Predictors of Taking Elective Music Courses in Middle School Among Low-SES,Ethnically Diverse Students in Miami

Abstract

Researchers attempting to show that music has positive effects on children need to understand and control for preexisting differences between those who do and do not select into musical participation in the first place. Within a large-scale, communitywide, prospective, longitudinal study of predominantly low-income, ethnically diverse students (N = 31,332), we examined characteristics of students who did and did not enroll in music elective courses (band, choir, orchestra, guitar, other) in public middle schools (sixth, seventh, and eighth grades) in Miami. Predictor variables included gender, ethnicity, poverty, special education, English language learner status, fifth-grade English proficiency, prior academic performance (fifth-grade grade point average [GPA], standardized math and reading test scores), and initial school readiness skills (social, behavioral, cognitive, language, and motor skills) at age 4. Only 23% of middle school students enrolled in a music class in sixth, seventh, or eighth grade, with band having the highest enrollment, followed by choir, orchestra, and guitar. Being male and having greater cognitive skills at age 4 and higher fifth-grade GPA and reading skills were related to later music participation. Black students, students in special education, and those not proficient in English were less likely to participate in middle school music classes. Results varied somewhat by type of music.

Keywords

music middle school selection electives ethnicity

Researchers have investigated the developmental value of music in children and found positive effects of music on cognition, behavioral, emotional, and social development (Brown & Sax, 2013; Moreno et al., 2011; Schellenberg, 2004; Williams et al., 2014; Winsler et al., 2011). In addition to these benefits, music has the potential to help children build confidence and drive for success through commitment, persistence, and motivation (Rosevear, 2010). Music is the most consistently available arts elective in the United States, but even music classes are not equally available to all students (Parsad et al., 2012). In the 2008–2009 academic year, schools where the majority of the students were eligible for free or reduced-price lunch (FRL) less consistently offered a music class (81%) compared to schools where only a small proportion of students were eligible for FRL (96%). Additionally, schools with the lowest poverty concentration more consistently offered five or more music classes (62%) compared to schools with a high poverty concentration (35%; Parsad et al., 2012). Students in poverty have limited access to music courses and fewer music course options.

The implementation of the No Child Left Behind Act led to a decline in budgets for music programs in the United States. Between 1999–2000 and 2003–2004, California schools experienced a 50% decrease in student enrollment in music classes and a 26.7% decrease in music teachers (Abril & Gault, 2006). This is especially unfortunate for low-income children of color, as several scholars argue that music and the arts can be especially helpful for children in poverty and have the potential to help reduce the achievement gap (Brown & Sax, 2013; Foran, 2009). The Turnaround Arts Initiative, for example, implemented high-quality arts education to low-income and underperforming school districts (Stoelinga et al., 2013). Students in Turnaround Arts schools increased more in reading and/or math proficiency compared to students in schools that received similar school improvement grants but allocated funding to other non-art-related areas (Stoelinga et al., 2013).

Arts programs in schools support student academic growth differently than do academic classes (Catterall, 1998). Catterall’s (1998) study compared children in poverty who did and did not have arts experiences. Students who were highly involved with the arts outperformed students who had little art involvement in English classes and on standardized tests, and this difference in academic performance became more pronounced between 8th and 10th grades. Students highly involved with the arts were less likely to be bored in school and less likely to drop out of school compared to students with little arts engagement (Catterall, 1998). Increased interest, persistence, and performance in school demonstrate the value of encouraging high arts involvement among all students and especially for students in poverty. Middle school is a critical time to examine student elective arts course enrollment because sixth through eighth grades is the first time students have the opportunity to choose full elective courses. Early participation is particularly important for music enrollment because, unlike high school, students may enter music classes in middle school with limited skills. High school music participation becomes more of a closed system because students often are expected to be equipped with certain musical skills in order to participate (i.e., auditions, ability to read music, basic technique; McNeal, 1998).

Methodological Issues

Research on the effects of music on children’s cognitive, social, and academic development is plentiful but limited because most of the research is correlational, which does not allow for causal inference. For example, there is correlational evidence that well-trained musicians show a variety of advanced cognitive skills, such as working memory, attention, executive functioning, and reaction time, compared to nonmusicians, presumably from years of practice and training (Bialystok & DePape, 2009; Pallesen et al., 2010; Zuk et al., 2014). However, it cannot be concluded that the musical training caused the enhanced cognitive function, because the students who went into music in the first place could have been different (e.g., higher initial IQ or greater initial executive-functioning skills) before the musical training took place, or a third variable, such as socioeconomic status (SES), parental education, or parental investment, could have led to both the cognitive enhancements and the pursuit of music.

Experimental research allowing causal inference through random assignment of students to either a musical training/experience group or a control group is rare but has provided causal evidence of ancillary benefits of musical training. Schellenberg (2004) randomly assigned 144 six-year-old children to four conditions. Two of the groups received music lessons, either piano or voice, for 36 weeks, and the other two groups received either drama lessons or no lessons for 36 weeks. Child IQ was assessed before beginning the lessons and after 36 weeks, and children in the music conditions had a greater increase over time in IQ than in the two control conditions (Schellenberg, 2004). Other experimental research also shows positive effects of musical training on cognition, executive functioning, and/or academic performance (Gardiner et al., 1996; Holochwost et al., 2017; Moreno et al., 2011).

Because randomly assigning students to musical experience groups is difficult, and often impossible (when examining in-school elective music courses, for example), quasi-experimental designs are critical in research on the effects of music on children. Quasi-experimental designs typically compare naturally existing groups of individuals who either are or are not exposed to musical experiences on cognitive or academic outcomes, with careful attention to measuring, and statistically controlling for, preexisting “selection factors”—the many ways that the children who do and do not get music exposure are initially different in terms of family demographics (ethnicity, SES, parental education) and/or child prior competence (earlier academic or cognitive skills). Numerous studies controlling for at least family demographics and in some cases many more covariates find that children involved in music perform better either in school or on various cognitive assessments compared to those without musical experience (Corrigall et al., 2013; Gerry et al., 2012; Miksza, 2010; Dos Santos-Luiz et al., 2016; Schellenberg, 2004; Southgate & Roscigno, 2009; Williams et al., 2014; Winsler et al., 2011).

Not all quasi-experimental studies, however, show significant “effects” of music, and the pattern appears to be that the more selection factors controlled for, and the stronger the statistical procedure used, the less likely significant effects of music are found. A recent study using data from the Panel Study of Income Dynamics Child Development Supplement (Foster & Jenkins, 2017) included many family and child covariates and found no evidence for a causal effect of musical experiences on child outcomes (math skill, verbal skill, and working memory; parent reports of social skills and behavior; and child reports of self-concept) after selection variables were included. Children who received musical experiences outside of school had parents with higher IQ, greater years of education, and more money and time to spend on their children, which explained the advanced academic achievement (Foster & Jenkins, 2017). These researchers, however, examined parent reports of student out-of-school musical training rather than middle school music enrollment.

Students who take music in school are also different from students who do not (Elpus, 2013; Kinney, 2010). Elpus (2013) found music students had significantly greater SAT scores than did nonmusic students. However, after controlling for factors such as SES, race, family composition, and academic achievement, the difference in SAT performance between music students and nonmusic students disappeared. It may be that students who enroll in music classes are those who are already well situated and advantaged and likely to excel in school regardless of their musical experience (Elpus, 2013). Additionally, music student performance on the SAT varied by the type of music elective. When background differences were not controlled for, instrumental students (band and orchestra) scored higher than nonmusic students on the SAT, whereas choir students scored lower than nonmusic students (Elpus, 2013). These studies highlight both the importance of understanding the selection effects at play in research on the potential benefits of musical experiences and the need to examine specific types of music exposure.

The Current Study

Using data from a large-scale, communitywide, prospective, longitudinal study of predominantly low-income, ethnically diverse students in Miami (N = 31,332), we examined characteristics of students who did and did not enroll in music elective courses (band, choir, orchestra, guitar, other) in public middle schools (sixth, seventh, and eighth grades). Predictor variables included gender, ethnicity, poverty, special education, English language learner (ELL) status, fifth-grade English proficiency, prior academic performance (fifth-grade grade point average [GPA], standardized math and reading test scores) and initial school readiness skills (social, behavioral, cognitive, language, and motor skills) at age 4. Prior research typically has not used longitudinal designs and is limited by small and relatively homogenous and advantaged samples. We followed a large group of ethnically diverse (mostly Hispanic and Black) students largely in poverty for 10 years from prekindergarten through eighth grade. We used official school transcript data to determine exposure to in-school music electives rather than relying on student or parent retrospective reports of more general types of in- and out-of-school musical experiences. In addition to examining common selection factors, such as SES, gender, and ethnicity, we examined a host of additional variables likely associated with middle school musical exposure, such as special education status, ELL status, English proficiency, initial school readiness in preschool (cognitive, social, behavioral, and motor skills), and prior elementary school academic performance.

We asked the following research questions: (1) What proportion of students in Grades 6 through 8 enroll in music classes, and what type of music classes are they taking (i.e., band, chorus, guitar, orchestra, other)? (2) What are the preexisting demographic and child-level differences (gender, ethnicity, poverty, special education status, ELL status, English proficiency, prior academic performance, school readiness at age 4) between students who do and do not select into music electives in middle school? and (3) To what extent are selection factors different for different types of music courses (i.e., band, choir, guitar, orchestra)?

Method

Participants

Participants of this study were children from the Miami School Readiness Project (MSRP; Winsler et al., 2008, 2019), a large-scale, prospective longitudinal study that followed five cohorts of children who attended either community-based childcare with subsidies or public school preK programs at age 4 between 2002 and 2007 in Miami, Florida. Not all cohorts in the sample reached seventh or eighth grade by the 2013–2014 academic year, when data collection stopped. Cohort-based attrition existed by design: There are sixth-grade data for five cohorts, seventh-grade data for four cohorts, and eighth-grade data for three cohorts. Thus, 30,413 students had sixth-grade data, 23,788 students had seventh-grade data, and 16,392 students had eighth-grade data at the time of this writing. Supplemental Table S1 (included with the online version of this article) shows the background characteristics, school readiness, and fifth-grade academic performance of the entire sample. Students were 51% male, and racial-ethnic diversity was 60% Hispanic, 33% Black, 6% White, and 0.7% Asian/Pacific Islander. A large majority of the sample was in poverty (81% received FRL in sixth grade). About 15% of the students received special education services in sixth grade. In kindergarten, 57% were categorized as ELLs, but by fifth grade, 95% of students were English proficient according to district criteria for no longer receiving English for Speakers of Other Languages (ESOL) services. The sample included children who repeated a grade in elementary or middle school (14.6%).

Procedure

School readiness was measured directly during the students’ preK year by well-trained outside assessors or by the student’s preK teacher, and parents and teachers reported on the child with the survey instruments described later (Crane et al., 2011; Winsler et al., 2008, 2019). School system student records were collected for each child from kindergarten through eighth grade. Administrative school records of student demographic information (e.g., student gender, ethnicity) were collected with consent and with appropriate deidentification procedures as approved by the institutions’ institutional review board procedures.

Measures

Child-Level Predictors

Gender

Females were coded as 0 and males as 1 using school record data.

Race-ethnicity

Child race-ethnicity was collected by the school district and coded into four categories: “Hispanic” (60%, n = 17,734) included individuals who identified as Hispanic or Latino; “Black” (33%, n = 9,635) included individuals who identified as African American, Black, Caribbean, or Black and some other racial group; “White/Other” (6%, n = 1,883) included individuals who identified as White or a mixture of other racial groups; and “Asian/Pacific Islander” (0.7% n = 199) included individuals who identified as Asian or Pacific Islander.

Cognitive, language, and motor skills at age 4

The Learning Accomplishment Profile–Diagnostic (LAP-D; Nehring et al., 1992) is a norm-referenced, developmental assessment administered individually to children at the beginning and end of the preK year (Time 1 [T1], September through October; Time 2 [T2], April through May). Scores from T2 were used to most accurately represent children’s school readiness at school entry, and if T2 scores were not available, T1 scores were used. Children who received subsidies to attend child care were administered the LAP-D by bilingual, trained assessors. Children who attended public school preK programs were administered the LAP-D by their preK teachers, who completed the same training by the publisher. The primary language (English or Spanish) used to administer the LAP-D was chosen by the assessor after discussion with the teacher and interacting with the student in both languages to determine the student’s strongest language. There are four scales with two subscales each: Cognitive (Counting, Matching), Language (Comprehension, Naming), Fine Motor (Writing, Manipulation), and Gross Motor (Body, Object). Internal consistency reliabilities ranged from .93 to .95 (Winsler et al., 2008).

Socioemotional skills and behavior at age 4

Children’s social skills and behavior were measured at age 4 using the Devereux Early Childhood Assessment (DECA; LeBuffe & Naglieri, 1999). Preschool teachers completed the DECA at the same two time points as described earlier. T2 scores were used when available, and if not, T1 scores were used. The DECA comprises three subscales (Initiative, Attachment, and Self-Control) that are combined to create an overall measure of children’s socioemotional strengths (total protective factors [TPF]), and there is a separate Behavioral Concerns subscale. A 5-point Likert scale indicates frequency of child behavior (0 = never to 4 = very frequently). Higher scores on TPF indicate greater strengths, and higher scores on Behavioral Concerns indicate more problems. Teachers could fill out the DECA in Spanish or English, and 34% filled out the Spanish form. Good internal consistency reliability was found among the sample for TPF (.94) and behavioral concerns (.81), and this did not vary by rater or language of form (Crane et al., 2011).

ELL status in kindergarten

ELL status was determined by the schools from parent-reported home language used at kindergarten entry. Those who reported predominantly speaking another language at home were considered ELLs in kindergarten by the school system.

English proficiency at the end of elementary school

ELL students were assessed by the district for English proficiency with the Miami-Dade County Oral Language Proficiency Scale–Revised (Abella et al., 2005). This test measures aural comprehension and oral production, and those students who were determined to be ELLs were provided with ESOL instruction. ESOL levels are marked 1 to 5: Levels 1 and 2 indicate beginning English for learners who still have much difficulty, Levels 3 and 4 are advanced stages of English learning, and Level 5 is considered sufficiently proficient in English to exit the ESOL program (Abella et al., 2005). Those at Level 5 (and those who were never considered ELLs) were considered English proficient (coded 1) and those with a level lower than 5 were coded 0.

Fifth-grade GPA

GPA is the average of grades each student received from all subjects in fifth grade. Grades were based on a 5-point A-to-F scale, where 5.0 = A, 4.0 = B, 3.0 = C, 2.0 = D, and 1.0 = F.

Fifth-grade standardized test scores

The state’s high-stakes standardized test (Florida Comprehensive Assessment Test; Human Resources Research Organization & Harcourt Assessment, 2007) is a mandatory exam given to 3rd through 12th graders in Florida. The test consists of a math and a reading portion, and questions are formatted using multiple-choice, short-answer, and detailed responses. The test is graded using a continuous score ranging from 100 to 500 with Cronbach’s alpha reliability for reading and math tests at .91 and .88 (Morrissey et al., 2014).

Special education status

Special education status in sixth grade was collected from administrative records in the form of the student’s primary exceptionality code. Students who had a code in sixth grade were coded 1 for special education, and others were coded 0. Disability types included intellectual disability, speech/language disorder, visually impaired, deaf or hard of hearing, specific learning disabled, dual-sensory impaired, autistic, emotionally disturbed, traumatic brain injured, and other health impaired. Gifted students were coded 0 and were considered nondisabled.

Poverty status

Children’s poverty status was based on student eligibility for FRL in sixth grade. Eligibility is determined at the beginning the year by the family meal application completed by the child’s primary caregiver. Children at 130% of the federal poverty line qualify for free lunch, and those who are 185% of the poverty line qualify for reduced-price lunch. This was coded 1 for receiving and 0 for not receiving free or reduced-price lunch.

Music exposure in middle school

Administrative data collected each year on each student for all grade levels included all course subjects taken on the end-of-year transcript (e.g., math, social studies, science, band). Music is an in-school arts elective that students had the option to take. Using whether a child’s transcript included a music class, we created the following variables denoting whether, when, and which music courses students took in middle school. If the course Band appeared in a given grade (sixth, seventh, or eighth), the student was flagged as having band experience in that grade (i.e., yes = 1 vs. no = 0). These were then aggregated across all grades (sixth, seventh, and eighth) to make a variable indicating whether (1) or not (0) that child ever experienced band at least once in middle school. The same was done for the other forms of music: chorus, orchestra, guitar, and other (music theory, keyboard, and general music). We also created the same set of variables aggregated across all forms of music to indicate whether the student elected to take any type of music course during middle school.

It is important to note that practically all (94.4%) of the children attended middle schools that offered music, although this varied somewhat by race-ethnicity (95.2% for White, 95.9% for Hispanic, 91.4% for Black, and 97.1% for Asian/Pacific Islander students, with Black students being slightly more likely to attend a school that did not offer music electives. Next, we report analyses including all students, but we note the rare cases where results changed slightly when analyses were rerun on just students who attended a school where music was offered. Also note that our data come from administrative records, and so we were not able to collect information on whether students had out-of-school musical experiences, either concurrently or earlier.

Results

Research Question 1: What proportion of students in Grades 6 through 8 enroll in music classes, and what type of music classes are they taking (i.e., band, chorus, guitar, orchestra, other)?

Table 1 shows overall music enrollment rates for each type of music elective for sixth, seventh, and eighth grades and across all of middle school. Out of the 31,332 students in the sample, 22% ever enrolled in a middle school music class. Student participation was lowest in eighth grade, at 12%; sixth graders had slightly more participation, with 14% enrollment; and seventh graders had the greatest participation in music classes, at 15%. Across all years, band had the highest enrollment, followed by choir, guitar, and orchestra. This pattern persisted except in eighth grade, where more students enrolled in orchestra than guitar. Only 3% of students enrolled in more than one music class, and this was most common in eighth grade, followed by seventh and then sixth grade. Only 4% of students enrolled in the rare courses of keyboard, general music, music theory, and jazz ensemble. Enrollment in these classes was highest for seventh grade, followed by eighth and then sixth grade.

Research Question 2: What are the preexisting demographic and child-level differences between students who do and do not select into music electives in middle school?

Table 1.

Proportion of Middle School Students Enrolled in a Music Class by Year and Type of Class.

	Grade 6 (n = 30,413)		Grade 7 (n = 23,788)		Grade 8 (n = 16,392)		Ever in Middle School (N = 31,332)
Type of Music Class	n	%	n	%	n	%	n	%
Any music class	4,379	14.4	3,613	15.2	1,992	12.2	7,040	22.5
Band	2,415	7.9	1,855	7.8	1,008	6.1	3,861	12.3
Choir	682	2.2	620	2.6	342	2.1	1,253	4.0
Guitar	497	1.6	480	2.0	237	1.4	968	3.1
Orchestra	468	1.5	369	1.6	272	1.7	687	2.2
Keyboard	266	0.9	225	0.9	160	1.0	594	1.9
General Music	94	0.3	150	0.6	84	0.5	327	1.0
Music Theory	118	0.4	96	0.4	43	0.3	249	0.8
Jazz Ensemble	0	0.0	4	0.001	1	0.001	5	0.001
More than one class	159	0.5	184	0.8	148	0.9	464	3.0

First, we examined variables individually using univariate analyses. Chi-square analyses were used for categorical variables and t tests for continuous variables. We then performed a series of developmentally informative, hierarchical, multivariate logistic regression analyses including all examined variables in blocks as predictors of participation in any type of music class in middle school. This process was then repeated separately for the four largest types of music classes: band, choir, guitar, and orchestra.

Any Music

Table 2 shows how enrollment in any music class in middle school varied by each of our demographic and prior-child-competence variables. Males were slightly more likely to enroll in music during middle school compared to females, χ²(1) = 5.31, p < .021. There were notable ethnic differences in enrollment in music overall, χ²(3) = 90.88, p < .001. Asian/Pacific Islander students were most likely to take a music class, followed by White students and then Hispanic students, with Black students being the least likely to take music classes in middle school. Students who received FRL were less likely to enroll in music classes overall compared to students not in poverty, χ²(1) = 15.28, p < .001, and students with disabilities were half as likely to enroll in a music class than typically developing students, χ²(1) = 269.89, p < .001. ELL students were slightly more likely to enroll in a music class than non-ELL students, χ²(1) = 21.23, p < .001, and students not fully English proficient in fifth grade were less than half as likely to take music, χ²(1) = 132.42, p < .001. Finally, as seen in Table 3, students who went on to select music elective courses in middle school performed better on every measure of prior academic performance in fifth grade (GPA, math, reading) and entered kindergarten 7 years earlier with stronger cognitive, language, motor, social, and behavioral skills than those who did not take music electives in middle school.

Table 2.

Demographic Correlates of Any Music Class Enrollment in Middle School.

Variable	Enrolled
Variable	n	%
Total	7,025	22.5
Gender*
Male	3,724	23.1
Female	3,301	22.0
Ethnicity***
White/Other	499	25.3
Hispanic	4,404	23.9
Black	2,058	19.6
Asian/Pacific Islander	62	30.2
Poverty status***
No free/reduced lunch	1,726	24.7
Free/reduced lunch	5,493	22.3
Disability status***
Nondisabled	6,249	24.5
Disability	678	13.8
English language learner (ELL)***
Non-ELL	2,831	21.2
ELL	4,208	23.4
English proficiency***
English proficient	6,669	23.6
Not English proficient	157	10.6

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

Table 3.

Early Skills and Academic Performance Bivariate Correlates of Any Music Enrollment in Middle School.

Variable	Enrolled			Not Enrolled
Variable	n	M	SD	n	M	SD	d
School readiness at age 4
Gross motor skills**	4,268	69.10	28.43	13,166	67.84	29.03	0.04
Fine motor skills***	4,852	59.32	27.76	14,959	55.75	29.17	0.13
Cognitive skills***	4,841	56.46	29.94	14,954	50.97	30.31	0.18
Language skills***	4,838	47.62	30.78	14,869	43.55	30.60	0.13
Social skills***	6,024	61.24	27.34	19,962	58.84	28.12	0.09
Behavioral concerns***	6,024	44.31	28.86	19,962	46.96	29.68	0.09
Fifth-grade academic performance
GPA***	6,750	4.19	0.52	22,204	4.06	0.59	0.25
FCAT Math***	6,751	303.73	67.50	21,822	279.13	68.01	0.36
FCAT Reading***	6,760	286.71	61.02	21,856	262.81	60.28	0.39

Note: GPA = grade point average; FCAT = Florida Comprehensive Assessment Test.

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

The preceding analyses are unadjusted bivariate associations. Here we report how the demographic, school readiness, and prior academic performance variables all combine to predict music course enrollment using logistic regression analyses that control for the intercorrelations between the predictor variables. Step 1 included the demographic variables (ethnicity, gender, FRL eligibility, disability status, ELL status, English proficiency) and school readiness skills at age 4. In Step 2, prior academic achievement (fifth-grade GPA, fifth-grade standardized test scores) was included. Model 2 shows the relationship between music enrollment and fifth-grade achievement, controlling for demographic and school readiness variables. Model 2 also shows whether the demographic and school readiness variables at age 4 remain associated with music enrollment in middle school after considering fifth-grade achievement. Thus, if an association established in Step 1 ceases to be significant in Step 2, the association is better explained by academic competence than demographics. Note that we did not include math scores in Block 2 because they were too highly correlated with reading (r = .88) to avoid multicollinearity.

Step 1

Table 4 shows the results of the logistic regression predicting selection into any music classes in middle school. Odds ratios (ORs) are provided, which indicate the extent to which the odds of taking a music elective course in middle school increase (greater than 1) or decrease (less than 1) as a function of being one level of the variable (i.e., male) compared to the other (female). For continuous predictors, the OR indicates how much the odds of music selection increase/decrease with a 1-point increase in the predictor (i.e., moving from the 39th to the 40th percentile in cognitive skills).

Table 4.

Logistic Regression Predicting Any Music Enrollment in Middle School.

Variable	Step 1		Step 2
Variable	Odds Ratio	SE(B)	Odds Ratio	SE(B)
Race-ethnicity
Hispanic/White	0.901	.086	0.952	.087
Black/White	0.722***	.088	0.825*	.089
Asian/White	0.942	.245	0.928	.247
Black/Hispanic	0.801***	.057	0.866**	.057
Male	1.199***	.041	1.257***	.042
Received free/reduced lunch	0.852**	.054	0.957	.055
Special education	0.633***	.077	0.761***	.078
English language learner	1.065	.053	1.017	.053
English proficiency	2.728***	.148	2.097***	.149
School readiness at age 4
LAP-D
Gross motor skills	1.000	.001	1.000	.001
Fine motor skills	0.998	.001	0.998	.001
Cognitive skills	1.005***	.001	1.004***	.001
Language skills	1.000	.001	0.999	.001
DECA
Social skills	1.000	.001	0.999	.001
Behavioral concerns	1.000	.001	1.001	.001
Elementary academic performance
GPA in fifth grade			1.249***	.045
Reading fifth grade			1.004***	.000

Note: To analyze the fourth ethnicity contrast (Black/Hispanic), we ran another regression model flipping the reference group from White to Black students. LAP-D = Learning Accomplishment Profile–Diagnostic; DECA = Devereux Early Childhood Assessment; GPA = grade point average.

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

In Step 1, ethnicity, gender, poverty, disability, English proficiency, and cognitive skills were unique predictors of music participation in middle school. Ethnicity, gender, disability, English proficiency, and cognitive skills remained unique predictors after including academic performance, but poverty ceased to be a unique predictor. Black students had 28% fewer odds than White/Other students and 20% fewer odds than Hispanic students of taking a music class in middle school. Males had 20% greater odds of taking music than females. Students with a disability had 37% fewer odds of enrolling in a music class than students without a disability. Students who were proficient in English in fifth grade had almost 3 times greater odds of enrolling in a music class than those who were not. Cognitive skills at age 4 were positively associated with later enrollment in a middle school music course. For every 1-point increase in cognitive skills at age 4, the odds of music enrollment increased by .005%. Thus, a 50-point increase in cognitive skills at age 4 results in 25% (e.g., moving from the 25th to the 75th percentile: .005 × 50) greater odds that the child will enroll in music in middle school. Recall that these are adjusted effects after controlling for other variables in the model. Thus, for example, even controlling for poverty status and school readiness, Black students were still less likely to take a music class in middle school than other groups.

Step 2

Step 2 included students’ prior academic competence variables (fifth-grade GPA and reading score). Prior achievement was significantly positively associated with later middle school music electives (controlling for demographic and school readiness skills). For every 1-point increase in prior GPA (moving from a B to an A), the odds of music enrollment increased by 25%. For every 1-point increase in fifth-grade reading, the odds of later enrolling in a music class increased by .004%. Thus, a 50-point increase in fifth-grade reading results in 20% (.004 × 50) greater odds that the child would later enroll in music in middle school.

It is important to note that once children’s fifth-grade achievement is taken into account, the demographic and school readiness effects (reported in Step 1) remained significant predictors except for poverty and ELL status. Once prior child performance in school was entered, there were no differences in the odds of taking a music class as a function for FRL, suggesting that the reason why poverty was associated with music selection before is because poverty leads to poorer academic performance in school, and it is school performance that is a more important predictor of music selection than poverty (within this mostly low-income sample). Also notable is that Black students were still less likely to take music classes compared to Hispanic and White students even after poverty and school performance were factored in; however, effect sizes reduced somewhat in magnitude after school performance was added. This means that part of the reason why Black students are less likely to take music is that they are not doing as well in school as other students. But it also means that the disparity in access to music in middle school for Black students is still true even for nonpoor students and students doing well in school.

The preceding analyses include all children in the sample; however, as mentioned in the Method section, a very small group (5%) of students attended a school that appeared to not offer music electives. To see whether that contributed to our findings, we reran the logistic regression limiting the sample to only those students who attended schools that clearly offered music options. Results were practically identical in almost all cases, with parameter estimates typically changing by only .02 points. However, when limiting the analyses to only children attending schools that offered music and reducing the sample size by 418 students, the formerly statistically significant effect of Black students being less likely to take music (compared to White and Hispanic students) reduced to marginal/nonsignificance. Additional exploration revealed that this change was mostly due to the race-ethnicity difference found in guitar enrollment specifically (see Supplemental Table S7). The race-ethnicity difference in guitar enrollment disappeared when analyses were limited to students who had the option of taking music (no changes were observed for the other forms of music). This suggests that Black students are more likely to attend schools where music (especially guitar) is not offered and that some (but not all) of the race-ethnicity effects involving Black students having less music participation can be explained by Black students being more likely to attend a school that does not offer music.

Univariate Analyses

Research Question 3: To what extent are selection factors different for different types of music courses?

Here we describe the unadjusted bivariate associations found for each type of music elective. Supplemental Tables S2, S4, S6, and S8 (included with the online version of this article) show how enrollment in band, chorus, guitar, and orchestra in middle school varied across the demographic and child competence variables. Overall, 12% of the sample enrolled in band, 4% enrolled in chorus, 3% enrolled in guitar, and 2% enrolled in orchestra at least once in middle school.

Gender

Similar to overall music discussed already, males were more likely to enroll in band, χ²(1) = 136.02, p < .001, and nearly twice as likely to enroll in guitar, χ²(1) = 46.69, p < .001, than females. Gender was oppositely related to chorus enrollment: Females were 3 times as likely to enroll in choir than males, χ²(1) = 371.69, p < .001. There was no gender difference for orchestra, χ²(1) = 3.69, p = .06.

Ethnicity

There was no association between race-ethnicity and band, χ²(3) = 1.78, p = .62, or chorus enrollment, χ²(3) = 4.97, p = .17. However, we did find associations between ethnicity and guitar, χ²(3) = 101.57, p < .001, and orchestra enrollment, χ²(3) = 182.64, p < .001. White/Other, Hispanic, and Asian/Pacific Islander students were more likely to take guitar than Black students. The same ethnic difference seen in guitar favoring White and Hispanic students over Black students was also found for orchestra. However, for orchestra, we saw that Asian/Pacific Islander students were much more likely to enroll than all other groups.

Poverty status

We found that the lower likelihood of music enrollment among students who receive FRL was driven mostly by guitar and orchestra. Students in poverty in sixth grade were less likely to take guitar, χ²(1) = 15.53, p < .001, and more than half as likely to take orchestra, χ²(1) = 157.39, p < .001, than students who were not in poverty. There was no difference between students who received FRL and students who did not for chorus, χ²(1) = 0.75, p = .39. Students who received FRL were actually slightly more likely to enroll in band than students who did not, χ²(1) = 9.87, p < .01. This suggests that poverty is not as large a barrier for involvement in band as it is for orchestra and guitar.

Disability status

As was the case for overall music enrollment, students who received special education services were half as likely to enroll in band, χ²(1) = 111.71, p < .001; chorus, χ²(1) = 51.28, p < .001; and guitar, χ²(1) = 52.26, p < .001, compared to typically developing students. There was an even greater difference in enrollment related to disability status for orchestra. Students with a disability were nearly 3 times less likely to enroll in orchestra than students without a disability, χ²(1) = 42.63, p < .001.

ELL status

The greater likelihood of ELLs enrolling in music overall was largely driven by guitar, χ²(1) = 68.62, p < .001, and orchestra, χ²(1) = 14.45, p < .001. Impressively, ELL students were nearly twice as likely to enroll in guitar than non-ELL students. Unlike guitar and orchestra, ELL students were slightly less likely to enroll in choir than non-ELL students, χ²(1) = 12.46, p < .001. We did not find a relationship between ELL status and band enrollment, χ²(1) = 0.00, p = .995.

English proficiency

Students who were proficient in English by fifth grade were more likely to enroll in band, χ²(1) = 54.04, p < .001; chorus, χ²(1) = 31.20, p < .001; guitar, χ²(1) = 14.39, p < .001; and orchestra, χ²(1) = 16.06, p < .001, compared to students not English proficient by fifth grade. Notably, students proficient in English were twice as likely to enroll in band and twice as likely to enroll in orchestra than students not proficient in English.

School readiness at age 4

Similar to overall music enrollment, students who later enrolled in band, t(17432.01) = −2.72, p < .01, and guitar, t(17432.01) = −4.16, p < .001, had higher gross motor skills at age 4 compared to students who did not choose to enroll in band or guitar in middle school. We did not find differences in gross motor skills between students who enrolled in chorus, t(809.08) = −1.15, p = .25, or orchestra, t(431.02) = 1.62, p = .11. Students enrolled in band, chorus, guitar, and orchestra on average had higher age 4 fine motor skills: band, t(19809.01) = −2.58, p < .01; chorus, t(937.42) = −5.52, p < .001; guitar, t(19809.01) = −3.49, p < .001; orchestra, t(19809.01) = −5.51, p < .001, compared to students who did not enroll in middle school. The same was true for cognitive skills in preK: band, t(3583.01) = −4.95, p < .001; chorus, t(934.18) = −6.00, p < .001; guitar, t(19793.01) = −5.97, p < .001; orchestra, t(491.15) = −7.04, p < .001); and for early language skills: band, t(3557.66) = −2.96, p < .01; chorus, t(19705.01) = −6.23, p < .001; guitar, t(19705.01) = −3.69, p < .001; orchestra, t(491.29) = −6.46, p < .001.

Similar to overall music enrollment, students enrolled in chorus, t(25984.00) = −5.44, p < .001; guitar, t(25984) = 1.751, p < .05; and orchestra, t(622.35) = −4.03, p < .001, had higher early social skills compared to students not enrolled in chorus, guitar, or orchestra, respectively. There was no relationship between social skills and band enrollment, t(4332.39) = −1.07, p = .28. The finding that middle school (overall) music students had fewer behavioral concerns was driven by chorus and orchestra. On average, students who enrolled in chorus, t(25984.00) = 4.54, p < .001, and orchestra, t(623.48) = 6.63, p < .001, had fewer behavioral concerns at age 4 compared to students not enrolled, respectively. No relations were found between behavior concerns at age 4 and enrollment in middle school band, t(25984.01) = 0.56, p = .58, or guitar, t(25984) = 1.75, p = .08.

Fifth-grade academic competence

Students enrolled in band, chorus, guitar, and orchestra, in all cases, had higher fifth-grade academic performance regardless of the performance indicator. Music students had higher fifth-grade GPA: band, t(28952.01) = −4.07, p < .001; chorus, t(28952.00) = −8.31, p < .001; guitar, t(28952.01) = −10.84, p < .001; orchestra, t(28952.01) = −14.08, p < .001); as well as math test scores: band, t(4876.26) = −15.30, p < .001; chorus, t(1317.85) = −7.38, p < .001; guitar, t(28571.01) = −17.17, p < .001; orchestra, t(702.97) = −12.84, p < .001; and reading test scores: band, t(4871.73) = −15.59, p < .001; chorus, t(1316.19) = −9.98, p < .001; guitar, t(28614.01) = −17.84, p < .001; orchestra, t(702.31) = −13.91, p < .001, compared to students not enrolled in those music types, respectively.

Multivariate Analyses

We just described the unadjusted bivariate associations found for each type of music elective. We now describe the multivariate results for each type of music elective and discuss how the findings relate to the pattern reported for overall music enrollment. Supplemental Tables S3, S5, S7, and S9 (included with the online version of this article) show the results of the multivariate logistic regression models for band, chorus, guitar, and orchestra, respectively. We discuss only the results of Step 2 (when all predictors were entered) but report the few cases where an effect of a demographic variable changed from Step 1 to Step 2 when elementary school competence was added.

Race-ethnicity

The decrease in odds for Black compared to White student enrollment found for overall music enrollment appears to be driven specifically by guitar and orchestra. Black students had 37% fewer odds of enrolling in guitar and 52% fewer odds of enrolling in orchestra compared to White students. Similarly, the decrease in Black student enrollment compared to Hispanic students found for overall music enrollment was also driven by guitar and orchestra. Black students had 40% fewer odds of enrolling in guitar and 40% fewer odds of enrolling in orchestra compared to Hispanic students. It is worth noting that Black student underrepresentation, specifically in guitar classes, reduced to marginal statistical significance when we reran the analyses limiting the sample to just those who were at schools known to offer music, which suggests that when given the opportunity, Black students sign up for guitar just as much as other groups. Black student underrepresentation in orchestra, however, was always present regardless of alternative model specification. There were no significant race-ethnicity differences found for band or chorus participation. A new and strong race-ethnicity effect was found for orchestra: Asian/Pacific Islander students had almost 3 times the odds of enrolling in orchestra compared to White students.

Gender

The increased odds of enrollment in music classes overall found for males compared to females was largely driven by band and guitar. Males had 74% greater odds of band enrollment and 71% greater odds of guitar enrollment compared to females. However, the gender effect was reversed for choir enrollment. Males had 66% fewer odds of enrolling in choir compared to females. No significant gender difference was found for orchestra enrollment.

Poverty status

The effect of FRL status changed markedly depending on the form of music examined and, in some cases, whether fifth-grade academic competence was controlled for. For band, FRL status was actually positively associated with enrollment: Those in poverty had 74% greater odds of taking band than those not in poverty. For chorus, although students in poverty initially had 23% less odds of enrolling, when fifth-grade academic performance was included, the poverty effect disappeared, meaning that school performance is more important than poverty status in predicting choir enrollment. A similar pattern was found for guitar: Those in poverty were less likely to enroll in guitar, but after prior elementary school performance was added to the model, the effect was no longer significant. The largest effect of income was found for orchestra. Controlling for all other variables, those receiving FRL at school were almost half as likely to be in orchestra compared to those not technically in poverty.

Disability status

The decrease in odds of overall music enrollment for students in special education was driven by guitar and band. Students with disabilities had 32% fewer odds of enrolling in band and 45% fewer odds of enrolling in guitar. No differences in enrollment in chorus or orchestra were observed for students with and without disabilities, after controlling for all other variables, including academic performance.

ELL status and english proficiency

Although no relationship between ELL status and overall music enrollment was observed, ELL status was significant for both chorus and guitar enrollment. ELL students had 22% fewer odds of being in chorus but 31% greater odds of being in guitar, compared to native speakers of English. ELL status was unrelated to band or orchestra enrollment. Recall that English proficiency was related to greater odds of overall music enrollment, as reported earlier. We found that this was driven mostly by band and chorus. English proficiency doubled the odds of band enrollment and more than tripled the odds of chorus enrollment. English proficiency was unrelated to both guitar and orchestra enrollment.

School readiness

For band, the only school readiness indicator related to enrollment after all other variables were included was cognitive skills at age 4: For each 1-percentile-point increase in cognitive skills, the odds of being in band 7 years later increased by .002. For chorus, none of the age 4 readiness scores were significantly related to enrollment in the final model. For guitar, gross motor skills were related to enrollment: Those with greater gross motor skills at age 4 were more likely to enroll in guitar in middle school while controlling for other factors. Similarly, cognitive skills at school entry were positively related to taking guitar in middle school. Interestingly, behavior problems at age 4 were associated with increased chances of being a guitar student later in middle school. For orchestra, gross motor skills and social skills were negatively related to later enrollment: Those with poorer social skills and weaker gross motor skills at school entry were actually more likely to sign up for orchestra later in middle school with all factors considered. Finally, behavior problems at age 4 were negatively associated with taking orchestra in middle school.

Elementary school academic performance

As seen in the online supplemental tables, math and reading test scores in fifth grade were consistently and strongly related to enrollment in all four types of music classes (band, chorus, guitar, and orchestra) in middle school. Those with better GPAs in fifth grade were also more likely to sign up particularly for orchestra and guitar. However, teacher-assigned grades in fifth grade were unrelated to enrollment in chorus and band with all other predictor variables included.

Discussion

Previous research shows that musical experiences provide multiple benefits (Brown & Sax, 2013; Moreno et al., 2011; Schellenberg, 2004; Williams et al., 2014; Winsler et al., 2011). Arts engagement is related to improved academic performance and socioemotional skills and appears especially helpful for children in poverty and ELLs (Catterall et al., 2012; Eisner, 1998). Yet schools in low-income communities are less likely to provide music electives than schools in richer communities (Parsad et al., 2012). The research on the effects of music participation is seriously limited because it tends to be quasi-experimental and correlational in nature rather than experimental, and often there is not sufficient control over selection effects—the ways that students who do and do not participate in music are initially different.

This study revealed many preexisting differences or selection effects that are present between students who are and are not involved in in-school middle school music elective courses in an ethnically diverse, large (N = 31,332) sample of children largely in poverty in Miami. Identifying preexisting differences between music students and those who do not take music is important in order to understand the effects music classes may have on students (Schellenberg, 2004). Authors of quasi-experimental research comparing existing music exposure groups must carefully control for selection effects when attempting to make causal inferences that music has positive “effects” on children. The first step, however, is to identify the many selection effects present. Previous research shows that children who receive musical experiences outside of school have parents with higher IQ, more education, more money, and more time to spend with their children (Foster & Jenkins, 2017). Here we examined many different selection effects associated with in-school music elective courses in public middle schools.

In our sample of ethnically and linguistically diverse students mostly in poverty, almost a quarter (22%) enrolled in music-related elective courses in middle school (Grades 6–8). This is less music participation than average for the southern U.S. region, which is 39% (Bell, 2014). This discrepancy is likely due to the low-income and urban nature of our sample, compared to the regional sample that includes students from all backgrounds. In order to provide students in poverty greater access to musical experiences, school systems need to focus resources on music programs, and educators must develop methods that target low-SES children and encourage their participation in musical experiences. Perhaps students in poverty are not encouraged to take music electives because music may not be seen as a “lucrative” skill or activity. Similar to other communities in the United States, practically all the Miami schools studied here offered music classes in sixth through eighth grades.

Importantly, we find that students who enroll in elective music classes in middle school are broadly and academically more competent than students who do not enroll in music classes as early as 7 years before they reach middle school. At age 4, students who later chose to take a music elective in middle school typically had greater motor, cognitive, language, social, and behavioral skills than students who did not take a music elective in middle school. Previous studies have shown that people who take band in middle school do better in school academically (Kinney, 2008), but we found that it is the students who are already doing better academically in elementary school who are later choosing to take music in middle school. Indeed, the strongest predictors of taking music classes in middle school in our study were the prior academic performance measures. It is possible that children with greater natural abilities self-select into music classes, where they can be challenged and their potential can be fostered (Elpus, 2013; Foster & Jenkins, 2017). Authors of previous studies sometimes have reported that after controlling for systematic differences between music students and nonmusic students, there are no associations left between music participation and improved cognitive outcomes (Elpus, 2013). Clearly, researchers trying to conclude that musical experience in middle school promotes later school outcomes need to understand and control for these preexisting selection differences between students who do and do not get exposure to middle school musical experiences. Another real possibility is that students with lower GPAs and/or those who score lower on standardized math and reading tests are made to take remedial classes, and these may take the place of elective courses, such as music.

It is notable that Black students in our sample had the lowest participation rate in music overall and across all music electives. After controlling for other variables, including poverty status and prior academic performance, Black students were still less likely than White and Hispanic students to take music electives, both overall and in the specific areas of guitar and orchestra. However, when our analyses were limited to the 95% of students who were known to attend a middle school that offered music electives, the finding of Black student underrepresentation in overall music and guitar reduced to marginal statistical significance. This suggests that some of the race-ethnicity effects involving Black students having less music participation can be partially explained by Black students being more likely to attend a school that does not offer music. However, even when available, Black students are less likely to enroll in orchestra compared to other racial-ethnic groups. Authors of other research examining ethnic differences in middle school arts participation more broadly, including music, visual art, drama, and dance, similarly find that Black students not only attend schools with fewer arts opportunities but that even when available, they are also less likely to select into arts opportunities (Winsler et al., 2019). Researchers need to investigate further why, even when given the opportunity, Black students are not choosing to enroll in school music and the arts compared to other groups.

Another robust ethnicity effect was observed for orchestra. Asian/Pacific Islander students were considerably more likely to take orchestra compared to other groups. The greater likelihood of orchestra enrollment among Asian/Pacific Islander students could be due to cultural values of discipline and sustained work within a focused domain. Concentration and commitment required to master a musical instrument are strongly valued in East Asian cultures (Morris & Leung, 2010; Yang, 2007). Additionally, artistic disciplines are used to support character formation within many Asian cultures (Lowry & Wolf, 1988). These values appear to be influential for immigrant students whose parents may encourage music enrollment to support exposure to mainstream cultural experiences (Yang, 2007).

Gender differences were also found, but they varied considerably by music type. Overall, males were slightly more likely to enroll in music electives. This was driven by band and guitar, where males were indeed more likely to be enrolled than females. Females, however, were more likely to enroll in chorus, whereas there was no gender difference for orchestra. With no bias for male or female enrollment, orchestra appears to be the most gender-neutral music elective. Unlike chorus, there is no social stigma against male participation in orchestra, nor is there widely accepted social support for female participation (Bennetts, 2013; Warnock, 2009). Regardless of gender, students appear to find a sense of belonging in orchestra.

Poverty effects were still often present for access to music courses even though our sample was predominantly low income, with 81% qualifying for FRL. The effect of poverty status on music enrollment differed by music type. Overall, students who qualified for FRL were less likely to take a music elective, specifically chorus, guitar, and orchestra, at least in bivariate relations. However, those in poverty were actually slightly more likely to enroll in band compared to students above the poverty threshold. Those in poverty were less likely to be in orchestra. Poverty thresholds do not appear to keep students from participating in band as they do for orchestra, where buying and renting instruments is more expensive. Schools may also give band more financial support (i.e., subsidized rental rates for instruments) than other music classes, which helps support student participation for low-income families (Albert, 2006; Kinney, 2008). Other research clearly shows that SES is positively related to childhood musical experiences (Elpus, 2013; Foster & Jenkins, 2017), but such research typically includes the full spectrum of SES, including much more wealthy participants, and has examined other musical experiences that happen outside of school. With our mostly low-income sample attending high-poverty, urban public schools in Miami, the negative effects of crossing federal poverty thresholds on public school music participation are reduced.

We are the first, to our knowledge, to examine disability status as a predictor of music participation in school. Importantly, students with disabilities were consistently less likely to take a music elective than typically developing students overall and for each type of music class. After including covariates, we found that differences in enrollment for students with and without disabilities went away for chorus and orchestra, which suggests that these two forms of music participation do not show access issues for children with disabilities. Disability status was still related to lower enrollment in band and guitar, however. Students with disabilities can benefit greatly from taking music classes in school, especially those in poverty (Nagel & Silverman, 2017). Perhaps additional training and support for band and guitar educators on best practices for recruiting and maintaining musical engagement for students with disabilities are needed.

Finally, we found relationships for ELL status and English proficiency for music enrollment in middle school. Overall, kindergarten ELL status was positively related to music enrollment, with ELL students being more likely to participate (especially true for guitar) compared to non-ELLs, but music participation was also strongly related to English proficiency in fifth grade. Those who were proficient in English (including native speakers and former ELLs) were more likely to take music in all of its forms. ELLs who were still working on their English proficiency were unlikely to be in music classes. A different pattern was observed for chorus, however: ELL students were less likely to select into the choral arts. Sufficient proficiency in English and having English as one’s native language appear to be important for middle school student choices to participate in choirs, where singing (mostly in English) is key.

It is worth noting that our longitudinal design and data analysis strategy involving the introduction of various predictors at several steps allowed us to observe not only unadjusted selection factors in a univariate fashion, which is common in other studies, but also whether such demographic differences disappeared and were better explained by other variables, like school readiness or elementary school achievement when included in the model. Several selection effects changed when other variables were included in the models, which highlights the importance of controlling numerous selection variables to explore what goes on beneath the bivariate correlational surface. For example, although higher performance on each of the school readiness measures was bivariately linked to overall music participation, when demographic variables, such as gender and poverty, were included, only cognitive and fine motor skills remained significantly related to later music enrollment. Similarly, poverty and ELL status were no longer related to taking music after controlling for prior academic performance. This suggests that how well students do in school is more important for predicting music course enrollment. Additionally, after controlling for prior academic performance, special education status was no longer related to chorus or orchestra enrollment.

Conclusions and Implications

We carefully examined important preexisting selection factors that differentiate those who do and do not take elective music classes in middle school in the unique, ethnically diverse community of Miami, Florida. We found many large differences between students who do and do not get exposure to music in middle school. Students who take music in sixth through eighth grades were more advantaged in terms of poverty and disability status, had stronger skills in a variety of different areas 7 years earlier, and were already doing much better in school before they got to middle school and had a chance to sign up for music electives. Understanding such selection effects is critical for scholarship exploring whether music has ancillary positive effects on other domains of child development and academic performance with quasi-experimental designs.

Our work contributes to the literature in many ways. Our long-term prospective, longitudinal design and very large sample are improvements compared to much of the existing research. We used official school record data to determine exposure to in-school musical courses rather than relying on student or parent retrospective reports of musical participation in general. In addition to commonly observed background variables (SES, gender, and ethnicity), we identified many other selection factors importantly related to middle school music selection, including disability status, ELL status, English proficiency, initial school readiness at kindergarten entry, and prior academic performance in the form of GPA and math and reading test scores. Our large sample uniquely included students from diverse ethnic and linguistic backgrounds, including many Hispanic students, an understudied group. However, our unique context of Miami, Florida, where Hispanic students are the majority and White students are a relatively small minority, and our predominantly economically disadvantaged sample raise challenges for generalizing to other communities. This should be kept in mind when interpreting and applying the results of the present study. Finally, rather than examining only a multimusic aggregate as is commonly done, we were able to identify selection factors that are specific to band, orchestra, choir, and guitar participation in middle school.

The present study has implications for educational practice and music education. We found that only a small portion of students chose to take a music elective in middle school, and students of color, students with disabilities, and those not doing as well in school had less representation in music electives. These students are not getting that same exposure to music in public middle schools as other groups are, which unfairly denies them the potential positive effects of art involvement (National Endowment for the Arts, 2013). Efforts to increase music program funding and offerings in middle schools and to increase student participation, especially in the groups found to be underrepresented here, are recommended. Given that unique background differences related to participation in different types of music classes (e.g., band vs. choir), music educators in this community and others can use this information in their student recruitment efforts. Although poverty may prevent students from taking instrumental music classes, like guitar and orchestra, it does not appear to keep students from participating in band or chorus. It is possible that band and chorus classes provide students with additional support that encourages participation among students in poverty, and if so, such methods might be used for these other music forms, as well.

In terms of implications for research, our results show that researchers interested in the “effects” of music participation on children need to understand that there are large preexisting selection factors, present years before the participation occurs, and appropriately control for these in their studies. Importantly, students who enroll in music classes in middle school are already performing better academically than their nonmusic classmates before reaching middle school, so one cannot conclude any causal effects of music in a correlational study that observes higher academic performance among music students. Armed with the new knowledge about these important selection effects and how they differ by specific types of music, researchers should carefully control for these selection factors while examining later academic outcomes for students in middle school music classes.

Supplemental Material

Supplementary_Tables_3.2.20 – Supplemental material for Predictors of Taking Elective Music Courses in Middle School Among Low-SES, Ethnically Diverse Students in Miami

Supplemental material, Supplementary_Tables_3.2.20 for Predictors of Taking Elective Music Courses in Middle School Among Low-SES, Ethnically Diverse Students in Miami by Alenamie Alegrado and Adam Winsler in Journal of Research in Music Education

Footnotes

Acknowledgements

We thank all the participating children, families, schools, and agencies.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported in part by an award from the Research: Art Works program at the National Endowment for the Arts (Grant 15-3800-7015). The opinions expressed in this article are those of the authors and do not represent the views of the Office of Research and Analysis or the National Endowment for the Arts. This work was also supported early on by the Early Learning Coalition of Miami-Dade/Monroe and the Children’s Trust. The Children’s Trust is a dedicated source of revenue established by voter referendum to improve the lives of children and families in Miami-Dade County.

Supplemental Material

The supplemental tables are available in the online version of the article at

Author Biographies

Alenamie Alegrado is a doctoral student at George Mason University in applied developmental psychology. Her research interests include music education, student outcomes from music participation, and education policy as it relates to arts education.

Adam Winsler is a professor of applied developmental psychology at George Mason University. His research examines the academic trajectories (preK–12) of low-income children of color and English language learners, bilingualism, executive functioning, private speech, and the effects of arts education (music, dance, drama, visual art) on students.

References

Abella

Urrita

Schneiderman

(2005). An examination of the validity of English language achievement test scores in an English language learner population. Bilingual Research Journal, 29(1), 127–144. http://doi.org/10.1080/15235882.2005.10162827

Abril

Gault

(2006), The state of music in the elementary school: The principal’s perspective. Journal of Research in Music Education, 54(1), 6–20. http://doi.org/10.1177/002242940605400102

Albert

(2006). Socioeconomic status and instrumental music: What does the research say about the relationship and its implications? Journal of Research in Music Education, 25(1), 39–45. http://doi.org/10.1177/87551233060250010105

Bell

. (May). Arts education in the South Phase I: Public school data and principals’ perspective. South Arts. http://programs.ccsso.org/link/aep/SouthArtsReport2014ALFLGAKYLAMSNCSCTN.p

Bennetts

(2013). Boys’ music? School context and middle-school boys’ musical choices. Music Education Research, 15(2), 214–230. http://doi.org/10.1080/14613808.2012.759550

Bialystok

DePape

A.-M.

(2009). Musical expertise, bilingualism, and executive functioning. Journal of Experimental Psychology, 35(2), 565–574. http://doi.org/10.1037/a0012735

Brown

E. D.

Sax

K. L.

(2013). Arts enrichment and preschool emotions for low-income children at risk. Early Childhood Research Quarterly, 28(2), 337–346. https://doi.org/10.1016/j.ecresq.2012.08.002

Catterall

J. S.

(1998). Involvement in the arts and success in secondary school. Americans for the Arts Monographs, 1(9), 1–10.

Catterall

J. S.

Dumais

S. A.

Hampden-Thompson

(2012). The arts and achievement in at-risk youth: Findings from four longitudinal studies. National Endowment for the Arts. http://arts.gov/sites/default/files/Arts-At-Risk-Youth.pdf

10.

Corrigall

Schellenberg

Misura

(2013). Music training, cognition, and personality. Frontiers in Psychology, 4(222). https://doi.org/10.3389/fpsyg.2013.00222

11.

Crane

Mincic

Winsler

(2011). Parent-teacher agreement and reliability on the Devereux Early Childhood Assessment (DECA) in English and Spanish for ethnically diverse children in poverty. Early Education and Development, 22(3), 520–547. https://doi.org/10.1080/10409289.2011.565722

12.

Dos Santos-Luiz

Monico

Almeida

Coimbra

. (2016). Exploring the long-term associations between adolescents’ music training and academic achievement. Musicae Scientiae, 20(4), 512–527. https://doi.org/10.1177/1029864915623613

13.

Eisner

E. W.

(2002). The arts and the creation of mind. Yale University Press.

14.

Elpus

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

15.

Foran

L. M.

(2009). Listening to music: Helping children regulate their emotions and improve learning in the classroom. Educational Horizons, 88(1), 51–58.

16.

Foster

Jenkins

(2017). Does participation in music and performing arts influence child development? American Educational Research Journal, 54(3), 399–443. https://doi.org/10.3102/0002831217701830

17.

Gardiner

M. F.

Fox

Knowles

Jeffrey

(1996). Learning improved by arts training. Nature, 381(6580), 284. https://doi.org/10.1038/381284a0

18.

Gerry

Unrau

Trainor

(2012). Active music classes in infancy enhance musical, communicative and social development. Developmental Science, 15(3), 398–407. https://doi.org/10.1111/j.1467-7687.2012.01142.x

19.

Human Resources Research Organization & Harcourt Assessment. (2007). Technical report for 2006 FCAT test administrations: Reading and mathematics. Harcourt.

20.

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. https://doi.org/10.1177/0022429409350086

21.

Kinney

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

22.

Holochwost

S. J.

Propper

C. B.

Wolf

D. P.

Willoughby

M. T.

Fisher

K. R.

Kolacz

Jaffee

S. R.

(2017). Music education, academic achievement, and executive functions. Psychology of Aesthetics, Creativity, and the Arts, 11(2), 147–166. https://doi.org/10.1037/aca0000112

23.

LeBuffe

Naglieri

J. A.

(1999). The Devereux Early Childhood Assessment. Kaplan Press.

24.

Lowry

Wolf

(1988). Arts education in the People’s Republic of China: Results of interviews with Chinese musicians and visual artists. Journal of Aesthetic Education, 22(1), 89–98. https://doi.org/10.2307/3332967

25.

McNeal

R. B.

Jr. (1998). High school extracurricular activities: Closed structures and stratifying patterns of participation. Journal of Education Research, 91(3), 183–191. https://doi.org/10.1080/00220679809597539

26.

Miksza

(2010). Investigating relationships between participation in high school music ensembles and extra-musical outcomes: An analysis of the education longitudinal study of 2002 using a bioecological development model. Bulletin of the Council for Research in Music Education, 186(Fall 2010), 7–25.

27.

Moreno

Bialystok

Barac

Schellenberg

E. G.

Cepeda

N. J.

Chau

(2011). Short-term music training enhances verbal intelligence and executive function. Psychological Science, 22(11), 1425–1433. https://doi.org/10.1177/0956797611416999

28.

Morris

M. W.

Leung

(2010). Creativity east and west: Perspectives and parallels. Management and Organization Review, 6(3), 313–327. https://doi.org/10.1111/j.1740-8784.2010.00193.x

29.

Morrissey

Hutchison

Winsler

(2014). Family poverty, school attendance, and academic achievement in early elementary school. Developmental Psychology, 50(3), 741–753. https://doi.org/10.1037/a0033848

30.

Nagel

J. J.

Silverman

M. J.

(2017). Experiences and perspectives of music therapists providing services to families experiencing poverty: A qualitative investigation. Voices: A World Forum for Music Therapy, 17(2). https://doi.org/10.15845/voices.v17i2.914

31.

National Endowment for the Arts. (2013). How a nation engages with art: Highlights from the 2012 survey of public participation in the arts (Research Report No. 57).

32.

Nehring

A. D.

Nehring

E. F.

Bruni

J. R.

Randolph

P. L.

(1992). Learning Accomplishment Profile Diagnostic (LAPD). Kaplan Press.

33.

Pallesen

K. J.

Brattico

Bailey

C. J.

Korvenoja

Koivisto

Gjedde

Carlson

(2010). Cognitive control in auditory working memory is enhanced in musicians. PLOS ONE, 5(6), 1–12. https://doi.org/10.1371/journal.pone.0011120

34.

Parsad

Spiegelman

Coopersmith

(2012). Arts education in public elementary and secondary schools: 1999-2000 and 2009-10 (NCES 2012-014). National Center for Education Statistics, Institute of Education Sciences, U.S. Department of Education.

35.

Rosevear

J. C.

(2010). Attributions for success: Exploring the potential impact on music learning in high school. Australian Journal of Music Education, 1, 17–24.

36.

Schellenberg

(2004). Long-term positive associations between music lessons and IQ. Journal of Educational Psychology, 98(2), 457–468. https://doi.org/10.1037/0022-0663.98.2.457

37.

Southgate

Roscigno

(2009). The impact of music on childhood and adolescent achievement. Social Science Quarterly, 90(1), 4–21. https://doi.org/10.1111/j.1540-6237.2009.00598.x

38.

Stoelinga

S. R.

Joyce

Silk

(2013). Turnaround Arts Initiative: Progress report 2013. President’s Committee on the Arts and the Humanities. http://www.pcah.gov/sites/default/files/Turnaround%20Arts%20Interim%20Evaluation_November%202013_1.pdf

39.

Warnock

(2009). Gender and attraction: Predicting middle school performance ensemble participation. Contributions to Music Education, 36(2), 59–78.

40.

Williams

Barrett

Welch

Abad

Broughton

(2014). Associations between early shared musical activities in the home and later child outcomes: Findings from the longitudinal study of Australian children. Early Childhood Research Quarterly, 88(2), 113–124. https://doi.org/10.1016/j.ecresq.2015.01.004

41.

Winsler

Ducenne

Koury

(2011). Singing one’s way to self-regulation: The role of early music and movement curricula and private speech. Early Education & Development, 22(2), 274–304. https://doi.org/10.1080/10409280903585739

42.

Winsler

Gara

Alegrado

Castro

Tavassolie

(2019). Selection into, and academic benefits from, arts-related courses in middle school among low-income, ethnically diverse youth. Psychology of Aesthetics, Creativity, and the Arts. Advance online publication. https://doi.org/10.1037/aca0000222

43.

Winsler

Tran

Hartman

S. C.

Madigan

A. L.

Manfra

Bleiker

(2008). School readiness gains made by ethnically diverse children in poverty attending center-based childcare and public school pre-kindergarten programs. Early Childhood Research Quarterly, 23(3), 314–329. https://doi.org/10.1016/j.ecresq.2008.02.003

44.

Yang

(2007). East meets West in the concert hall: Asians and classical music in the century of imperialism, post-colonialism, and multiculturalism. Asian Music, 38(1) 1–30. https://doi.org/10.1353/amu.2007.0025

45.

Zuk

Benhamin

Kenyon

Gaab

(2014). Behavioral and neural correlates of executive functioning in musicians and non-musicians. PLOS ONE, 9(6), e99868. https://doi.org/10.1371/journal.pone.0099868

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.56 MB