Guhn, M., S.D. Emerson and P. Gouzouasis. (2019). A Population-Level Analysis of Associations Between School Music Participation and Academic Achievement. Journal of Educational Psychology. Advance online publication: http://dx.doi.org/10.1037/edu0000376.

Abstract:

The present study employed population-level educational records from four public school student cohorts (n = 112,916; seventh through 12th grade 7-12) in British Columbia, Canada, to examine relationships between music education (any participation, type of participation, music achievement and engagement level) and mathematics and science achievement in 10th grade, as well as English achievement in 10th and 12th grade, while controlling for language/cultural background, seventh grade academic achievement and neighborhood socio-economic status. Effect sizes concerning scores among those with the highest level of instrumental music engagement compared to those without music were of a similar equivalent to, or surpassing, those observed for average annual gains for reading during the high school years. Findings suggested that multi-year engagement in music, especially instrumental music, may benefit high school academic achievement. Findings and implications are discussed within the broader inter-disciplinary literature on music learning.

Key Findings:

Researchers calculated effect sizes to demonstrate the differences in academic achievement between music and non-music learners. In practical terms, the effect sizes observed when comparing those in the very highly-engaged instrumental music group with the no music group (even after adjustment for numerous confounders) were greater than the average annual gains in reading, science and math that are seen during the high school years in the U.S. context. Students in the study that were highly engaged in instrumental music were, on average, more than one year ahead in their math, English and science skills, compared to those peers not engaged in school music.

Music participation was related to higher scores on all four subjects and these relationships were stronger for instrumental music than vocal music (Cohen’s d range: .28 to .44 [small-medium effect sizes] and .05 to .13 [null-small effect sizes]). School music achievement positively related to scores on all subjects; such relationships were stronger for achievement in instrumental music compared to vocal music. Higher levels of music engagement (number of courses) was related to higher exam scores on all subjects; this pattern was more pronounced for very high engagement in instrumental music (d range: .37 to .55; medium effect sizes) compared to vocal music (d range: .11 to .26; small effect sizes).

Researchers were surprised to see (1) such consistency across diverse domains (English, science, mathematics); (2) such substantive effect sizes even after controlling for important individual differences (i.e., gender, ethnicity, socioeconomic background and prior achievement in seventh grade numeracy and literacy); (3) evidence of higher engagement in music and higher academic achievement based on exposure; and (4) a clear trend showing more pronounced results with instrumental music (relative to vocal music).

Significance of the Findings:

This study extended beyond simple correlation analyses. It employs separate regression analyses to look at whether students’ (1) mean vocal school music grade or (2) mean instrumental music grade predicted academic achievement in 10th grade math, science and English; and 12th grade English, adjusting for previous academic achievement and socio-cultural confounders. The regression analyses reveal one-way, predictive relationships; In this study, music achievement (grades) predicts mathematics achievement (grades) in 10th grade, and not the other way around. (In this case, high achievement in math does not predict high achievement in music. The researchers also controlled for prior math learning using standardized seventh grade math grades.)

For vocal music, each one-unit increase in overall mean music vocal grade was associated with predicted increases in exam means in all four subjects (Math 10, Science 10, English 10, English 12), ranging from +0.19 (English 10) to + 0.38 (Math 10). Relationships between music grades and academic exam scores were significantly higher (i.e., had a larger coefficients) for overall mean instrumental music grades, with coefficients ranging from +0.26 (English 10) to +0.50 (Math 10).

Apart from the strength of the associations and how findings show that music achievement and engagement may predict academic outcomes, findings also remained consistent across subjects. It is possible that some skills learned in band, orchestra and conservatory music lessons transfer very broadly to adolescents’ learning in school.

Methodology:

The researchers examined school records for all students in British Columbia who started the first grade between 2000 and 2003; completed the last three years of high school; had completed at least one standardized exam for math, science or English (10th or 12th grade); and for whom they had appropriate demographic information (e.g., gender, ethnicity, neighborhood socioeconomic status).

Of the more than 112,000 student records studied, 13% of the students had participated in at least one music course in 10-12th grade. Qualifying music courses included concert band, conservatory piano, orchestra, jazz band, concert choir and vocal jazz. General music or guitar courses did not qualify as they required no previous music experience and, in the case of general music, did not require music-making or practice.

The study extended previous research in this area by using multi-year population-level data from British Columbia, Canada, to examine associations between public secondary school students’ sustained engagement in school music classes and their secondary school exam grades in math, science and English, while also taking into account demographic and socioeconomic background characteristics and controlling for elementary school achievement in numeracy and reading.

The researchers examined five research problems:

  1. School music participation versus no school music participation and academic achievement.
  2. Type of music engagement (i.e., instrumental vs. vocal ensemble music classes) and academic achievement.
  3. Grades in vocal music classes and instrumental music classes and academic achievement.
  4. Level of music engagement (cumulative number of courses) in vocal music and instrumental music and academic achievement.
  5. Associations between music education and academic grades by subject area.

Limitations of the Research:

The data set lacked information on parental educational achievement and family level socioeconomic status, so a proxy for family-level socioeconomic standing was employed by using neighborhood-level median household income and proportion of residents with post-secondary education. Aggregate SES data may mask within-group variability. Cultural background was approximated via language spoken at home (as listed on children’s educational files). Language can be an incomplete, marker of cultural background because children of diverse cultures may primarily speak English at home. As a result, misclassification likely occurred in some instances.

To control for children’s previous academic achievement, reading and numeracy scores on a standardized seventh grade assessment were used as covariates. Some children likely began music training prior to seventh grade and cognitive benefits of music training have indeed been documented (in other studies) among children prior to seventh grade. For those children, including these test scores in the analyses may have led to over controlling.

Data regarding music engagement were only available in terms of secondary school (10-12th grade) music course participation. Because of this, this study did not account for the participation and level of engagement in music outside of 10-12th grade courses. A proportion of students who took school music courses may have had additional out-of-school engagement in music, and some of the students who took no school music courses may have had formal exposure to music outside of the school context. Additionally, many studies in the related research literature, including this study, lack measures reflecting specific music teaching content or style. Various facets of what constitutes an outstanding music program and music environment were not controlled.

Omitted variables and observational nature of study. Given some conceptually important covariates were not included in the data set and the study represents an observational design, the findings cannot establish cause-effect relationships. Factors missing from the analyses may have confounded the relationship between school music and academic achievement.

The correlational nature of the study prevented testing specific cause-effect relationships, particularly regarding direction, temporal precedence and specific causal pathways relating music education to academic achievement. It should be noted, however, that the multiyear prior engagement in instrumental music does provide some index of temporal precedence, as most musical activity/training occurred prior to the English, math, and science exams taken in 10-12th grade.

Questions to Guide New Research:

Future studies may benefit from knowledge of participants’ level of music participation both within and outside of school to gain a more refined estimate of associations between music participation, music engagement and academic grades.

Additional measures of potential confounders in associations of music education with academic achievement would offer greater precision when estimating relationships. It may be useful to adjust analyses for variables such as household-level SES, parental involvement in students’ school life, students’ academic motivation and personal characteristics (e.g., conscientiousness).

Inclusion of factors theorized as pathways relating music education to academic achievement — such as, executive functioning, socioemotional functioning, audiation and brain imaging — would enable evaluation of how such factors may represent mechanisms relating music education to academic achievement.

In line with Bronfenbrenner’s bioecological model, and his propositions to examine the extent to which developmental effects of proximal processes vary as a function of their regularity, intensity and growing complexity, the analyses tested for some relevant interaction effects; for example, whether associations between music and academic achievement were moderated by the type and level of music engagement. A further extension of Bronfenbrenner’s propositions would be to not only control for previous academic achievement and sociodemographic factors (SES, gender, cultural background) — but also examine the extent of associations between the proximal process of music-making and developmental outcomes.