NEA Research Note #101
February 2011
Three NEA Monographs on Arts Participation: A Research Digest
Introduction
Every few years, the National Endowment for
◊
NEA Research Report #52, Arts
the Arts partners with the U.S. Census Bureau
Education in America: What the
to conduct the Survey of Public Participation
Declines Mean for Arts Participation,
in the Arts (SPPA), a study that tracks adult
by Nick Rabkin and E.C. Hedberg,
levels of involvement with arts activities.
NORC at the University of Chicago,
Since the early 1980s, social scientists within
http://www.arts.gov/research/2008-
the academic, government, nonprofit, and
SPPA-ArtsLearning.pdf
commercial spheres have supplemented the
NEA’s official findings with their own
◊
Arts Participation: A Case against
analyses of the SPPA data.
Demographic Destiny, by Mark J.
The most recent wave of the survey occurred
Stern, University of Pennsylvania,
in 2008. Soon afterward, the NEA
http://www.arts.gov/research/2008-
commissioned independent researchers to
SPPA-Age.pdf
mine the SPPA data for details on the
following topics: arts education; the personal
performance and creation of artworks; and the
NEA Research Report #53, Age and
◊
relationship between age and arts
participation. This Research Note presents key
findings from their investigations, which have
resulted in three research reports, now
available on the NEA website (arts.gov):
Three NEA Monographs on Arts Participation: A Research Digest
NEA Research Report #54, Beyond
Attendance: A Multi-Modal
Understanding of Arts Participation,
by Jennifer L. Novak-Leonard and
Alan S. Brown, WolfBrown, http://
www.arts.gov/research/2008-SPPABeyondAttendance.pdf
Working separately, but from a common data
and even what a full range of arts participation
source, these researchers brought original
opportunities might look like.
research hypotheses and methods to bear on
their own analyses. Collectively, the reports
challenge popular wisdom about which factors
Bonnie Nichols, NEA Office of Research &
Analysis, discusses their findings below.
are central to the future of arts participation in
America, who does or does not participate,
Summary
reported from 1982 to 2008,
1. Long-term declines in childhood arts
education have serious implications for
resulted partly from cuts in school-
the future of arts participation in
based arts instruction.
America. (Rabkin & Hedberg)
•
•
•
The relationship between arts
African Americans accounted for a
education and adults’ rates of arts
highly disproportionate share of all
participation has been consistently
adults who reported not having
strong throughout the survey’s
received arts education in
history.
childhood.
By 2008, only half of all 18-yearolds (49.5 percent, or 2.2 million)
had received any arts education in
2. Age is a poor predictor of arts
participation habits. (Stern)
childhood—a decline of 23 percent
since 1982.
•
From 1982 to 2008, Hispanics and
According to long-term patterns of
respondent recall, a “turning point”
in national access to arts education
likely occurred in the mid-1970s
and early 1980s. It seems
•
After accounting for other factors,
age predicted only 0.4 percent of
the variance in the total number of
arts events that U.S. adults
attended in 1982-2008. By
contrast, education predicted 15
percent.
reasonable to infer that the national
declines in arts education rates,
Three NEA Monographs on Arts Participation: A Research Digest
2
•
•
•
The strength of the relationship
3. A more comprehensive picture of
between age and arts attendance has
arts engagement—one not focused
waned over time. From 1982 to
exclusively on live arts attendance
2008, the age and generation group
rates—yields a narrative that is
of U.S. adults never predicted more
different from prior NEA reports
than 2 percent of the variance in the
about U.S. adult participation in the
total number of events attended.
arts. (Novak-Leonard & Brown)
The age distribution of arts-goers
•
Three out of four U.S. adults (74
now generally mirrors that of the
percent, or 166.4 million) did any
U.S. adult population. At jazz
single arts activity (exclusive of
performances, for example, Baby
literary reading) in the 2008 Survey
Boomers continue to dominate the
of Public Participation in the Arts,
audience population—just as they
inclusive of creating art or
did in the 1980s, when they were
participating via electronic media.
among the youngest age groups
This rate is more than double that
represented.
of attendance at “benchmark” arts
events.
“Cultural omnivores” have declined
as a share of the U.S. adult
•
One out of three adults (33 percent,
population. Also, these
or 74.2 million) both attended and
Americans—who typically attend a
created art. In contrast, 17 percent
variety of arts events, in many
of adults only attended arts, and 12
different art forms and settings—
percent only created or performed
curbed the average number of
art.
events they attended between 2002
and 2008. These two factors
accounted for 82 percent of the
overall decline in the total number
of “benchmark” arts events
attended over that period.
Three NEA Monographs on Arts Participation: A Research Digest
3
•
•
Relatively high rates of attendance
Arts education in childhood is one
at arts festivals—as well as
of the best predictors of both arts
attendance at schools and places
attendance and arts creation and
of worship—suggest the
performance later in life.
importance of venue to overall
arts-participation rates.
Relative Effects of Education and Arts Learning on Arts Participation
Previous NEA research reports have shown
least one “benchmark” arts activity, when a
the strong correlation between a person’s
variety of demographic and other variables
educational attainment—including a
have been held constant. Compared with
background in arts education—and his or her
adults who have only a grade-school
patterns of arts participation in adulthood.1
education, for example, adults with at least
Each of the three reports under discussion
some college are about 20 percent more likely
(#52, #53, and #54) reaches a similar
to attend a benchmark arts event, regardless of
conclusion with regard to the primacy of
their gender, age, race, income, or whether or
education and arts education in predicting
not they live in an urban/metro area.2 For
personal arts involvement.
adults with graduate degrees, the likelihood is
In their report, Novak-Leonard and Brown
examine the likelihood of adults attending at
more than 40 percent greater. (See the
following graph.)
“Benchmark” arts activities tracked since 1982 include jazz, classical music, opera, musical or
non-musical plays, ballet performances, and visits to art museums or art galleries.
Three NEA Monographs on Arts Participation: A Research Digest
4
Likelihood of Attending a Benchmark Arts Event,
by Educational Attainment
50%
40%
30%
20%
10%
0%
Some high
school
High school
graduate
Some college
Bachelor's
degree
Graduate
degree
Note: The basis for this comparison of attendance is adults with only a grade-school level of education
Source: Novak-Leonard and Brown, NEA Research Report #54
So much for the potential impact of education
performance. In the report by Novak-
on arts attendance. What about the role of arts
Leonard and Brown, arts education seems to
classes or lessons in fostering this behavior?
operate as a “leveler”—in effect, reducing the
Novak-Leonard and Brown show that even
potential impacts of socioeconomic status,
after we control for gender, age, race, and
including such variables as education and
other variables, adults who have taken art
income.
classes at any time in their lives are still more
than 20 percent more likely to attend
benchmark arts activities (compared with
Americans who have never taken art classes).
Let’s consider first the relationship between
general educational attainment and personal
arts creation and performance. Similar to the
findings for arts attendance, the likelihood of
Not only is arts education a key predictor
creating or performing art rises with
of adults’ attendance patterns; it has an
education.3 Adults with a bachelor’s degree
even stronger relationship with adults’
are 16 percent more likely to create or
levels of personal art creation or
perform art, compared with those whose
Three NEA Monographs on Arts Participation: A Research Digest
5
highest level of training is elementary school.
with adults who took no childhood art classes.
The likelihood rises to 23 percent for adults
Adults who took childhood classes in at least
with graduate degrees.
four art subjects were three times more likely
This outcome changes, however, when the
to attend the arts.
taking of art classes or lessons is introduced
Especially in light of this relationship, it is
into the model. After this adjustment,
disheartening to observe long-term declines in
education no longer predicts levels of arts
arts education as well as large differences in
creation or performance. Adults who have
the socioeconomic status of Americans who
taken art classes at any time in their lives are
have received an arts education and those who
32 percent more likely to create art of their
have not.
own.
Rabkin and Hedberg’s analysis reveals two
Arts education has a similar leveling effect on
telling characteristics of arts learning. First,
U.S. citizenship as a predictor of personal art-
the percentage of young adults taking
making. Before the taking of art classes is
childhood art classes, as captured by the
included in Novak-Leonard and Brown’s
SPPA, has declined.5 In 1982, nearly two-
statistical model, naturalized citizens and non-
thirds of 18-year-olds reported taking art
citizens are less likely to create or perform art
classes in their childhood. By 2008, that share
than native-born adults. But once art classes
had dropped to 50 percent.
are considered, citizen status also drops out as
a predictor of creation and performance.4
Access to Arts Learning
In their report, Rabkin and Hedberg reaffirm
the importance of arts education in predicting
arts participation rates. Combining data from
the 1982, 1992, 2002, and 2008 SPPA waves,
a statistical analysis reveals that adults who
took childhood classes in at least one art
form were about 50 percent more likely to
attend a “benchmark” arts event, compared
By tracking the rate of self-reported arts
education in childhood by the age of SPPA
respondents, Rabkin and Hedberg show that
childhood arts education likely grew
throughout much of the 20th century. A
turning point seems to have occurred in the
mid-1970s through early 1980s, however, as
the percentage of young adults who reported
having studied art as children began to fall.
This downward trend has continued into the
21st century.
Three NEA Monographs on Arts Participation: A Research Digest
6
According to the authors, this pattern stems
those survey years, most people who reported
from the expansion of arts education in public
having received music or visual arts
schools through the early 1970s, followed by
instruction said they did so in school.
declines in school-based arts instruction that
Therefore, describing 18-year-olds whose
are understood to have started in the mid-
childhood arts education was captured in the
1970s. Rabkin and Hedberg observe that in
SPPA data from 1985 through 2008, Rabkin
the late 1970s and throughout the 1980s,
and Hedberg note:
many school systems across the nation began
reducing arts instruction in response to budget
constraints and a stronger emphasis on
“basic,” non-art subjects.6
Declines were greatest in music and
visual arts, the two arts subjects taught
most in schools, while theater and
dance actually recorded small
In the authors’ view, the proportionately
increases. Given the mandatory status
greater declines in the rates of music and
of public education, there is good
visual arts education that 18-year-olds recall
reason to believe that the general
receiving in childhood, compared with the
decline in arts education participation
rates of decline reported for other types of arts
in childhood was in large measure the
learning, reflect those changes to school
result of cuts in school-based arts
curricula. Previous SPPA data (from 1992 and
instruction.7
2002) allowed researchers to identify where
survey respondents had received their arts
instruction—whether in or out of school. In
Three NEA Monographs on Arts Participation: A Research Digest
7
Percentage of U.S. Adults Who Received Any Arts
Education in Childhood Over the Past 77 Years
75%
50%
25%
0%
1930
1943
1956
1969
1982
1995
2008
Year respondent was 18 years old
Reported instances of childhood arts education recalled by survey respondents
Trend line
SPPA year
Source: Rabkin and Hedberg, NEA Research Report #52
A second key finding is that the decline was
Although sizable (15 percentage points), this
much sharper for Americans whose parents
drop pales in comparison with the decline for
were less educated. As the following chart
respondents whose parents were not as well-
indicates, childhood arts education rises with
educated. Between 1982 and 2008, the rate of
parents’ education levels, an accepted proxy
childhood participation in arts classes or
for the respondent’s socioeconomic status in
lessons, among young adults whose parents
childhood. In 1982, for example, nearly 90
were high school graduates, declined by more
percent of young adults taking childhood arts
than 36 percentage points—from 70 percent to
classes had parents with bachelor’s degrees or
34 percent.
higher levels of training. By 2008, that
proportion fell to 73 percent.
Three NEA Monographs on Arts Participation: A Research Digest
8
Even in families with low levels of education
childhood was a sizable 54 percent in 1982.
(less than a high school diploma), the
By 2008, however, few young adults from this
percentage of adults taking art classes in
group had taken classes—just 13 percent.
Percentage of U.S. Adults that Reported Taking Arts Classes
in Childhood, by Parent's Level of Education
(Ages 18-24)
73.2%
Bachelor's degree or higher
88.7%
51.8%
Some college
86.8%
2008
1982
33.7%
High school
Less than high school
70.0%
12.5%
54.0%
Source: Rabkin and Hedberg, NEA Research Report #52
Also telling are disparities in childhood arts
Americans had taken art classes when they
learning by race and ethnicity.8 As the
were children. By 2008, that percentage fell
following table shows, between 1992 and
to 28 percent—a 16-point decline.
2008 much of the decline in the percentage
of young adults taking art classes in
childhood was among African Americans. In
1992, nearly 44 percent of young African
Childhood arts learning also fell among
whites, but only by 7 percentage points—not
even half the drop reported by African
Americans. (The decline was also nearly 7
Three NEA Monographs on Arts Participation: A Research Digest
9
percentage points for young Hispanics.)
quarter of both African Americans and
Overall, however, the SPPA reveals a large
Hispanics experienced childhood arts
race/ethnicity gap in childhood arts learning.
learning.
In 2008, almost 60 percent of young white
adults reported having taken art classes as
children. By comparison, only about onePercentage of U.S. Adults Reporting Childhood Art Classes, by Race and Ethnicity
(Ages 18-24)
White*
African American*
Hispanic
1992
2008
64.8%
43.5%
34.8%
57.9%
26.2%
28.1%
Change
(pp)
-6.9
-17.3
-6.7
*Non-Hispanic
pp=percentage points
Source: Rabkin and Hedberg, NEA Research Report #52
Women tend to participate in the arts at higher
In 1982, for example, 59 percent of women
rates than men do, even after controlling for
18-24 took art classes in childhood, versus
arts classes or lessons received in childhood,
just under 55 percent of young men. After
as well as a variety of demographic
falling precipitously in 2002, the share of
characteristics such as age and education.9
young men taking childhood art classes
Coincidentally or not, as the following graph
climbed to 45 percent in 2008. The rate for
shows, higher percentages of women also
young women also fell sharply in 2002. But
report taking art classes as children.
in 2008 it continued to fall, narrowing the
gender gap in childhood arts education rates.
Three NEA Monographs on Arts Participation: A Research Digest
10
Percentage of U.S. Adults Who Received Arts
Education in Childhood, by Gender: 1982-2008
(Ages 18-24)
70%
65%
60%
Men
55%
Women
50%
45%
40%
1982
1992
2002
2008
Source: Rabkin and Hedberg, NEA Research Report #52
Arts Participation by Age
Age Composition of Arts Audiences
With the release of the 2008 SPPA data,
researchers noted not only long-term declines
for attendance at many types of arts events,
but also much less representation of younger
age groups in U.S. adult audiences for the
arts. For example, in 2008 the average adult
was 45 years old, six years older than in 1982.
population. However, once the age make-up
of the entire SPPA population is better
accounted for, the results are less dramatic.
Indeed, in 2008, the distribution of young
adults (18-29 years old) and older
Americans (60 years and older) in arts
audiences more closely matched each
group’s share of the general population.
Over this period, however, the average jazz
An “index of representativeness” shows the
concertgoer aged 17 years to reach 46; the
percentage by which the audience share for a
typical ballet attendee was also 46—up from
particular age group is larger or smaller than
10
an average age of 37 in 1982.
On the surface, these figures depict an arts
audience aging more rapidly than the adult
its share of the entire population.11 In 1982,
for example, the “benchmark arts index” was
11 for young adults (18 to 29 years old). This
means that compared with their share of the
Three NEA Monographs on Arts Participation: A Research Digest
11
U.S. adult population, young adults exceeded
As the following table shows, young adults
their share of arts audiences by 11 percent.
exceeded their share of arts audiences in 1982,
By 2008, the index for this age group was 0,
while much older adults (ages 60 and above)
which means their attendance share matched
were underrepresented. By 2008, attendance
their share of the adult population.
was generally closer to each group’s share of
In 1982, the benchmark index for adults 60
the U.S. adult population.
and above was -32 (they were 32 percent
below what we would expect, given their
share of the adult population). In 2008, the
index of representativeness rose to -13.
Although not 0, adults 60 and older were
better represented (more in line with their
share of the population) at benchmark arts
events in 2008 than in 1982.
Index of Representativeness for Benchmark Arts Activities: 1982-2008
Year
Age group:
1982
1985
1992
2002
2008
Under 30
11
2
3
-6
0
30-44
13
16
7
6
7
45-59
0
-1
6
15
4
60 and
older
-32
-24
-20
-21
-13
Source: Stern, NEA Research Report #53
These patterns differ somewhat for attendance
which is much closer to 0, indicating that their
at selected art forms. In 1982, for example,
share of the jazz audience was closer to their
adults under 30 were strongly overrepresented
share of the total adult population.
in jazz audiences—their index of
representativeness was 76. By 2008,
however, the index for young adults fell to -6,
A similar trend may be observed in adults 60
and older. The jazz attendance index for this
age group was -75 in 1982. By 2008, the
Three NEA Monographs on Arts Participation: A Research Digest
12
index climbed to -23—a 52-point
percent. This pattern likely reflects the aging
improvement in their share of jazz audiences.
of the Baby Boomers. Just as Boomers led
the jazz-going population when they were
Also notable is the representativeness of
young in 1982, their appetite for jazz
adults ages 45-59. In 1982, this age group
continued when they were middle-aged in
was 32 percent below its expected share of
2002 and 2008.
jazz audiences. By 2008, however, they
exceeded their share of jazz audiences by 29
Index of Representativeness for Jazz Attendance: 1982-2008
Year
Age group:
1982
1985
1992
2002
2008
Under 30
76
41
17
2
-6
30-44
1
28
21
12
-6
45-59
-32
-26
-5
17
29
60 and
older
-75
-68
-46
-40
-23
Source: Stern, NEA Research Report #53
With the exception of young adults, ballet
Audiences for classical music concerts, on
became more evenly represented by age. In
the other hand, became decidedly older. In
1982, the ballet index for adults under 30 was
1982, for example, the classical music index
-1, suggesting that their share of the ballet
for adults under 30 was -11. By 2008, the
audience was nearly on par with their share
index fell to -26. In other words, young
of the adult population. By 2008 the index
adults now make up an even smaller share of
for this age group fell to -17. For all age
the classical music audience. Americans 60
groups 30 and older, however, the index
years of age and older were also below their
improved over time. The index for adults
share of the population in 1982 (-18). But in
between 30 and 44 years of age was 34 in
2008, the index for this age group jumped to
1982, but by 2008 it was 4. Adults 60 and
22.
older were underrepresented in the ballet
audience in 1982 (an index of -33). In 2008,
the index for this age category was 9.
Three NEA Monographs on Arts Participation: A Research Digest
13
Age as a Predictor of Arts Participation
Despite the aging arts audience described
above, a more controlled analysis shows that
age and cohort are weak predictors of arts
participation. Once other characteristics are
considered—particularly education—the year
The overall model, including the effects of all
the variables, resulted in an “R square”
statistic of 20.5 percent. This means that the
combined variables predicted 20.5 percent of
the variance in average number of benchmark
activities attended.
a person was born plays only a marginal role
Adding age to the model yields little change
in predicting arts participation.
to the outcome. Independently, the
For example, combining data from the 19822008 SPPAs, Stern employed a regression
model that relates the average number of
benchmark arts activities attended to various
demographic and other characteristics,
including gender, marital status, educational
contribution from education dropped slightly
to 15.4 percent, while age predicted only 0.4
percent of the variance in attendance.
Moreover, adding age to the model increased
the R square value only slightly—from 20.5
percent to 21.1 percent.
attainment, and ethnicity. For this first
These results suggest that the effect of age on
model, age is excluded.
arts participation, though not zero, is
Of the variables modeled, education was the
stand-out predictor: on its own, it predicted
18.3 percent of the variance in number of
marginal. As the other SPPA research reports
have found, educational attainment is a far
better predictor of arts participation.
benchmark activities attended. The other
variables, independently, predicted no more
than 0.7 percent of the variance.
Three NEA Monographs on Arts Participation: A Research Digest
14
Percentage of Average Number of Arts Activities Attended,
as Predicted by Key Variables
Variable
Age
Education
Ethnicity
Gender
Marital status
1
Percent
0.4%
15.4%
0.5%
0.8%
0.4%
1
The percentage of the variation in average number
of activities attended over the period of 1982-2008,
predicted by each variable, independently.
This percentage is the "partial eta square."
This general linear model also included interactive terms,
including an interaction between age and education, which
predicted 0.3 percent.
Source: Stern, NEA Research Report #53
Age, Cohort, and Omnivorous Tastes in Art
Between 1982 and 2008, the percentage of
U.S. adults attending a benchmark arts event
declined from 39 percent to 34.6 percent. Of
course, any number of factors may have
played a role in that decline, including the
“cultural omnivore.” In his report, Stern
shows that young adults and those belonging
to the World War II and early Baby Boom
generations were more likely to be cultural
omnivores, compared with late Boomers and
members of Generation X12.
U.S. economic recession that was under way
As these generations aged, cultural
for six months when the 2008 SPPA was
omnivores declined as a share of the U.S.
conducted. But another likely contributor is
adult population. In 1982, for example,
that there are now fewer adults who are
when the early Baby Boomers were
characterized as “cultural omnivores,” those
considerably younger, cultural omnivores
who attend a variety of benchmark arts events,
made up 15 percent of all SPPA benchmark
and who attend the arts frequently.
respondents. By 2008, omnivorous arts
While age and cohort are weak predictors of
overall arts participation, they have a
somewhat stronger influence on shaping a
participants were 10 percent of the total. Over
the same period, adults who had attended zero
benchmark activities in the previous year rose
from 61 percent to 67 percent of the total.
Three NEA Monographs on Arts Participation: A Research Digest
15
Not only are there now fewer cultural
If we are correct that the cultural
omnivores, but the number of events that
omnivore is in decline, it may be
omnivores attend appears to be shrinking.
because the omnivore represented a
Between 2002 and 2008, the number of arts
transitional stage in our cultural
events attended per omnivore fell by more
development… Cultural participants
than one event per year. Stern estimates that
[are] no longer willing to let their
82 percent of the decline in the total number
social status define what cultural tastes
of benchmark activities attended between
were acceptable for them. Although
2002 and 2008 stems from this combination—
the omnivore — as measured by the
fewer cultural omnivores attending arts events
SPPA — may be foundering, this
less frequently.
quest for a more personal, flexible, and
Yet Stern is finally optimistic about what this
trend may bode for the future of arts
participation.
Three NEA Monographs on Arts Participation: A Research Digest
protean approach to cultural
engagement appears very much
alive.13
16
Race, Ethnicity, and Arts Participation14
In 2008, the benchmark arts attendance rate for white adults was roughly twice that of African
Americans and Hispanics.15
Percentage of U.S. Adults Attending at Least One Benchmark Arts
Activity, by Race and Ethnicity, 2008
39.7%
31.9%
21.5%
21.0%
Hispanic
White*
African American*
Other*
* Non-Hispanic
Source: Welch and Kim
Commissioned NEA research shows, however, that despite these visible discrepancies, race and
ethnicity are poor predictors of arts attendance. Once other characteristics (principally education) are correctly accounted for, race and ethnicity play virtually no role in predicting arts participation.
Using data from the 2008 SPPA, for example, a regression model predicting jazz attendance
shows initially that African Americans were 58 percent more likely than whites to go to a jazz
concert. Adding education to the model, however, renders race statistically insignificant.16 In
other words, it is not a person’s race, but rather his or her educational attainment that largely
predicts jazz concert attendance.17
Three NEA Monographs on Arts Participation: A Research Digest
17
Yet without controlling for education and other variables, racial/ethnic group disparities do
emerge within arts audiences. To illustrate, African Americans in 1992 were 11 percent of the
adult population, but 17 percent of the total jazz audience. In other words, the share of African
Americans attending jazz concerts exceeded this group’s share of the U.S. population by nearly
6 percentage points. By 2008, however, whites exceeded their share of the jazz audience (by
almost 9 percentage points). African Americans’ share of the total audience of jazz, as a result,
was much closer to this group’s share of the U.S. population.
It should be noted that over this timeframe, white adults’ rate of attendance at jazz concerts did
not rise to that of African Americans. Rather, the share of African Americans attending jazz
concerts fell from 16 percent in 1992 to the same rate reported for whites in 2008 (just under 9
percent).
Racial/Ethnic Composition of Jazz Concert-Goers, 1992-2008
1992
2008
Race/ethnicity
Percentage of
U.S. population
Hispanic
White*
African American*
Other*
8.4%
77.3%
11.2%
3.1%
Percentage of Difference Percentage of
jazz audience
U.S. population
4.8%
76.4%
17.1%
1.6%
-3.6
-0.9
5.9
-1.5
13.5%
68.7%
11.4%
6.4%
Percentage of Difference
jazz audience
6.8%
77.5%
12.5%
3.2%
-6.7
8.8
1.1
-3.2
* Non-Hispanic
Source: Welch and Kim
For a number of other art forms, white audience members continue to exceed their share of the
general population. In 1992, for example, whites were 84 percent of the total audience for musical plays, 7 points above their share of the adult population. By 2008, whites exceeded their
share of the audience by almost 14 points. This gap increased not because whites’ share of the
musical play audience rose, but because whites’ share of the U.S. adult population fell.
Three NEA Monographs on Arts Participation: A Research Digest
18
The Relationship between Arts Attendance and Personal Art-Making
Although most of the analysis discussed ear-
percent of U.S. adults attended arts events,
lier has focused on attendance at arts events,
created art, or experienced art via electronic
Novak-Leonard and Brown conceptualized a
media.18 By comparison, 34.6 percent of
much broader definition of arts participation.
adults (less than half the rate) attended the
For example, they note that in 2007-2008, 74
“benchmark” arts.
The Novak-Leonard and Brown analysis uses the following definitions of attendance and creation, based on questions from the 2008 SPPA:
Attendance
Music (jazz, classical music, opera, Latin or Spanish or salsa music, and outdoor performing
arts festivals); Theater (musical or non-musical plays); Dance (ballet or other dance); Visual
arts (art museums or craft fairs); Site visits for historic or design value.
Creation
Music (musical instrument-playing, performing opera, and singing with a choir or vocal group);
Theater (performing musical or non-musical plays); Dance (performing dance); Visual arts
(engaging in one or more of the following types of arts creation: pottery, ceramics, jewelry,
leatherwork, weaving, needlework, sewing, photography, films, videos, painting, drawing, or
sculpting); Creative writing; Arts curation (owning an original work of art).
Media
Internet-based arts activities (music, theater, dance, visual arts); Broadcasts and/or recordings
(jazz, classical music, opera, Latin or Spanish or salsa music, musical or non-musical plays,
dance, programs about artists and art works, and programs about books or writers).
Three NEA Monographs on Arts Participation: A Research Digest
19
The Novak-Leonard Brown report also inves-
activities do both. In 2008, 33 percent of
tigated the relationships between various
adults attended arts events and personally
“modes” of arts participation. Among their
performed or created art. Only 17 percent
most significant findings is the correlation
participated by attending only; 12 percent
between arts attendance and creation. The
participated only by creating or
report shows that most who engage in these
performing.
Distribution of U.S. Adults by Arts Participation Patterns, 2008
Attend only
17%
Create and attend
38%
Create only
Neither
33%
12%
Source: Novak-Leonard and Brown, NEA Research Report #54
Attendance rates among adults who create art
of the arts activities featured in the 2008
are two to five times higher than for those
SPPA was 2.3 times higher among adults who
who do not create art. For example, compared
created art. For dance attendance, the ratio
with adults who do no personal arts perform-
was almost five times higher.
ance or creation activities, attendance at any
Three NEA Monographs on Arts Participation: A Research Digest
20
Percentage of U.S. Adults that Attended Arts Events, by Whether They Created
or Performed Art, 2008
80%
70%
60%
50%
40%
30%
20%
10%
0%
Any
attendance
Music
Theater
Created or perf ormed art
Dance
Visual arts
Did not create or perf orm art
Source: Novak-Leonard and Brown, NEA Research Report #54
The strong relationship between attendance
Within art forms, the odds of attending are
and creation can also be demonstrated through
particularly high for adults who perform
19
the calculation of odds.
As the table below
dance and theater. The odds ratio for
shows, the chances that Americans who
performing and attending dance was 7.2; it
engage in creative activities will attend arts
was 5.7 for performing plays and theater-
events were almost six times better than for
going.
those who did not create art.
Odds of U.S. Adults' Participation via Attendance and Arts Creation, 2008
Odds ratios
Attend any
Creation:
Create in any form
Music
Theater
Dance
Visual arts
5.9
3.7
3.6
5.9
4.5
Attendance:
Music
Theater
4.4
2.7
2.9
5.2
3.2
Dance
Visual arts
5.3
2.6
4.9
7.2
3.6
5.7
2.8
2.6
4.0
2.8
4.1
2.4
5.7
2.6
2.7
Source: Novak-Leonard and Brown, NEA Research Report #54
Three NEA Monographs on Arts Participation: A Research Digest
21
•
These results suggest that successful
Providing artistic content and
audience-building strategies may consist of
instruction online and through
programs that combine art-making and
other media.
personal performance with live attendance
•
opportunities. Novak-Leonard and Brown
audiences to “enhance” arts
elaborate on this potential confluence in a
experiences by providing forms for
series of recommended “strategies for
conversation and context-building
engaging audiences and visitors.” These
activities.
strategies include:
•
Involving community artists in the
creation of artistic work within
professional arts organizations and
venues.
•
•
Providing opportunities for
Similarly, Novak-Leonard and Brown propose
a series of “strategies for engaging people in
the creation of artistic works.” Taken together
with their recommendations for researchers
and cultural policy-makers, the authors’ report
Allowing more interpretation and
“offers a unique context for understanding arts
interaction during exhibits and
participation [and] suggests that a more
performances.
expansive framework for the cultural ecology
Creating new program formats
is needed.”20
(e.g., the “mini-concert”).
Produced by Bonnie Nichols
Director, Sunil Iyengar
Senior Program Analyst, Sarah Sullivan
Office of Research & Analysis
National Endowment for the Arts
1100 Pennsylvania Avenue, NW
Washington, D.C. 20506
Three NEA Monographs on Arts Participation: A Research Digest
22
Endnotes
1
NEA Research Report #36, Effects of Arts Education on Participation in the Arts (1996). A
summary of this report is available at http://www.nea.gov/research/Researcharts/
Summary36.html
2
Other covariates include citizenship, marital status, and having children under age 18.
3
This model used the same covariates as the attendance model discussed earlier.
4
As a variable for analysis, U.S. citizenship status is one the few proxies available in the survey
for understanding respondents’ potential ties to other cultural heritages or traditions.
5
This analysis is restricted to young adults to improve the likely accuracy of recall of childhood
arts classes.
6
NEA Research Report #52, Executive Summary, p. 14.
7
NEA Research Report #52, Chapter Two, p. 42.
8
Over the years spanning the SPPA surveys, the U.S. Census Bureau has changed the way racial and ethnic categories were defined. In 1982, for example, Hispanic ethnicity was derived
by selecting White House Office of Management & Budget-defined categories from a list of
ethnicities (e.g., Mexican, Chicano, etc.). Beginning with the 1992 SPPA, however, Hispanic
ethnicity was a single, yes/no variable. In 1982, race categories were restricted to “white,”
“black,” and “other.” By 1992, the race categories were expanded to include American Indian,
Asian, and Hawaiian/Pacific Islander. In 2002, respondents could choose multiple race categories (e.g., White-Asian). The 1982 estimates reported in NEA Research Report #52 (Rabkin)
approximate the 2008 SPPA definitions of race and ethnicity. To make the estimates more
comparable, this Note reports trends in race and ethnicity between 1992 and 2008.
9
Please see Table 5 of Research Report #54 (Novak).
10
The median age for each benchmark arts attendee, 1982-2008, is reported in Arts Participation 2008: Highlights from a National Survey, available at http://www.nea.gov/research/NEASPPA-brochure.pdf.
11
The index of representativeness is calculated by dividing each audience share figure by that
age group’s share of the entire population.
12
Generations are defined by the following: World War II (born 1936-1945); Early Baby
Boomers (born 1946-1955); Late Baby Boomers (born 1956-1965); and Generation X (born
1966-1975).
Three NEA Monographs on Arts Participation: A Research Digest
23
13
NEA Research Report #53, Chapter Five, p. 66.
14
Race/Ethnicity and Arts Participation: Findings from the Survey of Public Participation in
the Arts, authored by Vincent Welch, Jr. and Yonghyun Kim, NORC at the University of Chicago, will be made available through the Education Resources Information Center (ERIC) via
the NEA website in 2011.
15
The white, African American, and “other” race categories exclude Hispanics.
16
Race was rendered statistically insignificant by adding an interactive term between educational attainment and race.
17
Although education was found to be the best predictor of arts participation, it cannot explain
all the variance in participation. For example, people of “other” races (the majority of whom
are Asian) are better educated than whites. Yet attendance rates for whites are higher than those
for people of other races.
18
This figure is inclusive of attendance rates reported for parks, monuments, buildings, and
neighborhoods visited for historic or design value. These events traditionally have not been reported by the NEA as “benchmark” arts activities—though they have been tracked for as long
as most arts-attendance activities.
19
An odds ratio of 1 would indicate that those who create and those who do not are equally
likely to attend the arts. For example, the odds ratio of dance to visual arts is 4, meaning that
the odds of visiting an art museum or craft fair are 4 times better for adults who perform dance,
compared with adults who do not personally dance.
20
NEA Research Report #54, Executive Summary, p. 15.
Three NEA Monographs on Arts Participation: A Research Digest
24
