Claims that music study makes you smart remain unconvincing

I take the latest claims that music study gives students an academic boost with the same pinch of salt with which I have accompanied previous claims.
A new study, published in the American Psychological Association's Journal of Educational Psychology this week, looks at public school students in British Columbia, and concludes that those kids who took musical instrument classes for several years had higher grades in math, science and English by the time of Grades 10-12.
I don't doubt that that was in fact the case. What I take issue with is the conclusions drawn from it. The study suggests a causal relationship that I find difficult to justify. One needs only look back to an earlier study by University of Toronto professor Glenn Schellenberg in 2004, which is often cited as proof of the same causal relationship. However, since that study was published, Prof. Schellenberg has renounced his conclusions, arguing that it is much more likely that pre-existing differences in the students' personalities and abilities (not to mention their parental support/pressure) explain why music students and non-music students perform differently in school. Basically, bright and engaged students are more likely to engage in music studies; it is not the music that is somehow, miraculously, making them smart, a notion that Prof. Schellenberg calls "just ridiculous".
The newer study claims to make some attempt to control for variables like gender, ethnicity, socio-economic status and prior academic achievement, and various convincing-sounding hypotheses for the mechanisms by which music improves academic performance  have been out forward, but the conclusions are still widely disputed, and the hypotheses are just that, hypotheses. Me, I remain skeptical.