Determinants, Effects, and Funding of Arts Consumption
60 Pages Posted: 24 Aug 2015
Date Written: August 12, 2015
We conducted a three-phase study of arts participation using SPPA (Survey of Public Participation in the Arts) data going back to 1982. Using statistical methods to correct for possible survey sampling bias (including clustering, stratification, and weighting), in phase 1 we confirm previous findings that arts participation as traditionally measured is declining, almost universally. We examined both individual art forms as well as aggregate measures. In phase 2 of the paper, we exploit the unique situation of free art museums in Washington, D.C., to quasi-experimental tests and found only mixed evidence to support the expected result that higher income leads to more arts participation. The more surprising and interesting finding, with stronger evidence to support it, is that arts participation actually leads to higher levels of family income.
In supplementary tests (phase 3), we found that treatment effects analysis using nearest-neighbor matching (on gender, natural log of income, education, ethnicity, urban location, and survey year) showed consistent and generally statistically significant evidence of an effect of the NEA budget cuts on arts education during the 1995/6 budget years. For every aggregated variable (combining 8, 7 or 6 of the individual categories) we saw findings consistent with the fact that students who were exposed to the reduction in arts funding (that is, survey respondents born after 1982, the “treated” group) had lower levels of arts involvement relative to their otherwise similar control matches. Finally, we examined whether arts participation is related in any way to state-level changes in opinions over time, as measured by presidential election results. We found that changes in arts participation over the 2002-2012 timeframe (the years for which SPPA data include survey respondents’ location) were robust to consideration of presidential election voting margins, although respondents from more Democratic-leaning states were more likely to attend art museums and live jazz, and to a lesser extent, plays and craft fairs.
Keywords: arts policy, arts participation, quasi-experimental methods, propensity score matching
JEL Classification: Z11, D19, H31, H44, C83
Suggested Citation: Suggested Citation