Does learning (to play) music increase IQ?

Question

In one fairly cited study by E. Glenn Schellenberg (~800 citations in Google Scholar) we find that

Compared with children in the control groups, children in the music groups exhibited greater increases in full-scale IQ. The effect was relatively small, but it generalized across IQ subtests, index scores, and a standardized measure of academic achievement.

The exact effect size computed by the authors was 0.35, and the raw increase in IQ score is summarized in the authors' graph below:

There is a paper by Kenneth M. Steele challenging the study on methodological grounds. This criticism received a reply from Schellenberg which in turn was answered by Steele.

So, has this finding, i.e., that music education affects IQ, been replicated in other studies?

Answer 1

Short answer
The recent literature shows mounting evidence for beneficial effects of music on cognitive abilities. The big 'but' in the issue is how specific those effects are and whether they will hold up in longitudinal studies.

Background
Great question.

Your citations provided reach back to 2006 so I thought to limit this answer using recent articles only. Overall, the beneficial effects of musical training seems quite overwhelming.

A number of papers, including reviews have shown that in kids learning to play a musical instrument

• Raises their IQ (Protzko, 2017)
• Enhances musically relevant auditory skills (Habibi et al., 2018
• Enhances working memory (but not other outcome measures) (Guo et al., 2018)

In the general adult population review reports conclude that musical training

• Enhances cognitive skills spanning executive control to creativity, albeit sometimes limited to the auditory domain (Benz et al., 2016)
• Is correlated with positive outcomes on cognitive ability (Schneider et al., 2018)

Lastly, in the elderly musical training has been shown to

• Better attention, episodic memory and executive functions (Mansens et al., 2016)

However, virtually all review studies conclude that more data is needed, especially from carefully controlled longitudinal studies. Many papers are retrospectively organized, and many studies find benefits on some outcome measures, but not all. Indeed, Hille et al. (2011) say:

It is no longer the question whether or not musical training is associated with [better] cognitive abilities, because there is growing evidence that it is. An unresolved issue however, is the nature and specificity of the link.

Indeed, many primary research articles focus e.g. on auditory tasks, and not on others.

<sub>References
- Benz et al, Front Psychol (2016)
- Guo et al., Front Psychol (2018)
- Habibi et al., Annals NY Acad Sci (2018)
- Hille et al. Adv Cogn Psychol (2011); 7: 1–6
- Mensens et al., Aging & Mental Health
- Protzko, Developmental Rev (2017); 46: 81-101
- Schneider et al., J Appl Gerontology (2018)</sub>

Answer 2

I found one 2017 meta-analysis by Sala and Gobet which is probably superior methodologically to the few non-systematic reviews found by AliceD. I say "probably" because the meta-analysis includes quite a few not-directly-IQ transfers, e.g. to math. Nevertheless, they do describe IQ transfers separately... and overall Cohen's d they found for that is basically equal to Schellenberg's... 0.35.

Of note that among the studies on music training and IQ (besides Schellenberg 2004) there's:

• one successful replication in Iran: H. Kaviani, H. Mirbaha, M. Pournaseh, O. Sagan Can music lessons increase the performance of preschool children in IQ tests? Cognitive Processing, 15 (2014), pp. 77-84

• one successful replication in Israel (for disadvantaged children). A. Portowitz, O. Lichtenstein, L. Egorova, E. Brand, "Underlying mechanisms linking music education and cognitive modifiability", Research Studies in Music Education, 31 (2009), pp. 107-128,

Neither of these had an active control group (unlike Schellenberg 2004).

• one failed replication in Australia: N.S. Rickard, C.J. Bambrick, A. Gill "Absence of widespread psychosocial and cognitive effects of school-based music instruction in 10-13-year-old students", International Journal of Music Education, 30 (2012), pp. 57-78.

The latter did have active controls.

There are a few other papers (mostly pre-2004) subtests of which were used in the meta-analysis in the intelligence section, but the above two were the only more-or-less full-scale replication attempts I could find.

On the issue of publication bias, the meta-analysis did not find one, but this is for the whole shebang, there was no separate test for each dependent variable (like just for IQ).

The authors of this meta-analysis do note (in sections 4.1-4.2) that the observed IQ improvements failed to translate into academic achievements in other areas... with math being the closest thing were something was observed (Cohen d = 0.17). Their conclusion is less optimistic than the interpretation others have given (pretty much to the same numbers/studies):

The results of this meta-analysis fail to support the hypothesis that music skill transfers to cognitive or academic skills in the general population of children and young adolescents. Together with previous findings in psychology and education, these results suggest a sobering conclusion: when the potential occurrence of far-transfer is tested rigorously, the results are often, if not always, disappointing.

Eye of the beholder.

Answer 3

Just a few comments. The sample size in that study was quite small (n = 144) and there were four conditions.

A few fundamental principles of psychology suggest that learning an instrument would not increase IQ in any meaningful way.

• Within a given society, IQ is heavily influence by genetics, and generally not readily attributable to individual environmental interventions. For example, almost no variance is attributed to shared family environment.
• Broad fundamental traits like intelligence require massive interventions to change. Simple interventions to improve IQ have rarely proved successful, although some evidence suggests that the difference between those in impoverished versus enriched adopted environments and the difference between people who did and did not miss out on a year of education might be in the ball park of 3-5 IQ units, but even there some people argue that the effects aren't there or are not lasting. The point is that even massive interventions as dramatic as fundamentally changing a child's academic environment only has small effects on IQ. For example, the Flynn effect shows that in many populations IQ has been increasing across generations. The causes of this are somewhat unknown, but we might say that for example a 3 point IQ increase over 25 years might be result of a range of fundamental changes related to nutrition, medical care, pregnancy habits, parenting principles, and educational practices. This all suggests that simple interventions like learning an instrument would not be enough to change IQ, or if there was a change, it would be tiny (e.g., like 0.5 IQ units), and certainly not in the 3+ units reported in the study.
• Transfer from skill acquisition is narrow. Presumably, practicing an instrument would teach a lot about how to play that instrument, and would yield other generalisable skills around pitch perception, rhythm, reading music, and so on.