Stephen Krashen writes:
According to a recent column in Science Daily, "A new study in the journal Social Science Quarterly reveals that music participation, defined as music lessons taken in or out of school and parents attending concerts with their children, has a positive effect on reading and mathematic achievement in early childhood and adolescence."
This kind of announcement deserves a close look at the actual data. As usual, the cheerful announcement of the benefits of music was not quite accurate, and some very important results were not mentioned.
Fortunately, the analysis was multivariate, which means that important factors such as socio-economic status were controlled. Unfortunately, it is difficult to quantity the impact of all the predictors, as many were simply coded as "present" or "absent" (score of 1 or zero), so my statements about strong and weak effects are somewhat imprecise.
Music lessons outside of school, it turns out, had no impact on math scores, and was actually negatively correlated with children's reading scores. It had a small positive effect on adolescents' reading scores.
Music courses taken between grades eight and ten had a small positive effect on adolescents test scores.
Music participation in school (at least once a week) had a modest effect on both reading and math for children, and a much weaker effect for adolescents in reading and was not significant for adolescents in reading.
Parents attending concerts had no effect on reading at all, no effect on adolescent math scores and a weak positive effect for children and math. It is also not clear from the paper what this variable means, whether it means attending concerts with or without their children, or concerts in which children are performing.
In other words, not all these predictors counted. More important, those that counted were not very strong.
The most dramatic case is adolescent reading: Adolescents who do music both inside and outside of school are predicted to score 1.32 points higher in reading. In contrast, the study also reports that having more than 50 books in the home, and higher socioeconomic status predicts a score of nearly seven points higher (6.97). Higher socio-economic status, as has been pointed out, means, among other things, more books available in the community and at school, as well as at home. A reasonable interpretation is that access to books is a much stronger factor that music.
I wonder if some people will conclude from the Science Daily summary that music classes are all we need: Since we have music, we don't need to worry about school library quality. There are plenty of good reasons to include music in the school curriculum, but ironically the article provides more evidence for supporting libraries than for music when it comes to reading as well as math.
Southgate, D. and Roseigno, V. 2009. The Impact of Music on Childhood and Adolescent Achievement. Social Science Quarterly 90, 1:4-21.
Adolescents Involved With Music Do Better In School. (2009). http://www.sciencedaily.com/releases/2009/02/090210110043.htm