Music education promotes lifelong engagement with the arts

Abstract

The purpose of this study was to understand the effects of school-based music education on later adult engagement with the arts using nationally representative data from the U.S. National Endowment for the Arts 2012 Survey of Public Participation in the Arts. The probability of adult arts engagement as performer/creator and patron/consumer was estimated as functions of prior school-based music and arts education participation with statistical covariate control applied for demographic variables known to vary with music education status. Results suggest that both music performance and music appreciation courses are strongly associated with later arts participation as patron/consumer and performer/creator, even after controlling for socioeconomic status, sex, and race/ethnicity. Former music appreciation students were 93% more likely to attend classical music or opera performances as adults and 255% more likely to play a musical instrument as adults than were non-participants. Former music performance students were 342% more likely to play a musical instrument, 258% more likely to sing, and 186% more likely to take photographs as an artistic endeavor than were non-participants. Results of this study suggest that lifelong engagement with music and the arts is one measurable outcome of school-based music education in the United States.

Keywords

adults audiences engagement lifespan live performance performance school education

Many music educators assert that fostering a lifelong engagement with music – a connection to musical understanding that extends far beyond the years of schooling – is one of their central intended outcomes for their students (Zakaras & Lowell, 2008). While researchers have sought to describe the populations of music students and explore reasons for elective uptake of music education – for example, Elpus and Abril (2011) in the United States and Bray (2000) in the United Kingdom – to date, relatively little research has directly examined the question of whether, and how, students of school-based music education continue their involvement with music and the arts beyond schooling and into their adult lives.

Music education philosophers contend that the value of music education to society is directly and inextricably linked to the value that society and individuals within society place on music itself (Elliott & Silverman, 2014; Reimer, 2003). Reimer writes that one must understand “the comprehensive and satisfying incorporation of music into people’s lives” (Reimer, 2003, p. 12) to understand why aesthetic education in music has value beyond the school walls. Elliott and Silverman fundamentally agree that music education provides value outside of the schoolyard gate: “music, music education, and [community music] are valuable sources of human insight … but if we make music only in school … then music making and listening fall short of their enormous potential to improving communities and societies” (Elliott & Silverman, 2014, p. 52).

Key to this rare point of agreement is a linkage – the connection between the making, doing, and consumption of music and art in the world with the making, doing, and consumption of music and art in educational contexts. But what is the directional nature of this linkage between the real-world experience of music and the arts and school- and community-based music education? Does society value music and thereby make provision for music education in schools and communities? Are students educated in music and the arts more likely to become artistically engaged members of society later? Should the active making of music, as Elliott might suggest, be the foremost aim of music education, or, as Reimer might have suggested, is learning “within” and “through” music, in less performative ways, a better way to encourage lifelong musical engagement? While each of these questions may seem suitable for philosophical inquiry, empirical research can examine the linkage between various types of in-school music education and later participation in the arts as creator or consumer. Fortunately, at least in the United States context, a relevant rich source of national data exists that can be used to research the arts engagement of adults who had studied the arts in their youth – the Survey of Public Participation in the Arts (SPPA), a recurring project of the U.S. National Endowment for the Arts (NEA).

Existing research using data from the National Endowment for the Arts Surveys of Public Participation in the Arts (SPPA) has suggested two important conclusions for the value and status of arts education in the United States: first, that arts education experience is the strongest predictor of future engagement with the arts and second, that the proportion of adult Americans who have experienced arts education has been declining over the past 30 years (Bergonzi & Smith, 1996; Rabkin & Hedberg, 2011). Taken together, these conclusions are troubling, both for arts educators and the art making community at large. However, in recent empirical work using data from the National Center for Education Statistics, Elpus (2014) found that the proportion of students pursuing music coursework in American high schools had not declined, but instead remained flat, during the period between 1982 and 2009. The discrepancy implied by the Elpus (2014) results as compared to the earlier analyses of SPPA data (e.g., Rabkin & Hedberg, 2011) may be due to differences in data collection methods: Elpus (2014) used administrative high school transcript data for nationally representative cohorts of students in American schools, while the SPPA data analyzed by Rabkin and Hedberg (2011) asked adults to recall and self-report their childhood arts education experiences.

Prior research does suggest that arts education, writ large, is somewhat linked to later engagement in certain arts activities. Bergonzi and Smith (1996) sought to determine the effect of childhood arts education on later engagement with classical music, jazz, opera, musical plays, non-musical plays/dramas, ballet and other dance, poetry, novels or short stories, and the visual arts. These particular arts modalities were chosen because they are generally referred to in publications of the National Endowment for the Arts (NEA) as the “benchmark” arts activities – those arts modalities for which the NEA has consistently gathered data on Americans’ participation as consumer, creator, or performer since at least the 1982 wave of the SPPA. Using SPPA data, Bergonzi and Smith (1996) found that childhood arts education was a strong predictor of later engagement in these benchmark arts activities, and that prior engagement with arts education moderated more than half of the differences in arts attendance predicted by socioeconomic status. That is, adjusting for arts education reduced the often-observed association between socioeconomic status and benchmark arts attendance. Although arts attendance and art creation in various forms were strongly predicted by childhood arts education, Bergonzi and Smith found that performing (as a musician, actor, or other performing artist, whether for recreation or occupation) was not predicted by arts education. As may be reasonably assumed, Bergonzi and Smith found that socioeconomic status was strongly, positively correlated with the amount of arts education one received in childhood, and this relationship persisted both for school-based and community-based arts education, though the relationship was stronger for community-based arts education.

Similar to Bergonzi and Smith (1996) before them, Rabkin and Hedberg (2011) found that childhood or adult arts education was associated with attending at least one benchmark arts event, with the effect of arts education on the likelihood of benchmark arts attendance increasing with the number of distinct art forms studied. In their analyses, Rabkin and Hedberg did not differentiate among the various disciplines within the visual and performing arts, leaving open the question of whether there are differential effects of arts education on later arts consumption or creation based on the kind of arts education pursued, nor did they differentiate among the various kinds of benchmark arts activities, leaving open the question of whether there are differential effects of arts education on later arts participation based on the various arts modalities making up the benchmark arts activities.

Whereas previous research using SPPA data has found a link between arts education and later artistic involvement with benchmark arts events, Mantie and Tucker (2008) have suggested that the link between school-based music education and later music making is more tenuous. Arguing that “traditional performing ensembles in secondary schools were not found to encourage lifelong involvement in music” (p. 217), Mantie and Tucker explored how avocational adult musicians who were members of community-based ensembles described their music participation. They concluded that connections between school music – particularly, traditional wind band ensembles – and life were not readily apparent and note that they “can only speculate on how many instruments are currently sitting in closets collecting dust” (p. 225).

Outside of music and arts education researchers, economists have addressed questions regarding demand for the arts as an economic phenomenon. Much of this work is summarized by Seaman (2006), who concludes that the research literature offers few concrete answers despite a four-decade history and the recent incorporation of more sophisticated analyses. Seaman contends that part of the opaqueness of estimating demand for the performing arts stems from the relationship of arts participation and consumption with covariates that are not typically examined in other economic analyses: arts education, as separated from other formal education, gender, socialization into arts appreciation, and sexual orientation among them. Seaman notes other limitations in the literature – mainly, that most of the studies use as data sources either surveys of audiences as members (thus suffering from self-selection biases) or surveys of the broader population, which may not adequately capture the effect of pricing on participation and attendance. Seaman suggests that the collinearity between educational level and income is another key limitation facing econometric analyses of arts demand – it is difficult for economists to disentangle the individual contributions of education and income as predictors of arts participation because they are so closely related. The challenge notwithstanding, there is a consensus that educational level, rather than income, is the most predictive characteristic determining arts participation, though, as Seaman reports, some extant studies have in fact come to the opposite conclusion.

One study in economics (Borgonovi, 2004) used national data from the 2002 wave of the SPPA to estimate whether and how the likelihood that members of the population would attend various performing arts events varied as a function of certain characteristics. Borgonovi’s chief innovation was the linking of data from the National Center for Charitable Statistics external to the SPPA data on arts attendance. Borgonovi theorized that greater support from donations would tend to reduce per person admission prices to these events, and her estimates therefore hinge on how much credence is given to this theory. Borgonovi found that previous arts education is the primary personal characteristic predictive of performing arts attendance, even when controlling for her proxy measures for cost of attendance. Interestingly, Borgonovi reports that controlling for cost of attendance had the net effect of reducing the influence of socioeconomic status, race/ethnicity, sex, and occupation on performing arts attendance. She notes, however, that music/arts education is not equally distributed among the population and that since arts education is art-specific, but complementary in that participation in one kind of arts education generally increases consumption of other art forms, a wide range of arts education experiences should be offered in schools.

Corroborating Borgonovi’s assertion about the unequal distribution of music and arts education in the U.S., recent research in American music education has suggested that there are many socioeconomic and demographic differences between the population of students who elect nonmandated music education and those who do not (Elpus, 2013, 2014, 2015; Elpus & Abril, 2011; Fitzpatrick, 2006; Kinney, 2008). Synthesized, this line of research suggests that music students in the United States are more likely to be female, more likely to be White (and, commensurately, less likely to report Hispanic or Latino origin), more likely to come from the highest socioeconomic status quartile, more likely to speak English as a native language, more likely to have parents with advanced educational attainment, and more likely to demonstrate higher academic achievement prior to the age at which their music study commences than are students who do not pursue elective music study in schools. As many of these same sociodemographic differences tend to manifest in national profiles of adult engagement in the creation and consumption of art (National Endowment for the Arts, 2013), there is a need for updated research examining the link between music education and later arts engagement. The present study was designed to fill this need.

Purpose of the study and research questions

The purpose of this study was to examine the association between school-based music education in childhood and adult engagement with arts as creator or consumer. The following research questions guided the investigation:

What is the effect of past school-based music education on the likelihood of adult attendance at (a) live classical music or opera, (b) live jazz, (c) live theatre, (d) live ballet, or (e) live dance performances other than ballet?

What is the effect of past school-based music education on the likelihood of adult creation or performance of (a) music (singing or playing an instrument), (b) performing in musical or non-musical theatrical productions, (c) taking photographs as an artistic outlet, or (d) creating other non-photographic visual art?

What are the differential effects, if any, on these outcomes among performance-based music education, non-performance based music education, and other types of arts education?

Method

Data and participants

For the present study, I analyzed data from the 2012 Survey of Public Participation in the Arts (SPPA). Since 1982, the U.S. National Endowment for the Arts, a federal government agency charged primarily with advancing culture and the arts in America, has sponsored a periodic nationwide data collection effort to ascertain the extent and nature of artistic participation, patronage, and creation by the American people. That effort, the SPPA, is administered by the U.S. Census Bureau as a supplement to the monthly Current Population Survey (CPS). In 2012, questions comprising the SPPA were asked of a sample of 35,735 U.S. adults, aged 18 and over. When weighted using the sampling weights provided by the Census Bureau, respondents to the SPPA are representative of the entire U.S. non-institutionalized adult population. Each respondent answered one of two “Core” questionnaires, the first of which was designed to permit trend analysis with previous waves of the SPPA and the second of which was designed to ask updated questions reflecting advances in the state of the arts. In addition to answering one of the two Core questionnaires, respondents were randomly assigned to answer two of the five possible questionnaire modules, each of which was focused on a specific area of interest. This resulted in each respondent answering the questions on either Core 1 or Core 2 plus two additional modules. The modular design of the SPPA permits greater data collection from larger numbers of respondents – ensuring data are nationally representative – while simultaneously reducing the overall response burden for each individual CPS participant.

Outcome variables

For the present study, I consider two forms of arts “engagement” – namely, art creation/performance or arts event patronage – both of which are observable in the SPPA data. For active art making, I use respondents’ self-reported answers to questions asking them if, at any point in the 12 months prior to the survey, they had played a musical instrument (alone, or with other people), had performed or practiced any singing (alone or with other people), had performed in a theatrical production (musical or non-musical), had taken photographs specifically as an artistic outlet (as opposed to taking “snapshots” or photos more for mementos than as art), or had created any other non-photographic visual art. For arts patronage, I use respondents’ self-reported answers to having attended, within the 12 months prior to the survey, live performances of classical music or opera, live jazz, live theatre (musical or non-musical), live performances of ballet, or live performances of dance genres other than ballet.

These specific arts attendance outcome measures for the present analysis were chosen based primarily on their consistent availability in the various waves of SPPA data collection. In the research literature using SPPA data, these have come to be known as the “benchmark” arts events. It is clear that attendance at the benchmark arts events of classical music/opera, live jazz, live theatre, ballet, and other non-ballet live dance most definitely do not make up the entirety of the arts patronage experiences of American adults. However, given that the various waves of the SPPA have measured these experiences for over 30 years, much of the prior research using SPPA data has focused on attendance on at least one of these benchmark arts events. For the present study, I take a somewhat novel approach and consider each benchmark arts event as an outcome distinct from the others rather than using attendance of “at least one” benchmark arts event as the outcome measure. Additionally, by considering arts creation or performance outcomes in addition to attendance at benchmark events, the present study treats arts making experience as a form of adult arts engagement distinct from arts attendance. The present study, however, must be considered limited in examining only these few outcomes – at least to the extent that American adults do engage with art making and performance in modalities other than those investigated here.

Music and arts education participation variables

I drew a set of dichotomous school music and arts education participation variables, coded one for participants and zero for nonparticipants, from Module E of the SPPA, which was administered to 9,482 sample members. The data are drawn from questions asked of the adult respondents about their own childhood experiences with music and arts education. Similarly coded dichotomous adult arts patronage and creation variables were drawn from either Core 1, to which 18,011 sample members responded, or from Core 2, to which 17,635 sample members responded. To increase the sample sizes for the analyses reported here, if a similar outcome question was asked of Core 1 and Core 2 respondents, respondents who answered either version of the question were included in the analysis. Although there were slight differences between the Core versions in the wording of the questions, the benefit of increased sample size was considered an acceptable trade off in that it outweighed the potential confounding nature of slightly different question wording. Although the 2012 wave of the SPPA asked adults about different arts modalities than the “benchmark” arts activities that have been consistently included since the 1982 SPPA, only the benchmark arts activities were included in both Core 1 and Core 2 questionnaires. This means that to gain the benefit of increased sample size in the present study (i.e., using all 9,482 Module E respondents, regardless of which Core questionnaire those respondents received), the outcomes available for analysis here are necessarily limited to the benchmark arts activities.

Covariates

There is an active line of research within music education that has established the existence of key demographic differences between the populations of people who do and do not elect to participate in school-based music education (e.g., DiMaggio & Mukhtar, 2004; Elpus, 2013, 2014; Elpus & Abril, 2011; Fitzpatrick, 2006; Kinney, 2008, 2010). I drew a set of covariates from the SPPA data to adjust for these documented differences in arts education uptake. The covariates included were: respondent race/ethnicity, self-identified gender, and parental educational attainment. Additionally, since the outcomes examined may tend to differ by the respondent’s own socioeconomic status, level of educational attainment, and the region of the country in which they resided, the covariates employed also included measures for the respondent’s level of educational attainment, whether the respondent’s household income was above or below $50,000 per year, and whether the respondent resided in the Northeast, Southeast, Midwest, or Western regions of the United States.

Empirical approach

The present study was conceived as a quasi-experimental study, using observable covariates as described above to adjust for selection bias (Murnane & Willett, 2011) in an effort to estimate the association of music and arts education and later arts engagement distinctly from the effects of the observable covariates. I identified 2012 SPPA sample members with music and arts education experience based on their self-reported responses to the SPPA Module E questions about their past school-based music and arts education experiences.

Estimates of the effect of arts education on later arts participation were generated using direct entry of covariates into regression models, using disaggregated (i.e., “music student,” “visual art student,”) indicators of arts education experiences.

Results were generated through estimation of the following theoretical model:

\begin{array}{l} O u t c o m e_{i} = α + β_{1} M u s i c S t u d y_{i} + β_{2} V i s u a l A r t S t u d y_{i} + β_{3} D a n c e S t u d y_{i} + β_{4} T h e a t e r S t u d y_{i} \\ + β_{5} F i l m S t u d y_{i} + γ C o v a r i a t e s_{i} + ε_{i} \end{array}

(1)

In Equation (1), Outcome_i represents the arts participation outcome of interest for the ith respondent, each type of arts discipline represents a dichotomous variable indicating the arts education status of the ith respondent (0 = studied the arts discipline in school, 1 = did not study the arts discipline in school), and Covariates are the vector of covariates described above. Each β represents the unique contribution of each form of arts education, net of the other arts education types (for those students who pursued more than one form of arts education) and the covariates described above.

Since the SPPA uses the complex sampling strategy devised by the U.S. Census Bureau for its Current Population Survey (to which the 2012 SPPA was a supplement), all estimates were computed using the appropriate weights for the SPPA modules involved in the analyses. For the present study, data documentation from the Census Bureau and the NEA indicate that the variable labeled “PWNWGT” in the SPPA dataset is the appropriate weighting variable to generate nationally representative estimates from respondents who answered questions from the arts education module (Module E). Weighting the sample ensured that all reported estimates are representative of the population of non-institutionalized adults in the U.S., the target population for the SPPA and the CPS – even as the sample size for individual analyses shifts due to the modular design of the SPPA. Due to the complex sampling structure of the Current Population Survey, variance and standard error estimation are not as straightforward as they would be using data arising from a simple random sample. To address the non-random nature of the sample in SPPA, the Census Bureau provides a set of 160 replicate weights of the PWNWGT variable which are suitable for “balanced repeated replication” (BRR) estimates of variance and standard errors (Heeringa, West, & Berglund, 2010). The analyses reported here made use of the replicate weights for BRR variance estimation.

Results

Descriptive statistics

Table 1 presents unweighted descriptive statistics for music and arts education variables, covariates, and outcome variables for the entire SPPA sample. The percentages reported in Table 1 represent the proportion of SPPA sample members whose value for each variable was 1, indicating that they possessed the characteristic represented by the variable. As the proportions presented in Table 1 are unweighted, they are representative only of the sample members themselves and are not nationally representative estimates. The various sample sizes reported in Table 1 demonstrate the effect of the modular design of the SPPA in that not all respondents were asked all questions.

Table 1.

Sample sizes and proportion of respondents reporting each characteristic, 2012 SPPA Data (unweighted).

Variable	(1)	(2)
	N	% in category
Predictor variables
School-based music education	9,411	28%
School-based photography/film	9,406	5%
School-based visual art	9,393	13%
School-based theatre	9,395	6%
School-based dance	9,385	7%
School-based art appreciation	9,330	9%
School-based music appreciation	9,320	9%
Any school-based arts education	9,444	38%
Covariates
White	35,735	84%
Black or African American	35,735	9%
Native American Alaska Native	35,735	1%
Asian American	35,735	4%
Hawaiian/Pacific Islander	35,735	†
Multiracial	35,735	1%
Other Race	35,735	†
Hispanic/Latino origin (may be any race)	35,735	11%
Female	35,735	53%
Male	35,735	47%
Parent did not graduate high school	8,456	22%
Parent holds only H.S. diploma	8,456	38%
Parent attended some college	8,456	14%
Parent holds bachelor’s degree	8,456	17%
Parent holds postgraduate degree	8,456	9%
Respondent did not graduate high school	35,735	11%
Respondent holds only H.S. diploma	35,735	30%
Respondent attended some college	35,735	19%
Respondent hold’s associate’s degree	35,735	10%
Respondent holds bachelor’s degree	35,735	20%
Respondent holds postgraduate degree	35,735	11%
Household income $50,000 or more	35,735	52%
Outcomes
Attended live classical music/opera	23,634	13%
Attended live jazz	23,693	11%
Attended live theatre	35,425	20%
Attended live ballet	35,361	3%
Attended live dance (other than ballet)	35,334	6%
Donated money to arts organization	4,917	11%
Plays an instrument (as an adult)	13,776	13%
Sings (as an adult)	13,755	10%
Dances (as an adult)	13,769	5%
Performs in live theatre (as an adult)	13,787	2%
Creates Photography (as an adult)	9,510	13%
Creates Other Visual Art (as an adult)	9,503	6%

Note. Unweighted percentages of sample members reported here are not nationally representative.

† = rounds to zero.

Using the sampling weights, it is possible to generate estimates of certain variables of interest that are nationally representative. Weighted estimates indicate that 36.95% of adults in the United States reported having participated in some school-based arts education. By arts discipline, estimates from the SPPA suggest that 27% of adults were in school music performance or appreciation, 12.72% of adults were in school visual art creation or appreciation, 4.39% of adults were in school photography or film, 5.5% of adults were in school theatre, and 6.5% of adults were in school dance. It is important to note that these estimates are based on self-reported recall and not on school transcript or administrative data, which has been used commonly in the prior literature to estimate national music education participation and enrollment rates. These “recall” estimates are therefore not directly comparable to research reporting participation and enrollment rates using school transcript data.

Adult creation or performance of music and art

When asked about art making within the past 12 months, across the sample of SPPA respondents (unweighted), 13% reported that they had played a musical instrument, 10% reported that they had sung, 5% reported that they performed dance, 2% reported they that had performed in musical or non-musical theatre, 13% reported that they had taken photographs as an artistic endeavor, and 6% reported that they had created non-photographic works of visual art. Table 2 presents the results of a series of logistic regressions for the effect of prior school-based music and arts education experiences on later adult arts creation or performance, controlling for the covariates discussed above.

Table 2.

Logistic regression for effects of arts education disciplines on active arts creation outcomes.

	(1)	(2)	(3)	(4)	(5)
	Play an instrument	Sing	Perform in theatre	Take artistic photographs	Create other visual art
Arts education
School music (performance)	3.42***	2.58***	2.36	1.86**	1.17
	(0.69)	(0.61)	(1.15)	(0.35)	(0.39)
School visual art (creation)	1.25	1.28	2.63	1.79*	4.25***
	(0.33)	(0.41)	(1.63)	(0.42)	(1.36)
School photography/cinema	1.37	0.79	0.76	3.23***	4.64***
	(0.53)	(0.34)	(0.49)	(1.06)	(1.73)
School theatre	0.87	1.82*	3.66**	1.21	1.36
	(0.32)	(0.47)	(1.57)	(0.34)	(0.58)
School dance	1.65	1.23	1.48	1.24	1.33
	(0.55)	(0.43)	(0.93)	(0.39)	(0.52)
School visual art appreciation	0.67	1.20	1.62	1.90*	1.59
	(0.21)	(0.33)	(0.91)	(0.46)	(0.59)
School music appreciation	2.55***	2.02**	1.09	1.29	1.08
	(0.59)	(0.50)	(0.52)	(0.33)	(0.42)
Male (vs. Female reference category)	1.80**	0.89	1.22	0.84	0.59*
	(0.34)	(0.18)	(0.59)	(0.15)	(0.15)
Race (vs. White reference category)
African American	0.33*	1.38	2.84	0.76	0.36
	(0.14)	(0.45)	(1.69)	(0.25)	(0.19)
Native American/Alaska Native	1.00	0.88	–†	1.84	5.42*
	(0.78)	(0.65)		(1.60)	(4.39)
Asian	1.21	1.17	1.94	0.87	0.35
	(0.49)	(0.71)	(1.29)	(0.35)	(0.28)
Hawaiian/Pacific Islander	12.65*	0.57	–†	2.23	0.70
	(13.69)	(0.62)		(3.22)	(0.54)
Multiracial	0.91	1.08	–†	0.74	3.45*
	(0.62)	(1.00)		(0.55)	(2.09)
Hispanic ethnicity	1.20	0.89	0.54	0.59	0.98
	(0.36)	(0.35)	(0.50)	(0.17)	(0.41)
Household income $50,000+	1.04	0.67	0.83	1.29	1.10
	(0.20)	(0.15)	(0.39)	(0.23)	(0.31)
Educational attainment (vs. no H.S. diploma)
Respondent holds H.S. diploma only	2.25	1.16	9.02	1.05	0.54
	(1.00)	(0.45)	(10.13)	(0.41)	(0.27)
Respondent attended some college	3.26*	1.07	3.94	2.54*	1.00
	(1.48)	(0.50)	(5.06)	(1.04)	(0.56)
Respondent holds associate’s degree	2.84*	1.03	8.02	2.81*	0.61
	(1.47)	(0.56)	(10.07)	(1.28)	(0.41)
Respondent holds bachelor’s degree	2.60	1.29	4.23	2.85**	0.85
	(1.36)	(0.61)	(5.21)	(1.13)	(0.48)
Respondent holds postgraduate degree	2.61*	1.29	3.91	4.23***	1.80
Respondent holds postgraduate degree	(1.20)	(0.56)	(5.45)	(1.77)	(1.04)
Respondent’s parent holds H.S. diploma only	1.05	1.30	1.11	1.10	1.05
Respondent’s parent holds H.S. diploma only	(0.31)	(0.39)	(0.81)	(0.29)	(0.41)
Respondent’s parent attended some college	2.05*	2.10*	3.43	1.11	0.89
Respondent’s parent attended some college	(0.67)	(0.78)	(2.63)	(0.33)	(0.52)
Respondent’s parent holds bachelor’s degree	2.17*	0.87	0.47	0.99	1.63
Respondent’s parent holds bachelor’s degree	(0.70)	(0.38)	(0.50)	(0.31)	(0.80)
Respondent’s parent holds postgrad degree	2.68*	1.42	0.64	1.13	1.57
Respondent’s parent holds postgrad degree	(1.08)	(0.58)	(0.61)	(0.34)	(0.75)
Region of Country (vs. East)
Resides in Midwest	0.71	0.71	0.49	0.95	1.36
	(0.19)	(0.23)	(0.33)	(0.23)	(0.47)
Resides in South	0.57*	0.89	1.50	0.83	1.53
	(0.14)	(0.25)	(0.93)	(0.20)	(0.57)
Resides in West	0.74	1.37	1.66	1.09	1.88
	(0.16)	(0.39)	(1.12)	(0.27)	(0.66)
Constant	0.02***	0.05***	0.00***	0.04***	0.03***
	(0.01)	(0.02)	(0.00)	(0.02)	(0.01)

Note. Coefficients are reported as odds ratios. Standard errors in parentheses. Reported estimates and standard errors employ the SPPA survey weights and are adjusted for the stratified sampling.

†

indicates category was eliminated from this analysis due to collinearity/lack of variation in the outcome.

As seen in Table 2, controlling for other factors related to arts participation, adults who report having participated in certain forms school-based arts education were considerably more likely than adults without those forms of school-based arts education to play an instrument, to sing, to perform in the theatre, to take photographs as an artistic endeavor, or to create other visual art. It is perhaps unsurprising that there is a general congruence between the kind of art studied in school and the kind of art created as an adult. Former students of music performance and appreciation were more likely to play an instrument (3.42 times more likely for former music performance students; 2.55 times more likely for former music appreciation students) or to sing (2.58 times more likely for former music performance students; 2.02 times more likely for former music appreciation students). Former theatre students were 3.66 times more likely than those without school theatre education to perform in a theatrical production. Those adults who had pursued school-based visual art, photography or cinema, and visual art appreciation were significantly more likely than those without those educational experiences to pursue photography as an artistic endeavor (as opposed to taking “snapshots”). Former visual art creation and photography/cinema students were also significantly more likely to create other, non-photographic forms of visual art as adults.

Interesting results in the demographic and socioeconomic covariates are evident in Table 2. Males are about 80% more likely than females to report playing an instrument, a result that is statistically significant. African American adults were significantly less likely (67% less likely) than White respondents, the reference category for this regression, to report that they played an instrument. Future research into the possible mechanisms underlying this disparity might help establish a greater representation of African American adults among the population of those playing musical instruments. Hawaiian/Pacific Islander respondents were considerably more likely to report that they played a musical instrument than were White respondents. This is possibly due to the cultural integration of musical instruments in Hawaiian and certain other island cultures, though future research is needed to confirm this speculation.

Perhaps most interestingly, this analysis finds no relationship between household income (dichotomized here to above or below $50,000 per year) and artistic creation in adulthood – once the other variables in this analysis are statistically controlled. Prior research exploring predictors of arts engagement has noted a relationship between family income and attendance at the arts benchmark events, but here it seems that the effect of prior educational experience within an art nullifies the possible association between income level and artistic creation or performance. While income may be related to certain kinds of benchmark arts attendance, as described below, it would seem that lower income-earning adults in the U.S. are no less likely than their higher income-earning counterparts to engage in art making.

While this analysis found no association between income status above or below $50,000 per year and art making, the same is not true for educational attainment and various forms of artistic creation or performance. Respondents with higher levels of education were significantly more likely to play an instrument and to take artistic photographs. Adult playing of an instrument was also significantly associated with the respondent’s highest-achieving parent’s level of educational attainment. Taken together, the association between prior school-based music performance education and parental educational attainment for musical instrument playing seem to correspond with the notion explicated in past research (e.g., Elpus & Abril, 2011; Elpus, 2013; Kinney, 2008) that children of higher socioeconomic status (SES) parents are more likely to pursue instrumental music education, with parental education serving in this analysis as a blunt proxy for respondents’ family SES in childhood.

Predicted probabilities from the logistic regression models for arts creation are presented in Table 3. These predicted probabilities can help concretize the odds ratio results reported above (Long & Freese, 2014). Former music students had a .24 probability of playing an instrument as an adult, compared to a .15 chance for the total population. Similarly, adults with who had former music education experience had a .15 probability of continuing to sing as adults, compared to a .09 probability for the total population. Interestingly, former students of music performance also had a greater probability than the population average of taking photographs as an artistic activity, with a .19 probability for former music performance students compared to a .14 probability for the population.

Table 3.

Predicted probabilities for creating/performing art by educational attainment and arts education status.

	(1)	(2)	(3)	(4)	(5)
	Play an instrument	Sing	Perform in theatre	Take artistic photographs	Create other visual art
Population average	.15	.09	.02	.14	.06
Respondent’s educational attainment
Less than H.S. diploma	.07	.08	–	.07	.07
H.S. diploma only	.13	.09	.02	.08	.04
Attended some college	.17	.09	.01	.16	.07
Holds associate’s degree	.16	.08	.02	.17	.05
Holds bachelor’s degree	.15	.10	.01	.17	.06
Holds postgraduate degree	.15	.10	.01	.22	.11
Arts education by discipline
School music (performance)	.24	.15	.02	.19	.07
School visual art (creation)	.16	.11	.03	.20	.14
School photography/cinema	.18	.08	.01	.29	.17
School theatre	.13	.14	.04	.16	.08
School dance	.20	.11	.02	.17	.08
School visual art appreciation	.11	.10	.02	.21	.09
School music appreciation	.25	.15	.02	.17	.08

Note. Predicted probabilities, other than the population average, in this table derive from the logistic regression models reported in Table 2 above, and hold all other covariates constant.

Attendance/patronage of “benchmark” artistic events

Across the sample of SPPA respondents (unweighted), 13% attended live classical music or opera, 11% attended live jazz, 20% attended live musical or non-musical theatre, 3% attended live ballet, and 6% attended live dance other than ballet. Table 4 presents the results of a series of logistic regressions for the effect of school-based arts education on later adult attendance at these arts events, controlling for the same series of covariates previously discussed. Prior research has shown socioeconomic status and demographics to be related to arts enrollment and benchmark arts attendance; region of the country is included to adjust for the potential of uneven distribution of access to live arts.

Table 4.

Logistic regression results for effect of disaggregated arts education on attendance at live arts events.

	(1)	(2)	(3)	(4)	(5)
	Classical music/Opera	Jazz	Theatre	Ballet	Other dance
Arts education
School music (performance)	1.35*	1.46**	1.30**	1.48*	1.37*
	(0.16)	(0.18)	(0.12)	(0.29)	(0.17)
School visual art (creation)	0.87	1.25	1.03	1.26	1.58**
	(0.13)	(0.18)	(0.11)	(0.29)	(0.25)
School photography/cinema	0.97	1.25	0.97	0.85	1.42
	(0.19)	(0.25)	(0.14)	(0.26)	(0.29)
School theatre	1.39*	0.94	1.29	1.17	1.21
	(0.21)	(0.17)	(0.18)	(0.35)	(0.21)
School dance	1.16	1.63**	1.73***	1.10	1.60**
	(0.18)	(0.28)	(0.21)	(0.28)	(0.25)
School visual art appreciation	1.28	0.92	1.71***	1.41	1.24
	(0.19)	(0.18)	(0.20)	(0.37)	(0.22)
School music appreciation	1.93***	1.80***	1.16	1.29	0.99
	(0.26)	(0.24)	(0.14)	(0.28)	(0.16)
Male (vs. Female reference category)	0.84	1.04	0.77**	0.57**	0.67***
	(0.10)	(0.10)	(0.06)	(0.10)	(0.07)
Race (vs. White reference category)
African American	0.48**	2.44***	0.93	0.43*	0.98
	(0.13)	(0.46)	(0.15)	(0.15)	(0.21)
Native American/Alaska Native	0.27	0.65	0.59	0.52	1.97
	(0.23)	(0.39)	(0.30)	(0.41)	(0.89)
Asian	1.10	0.58	0.52***	0.89	1.00
	(0.30)	(0.18)	(0.10)	(0.36)	(0.31)
Hawaiian/Pacific Islander	0.88	0.84	1.05	0.52	2.53
	(0.87)	(0.68)	(0.55)	(0.52)	(1.58)
Multiracial	1.76	1.52	0.78	0.57	1.25
	(0.72)	(0.57)	(0.22)	(0.46)	(0.54)
Hispanic ethnicity	0.86	1.07	0.75	1.20	1.01
	(0.21)	(0.26)	(0.12)	(0.39)	(0.23)
Household income $50,000+	1.19	1.33*	1.58***	1.35	1.14
	(0.13)	(0.16)	(0.12)	(0.25)	(0.12)
Educational attainment (vs. no H.S. diploma)
Respondent holds H.S. diploma only	1.58	1.96*	2.04**	2.36	1.97
Respondent holds H.S. diploma only	(0.54)	(0.66)	(0.45)	(1.56)	(0.73)
Respondent attended some college	2.63*	2.27*	3.61***	3.03	2.72**
	(1.02)	(0.77)	(0.79)	(1.81)	(0.98)
Respondent holds associate’s degree	2.62*	3.35***	3.08***	3.50	3.77**
Respondent holds associate’s degree	(1.05)	(1.15)	(0.72)	(2.35)	(1.47)
Respondent holds bachelor’s degree	5.47***	3.43***	5.69***	6.39**	3.80***
	(2.07)	(1.20)	(1.23)	(4.16)	(1.47)
Respondent holds postgraduate degree	8.66***	5.21***	6.38***	9.98***	5.82***
	(3.52)	(1.85)	(1.51)	(6.71)	(2.33)
Respondent’s parent holds H.S. diploma only	1.06	1.13	1.09	1.07	1.00
Respondent’s parent holds H.S. diploma only	(0.21)	(0.24)	(0.13)	(0.31)	(0.21)
Respondent’s parent attended some college	1.25	1.50	1.10	1.36	1.36
	(0.29)	(0.31)	(0.15)	(0.43)	(0.29)
Respondent’s parent holds bachelor’s degree	1.32	1.51	1.32*	1.53	1.55
Respondent’s parent holds bachelor’s degree	(0.27)	(0.33)	(0.18)	(0.48)	(0.36)
Respondent’s parent holds postgrad degree	1.63	1.47	1.68***	1.58	1.68*
	(0.40)	(0.33)	(0.25)	(0.55)	(0.40)
Region of Country (vs. East)
Resides in Midwest	1.03	0.89	0.79*	0.79	1.10
	(0.16)	(0.15)	(0.08)	(0.20)	(0.21)
Resides in South	0.99	0.77	0.68***	0.81	1.20
	(0.15)	(0.15)	(0.07)	(0.17)	(0.22)
Resides in West	1.42*	0.80	1.04	0.63*	1.04
	(0.23)	(0.13)	(0.11)	(0.14)	(0.23)
Constant	0.03***	0.02***	0.05***	0.01***	0.01***
	(0.01)	(0.01)	(0.01)	(0.00)	(0.00)

Note. Coefficients are reported as odds ratios. Standard errors in parentheses. Reported estimates and standard errors employ SPPA survey weights and are adjusted for complex survey sampling.

***

p < .001. **p < .01. *p < .05.

As one might expect, school music performance classes do significantly predict adult attendance at classical music or opera, yet school music appreciation classes are considerably more predictive of adult attendance at classical music or opera. Adults who had taken music appreciation were 93% more likely than those who had not to attend live classical music or opera, while adults who had taken music performance in school were 35% more likely to attend classical music or opera than those who had not studied music performance. Along with former students of music performance (46% more likely) and former students of music appreciation (80% more likely), former students of dance were significantly more likely (63% more likely) than those without dance education to attend live jazz. Adult attendance at live theatre was predicted by previous school-based coursework in dance (73% more likely), visual art appreciation (71% more likely), and music performance (30% more likely). Adults who had taken school-based music performance classes were 48% more likely to have attended ballet than those who had not taken school music performance courses, and those who had taken music performance, visual art creation, and dance courses in school were significantly more likely as adults to report attending non-ballet live dance.

In terms of demographics, sex and race were both related in various ways to benchmark arts attendance, even after controlling for prior arts education experiences and the other covariates in the model. Males were significantly less likely than females to report attending live theatre, ballet, or dance. African American adults were significantly less likely to attend classical music or opera and ballet performances, but were 2.44 times more likely than people of other races or ethnicities to report attending live jazz events. Asian Americans were significantly less likely to report attending live theatre than adults of other races or ethnicities. There were no other significant associations between race and benchmark arts attendance.

Clear relationships were found between the two variables serving as proxies for socioeconomic status – income and educational attainment, regardless of the respondents’ prior arts education experiences in any discipline. Unlike the earlier results for arts creation/performance, income was related to two forms of benchmark arts attendance. Those with incomes at $50,000 per year or more were significantly more likely than those in households with lower incomes to attend live jazz (33% more likely) and live theatre (58% more likely) events. Even more striking than the associations with income are the associations between benchmark arts attendance and educational attainment. Respondents with earned bachelors’ and postgraduate degrees were consistently, and statistically significantly, more likely than those without high school diplomas to attend all five benchmark events, and the effect was especially pronounced for classical music/opera and ballet. However, apart from ballet, even those with only some college attendance (but no degree) or associates’ degrees only were more likely than those without high school diplomas to attend benchmark arts events.

Table 5 presents predicted probabilities of arts attendance/patronage estimate from the logistic regression results reported above. Across all adults in the United States, regardless of arts education status, the probability of attending live classical music or opera was .12. Those with prior school music performance had a .14 probability of attending classical music or opera while those with music appreciation classes in school had a .18 probability. For theatre attendance, the population probability was .20; among those with theatre arts education, the probability of attending live theatre was .23 and for those with dance education, the probability was .28. For ballet attendance, the probability of attendance was small for all groups: across the entire population, the probability was .03 and there was little variation by arts education experience, though the probability for those with advanced educational attainment was considerably higher at .06.

Table 5.

Predicted probabilities of attending benchmark arts events by educational attainment and arts education status.

	(1)	(2)	(3)	(4)	(5)
	Classical music/Opera	Jazz	Theatre	Ballet	Other dance
Population average	.12	.10	.20	.03	.06
Respondent’s educational attainment
Less than H.S. diploma	.04	.04	.07	.01	.02
H.S. diploma only	.06	.08	.13	.01	.04
Attended some college	.10	.09	.20	.02	.05
Holds associate’s degree	.10	.12	.18	.02	.07
Holds bachelor’s degree	.18	.12	.28	.04	.07
Holds postgraduate degree	.25	.17	.30	.06	.10
Arts education by discipline
School music (performance)	.14	.12	.22	.03	.07
School visual art (creation)	.11	.12	.23	.03	.08
School photography/cinema	.12	.12	.20	.02	.08
School theatre	.15	.10	.23	.03	.07
School dance	.13	.15	.28	.03	.08
School visual art appreciation	.14	.10	.27	.04	.07
School music appreciation	.18	.16	.22	.03	.06

Note. Predicted probabilities, other than the population average, in this table derive from the logistic regression models reported in Table 4 above, and hold all other covariates constant.

Discussion

Affirming many of the results of the previous Bergonzi and Smith (1996) analysis, and contrary to the main conclusions of Mantie and Tucker (2008), the present study provides evidence that school-based music and arts education in the United States has measurable, positive arts-related outcomes in adulthood – at least in terms of attending the “benchmark” arts events that have been consistently included in the Survey of Public Participation in the Arts and in certain kinds of arts creation or performance. The present study provides evidence that school-based music education, in particular, appears to be a worthwhile investment in promoting a lifelong participation with the arts. Adults who had experience with school-based music education are considerably more likely than those with no arts education to attend or patronize arts performance and pursue art creation and performance as adults, even after controlling for respondents’ household income, educational attainment, sex, race, and parental educational attainment. If one aim of music education, as many music educators report (Zakaras & Lowell, 2008), is to engender a lifelong connection with the arts, the results of this study suggest that music (and arts education more broadly) is achieving this aim for many alumni.

Former chair of the U.S. National Endowment for the Arts Bill Ivey has argued that “arts education in the United States exhibits fundamental flaws that shortstop engagement in art broad enough to make art part of the life of every citizen” (Ivey, 2010, p. 109). Evidence from the present study, however, may undercut Ivey’s assertion. Rather than disengage from art-making and arts attendance upon graduation, the present study provides evidence that students of school-based music and arts education were significantly more likely to create art in their own lives and to patronize arts events than were adults who lacked school-based music or arts education. In other words, in this study I find clear, consistent relationships between school-based arts education in various disciplines and later adult engagement as a performer or creator of the studied art. Thus, any cultural policy or program that seeks to increase lifelong engagement with the arts must seriously contend with the improvement of arts education in the locality to be served. The National Association for Music Education has long argued that the improvement of local arts education should be focused on the high quality delivery of a sequential, rigorous program taught by a qualified and credentialed teacher (National Association for Music Education, 1999). Cultural policy and programs that emphasize audience development, and by extension, the arts as drivers for economic development, should be examined for connections to and intersections with enhancements to school-based, sequential music and arts education – which may currently serve (in one capacity) as the largest audience development enterprise in the nation. Thus, if Mizell’s (2005) and Rabkin and Hedberg’s (2011) conclusions that music and arts education participation in the United States is in a state of decline – a characterization disputed by Elpus’s (2014) analysis of high school transcript data – then the present analysis suggests the shoring up of music and arts education within the nation’s schools will have beneficial impacts on the participation in and patronage of music and arts organizations in the future.

Rationales for arts education in the schools often try to echo the utilitarian value of the arts to whatever other aims are the then-present focus of the educational system (Elpus, 2007). Yet, when pressed, most arts educators would readily admit that, philosophically, they entered the profession primarily to share and develop the art that had become so central to their own lives with their students (Eisner, 2002; Reimer, 2003). Results of the present study suggest that these efforts have not been in vain – school-based arts education does lead to future engagement with the arts and does have measurable arts outcomes. While educators should be heartened to learn that arts education uptake is potentially far greater than they might imagine, they should take the increase of persistence rates as a personal charge. While one year of arts education is frequently mandated, future years of arts education are clearly elective, and in many cases the result of the experiences provided to students by the arts educators with whom they work. Arts educators often serve as important, influential non-parental adults in the lives of the students they teach (Beam, Chen, & Greenberger, 2002; Ivaldi & O’Neill, 2008), and deepening the connection of students to the making and consumption of arts beyond one year of a mandatory course and throughout their lives is often the result of teacher-driven inspiration. Arts educators should not take this charge lightly, and recognize that their daily work, if done well, can help their students develop into an arts-engaged citizen.

The present study has several implications for future research in the field as well. Whereas much of the literature examining outcomes related to music education participation and non-participation has focused on so-called “extramusical” or non-artistic outcomes (e.g., Costa-Giomi, 2004; Dege, Wehrum, Stark, & Schwarzer, 2014; Schellenberg, 2005), the present study suggests that at least some arts-related consequents of music and arts education are viable for research. Future research should continue to examine arts-related consequents of music instruction, particularly those that are more democratic and universally available than are the benchmark arts activities explored in the National Endowment for the Arts surveys. Additionally, the present study affirms prior research (e.g., Seaman, 2006; Rabkin & Hedberg, 2011) showing that there are clear, high magnitude relationships between educational attainment and benchmark arts event attendance. However, the precise nature or mechanism underlying this relationship has yet to be fully explained. Myriad possible explanations exist. For example, those who have earned college and postgraduate degrees may have had greater opportunities to be introduced to the benchmark arts during general education requirements in their college courses of study and discovered an affinity for them, prompting them to continue to patronize the arts beyond their schooling years. Alternatively, and perhaps less kindly, it may be that those with the cultural capital afforded by postsecondary education may see attendance at benchmark arts events as important parts of maintaining social status and patronize them (or report patronizing them) without having developed any particular affinity or preference for the arts. Regardless of the reason for the observed association between educational attainment and benchmark arts attendance, further research is needed to disentangle these possibilities from among the many potential causal mechanisms.

Clearly, arts advocates, working artists, arts administrators, and arts educators in the United States should come to recognize the central importance of school-based arts education in creating an arts-engaged populous. Nearly two decades ago, Charles Fowler (1996) argued that “the arts, like other subjects in the curriculum of American schools, are affiliated with the schools’ important responsibility to pass on the civilization – in this case, our rich and rewarding cultural heritage – to the next generation” (p. 3). Results from the present study should suggest to stakeholders that improvements in the status, availability, and uptake of arts education in the nation’s schools will likely lead to greater democratized art making, greater support for artists and arts organizations, and greater engagement in the arts by adults. The increasing uptake rates of high school arts courses over the three-decade period between 1982 and 2009 are indeed heartening, but as school-based arts education appears to be a worthwhile investment in the future artistic engagement of the nation’s youth, the important place of the arts in American education should continue to be celebrated, protected, and advanced.

Footnotes

Acknowledgements

The opinions expressed in this paper are those of the author and do not represent the view of the Office of Research & Analysis or the National Endowment for the Arts. The NEA does not guarantee the accuracy or the completeness of the information included in this report and is not responsible for any consequence of its use.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This analysis was supported in part by an award from the Research: Art Works program of the National Endowment for the Arts: Grant #14-3800-7020.

References

Beam

M. R.

Chen

Greenberger

(2002). The nature of adolescents’ relationships with their “very important” nonparental adults. American Journal of Community Psychology, 30(2), 305–325. http://doi.org/10.1023/A:1014641213440

Bergonzi

Smith

(1996). Effects of arts education on participation in the arts (No. 36). Washington, DC: National Endowment for the Arts. Retrieved from http://arts.gov/sites/default/files/NEA-Research-Report-36.pdf

Borgonovi

(2004). Performing arts attendance: An economic approach. Applied Economics, 36, 1871–1885. http://doi.org/10.1080/0003684042000264010

Bray

(2000). An examination of GCSE music uptake rates. British Journal of Music Education, 17(1), 79–89.

Costa-Giomi

(2004). Effects of three years of piano instruction on children’s academic achievement, school performance and self-esteem. Psychology of Music, 32(2), 139–152. http://doi.org/10.1177/0305735604041491

Dege

Wehrum

Stark

Schwarzer

(2014). Music lessons and academic self-concept in 12- to 14-year-old children. Musicae Scientiae. http://doi.org/10.1177/1029864914523283

DiMaggio

Mukhtar

(2004). Arts participation as cultural capital in the United States, 1982–2002: Signs of decline? Poetics, 32(2), 169–194. http://doi.org/10.1016/j.poetic.2004.02.005

Eisner

E. W.

(2002). The arts and the creation of mind. New Haven, CT: Yale University Press.

Elliott

D. J.

Silverman

(2014). Music matters: A philosophy of music education (2nd ed.). New York, NY: Oxford University Press.

10.

Elpus

(2007). Improving music education advocacy. Arts Education Policy Review, 108(3), 13–18. http://doi.org/10.3200/AEPR.108.3.13–18

11.

Elpus

(2013). Is it the music or is it selection bias? A nationwide analysis of music and non-music students’ SAT scores. Journal of Research in Music Education, 61(2), 175–194. http://doi.org/10.1177/0022429413485601

12.

Elpus

(2014). Evaluating the effect of No Child Left Behind on U.S. music course enrollments. Journal of Research in Music Education, 62(3), 215–233. http://doi.org/10.1177/0022429414530759

13.

Elpus

(2015). National estimates of male and female enrolment in American high school choirs, bands and orchestras. Music Education Research, 17(1), 88–102. http://doi.org/10.1080/14613808.2014.972923

14.

Elpus

Abril

C. R.

(2011). High school music ensemble students in the United States: A demographic profile. Journal of Research in Music Education, 59(2), 128–145. http://doi.org/10.1177/0022429411405207

15.

Fitzpatrick

K. R.

(2006). The effect of instrumental music participation and socioeconomic status on Ohio fourth-, sixth-, and ninth-grade proficiency test performance. Journal of Research in Music Education, 54(1), 73–84. http://doi.org/10.1177/002242940605400106

16.

Fowler

(1996). Strong arts, strong schools: The promising potential and shortsighted disregard of the arts in American schooling. New York, NY: Oxford University Press.

17.

Heeringa

West

B. T.

Berglund

P. A.

(2010). Applied survey data analysis. Boca Raton, FL: Chapman & Hall.

18.

Ivaldi

O’Neill

S. A.

(2008). Adolescents’ musical role models: Whom do they admire and why? Psychology of Music, 36(4), 395–415. http://doi.org/10.1177/0305735607086045

19.

Ivey

(2010). Arts, inc.: How greed and neglect have destroyed our cultural rights. Berkeley: University of California Press.

20.

Kinney

D. W.

(2008). Selected demographic variables, school music participation, and achievement test scores of urban middle school students. Journal of Research in Music Education, 56(2), 145–161. http://doi.org/10.1177/0022429408322530

21.

Kinney

D. W.

(2010). Selected nonmusic predictors of urban students’ decisions to enroll and persist in middle school band programs. Journal of Research in Music Education, 57(4), 334–350. http://doi.org/10.1177/0022429409350086

22.

Long

J. S.

Freese

(2014). Regression models for categorical dependent variables using Stata (3rd ed.). College Station, TX: Stata Press.

23.

Mantie

Tucker

(2008). Closing the gap: Does music-making have to stop upon graduation? International Journal of Community Music, 1(2), 217–227. http://doi.org/10.1386/ijcm.1.2.217/1

24.

Mizell

(2005). Arts Education in the US: 1982–2002 (Research Report No. C02-94). Washington, DC: National Endowment for the Arts.

25.

Murnane

R. J.

Willett

J. B.

(2011). Methods matter: Improving causal inference in educational and social research. New York, NY: Oxford University Press.

26.

National Association for Music Education. (1999, January). The value and quality of arts education: A statement of principles. Retrieved from http://www.nafme.org/about/position-statements/the-value-and-quality-of-arts-education-position-statement/the-value-and-quality-of-arts-education/

27.

National Endowment for the Arts. (2013). How a nation engages with art: Highlights from the 2012 Survey of Public Participation in the Arts (No. 57). Washington, DC: Author.

28.

Rabkin

Hedberg

E. C.

(2011). Arts education in America: What the declines mean for arts participation (No. 52). Washington, DC: National Endowment for the Arts.

29.

Reimer

(2003). A philosophy of music education: Advancing the vision (3rd ed.). Upper Saddle River, NJ: Prentice Hall.

30.

Schellenberg

E. G.

(2005). Music and cognitive abilities. Current Directions in Psychological Science, 14(6), 317–320. http://doi.org/10.1111/j.0963–7214.2005.00389.x

31.

Seaman

B. A.

(2006). Empirical studies of demand for the performing arts. In Ginsburg

Victor A.

Throsby

David

(Eds.), Handbook of the economics of art and culture (Vol. 1, pp. 415–472). Amsterdam, the Netherlands: Elsevier. http://doi.org/10.1016/S1574–0676(06)01014–3

32.

Zakaras

Lowell

J. F.

(2008). Cultivating demand for the arts: Arts learning, arts engagement, and state arts policy. Santa Monica, CA: RAND.