Elpus, K. (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.


Previous studies have found evidence of a link between high school music participation and higher standardized test scores. These studies, however, typically fail to account for the existence of pre-existing systematic differences between students who choose to take music courses in high school and those who do not. This study compared college entrance exam scores for high school students who participated in music classes and those who did not while accounting gender, socioeconomic status (SES), prior academic achievement, attitudes towards school, and receipt of special education services. The study found that music participation in itself did not appear to have a statistically significant effect on test scores. The researcher posited that there must be factors that affect a student’s choice to participate in music that have caused the higher test scores reported in previous studies.

Key Findings:

The researcher found that after taking covariates into account, participation in music in high school did not have any statistically significant effect on either students’ SAT scores or on other standardized test scores. Type or amount of music studied did not influence results. Factors that did predict higher SAT scores were high socioeconomic status and prior academic achievement.

Significance of the Findings:

These findings suggest that the positive influence of music education on test scores seen in previous studies might actually be the result of other pre-existing systemic differences between students who elect to study music in high school and those who do not. They provide an important reminder for advocates of arts education that, although music and other arts programs often appear to be positively correlated with measures of academic success, the intrinsic benefits of arts experiences should also be examined and trumpeted.


Multiple stages of statistical analysis were performed on data from the Educational Longitudinal Study of 2002 (ELS), an ongoing study of a nationally representative cohort of students who were sophomores in United States public and private high schools in 2002; data collected from a second wave of ELS surveys in 2004 were also utilized in the present study. From the ELS data sets, the researcher was able to observe students’ college entrance exam scores, SES, race and ethnicity, family composition, academic achievement as measured by an ELS-specific standardized test as well as grade point average (GPA), and a variety of self-reported time use and attitudinal measures related to student engagement. All of these factors are associated with standardized test scores and have been identified in previous research as pre-existing systematic differences between students who choose to study music and those who don’t. Music and non-music students were identified through a close examination of high school transcripts. The first stage of analysis compared SAT and ACT scores for music and non-music students. The second stage examined the amount of music studied, to see if more music participation might have had a different effect on test scores than less or none at all. The third stage looked at type of music studied by disaggregating instrumental music students from choral students. The next stage compared scores on an ELS-specific standardized math test in order to include a wider swath of students, including those who did not elect to take a college entrance exam. Finally, the researcher examined music as a quasi-experimental intervention, by analyzing test scores of students who only took music classes in between a pre- and post-test. All equations used for statistical analysis controlled for the presence of the previously identified factors impacting standardized test scores in order to obtain unbiased results.

Limitations of the Research:

This study considered music students to be only those who had taken music classes in high school; other students who might have participated in extracurricular or personal music activities were not accounted for in the analysis. The researcher also did not look at music participation prior to high school to see if early childhood music experiences might have influenced high school music participation or test scores.

Questions to Guide New Research:

Why does it appear to be the case that students who are predisposed to academic success tend to enroll in music classes more than their peers? What factors actually motivate a student to participate in music in high school? Can effects of music participation on subgroups of students identified in this study as less likely to score highly on standardized tests (students with low SES, students who have not previously been academically successful, and students with special needs) be isolated and studied further? Does music participation in younger grades have an effect on test scores? What other benefits not measured by standardized test scores might high school music participation offer to students?