contribution of each cohort to the arts audience as it has
changed over time; focus now turns from the question of "attendance" to those
who attend the arts. Do members of all cohorts, on average, attend the same num-
ber of times per year? This question is of importance because some commentators
have observed that those in older cohorts, who do attend, attend more often than
do attendees in the more recent cohorts. Thus, to draw one hypothetical example,
it may be that those in a pre-World War II cohort are likely to attend five times a
year while baby boomers, on average, attend twice a year. If this were the case two
fifty-year olds would account for (2x5) =10 attendances, while it would take five
twenty year olds to account for 10 attendances (5x2) =10.
To quickly show the degree, to which a few attendees account for the lion's
share of the attendance of a cohort, the proportion of attendees of a given cohort
is compared with their contribution to the entire audience. These comparisons for
the cohorts in each benchmark art form are shown in Tables 3.8 to 3.14. These
tables are parallel to Tables 2.8 to 2.14 analyzed in Chapter 2. The figures in the
tables are "difference scores."16 The (+) sign in the upper-right-most cell of Table
3.8, for example, means that in 1997 attendees in this cohort, on average, attended
more often than did attendees in the sample as a whole. In other words, a relatively
few in the youngest cohort accounted for much of the attendance of this cohort at
classical music concerts. The (-) sign in the cell just below in the same table means
that in 1997 those respondents under twenty years of age who attended did so less
often than did all attendees that year. The size of the numbers suggests the degree
to which a few or many people of a cohort contributed to the audience.17 The
numerous no-zero differences scores in Tables 3.8 through 3.14 show that, indeed,
attendees in different cohorts often do not all attend equally often.
Focusing briefly on the youngest cohort who were old enough to be surveyed
only in 1997, those born between 1976 and 1980, there is no clear pattern in
the frequency of their attendance, since a few attendees accounted for the
cohort's attendance in about half the art forms, while a larger number of attendees
38 Rge and firts Participation: 1Q82-1QQ7
Table 3.8.
Classical Music - Difference between % of Total Times Attended and
% of People Attending
1982 1992 1997
1976-1980
+ 1.1*
1966-1975
-1.0
-3.7**
1956-1965
+ 1.4
-1.4
-3.2**
1946-1955
+0.3
-2.8*
+0.0
1936-1945
+0.6
-0.2
+0.7
1926-1935
-0.9
+ 1.6
+2.8**
1916-1925
-1.4*
+3.2**
+ 1.8*
Before 1916
-0.1
+0.5
+0.6
These differences are statistically significant where indicated (* p<.05; **p<.01). A statistically signifi-
cant difference means that the probability of this difference occurring merely by chance is less than 5%
(for p<.05) or less than 1 % (for p<.01), based on this sample. Therefore, we can reasonably conclude
that this difference actually exists.
participated less often in the rest. A pattern is, however, much clearer for the early
Generation Xers, those born between 1966 and 1975. Only for jazz did a few afi-
cionados account for the cohort's attendance in 1992 and in 1997. For all the rest
of the art forms, a larger number of Generation Xers attended less often.
Now focus turns to the baby boom and pre-baby boom cohorts that partici-
pated in all rounds of the SPPA survey. Here, rather than focusing on each cohort
individually, a contrast is made between the two baby-boom cohorts on the one
hand and the three cohorts that preceded them, because these two groups of
cohorts show quite different patterns.18 In order to get an impressionistic summary
measure of these difference scores, the number of times attendees of a particular
cohort attended more or less often than would be expected from the total sample
rate is examined for the seven benchmark arts in 1982 and 1997.
Focusing on the two baby boom cohorts, very interesting variations are found.
A larger number of attendees tend to participate infrequently, and this trend
became greater from 1982 to 1997. This means that baby boom arts attendees, at
least over the span of years covered in the survey, tend to sample an art form rather
than showing a strong commitment to arts participation as shown by their infre-
quent attendance.
Turning to the three pre-baby boom cohorts, there is a clear trend, and it is the
opposite of the pattern for baby boomers. While in 1982 more infrequent samplers
were found, by 1997 the attendance for these cohorts was contributed by relatively
few attendees who tended to go more often. While this trend is apparent in all three
Chapter 3 39
pre-boomcr cohorts, it is shown most dramatically by those in the cohort born
between 1926 and 1935, since across all art forms, attendees in this cohort tended
to be samplers in 1982 and aficionados in 1997. There is a better chance to under-
stand why the frequency of attendance varies so dramatically between baby
boomers and pre-boomer cohorts when, in later chapters, the predictors of arts
attendance across cohorts are examined.
Table 3.9.
Opera Music - Difference between % of Total Times Attended and %
of People Attending
1982 1992 1997
1976-1980
-1.7
1966-1975
-2.5
-3.6*
1956-1965
-3.0
-1.1
-0.7
1946-1955
-1.3
-2.5
+2.0
1936-1945
+5.6"
+3.6
+ 1.1
1926-1935
-1.4
+2.0
+ 1.5
1916-1925
-0.5
+0.2
+ 1.1
Before 1916
+0.8
+0.1
+0.1
See footnote on Table 3.8.
Table 3.10.
Musical - Difference between % of Total Times Attended and % of
People Attending
1982 1992 1997
1976-1980
+0.0
1966-1975
1.3
-0.8
1956-1965
-0.2
-2.1*
-0.4
1946-1955
+0.3
-1.2
-0.9
1936-1945
-0.8
+ 1.1
+ 1.4*
1926-1935
-0.1
+0.8
+0.8
1916-1925
-0.7
+2.7**
-0.1
Before 1916
+1.5**
+0.2
+0.3
See footnote on Table 3.8.
40 fl g e and Arts Participation: 19B2-1QQ7
Table 3.11.
Jazz - Difference between % of Total Times Attended and % of People
Attending
1982 1992 1997
1976-1980
-0.3
1966-1975
+0.2
+3.7**
1956-1965
+4.0**
+3.2*
-0.1
1946-1955
+ 1.2
+ 1.8
-2.8*
1936-1945
-2.5**
-2.9**
-1.2
1926-1935
-0.9
-1.3
+0.0
1916-1925
-2.0**
-1.2*
+0.6
Before 1916
+0.2
+0.0
-0.1
See footnote on Table 3.8.
Table 3.12.
Theater - Difference between % of Total Times Attended and % of
People Attending
1982 1992 1997
1976-1980
-0.6
1966-1975
-2.7**
-0.1
1956-1965
+2.8**
-0.5
-0.9
1946-1955
-1.5
-0.5
-0.1
1936-1945
+ 1.5
+ 1.4
+ 1.4
1926-1935
-2.0*
+2.1*
+0.0
1916-1925
-0.2
-0.1
+0.2
Before 1916
-0.6
+0.3
+0.2
See footnote on Table 3.8.
Table 3.13.
Ballet - Difference between % of Total Times Attended and % of
People Attending
1982 1992 1997
1976-1980
+0.3
1966-1975
-0.8
-0.4
1956-1965
+5.0**
-0.7
-2.4
1946-1955
-2.4
-4.6*
+ 1.7
1936-1945
-1.9
+2.3
+0.4
1926-1935
-1.3
-1.1
+ 1.7
1916-1925
-1.5
+5.1**
-1.0
Before 1916
+ 1.9
+0.1
-0.5
See footnote on Table 3.8.
Chapter 3
41
Art Museum
Table 3.14.
Difference between % of Total Times Attended and %
of People Attending
1982
1992
1997
1976-1980
+0.4
1966-1975
-0.5
-0.6
1956-1965
-0.2
-1.3
-2.4**
1946-1955
-1.9*
-1.1
+ 1.0
1936-1945
+ 1.5*
+0.4
+2.8**
1926-1935
-1.4*
+2.0**
+0.1
1916-1925
+ 1.3*
+0.4
-0.9**
Before 1916
+0.6
-0.0
-0.3
See footnote on Table 3.8.
Chapter 4 THE IMPORTANCE OF R GE AS H DETERMIHHHT OF
HRTS PHRTI CIPHTI OH
In Chapters 2 and 3 we looked at the changing age and cohort composition of
the audience for each of the seven benchmark arts in the United States over the
years from 1982 to 1997. Here, focussing on the data of the 1997 survey, we seek
to find the importance of age in the context of other factors as a determinant of
arts participation. First we ask how well age predicts participation for each of the
art forms; next, we ask the importance of age when other measured factors that
have been shown to influence arts participation are taken into account; and which
of the other variables are the most important in predicting participation in addi-
tion to age. Finally, we ask the same set of questions when considering a summary
measure of participation in all of the seven benchmark arts together.
The reason for controlling for the effects of other variables when looking at the
importance of age is to show the effects of age, per se. after statistically eliminat-
ing the effect of these other factors. As will be shown, the influence of nine other
factors is considerable. For example, as seen in Chapters 2 and 3, persons in their
middle years tend to go more often than those who are younger or older. The data
presented in this chapter, however, show that when the other factors are taken into
account, arts participation rises significantly with age. The meaning of these results
is discussed in detail in the concluding section of the chapter.
METHODS
A series of ordinary least squares regression analyses are used to compare the
relationship between arts participation and age, with the relationship between par-
ticipation and age after controlling for the effects of nine other relevant variables.
Following our focus in Chapters 2 and 3 on both attendance and attendees, we are
interested in explaining both whether respondents attend an art form in the prior
twelve months and also their frequency of attendance. Thus, parallel sets of regres-
sion analyses are presented, one with the frequency of attendance as the dependent
variable, and the other with whether or not the respondent is an attendee as the
dependent variable.19
Independent Variable
In this chapter, "Age" is measured as the respondent's year of age at the time
they took the survey.
Chapter 4 43
Dependent Variables
"Attendances" is measured as the number of times R attended an art form dur-
ing the last year. A few respondents attended one or another art form 50 to 150
times in a year. While it is perfectly possible to attend this many times in a year,
such persons are not ordinarily what we think of as audience members in the usual
sense of the term. In all likelihood these very frequent attendees are art critics, art
teachers, or managers. Accordingly, to correct for the skewing effect of these few
respondents, we set 24 as the highest frequency, so that the top of the attendance
measure scale is "24 or more."
"Attendee" is simply whether the respondent attended the art form during the
last year (0=No, l=Yes).
"Summary Arts Participation" is the number of art forms that the respondent
attended during the last year (range: 0 to 7). This measure does not take into
account how often the respondent attended arts events. The measure here is the
respondent's range of arts attendance and not their total number of attendances at
a given form. Thus the person who attended seven jazz concerts and no other art
form gets a score of 1 while the person who participated in each form just once
attains a score of 7. Since the age composition of the audience for jazz is younger
than that for the other arts, and since its audience differs in other ways as well,
excluding jazz from the summary measure was considered, but the results were vir-
tually the same as when all seven art forms were included in the summary measure.
Control Variables
Gender was measured as Female=l, Male=0 because prior research has shown
that women more often attend the arts than do men. Accordingly, the variable is
designated "Female" in Tables 4.1, 4.2 and 4.3.
Race was coded as Black=l, Other=0 because "Other" is a mixture of several
different categories including white, Asian, and non-white Hispanic. Accordingly,
the variable is designated "Black" in the tables.
Marital status was reported as Never Married, Married, Divorced, Separated,
or Widowed. The rates of arts participation of each of these groups was compared
in a preliminary analysis and the groups most alike in their participation were
grouped together. Never Married, Divorced, and Separated had similarly high rates
of participation and accordingly were grouped together and called "Not Married."
They were coded =1. Those who were Married or Widowed generally had lower
rates of participation and were coded =0.
44 Age and Arts Participation: 1 9 B 2-1 997
Household income is designated "Income" in the tables. It was coded 1-7 on
the survey responses, and these correspond to the following income ranges:
l=$10,000orless
2=$10,001 to $20,000
3=$20,001 to $30,000
4=$30,001 to $40,000
5=$40,001 to $50,000
6=$50,001 to $75,000
7=$75,001 to $100,000
8=Over $100,000
"Education" indicates the highest level completed by the respondent. It has a
range of 1-13 as follows:
l=7th grade or less
2=8th grade
3=9th to 11th grade
4= 12th grade but no diploma
5=High school diploma/equivalent
6=Vocational/Technical program after high school
7=Some college but no degree
8= Associate's degree
9=Bachelor's degree
10=Graduate or professional school but no degree
1 l=Master's degree
12=Doctorate (PhD, EDD)
13=Professional (Law LLB or JB; Medicine/MD; Dentistry/DD)
The ordering is the same as in the questionnaires. Initially it was thought that
respondents with a Doctorate should be ranked higher those than with a profes-
sional degree, so a number of preliminary runs were made, and, in the case of most
art forms, respondents with professional degrees were more likely to attend than
were those with a Doctorate.
As in the case of the respondent's education, "Father's Education" was meas-
ured as the highest grade or degree earned by the respondent's father, using the
same scale ranging from 1 to 13. Before selecting Father's education as the best
measure of the educational atmosphere in the home, a number of preliminary
analyses were made. Father's education was found to correlate somewhat better
with the respondent's arts attendance than did Mother's education, and including
both did not appreciably increase the ability to predict respondent's arts participa-
tion over using Father's education alone.
Chapter 4 45
"Children" comprises a continuous measure of the number of children 18 or
under in household. Earlier SPPA surveys had shown that having children six and
under in the household was a much better predictor of lower arts participation, but
alas, the 1997 survey did not ask about children six or younger.
"Health" is a five point ordinal scale of self-reported health status with the fol-
lowing levels:
1 =Poor
2=Fair
3=Good
4= Very Good
5=Excellent
Several more focussed measures of health status were available in the survey.
These included self-reports of eyesight, hearing, and the ease of walking. Hearing
ability was a good predictor of Classical Music concert attendance, and the ease of
walking was a good predictor of art museum attendance, but overall none of these
singly or in combination significantly increased the predictive value of the global
health measure alone.
"Metro" is the dichotomous indicator of the size of the respondent's place of
residence. It is approximately equivalent to the "size of place" variable in the
three earlier surveys. Metro was coded 1 if the respondent lives in any of the
following eleven major metropolitan areas as operationalized in the survey, and
0 if they do not:
Boston-Worchester-Lawrence,
Chicago-Gary-Kenosha,
Dallas-Ft. Worth,
Detroit-Ann Arbor-Flint,
Houston-Galveston-Bazoria,
Los Angeles,
Miami-Ft. Lauderdale,
New York-Northern New Jersey-Long Island,
Philadelphia-Wilmington-Atlantic City,
San Francisco-Oakland-San Jose,
Washington DC-Baltimore
In a preliminary analysis we also tried to use the "state and counties" variable
in the 1997 survey to approximate the "region" variable of earlier surveys. The
level of detail provided in the 1997 data made it possible to see that the range of
46 Rge and Arts Participation: 1982-1997
variation within regions and states was as wide as the variation between regions.
For example, both Florida and Texas proved to be quite different from the rest of
the South, and the rates of arts participation in the Central Valley of California was
among the lowest in the nation, while those of the San Francisco and Los Angeles
areas were among the highest.
No conceptually justifiable division among geographic areas was any clearer
than the metropolitan area measure just described, so it is used in the regression
analyses.
RESULTS
The three tables below show the results of the regression analyses. Just as in
Chapters 2 and 3, we are interested in both whether the respondent attended an
art form in the previous year and the respondent's frequency of attendance.
Accordingly, the dependent variable in Table 4.1 is the frequency of attendance,
while the dependent variable in Table 4.2 measures whether the respondent
attended a particular art form during the prior year or not.
Reading the Tables
A good deal of information is condensed in Tables 4.1, 4.2 and 4.3. Take the
left-hand portion of Table 4.1 for example. This shows the results for Classical
Music attendance. The "Beta" is the standardised coefficient so its range is from
+1.0 to -1.0. The focus is on the Betas because their relative strength is compara-
ble across variables and across models as well. The relationships are all positive
unless a negative sign precedes them. Thus, for example both being black and hav-
ing children 18 or under in the household tends to reduce the level of Classical
Music attendance.
The measure "R-square" refers to the amount of variance in the dependent vari-
able that is explained by the independent and control variables included in the
model. The one or two stars (* or **) indicate whether the variable in that row of
the table contributes significantly to predicting the dependent variable with the dif-
ference being obtained by chance in less than one case in fifty (p<.05) or one chance
in a hundred (p.<.01) respectively. Unless otherwise noted, just the statistically sig-
nificant relationships will be discussed. Finally, note that the smaller top part of the
column represents the bivariate relationship between arts attendance and age,
while the longer, bottom part of the column shows the predictive value of age when
the contribution of the other nine control variables is taken into account.
The corresponding right-hand part of Table 4.2 shows the same set of informa-
tion, in this case for being an "attendee" rather than the frequency of "attendance"
Chapter 4 47
shown in Table 4.1. As an inspection of the three tables shows, the pattern of rela-
tionship for attendance (shown in Table 4.1 ) and attendee (Shown in Table 4.2) are
very similar. Accordingly, the focus will be on attendance except where noted.
Table 4.3 presents the results for the "Summary Arts Participation" measure
defined above.
The Effect of Age on Arts Participation With and Without Controls
Here the figures in the top two rows of Table 4.1 are compared with the age and
R-square figures in the bottom part of the table. Focussing first on the top row all
the way across the table, one can see that when the influence of other factors is not
taken into account the relationship between age and arts attendance is positive for
classical music and opera, and negative for jazz and art museum attendance. As
indicated by the lack of a significance, however, there is no apparent relationship
between the age of the respondent and attendance at musicals, theater, and ballet.
The R-squares are all less than one percent (.01) suggesting that, while there is a
statistically significant association between age and participation in classical music,
opera, jazz, and art museums, the relationships are substantively trivial.
With the exception of jazz, the Betas for age are positive and significant for all
the art forms when the effects of the nine control variables are taken into account.
Taken together, these results suggest that the direct effects of age on arts participa-
tion is masked by the effects of other factors including those measured by the
control variables. When these variables are taken into account, arts attendance
increases with age. This means that older persons attend the art forms more often
than do younger people of the same gender, education, income, etc.
Contribution of Controls to Predicting Participation in Each Benchmark Art
In this section we look at Tables 4.1 and 4.2 in order to identify the control vari-
ables that are most important in predicting participation in each of the benchmark
arts. Please note that each of the statements made in this section is predicated on
controlling for the effects of age and the other measured variables.
The figures for Classical Music are shown in the first column of these two tables.
On inspection, they show that age is the second-most important variable in pre-
dicting arts attendance. The respondent's amount of formal education is the most
potent predictor of classical music attendance. What is more, father's education is
the third most important predictor. Interestingly, respondent's health, race, gender,
and number of children do not importantly affect classical music attendance. The
relative magnitude of the figures in Table 4.2 is very much like those of Table 4.1
with the exception of household income. Income is a better predictor of whether
one attends classical music concerts (Table 4.2) than of how often one attends
48
flge and Arts Participation: 1982-1997
Q)
O 3"
5. O CD
— C CO
5- 3. CD
CO' 3 CD
(J(0 ?
■* — CD
® 3 2-
O CD w
O ^ CD
= o a
-• 5 o
CO Q ft)
» ® ve
W CO
CD CO
CO 3
CO 5-
oi S
#f
ft
~« 3
TJ O-
A O
b Si
Ol CD
*— Q.
O — ~
-i *
ST73
CD A
CO •
CO O
,» Ol
3
0) »
-X
c£ o
A
CO
b
CO
^^
CO
M
o~
o'
CD
2L
CO
CD
*<"
D.
CO
O
CO'
3
3_
■^
i-+
3"
o
co'
03
3
CO
~**
03
CD
3
3:
■D
8
CD
3
-1
CD
3-
Q>
CD
3
CO
o"
3"
-i
03
CD
.-*•
m
— .
CD
3"
CD
O
13
0)
O
3
O"
-^
CD
CD
CT
03
g a
#|
o w
O CD
3 3:
O CD
c 2.
Q.
CD
CO
.Q
c
03
-1
CD
s
I
o
Tl
m
3
z
DD
-n
>
^
XI
>
CD
CD
3"
CO
CL
O
o
03
CD
CQ
c
CO
CO
«-*■
9L
Si
3
c
o
03
o
^*
o
3
CD
CD
o
3"
Q.
CD
3
s
*r
03
<■
c
CD
-
CD
0)
CD
03
03
3
CO
m
6'
3
©'
Q.
co'
CD
Q.
CD
c
o
2
CD
3>
<
03
£.'
CD
ooooooooooo
00 0 00^0000^
•JUMOMMOIOIMIOA
(5>OAO)0-'AO)*.0)CO
o o
b b
o -o
o> Ol
co
(A
(A
O
03
OOOOOOOOOOO
bbbbbbbbbbb
COAIO-'COCOAIk-'KJO)
O^JOIOI-'^JCOCOCO-OIO
o o
b b
o to
o o
»
O
ooooooooooo
bbbbbbbbbbb
cnuoiAUCocD^-'cna)
OOOfOOOOrorOIOCOOl
O O
b b
o -*■
O CO
c
(A
o
03
ooooooooooo
bbbbbbbbbbb
CO-'OOJCOCDOIMCJI-'K)
^J0)-^-^-0D00O>J-U'>0-*».
******
**»*»*
o o
b b
o Ol
co ao
09
ooooooooooo
bbbbbbbbbbb
(Oioco(OMUi(oic-'ua)
**»»»** »»
» • *
o o
b b
o -■•
o 4*
a
0)
ooooooooooo
o o o o
K3 A M 5
_L _l _l Ji.
o o o o o o o
.»>. O) -&. CO -* A Ol
ai ^ ^ a) ui O) o
* * * » * »
* * » * * *
o o
8 8
O Ol
03
OB
7)
a>
CO
-«
(D
(A
(A
33
CD
V)
CO
O
3
a
<D
>
3 H
CD Q)
IS
0J CD
o r
<D -*
(A "
>
CO
CD
<^
(A
0)
3
a
0)
N
(D
a
o
o
CD
=*
O
5"
3
sr
ooooooooooo
bbbbbbbbbbb
(OSIIOUCOOOCDCO-'UA
-OCOO)030101COK30N30>
....... •»
* , . . » * •»
o o
b b
o co
->• co
»
>
3.
c
3
Chapter 4
49
CO I 3D ^
o
o
5*
o
S"
o
CD
c
CD
I
O
T1
m
3
Z
DD
-n
>
2
3D
>
CD
I
3"
W
Q.
o
o
0)
CD
(Q
c
CO
<'
B
=^
5
C
o
^
o
3
CD
CD
CD
Er
CD
CD
o
03
3
CD
0)
?r
CD
<"
CD
03
3
CO
m
Q.
c
o
o
3
<d'
Q.
03'
CD
o
CD
CD
o
Ql
CD
ooooooooooo
'-•■ b b b b ro b b b b '-•■
MIOti)(0(D-'000)MO)Q
^IO(DOIAOIIOOIS1(DQ
OOOOOOOOOOO
bbbbb^bbbbb
(OWIO(DO)(D01Jk^Cjl(0
OOOOOOOOOOO
^•bbbb'-^^bbbb
-'^OlUCJCOA^fOlOO)
0)^00>JO)0(J)-'-'0)A
o o
b b
o en
CO oo
o
b
o
o
o
b
o
o
o o
b b
o o
o -O
o
o
o
O
o
O
o
o
o
o
o
o
o
b
b
b
b
b
_L
!_L
!_i
b
b
b
b
b
CD
— ^-
— ^
Ol
O)
-&•
o
o
00
o
o
o
00
— '
o
CD
CJl
*
-•0
*
o
o
»
*
co
*
.&.
CO
-«j
CO
*
o
o
o
o
o
p
o
o
o
o
o
o
o
b
b
b
b
b
Li.
b
b
b
b
b
b
b
00
fO
-t*.
-t*
-&.
co
•s]
-0
ro
-p».
■n]
o
o
■t-
Ol
o
NO
*.
CO
CO
— i-
■^
Ol
00
o
— ».
*
*
*
*
*
*
*
*
*
•
o
o
o
o
o
p
o
o
6
o
o
o
6
CD
5L
b
o
b
rO
co
b
ro
(O
b
CO
b
Ol
co
o
Ol
b
Ol
00
b
ro
b
V]
b
00
CO
b
Ol
b
o
o
b
o
Ol
ST
p
o
o
o
o
p
p
o
o
o
o
o
o
— i.
b
b
b
b
ro
Lt
b
b
b
b
b
b
Ol
Ol
4*.
CO
-j
O)
o
Ol
— L
Ol
— t
o
-»j
CO
■>o
-o
o
■t*.
ro
O)
_ i
Ol
o
00
CO
00
*
*
*
*
•
*
*
*
*
o
(A
o'
CD
30
CD
CO
(D
(A
</>
o
JO
»
O
o
CD
2
c
(/>
cV
03
(D
>
3 O"
O *
- K9
03
N
N
CD
03
rf
CD
>
3-
c
CD
c
3
>
co
<D
</>
0)
3
a
0)
N
O
a
o
o
(D
O
<D
3
(A
50 Age and Arts Participation: 1982-19Q7
(Table 4.1). This attendee/attendance difference is discussed at greater length
below.
Looking at the Betas for Opera in Tables 4.1 and 4.2, we find again that age is
the second most important predictor of attendance, and education is the most
important. Not having children in the home is the third most important predictor
of opera attendance. Residence in one of the largest metropolitan areas has more
to do with predicting opera attendance than with any other art form, with the
exception of visiting art museums.
Looking at the Betas for Musicals in Tables 4.1 and 4.2, the respondent's edu-
cation and household income are again most important in predicting attendance.
Unlike the art forms discussed so far, however, gender is the third most important
variable in predicting arts attendance, women going to musicals significantly more
often than men.
Again, education is the most important predictor of attending Jazz events as
seen by the Betas for jazz in Tables 4.1 and 4.2, but beyond that the pattern is dif-
ferent from the other arts forms. The second most important predictor is marital
status, with those not currently married more likely to attend. The third most
important variable in predicting jazz attendance is the lack of children living in the
respondent's home. The fourth most important is being African- American. Neither
age nor gender are significant predictors, but they are interesting, nonetheless,
because unlike their coefficients for any of the other art forms, attendance tends to
be more frequent for men than for women and tends to go down rather than up
with age.
The pattern of best predictors of Theater attendance is the same as for opera
going as can be seen by looking at the relevant Betas in Tables 4.1 and 4.2.
Education is the most important predictor, age is second, being unmarried is third,
and residence in one of the larger metropolitan areas is fourth. Being female fol-
lows education and age as the most important predictor of Ballet attendance.
Income is the fourth. As for all the other art forms, education is the most impor-
tant predictor of Art Museum attendance, and as for classical music, father's
education is the third most important. The second most important predictor is
being unmarried. Residence in a metropolitan area, income, and age follow in that
order. It is notable that art museum attendance is less dependent on age than is
attendance at any of the other benchmark arts.
The Relative Contribution of Each of the Control Variables
Here looking across the rows of Betas in Tables 4.1 and 4.2 focus is put on the
contribution of each of the control variables in turn. The goal is to predict partic-
Chapter 4 51
ipation in each of the various art forms, when the effects of age and each of the
other measured control variables are taken into account. Again, the results for
attendance shown in the Betas of Table 4.1 and attendee shown in Table 4.2 are
very similar, and so they will be considered separately only in the case of Income
where the figures are clearly different.
Looking first across the row for Gender ("Female") it can be seen that with the
other variables taken into account, women are significantly more likely to attend
each of the art forms, except for jazz. The predominance of females is most pro-
nounced for ballet and musical theater attendance. As for jazz, men attend
somewhat (but not significantly) more often (as indicated by the negative but not
significant sign of the Beta -0.017).
Of all the control variables, Race is least often a significant predictor of arts
attendance. Only in the case of classical music where African-Americans attend less
often and jazz where they attend more often than others, is race a significant pre-
dictor of attendance.
Being single, divorced, or widowed, that is to say, Not Married is a significant
predictor of attendance at each of the seven art forms. What is more, currently
being without a companion is the second most important predictor of both jazz
and art museum attendance.
Higher Income significantly influences attendance at each of the forms, but it is
the second most important predictor for only one of them, attending musicals. In
the common stereotype, wealth is associated with classical music attendance,
opera, and art museum going, but income is only the fifth most important control
variable for these art forms. As noted earlier, the relative magnitude of the Betas in
Table 4.2 is very much like that of Table 4.1 with the exception of household
income. Income is a better predictor of whether one attends each of the seven art
forms (Table 4.2) than of how often one attends (Table 4.1 ). This suggests that hav-
ing sufficient money is important to arts attendance, but, after this threshold is
reached, other factors are more important in predicting how often one attends. In
other words, respondents with higher levels of income are more likely to attend the
arts, but not necessarily to participate more often than respondents with lower
incomes, all other things being equal.
Education is far and away the best predictor of attendance at every one of the
seven benchmark arts. It is the only control variable that is consistently more
important than age. The Beta coefficients in Table 4.1 show that education is most
important for art museum attendance, classical music, and theater attendance in
that order. The general importance of education in fostering arts attendance is fur-
ther underlined by the fact that, even controlling for respondent's education,
52 Age and Arts Participation: 1982-1997
Father's Education is a significant predictor for all the art forms, and it is the third
most important predictor for classical music and art museum going.
Earlier SPPA reports have shown the importance of having young children in
reducing arts participation. The measure inserted in the 1997 survey, Children 18
and Under in the household, has proved a less potent depressant. Children signif-
icantly reduce attendance at musicals, theater, and art museums. Only in the case
of jazz does children in the household become the third most important predictor
coming just behind education and being unmarried.
While it is not among the leading predictors of arts attendance, Good Health
understandably contributes significantly to arts participation. This holds true in all
the art forms except jazz and ballet. The audiences of these forms are younger, on
average, but since the effects of age has been taken into account, it must be that the
older people who do attend these art forms must be, on average, in better health
than are older participants in the other art forms.
Finally, Metropolitan Place of Residence adds significantly to predicting partic-
ipation in each of the art forms with the exception of jazz, yet jazz would seem to
be as much an urban form as the others. It may be that the costs of production and
the number of people involved in production are lower than for the other art
forms, so it is easier to tour outside the metropolitan areas. In addition, it may be
that the younger audience for jazz is more likely to be college students and students
are more likely to be concentrated in a few places such as Madison, Wisconsin;
Chapel Hill, North Carolina; Austin, Texas; and western Massachusetts, outside
metropolitan areas. Such places present greater opportunities to attend jazz con-
certs than do most other non-metropolitan places who include fewer young adults
in school.
The Summary Arts Participation Measure
As noted above, in order to get a picture of the contribution of age to arts par-
ticipation in general, the variable "Summary Arts Participation" was created by
summing the number of benchmark arts attended by the respondent. Then an
analysis parallel to those discussed so far in this chapter was performed. At one
extreme of the summary measure are those who have attended none of the art
forms in the prior year. At the other are respondents who have attended all seven
in the prior year.
To measure the contribution of age to predicting general arts participation with
and without controls, an OLS regression analysis parallel to the attendee measures
shown in Table 4.2 were performed. Table 4.3 shows the results of this analysis.
The left-hand side of the table shows the results when all seven benchmark arts are
Chapter 4 53
considered together. The right side shows the same analysis made without taking
into account attendance at jazz concerts. We inspected this latter measure because,
as shown in Tables 4.1 and 4.2, many of the predictors of jazz participation are dif-
ferent from those of the other art forms.
Table 4.3.
Regression Results of Summary Measure of Attendance on Age
(standardized coefficients)
Sum of 7 Arts Sum of 6 Arts (w/o Jazz)
Bivanate Model
Age -0.03050957" -0.014
R-square 0.001 0
Multivariate Model
Age 0.090" 0.102
Female 0.094" 0.106
Black 0.021* 0.003
Not Married 0.098" 0.084
Income 0.153" 0.147"
Education 0.302" 0.304"
Father's Educ 0.095" 0.090*
Children -0.052" -0.044"
Health 0.059" 0.063"
Metro 0.057" 0.061
R-square 0.236 0.233
*♦
**
*+
See footnote on Table 4. 1 .
The figures for arts participation with and without jazz prove to be remarkably
similar. There are only two differences of note. The first has to do with the bivari-
ate relationship between arts participation and the age of the respondent. As seen
in the top row of Table 4.3, age does not significantly predict arts participation
when jazz attendance is not included in the measure. When jazz is included in the
arts participation measure, however; there is a significant negative relationship
between arts participation and age.
The only control variable showing a significant difference is the race measure,
Black. When jazz is not included in the measure, there is no difference between the
number of forms attended by African-Americans and attended by others. When
jazz, which has a significantly higher attendance rate among African-Americans
than among others is included in the summary arts participation measure, blacks
attendance is significantly higher than for others in the sample. Since, with this one
54 Age and Arts Participation: 1982-1997
exception, the results are so similar, only the results for the seven arts together will
be considered in the following discussion.
The importance of controlling for the influence of other variables when consid-
ering the relationship between age and general arts attendance is dramatically clear
in comparing its direct relationship with its relationship after the effects of the con-
trol variables are taken into account. As the left side of the top line of Table 4.3
shows, the slight but statistically significant Beta is negative, meaning that atten-
dance tends to go down as age goes up. Taking the nine control variables into
account, however, the relationship between arts participation and age becomes
positive and significant. This means that, other variables taken into account, the
number of art forms attended tends to go up as people get older.
Examining the Beta coefficients in the left-hand column can see the relative
importance of age in predicting arts participation. As we have seen for each of the
art forms considered separately, education, not age, contributes the most to pre-
dicting the number of art forms attended. Interestingly, household income is the
second most potent predictor. Next gender, marital status, father's education, along
with age are all about equally important predictors. Then children in the home,
health, and residence in a metropolitan area are all about equally important, and
finally race, while being significant, is the least substantively important predictor
of the number of art forms attended.
CHAPTER 5 CORRELATES OF BABY BOOMER. PRE-BOOMER.
AND POST BOOMER PARTICIPATION
Here in Chapter 5 is concluded the exploration of the causes of arts participa-
tion begun in Chapter 4. There the focus was on the whole 1997 sample, here the
focus is on each of three major birth cohorts, baby boomers, those born before
them, and those born after the boomers.
The sample is divided into three parts because some of the other measured fac-
tors are more likely to affect young people, and some are more likely to affect those
who are older. For example, adults under the age of forty are more likely than their
elders to have young children in the home, so the presence of young children in the
home is more likely to depress the arts participation of young people, likewise,
older people are more likely to have their arts participation reduced by ill health.
In addition, the same variable may mean quite different things depending on the
respondent's age. For example, for young people "not married" means being sin-
gle, separated, or divorced. Such people are more likely to actively seek partners
and engage in arts participation, while their married age-mates are likely to have
their leisure activities taken up in family activities. For older cohorts, "not mar-
ried" is likely to mean widowed which may mean these individuals have no one to
accompany them to arts events.
METHODS
Birth Cohorts
As noted in Chapter 1 , birth cohorts consist of all those persons born within the
same span of years. The number of years to be included in a cohort depends on the
research question at hand. In Chapter 3 cohorts were divided into persons born in
the same ten-year period because of an interest in the differential experience of
those born at historically specific moments. Focal here are the factors differentially
influencing the arts participation of those in different stages of their lives, with a
special interest in baby boomers. Consequently, the respondents are divided into
three groups: baby boomers — those born between 1946 and 1965, pre-baby
boomers — those born before 1946, and post-baby boomers — those born after
1965. The number of respondents in each of these three groups is as follows: pre-
boomers: 4076, boomers: 5317, and post-boomers: 2653.
56 Hge and Arts Participation: 1 Q 8 2-1 997
Independent and Dependent Variables
As in Chapter 4, the independent variable is "Age" measured as the respon-
dents' years of age at the time they participated in the survey.
The dependant variable is "Attendances." It is measured as the number of times
R attended an art form during the last year — a few respondents attended one or
another art form 50 to 150 times in a year. While it is perfectly possible to attend
this many times in a year, such persons are not ordinarily thought of as audience
members in the usual sense of the term. In all likelihood these outliers are art crit-
ics, art teachers, or managers. Accordingly, to correct for the skewing effect of these
few respondents, 24 was set as the highest frequency, so that the top of the atten-
dance measure scale is "24 or more."
"Summary Arts Participation" is the number of art forms that the respondent
attended during the last year (range: 0 to 7). This measure does not take into
account how often the respondent attended arts events.
Control Variables
Gender, Race, Marital Status, Household Income, Education, Father's
Education, Children in the Home, Health, and Metropolitan Place of Residence are
measured in exactly the same way as they were in Chapter 4. The reader is referred
to that discussion.
One additional control variable, "Student Status," is added for those respon-
dents in the post-boomer sample. Since a goodly number of post-boomer
respondents are still in school, it was good to be able to take this into account in
evaluating the impact of education on arts participation. SPPA surveyors were
instructed to ask all respondents 18 to 25 years old: "During the last 12 months,
were you enrolled in a high school, college, or university?" Since the oldest per-
sons asked this question were 25, those post-boomers born between 1966 and
1971 were not asked this question. Consequently, this measure misses the fact that
some of the older post-boomers have not completed their formal education.
FINDINGS
Three lines of findings are discussed in searching for differences in the predic-
tors of arts attendance of boomers, pre-boomers and post-boomers. First, how
successfully the measured variables in aggregate predict arts participation in the
three age groups is examined. Second, the best predictors of arts participation in
each art form and in the summary measure of arts participation are found.
Finally, looking across the tables, the focus will be on the relative importance of
age and each of the control variables in predicting arts attendance at all of the
benchmark arts.
Chapter 5 57
The results of the analyses are summarized in Tables 5.1 through 5.8. They can
be read in exactly the same way as the three tables in Chapter 4, so refer to the dis-
cussion "Reading the Tables" offered there.
Table 5.1.
Regression Results of Classical Music Attendances on Age by Cohort
(standardized coefficients)
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
Boomers
Boomers
Post-Boomers
0.088**
0.093"
-0.019
0.080"
0.011
-0.017
-0.022
-0.019
-0.036
0.027
0.077**
0.026
0.126"
0.043*
-0.012
0.192"
0.172"
0.132"
0.065"
0.083"
0.047
0.012
0.013
-0.015
0.028
0.026
-0.006
0.026
0.052**
0.129**
-0.003
0.100
0.082
0.049
These effects are statistically significant where indicated (* p<.05; **p<.01). A statistically signifi-
cant effect means that the probability of this effect occurring merely by chance is less than 5%
(for p<.05) or less than 1 % (for p<.01), based on this sample. Therefore, we can reasonably
conclude that this effect actually exists.
The Aggregate Explanation of Arts Attendance
The row of numbers across Tables 5.1 through 5.8 gives the R-square for each
of the twenty-four OLS regression analyses. R-square, as noted above in Chapter
4, is a measure of how well the variables together predict arts attendance. If the
variables together perfectly predicted attendance then the R-square would be 1.00.
The bottom line of Table 5.1 shows that the measured variables account for 10
percent of the variance in the classical music attendance of pre-boomers, 8.2 per-
cent of the variance for baby boomers and just 4.9 percent of the variance for
post-boomers. This means that the set of demographic variable was twice as suc-
cessful in predicting attendance at classical music concerts for pre-boomers as for
post-boomers.
0.052*
0.008
0.000
0.093**
-0.002
0.006
0.012
-0.018
-0.045*
0.047*
0.047**
0.025
0.124**
0.020
-0.010
0.080**
0.115**
0.076**
0.067**
-0.001
0.064*
0.007
-0.022
0.008
0.014
0.026
0.029
0.054*
0.039*
0.072*
0.066**
0.055
0.025
0.029
58 Hg e and Arts Participation: 19B2-1QQ7
Table 5.2.
Regression Results of Opera Attendances on Age by Cohort
(standardized coefficients)
Pre-Boomers Boomers Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
See footnote on Table 5.1.
Table 5.3.
Regression Results of Musical Theater Attendances on Age by
Cohort (standardized coefficients)
Pre-Boomers Boomers Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Hearth
Metro
Student
R-square
See footnote on Table 5.1.
Quickly inspecting the other seven tables, shows the same pattern, better pre-
diction for the older age group and considerably attenuated prediction for younger
respondents, with boomers intermediate, is found also for attendance at musicals,
theater, art museums, and also for the summary arts measure. Like the forms dis-
cussed so far, opera attendance is best predicted for the older age group but the
0.044
0.035*
0.003
0.114**
0.051**
0.057*
-0.025
0.039*
-0.004
0.012
0.050**
0.038
0.122**
0.082**
0.046
0.126**
0.161**
0.090**
0.021
0.016
0.058*
-0.016
-0.021
-0.053*
0.095**
0.050**
-0.020
0.030
0.066**
0.032
-0.010
0.081
0.073
0.034
Chapter
59
prediction for the other two age groups is equally poor. The pattern is similar for
ballet attendance but the degree of explanation is so low that it is not worth con-
sidering. Only in the case of jazz is the pattern reversed. Here the R-square for the
pre-hoomers is very low, but is somewhat higher for the other two age groups.
Table 5.4.
Regression Results of Jazz Attendances on Age by Cohort
(standardized coefficients)
Pre-Boomers
Boomers
Post-Boomers
Age
0.007
-0.012
0.050
Female
-0.019
-0.011
-0.027
Black
0.037
0.075"
0.031
Not Married
0.012
0.102"
0.068"
Income
0.057*
0.058"
0.045
Education
0.081"
0.095"
0.084"
Father's Educ
-0.001
0.047"
0.035
Children
-0.003
-0.054"
-0.074"
Health
0.044
-0.004
-0.018
Metro
0.022
0.013
0.017
Student
0.067*
R-square
0.023
0.045
0.042
See footnote on Table 5.1.
Table 5.5.
Regression Results of Theater Attendances on Age by Cohort
(standardized coefficients)
Age
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
Pre-Boomers
Boomers
Post-Boomers
0.007
-0.012
0.050
0.079"
0.020
0.014
0.060"
0.004
0.037
0.005
-0.006
0.003
0.033
0.071"
0.070"
0.117"
0.032
-0.031
0.185"
0.145"
0.057*
-0.003
0.038*
0.094"
0.005
-0.037*
-0.026
0.044*
0.027
-0.035
0.048*
0.052"
0.015
0.058
0.080
0.051
0.031
See footnote on Table 5.1
60
Hge and Arts Participation: 1 Q B2-1 99 7
Table 5.6.
Regression Results of Ballet Attendances on Age by Cohort
(standardized coefficients)
Pre- Boomers
Boomers
Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
0.039
0.041*
-0.031
0.060"
0.060**
0.002
-0.020
-0.009
-0.023
0.045*
0.040*
-0.012
0.094**
0.029
0.012
0.033
0.080**
0.052
0.077**
0.047**
-0.008
0.034
0.008
-0.028
0.019
0.027
0.000
0.074**
0.052**
-0.023
0.026
0.038
0.028
0.007
See footnote on Table 5. 1 .
Table 5.7.
Regression Results of Art Museum Attendances on Age by Cohort
(standardized coefficients)
Pre-Boomers
Boomers
Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
0.003
0.030
-0.002
0.095**
0.022
-0.009
0.023
0.004
-0.026
0.057**
0.093**
0.090**
0.150**
0.053**
-0.006
0.194**
0.180**
0.168**
0.074**
0.103**
0.062*
0.007
-0.037*
-0.027
0.047*
0.009
0.022
0.070**
0.064**
0.094**
0.037
0.142
0.096
0.073
See footnote on Table 5. 1 .
The difference in predictive power seen across all birth cohorts means that those
interested in increasing arts participation cannot focus on the same set of variables
across all age groups. In most art forms the usual list of predictors of arts atten-
dance for the population born before World War II, variables including education,
Chapter 5 61
income and gender, are no longer as important in determining attendance. The
findings reported in Research Report #37 suggest that this difference is due in part
to differences between cohorts and is not simply a function of the age of respon-
dents. If Research #37 showed that baby boomers are different from the cohorts
born before them, the findings reported here suggest that the post-boomer cohorts
are different again.
THE RELATIVE CONTRIBUTION OF AGE AND THE CONTROL
VARIABLES TO PARTICIPATION IN EACH BENCHMARK ART
FORM
Here the predictors of arts attendance in each of the art forms are examined
in turn. These results are found in Tables 5.1 through 5.7. As in Chapter 4, to
facilitate comparisons, the focus here in Chapter 5 is on the Betas (standardized
coefficients).
Classical Music Looking first at the Betas across the top line of Table 5.1, even
after separating the whole sample into the three age groups (pre-boomers,
boomers, and post-boomers), age is still an important predictor of classical music
attendance for the older two groups. This means that even within these age groups,
older pre-boomers and older boomers are more likely to attend than are their
somewhat younger colleagues. Age is the third most important predictor for pre-
boomers, and the second most important predictor among boomers.
Reflecting the findings of Chapter 4, the respondents' years of education is by
far the most important predictor of classical music attendance. Among pre-
boomers and boomers the importance of the respondents education and,
independently, by their father's education as well. Among post-boomers education
along with student status are the only two significant predictors of classical music
attendance.
Household income is the second most important predictor for pre-boomers but
falls to sixth most important for boomers. Only among pre-boomers are women
more likely to attend classical music concerts than are men. Finally, being married,
and metropolitan residence, are the only significant predictors of classical music
attendance among boomers.
Opera Only among pre-boomers, as is seen in the Betas of Table 5.2, is age a sig-
nificant predictor of opera attendance. Income is by far the best predictor among
pre-boomers but is unimportant among the younger age groups.
62 Age and Arts Participation: 1982-199?
Education is the most important predictor of boomer and post-boomer opera
attendance. Metropolitan residence is the second most important predictor for
post-boomers and is also a significant predictor among the older age groups. As
with classical music, gender is important only for pre-boomers. Being married is a
significant predictor of not attending the opera among pre-boomers and boomers,
as is being black among post-baby boomers.
Musical Theater As seen in the Betas of Table 5.3, education is far and away the
best predictor of musical theater attendance for all three age groups, and father's
education is also important for post-boomers.
For pre-boomers and boomers, income is the second most important predictor
of attending musicals but is not significantly important among post-boomers.
Women of all ages are more likely to go to musicals than are men. Among baby
boomers, blacks, and those living in metropolitan areas are more likely to attend
musicals than are whites and those living outside the metropolitan areas.
Jazz The Betas in Table 5.4 show that among pre-boomers, few of whom attend
jazz concerts, higher education and income are the only significant predictors of
attending jazz concerts. Education is of prime importance among boomers and
post-boomers as well.
For boomers and post-boomers, both being married and having young children
in the home significantly reduce jazz attendance. Finally, the respondents being
black is significantly correlated with jazz attendance, but only among baby
boomers.
Theater Education is the most important predictor of theater attendance for pre-
boomers and boomers alike as shown by the Betas. But beyond this point the
predictors of theater attendance are quite different across birth cohort groups.
Among pre-boomers, family income, age, and gender follow in importance. In con-
trast, not being married and living in a metropolitan area are the second and third
most important predictors for boomers.
Among post-boomers, education is the all-important predictor, but interestingly,
the best single predictor is father's education, followed by student status and
respondents education. The only other significant predictor of theater attendance
among post-boomers is not being married.
Ballet In contrast to the art forms discussed so far, income is the best predictor of
ballet attendance among pre-boomers, as can be seen by the relevant Betas in Table
Chapter 5 63
5.6. Curiously, father's education is the second best predictor and the respondent's
education is not a significant predictor. Being female is an important predictor, and
it may well he that many of these older female ballet fans born before the Second
World War came from well educated families but, because of their gender, did not
so often have educational opportunities themselves. Living in a metropolitan area
and not currently being married are the two other variables important in predict-
ing ballet attendance among pre-boomers.
Fathers education is still important for baby boomers but respondent's educa-
tion is the most important predictor followed by being female and living in a
metropolitan area. Unlike several of the other art forms already examined, having
children does not depress boomer attendance. This may be because many young
people, especially girls, are regularly taken to ballet performances.
The results for post-boomers are strikingly different, because none of the meas-
ured variables contribute significantly to ballet attendance.
Art Museum As Table 5.7 shows, the predictors of art museum attendance are
roughly the same across all three age groups. These include both respondent's and
father's education, being unmarried, and metropolitan area residence. In addition
to these four, household income is important for boomers and pre-boomers, gen-
der and health are important just for pre-boomers, and having children in the home
is important just for boomers.
Summary Arts Participation Recall that summary arts participation measures the
number of benchmark forms that the respondent has participated in during the
prior year. The results for the summary arts participation measure shown in Table
5.8. Considerably more variables prove to be significant predictors than seen in the
earlier tables for each of the art forms alone. This is not surprising because the
range of this measure is from 0 to 7 while the range of all of the single discipline
measures is 0 and 1. Respondent's education, household income, gender, father's
education, and metropolitan residence are all significant for all three age groups.
Redolent of the findings for several of the individual art forms, age is a sig-
nificant predictor of attendance for pre-boomers and boomers but not for
post-boomers. Finally, reflecting stage-of-life exigencies, health is a predictor of
summary arts participation for pre-boomers and boomers; children in the home
are significant for boomers and post-boomers, and student status is important
for post-boomers.
64
Me and Arts Participation: 1082-199?
Table 5.8.
Regression Results of Summary Arts Attendances on Age by Cohort
(standardized coefficients)
Pre-Boomers
Boomers
Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
0.057**
0.054**
0.019
0.145**
0.092**
0.049*
0.018
0.024
0.014
0.043*
0.096**
0.107**
0.208**
0.163**
0.057**
0.308**
0.311**
0.273**
0.063**
0.082**
0.137**
-0.01 1
-0.035*
-0.060**
0.101**
0.043**
0.022
0.054**
0.063**
0.051*
0.152**
0.273
0.246
0.217
See footnote on Table 5. 1 .
Table 5.9.
Regression Results of Summary Arts Attendances (without Jazz) on
Age by Cohort (standardized coefficients)
Pre-Boomers
Boomers
Post-Boomers
Age
Female
Black
Not Married
Income
Education
Father's Educ
Children
Health
Metro
Student
R-square
0.066**
0.060**
0.007
0.154**
0.107**
0.059**
0.007
0.005
-0.008
0.040*
0.074**
0.095**
0.203**
• 0.153**
0.054*
0.305**
0.318**
0.268**
0.059**
0.077**
0.138**
-0.010
-0.029
-0.049*
-0.103**
-0.047**
-0.024
0.061**
0.268
0.067**
0.245
0.053*
0.141**
0.204
See footnote on Table 5. 1 .
Chapter 5 65
THE RELATIVE IMPORTANCE OF AGE AND INDIVIDUAL
CONTROL VARIABLES IN PREDICTING PARTICIPATION
ACROSS THE BENCHMARK ARTS
Looking again at Tables 5.1 through 5.7, now the focus is on the predictive
importance of each variable in turn across all seven forms. Before examining the
individual variables, however, it is worthwhile noting the consistency in the results
across the tables. Among the 217 coefficients,2" across all seven art forms the sign
of significant relationships is consistent with one single exception.21 This consis-
tency attests the reliability of the measures in predicting arts attendance. Thus,
variability across tables has to do with differences in the strength of the predictive
value of variables by art form and by age, but they do not show any difference in
the directions of predictions.
Turning now to the individual variables, we see that having divided the sample
into three parts by age, the predictive power of age has been attenuated, but not
completely. Older pre-boomers and baby boomers are still more likely to attend
several of the art forms than are the younger members of their cohorts.
The importance of education for socialization to arts participation is very
apparent in the figures in Tables 5.1-5.7. Among pre-boomers, the respondent's
education is the most important predictor of arts participation for all forms except
ballet. It is the most important among boomers for all the arts, and among post-
boomers for five of the forms. Likewise, father's education is also independently
important in four arts forms for each of the age groups, and finally, student status
is important among post-boomers.
Household income is a significant predictor for all seven of the art forms among
pre-boomers, and it is also important for four of the seven among baby boomers.
Yet it is without significance for post-boomers. It may be that differences in wealth
are becoming less important over these birth cohorts. It may also be that the vari-
able "household income" is a poor measure among young adults in any time
period because the current income of those who went on the job market early,
and rarely attend arts events, is still as high as that for those who spend many
more years in formal schooling and are more likely to attend. The importance
of education, father's education, and student status gives credibility to this lat-
ter life-cycle interpretation.
The two ascribed status variables, race and gender, show contrasting results.
Controlling for the other measured factors, Black respondents generally show the
same pattern of arts attendance as do non-Blacks. Only in the case of jazz atten-
dance among baby boomers, are Black rates higher than for the rest, and only for
post-boomers is attendance at opera lower than for non-Blacks.
66 Age and Brts Participation 1982-119?
The results for gender show a pattern that is clearly consistent with changes
between cohorts. Controlling for the other factors, pre-boomer women are
more likely to attend every art form more often than are men except jazz.
Among baby boomers, women are more likely to attend only two forms, musi-
cals and ballet, and among post-boomers the attendance rates for women and
men are virtually the same for every form except musicals, which women attend
more often than men.
Four of the variables measure the effects of exigencies that affect arts participa-
tion. The first is marital status. Those not married that is to say the never married,
the divorced, and the separated — show higher rates of arts participation in four art
forms among pre-boomers, in all seven among boomers, and in three forms among
post-boomers as compared with those who are married or widowed. These find-
ings suggest that those married persons who have available companionship and
those who are involuntarily alone, are less likely to seek out the arts than are those
who are unmarried or otherwise voluntarily living alone.
Large cities provide more arts participation opportunities than do smaller cities
and towns, so it is reasonable to expect that metropolitan place of residence should
equally influence all age groups. Metropolitan residence is important for partici-
pation in four of the arts forms for pre-boomers, and six forms for baby boomers.
But among post-boomers, it is important for only two art forms, opera and
museum attendance. It may be that due to wider travel, more education, and
greater exposure to the arts via the media, metropolitan residence is becoming less
important. If this is the case, this is a clear cohort effect. It may also be that more
of the well-educated and arts-oriented younger people living outside metropolitan
areas are located in college towns and other high-technology towns.
Finally, the participation of two variables, children in the household and health,
should be correlated with their stage of life, and, indeed, that is seen in the data.
Presence of children in the home is an irrelevant consideration among most of the
older pre-boomers, but significantly depresses participation in several art forms for
boomers and post-boomers cohorts, many of whose members were in 1997 in the
midst of child raising. In parallel but opposite direction, health is irrelevant for
post-boomers, is somewhat important for boomers and is even more important for
the more elderly pre-boomers. These patterns are reflected exactly in the summary
arts participation shown in Table 5.8. The presence of children in the home is
important only among the two younger cohorts and (poor) health is important
only among the two older cohorts. Thus the findings for these two variables reflect
changes associated with life-stages.
NOTES
1 As a measure of "average age" we use the "median." The median value of a
set of measures means that half the individuals in the sample are older and half are
younger than the median age. Another way to measure "average age" is to use the
mean values. The mean takes into account the distance of all the individual ages
from the mean. The expected mean age for 1982 is 43.1 years, for 1992 45.4, and
for 1 997 46. 1 . These mean ages are three and a fraction years older than their cor-
responding medians for each of the survey years reported in Table 1.1 because
those below the mean cannot be younger than 18 while those above the mean can
be in their 80s or older. Therefore, we find it more appropriate to use the median
as the measure of central tendency, of "average."
2 SPPA respondents are older on average than the United States civilian popula-
tion because only those 18 years of age or older were sampled (NEA 1998a).
3 It is worth noting that in all cases the ages of art-form audiences in 1 992 is
between, or equal to, those for 1982 and 1997, giving support to the assertion that
the comparison of proportions method used here makes the 1997 figures compa-
rable with those of earlier survey years.
4 It is also worth noting at the outset that those in the sample who are 60 and
over have gone from 21.2 percent in 1982 to 23.1 percent in 1997 of survey
respondents. This is an increase of just under two percent, but as we will see in
Chapter 2, the proportion of these elders in arts audiences has risen considerably
more.
5 They have, for example, linked age with the changing size of the audience and
the influence of age relative to other factors on the rate of audience attendance over
time.
6 A calculation was made of the relative percentage that each age group and
birth cohort represented in the total number of "attendances" (the number of times
each respondent attended the art form) in each sample. A variable in each data set
represented the number of times the respondent attended that benchmark art in the
last twelve months (with the exception of 1982, in which the question covered only
the last month. This value was assumed to represent an "average" month and was
multiplied by twelve to reflect the number of attendances in a year).
7 Figures for attendees under twenty years of age are shown in all the tables in
this Chapter, but they will not be discussed here because the numbers are based just
on those respondents 18 and 19 years of age while all the other age groups (except
for the eldest) span ten years.
68 | Rge and Arts Participation: 1 Q B2-1 QQ7
8 These figures are reached by adding 8.5 and 7.1 percent = 15.6 percent and
14.4 and 15.7 percent = 30.1 percent. The computations for other combined age
groups will be made in the same fashion without note.
9 If you wish to reconstruct a table of the proportion of the audience of any par-
ticular age for any of the art forms who are attendees, simply subtract, if positive,
or add, if negative, the appropriate number in the difference tables (Tables
2.8-2.14) from the corresponding cell in the attendance tables (Tables 2.1-2.7).
Thus, for example, to find the proportion of the symphony orchestra audience that
were teens in 1982 subtract 1.0 (found in the upper left-hand cell of Table 2.1 )from
5.4 (found in the upper left-hand cell of Table 2.8). The resulting 4.4 is the pro-
portion of all classical music attendees in 1982 who were teens.
10 The relative percentage that each birth cohort contributed to the total num-
ber of "attendances" (the number of times one attended a benchmark art) in each
benchmark art form for each year were calculated. A variable in each 1992 and
1997 data set represents the number of times the respondent attended that partic-
ular benchmark art in the last twelve months. In 1982, the question covered only
the last month, so this value was assumed to represent an "average" month and
was multiplied by twelve to reflect the attendances in a year. To calculate the atten-
dance figures shown in Tables 3.1 through 3.7 cross-tab frequencies of the cohort
by the number of attendances were made in order to know how many respondents
in each cohort attended that benchmark art form every possible number of times.
Using the example of jazz attendance mentioned in Note 2 of Chapter 2, in 1997
the 20-29 age group, 95 respondents attended jazz one time, 75 attended two
times, 48 attended three times, etc., up to a maximum possible value of 72 times
in one year. Next, each value in the frequency cell was multiplied by corresponding
the number of attendances for each cohort (in the same example, 95 times 1, 75
times 2, 48 times 3, etc.). Then these products were summed for each cohort to rep-
resent the total number of times that respondents in this cohort attended that
benchmark art in the previous year (e.g., the 20-29 year old respondents attended
jazz a total of 1,084 times in 1997). Finally, the sum for each age group or cohort
was divided by the summed total of every age group or cohorts' attendances (the
global number of attendances by all respondents in that year) to reflect the pro-
portion of attendances reported by each cohort in relation to the others within each
sample year (e.g., the 1,084 total for the 20-29 group was divided by the grand
total of 5,123 attendances for the whole sample in 1997, showing that 21 percent
of total attendances at jazz concerts were 20 to 29 years old).
11 Unlike the other decade-long cohorts, this one includes cohorts, this one
include just those born over a five-year period. This is because in the survey year,
dotes 69
1997, those horn in 1981-1985 were under 18 years of age and thus were too
young to be part of the SPPA survey sample.
u The observed participation for each cohort is found in the appropriate Table
from 3.1 to 3.7. The expected participation for each cohort is expected in the
appropriate row of Table 1 .2. The difference figures in parentheses are obtained by
subtracting the expected participation from the observed participation.
" In this instance, the obtained percent is 5.2 and the expected percentage is 4.5,
and since the observed is larger than the expected, the sign of the figure is positive.
14 The data for 1992 are available, but they are not explicitly examined here.
The 1992 values are intermediate between those of 1982 and 1997 in 20 of the 42
possible comparisons (six cohorts x seven art forms). And in most of the rest of the
cases where a figure for 1992 was not intermediate, it was roughly equal to that of
1982 or 1997. The only notable exceptions are the audiences for theater and for
musical theater. The 1 992 audience for theater tended to be older, and the audience
for musical theater tended to be younger, in 1992 than in the survey years before
and after 1992. These variations from the norm probably have to do with changes,
beyond the scope of this monograph, taking place in these art forms during the
early 1990s.
15 To be sure, that report found that the participation rate of these earlier
boomers was low compared to earlier cohorts after their relatively high education,
income, and other factors were taken into account. That is to say, their participa-
tion rate was low given their relatively advantaged status.
16 Scores in Tables 3.8-3.15 reflect the proportion of attendance minus the pro-
portion of attendees.
17 If you wish to construct a table of the proportion of the audience of any par-
ticular age for any of the art forms who are attendees, simply subtract, if positive,
or add, if negative, the appropriate number in the difference tables (Tables
2.8-2.14) from the corresponding cell in the attendance tables (Tables 2.1-2.7).
Thus, for example, To find the proportion of the symphony orchestra audience that
were teens in 1982 subtract 1.0 (found in the upper left-hand cell of Table 2.1)from
5.4 (found in the upper left-hand cell of Table 2.8). The resulting 4.4 is the pro-
portion of all classical music attendees in 1982 who were teens.
18 The earliest cohort, the one including all those born before 1916 will not
be considered here because, while its participants tend to attend less often than
average, its rates of participation do not show any coherent pattern across all
the art forms.
19 We chose to use simple OLS regression with the dichotomous dependent vari-
able "attendee" rather than logistic regression in order to facilitate interpretation
70 Rg e and Arts Participation: 1902-1997
of the results and simplify comparisons across all three sets of models. For a justi-
fication of this choice see Davis (1994). Davis, James A. 1994 "What's Wrong with
Sociology?" Sociological Forum 9: 179-197.
20 This number of measures is obtained by recognizing that for each of the seven
art forms there are ten variables for each of the three age groups and there is an
additional variable (student status) for post-boomers. This means (7 x 10 x 3) + 7
= 217 measures.
21 The single reversal in this finding is that opera-going among black post-
boomer respondents is significantly lower than expected by chance.
Bibliography
Balfe, Judith H. (1989). "The Baby-boom Generation: Lost Patrons, Lost
Audience?" Pp. 9-25 in Margaret Wyszomirski and Pat Clubb (eds.) The
Cost of Culture. New York: ACA Books.
Davis, James A. (1994). "What's Wrong with Sociology?" Sociological Forum 9:
179-197.
DiMaggio, Paul (1982). "Cultural Entrepreneurship in Nineteenth-century Boston:
the Creation of an Organizational Base for High Culture in America."
Media, Culture, and Society 4: 4-33; 303-320.
. (1991). "Constructing an Organizational Field as a Professional Project:
The Case of Art Museums, 1920-1940." In Walter W. Powell and Paul
DiMaggio, (eds.) The New Institutionalism in Organizational Analysis.
Chicago: University of Chicago Press.
Esterlin, Richard A. (1987). Birth and Fortune. Chicago: University of Chicago
Press.
Glenn, Norval D. (1989). "A Flawed Approach to Solving the Identification
Problem in the Estimation of Mobility Effect Models." Social Forces
69:789-795.
Loomis, Laura and Mary Collins (1998). "Changes in Survey Procedures and their
Potential Effects on Estimates of Arts Participation: 1997 Survey of Public
Participation in the Arts." Westat, unpublished.
Mather, John (1993). "A Challenging New Audience — 'Baby Boomers.'" Bulletin
of the Association for Performing Arts Presenters. March: 3.
Miller, Judith (1996). "As Patrons Age, Future of Arts is Uncertain." New York
Times p Al, C12 February 12.
National Endowment for the Arts (1993). Research Division Report #27. Arts
Participation in America: 1982-1992. National Endowment for the Arts:
Washington D.C.
National Endowment for the Arts (1996). Research Division Report #37. Age and
Arts Participation: With a Focus on the Baby Boom Cohort. Richard A.
Peterson, Darren Sherkat, Judith Balfe, and Rolf Meyersohn. Santa Ana, CA:
Seven Locks Press. National Endowment for the Arts: Washington D.C.
National Endowment for the Arts (1998a). Research Division Report #39. 1997
Survey of Public Participation in the Arts: Summary Report. National
Endowment for the Arts: Washington D.C.
72 fl g e and Arts Participation HB2-IH7
National Endowment for the Arts (1998b). Research Division Note #70.
"Technical Appendix" to "Comparing 1997 Survey of Public Participation
in the Arts (SPPA) Results with Prior SPPAs (1992,1985,1982)." National
Endowment for the Arts: Washington D.C.
National Endowment for the Arts (1999). Research division Note #71.
Demographic Characteristics of Arts Attendance: 1997." National Endow-
ment for the Arts: Washington D.C.
Newman, Katherine S. (1993). Declining Fortunes: The Withering of the American
Dream. New York: Basic Books.
Peterson, Richard A. (1992). "The Battle for Classical Music in the Air." Pp.
271-286 in Judith H. Balfe (ed.) Paying the Piper: Causes and Consequences
of Arts Patronage. Urbana, Illinois: University of Illinois Press.
. and Albert Simkus (1992). "How Musical Taste Groups Mark
Occupational Status Groups." Pp. 152-168 in Michele Lamont and Marcel
Fournier, editors, Cultivating Differences: Symbolic Boundaries and the
Making of Inequality. Chicago: University of Chicago Press.
Rodgers, Willard L. (1982). "Estimable Functions of Age, Period, and Cohort Ef-
fects." American Sociological Review 47: 774-787.
*to
NATIONAL
ENDOWMENT
FOR THE ARTS
National Endowment for the Arts
Seven Locks Press
Santa Ana, California
ISBN# 929765-86-9
51195
9 780929"76586(