Art for Art’s Sake | The Music Trust

Art for Art’s Sake?
The Impact of Arts Education
Educational Research and Innovation, OECD
Ellen Winner, Thalia R. Goldstein and Stéphan Vincent-Lancrin
This is a rough summary by Richard Letts of a very broad 270-page survey by the OECD. Readers are urged to go to the survey itself to consider the evidence and the full argument. The summary begins with an overview and then looks at sections dealing specifically with music education.

The survey can be read at

This summary looks firstly at the study as a whole and then at the sections dealing specifically with music.

This is a study by the OECD assessing a very large body of research exploring the benefits of arts education. While it supports the cause of arts education, it assesses the research on its merits. Research that shows no benefit gets treatment equal to that of research that is more helpful to our cause.

Most such reviews cover only studies published in English. “This new enquiry involves the systematic investigation of research databases in education and psychology in the following languages: Dutch, English, Finnish, French, German, Italian, Japanese, Korean, Portuguese, Spanish and Swedish.” It attempts to cover all empirical studies published at least since the 1980s, and makes fresh use of studies unearthed in former meta-analyses (from 1950 on).

Chapter 1 gives a great, easily comprehended summary of the research. It deals, importantly, with the issue of causality. Most studies show correlations and correlations unless they control variables very rigorously do not prove causality. Did the members of the experimental group show superior outcomes as compared with the members of the control group because of the visual arts education they received, or because they were not randomly selected or there was insufficient control of variables. Can there be some other explanation of their superior outcomes? – eg they have higher average IQ or they brought greater knowledge of art or they self-selected for the experimental group because of their aptitude for the artform….

The study distinguishes between research that shows correlations – the great majority of the studies because, as it says, in the messy school environment it is difficult to set up other than correlational studies – and research that is quasi-experimental or experimental and sets out to prove causality. The research is divided into a meta-analysis of studies completed by 1998 and coverage of studies done post-1998. Causality certainly is shown less often than correlation and the study takes the position that a case has not really been made unless causality is demonstrated.

It notes that the arts are peripheral in most school systems and that much of the research is motivated by the desire to have them more firmly and centrally established. So we try to show the instrumental value of the arts in terms that might appeal to decision makers. This leads us probably to look at the research outcomes as a glass half full. But anyone hostile to our position, reading this paper could point to many studies which do not show a positive connection between arts education and eg academic outcomes.
I thought one very interesting piece of information is shown in Figure 1.1: percentages of tertiary graduates from specific fields holding a highly innovative job. There are 12 fields from which grads come: the arts, and others like engineering and computing, sciences and maths, agriculture, law and so on. In rank order, arts grads were second in product innovation, fifth in technology innovation, seventh in knowledge or method innovation, fifth in any type of innovation. (Law is last in all categories!).

In the last chapter there is a strong argument made for promoting the importance of arts education on intrinsic grounds. We are informed that of course, arts education brings both depth of satisfaction in arts experience and jobs in the arts – but also, nearly half of arts graduates work outside the arts, including in product innovation etc.

There are chapters for research showing cognitive outcomes of education in each of our four artforms dance, music, theatre and visual arts, and for outcomes in creativity, motivation, social skills and brain development. The amount of research in music and visual arts education far exceeds that for theatre or dance and perhaps that suggests an agenda for the latter. There is also a section briefly addressing the arguments for the intrinsic value of arts education.

Why arts education? Summary and conclusions
This summary includes many direct quotations from the study and they are shown in italics.

One of the key messages of the OECD Innovation Strategy is to “empower people to innovate” (OECD, 2010). How does arts education contribute to this agenda? In this concluding chapter, we summarise the methodology and main findings of the report, propose an agenda for future research and explore some policy implications of our findings. We begin by setting the policy context and providing a brief overview of the skills needed in innovation-driven societies. We then summarise the main findings of our review of the impact of arts education. Next we propose an agenda for future research on arts education, followed by a policy agenda. A key argument we make is that the main contribution of arts education to innovation societies lies in its development of broad and important habits of mind. We conclude by arguing that the value of the arts for human experience is a sufficient reason to justify its presence in school curricula whether or not transfer results from arts education.

The final chapter offers a lot of valuable discussion.

The first section is titled “Skills and education for innovation”.

Obviously, the arts are about creativity and innovation. But arts education is not necessarily about innovation and this research anthology found very few studies that demonstrated that creativity acquired through arts education transfers to creativity in non-arts activity. So it is rather curious that the final chapter puts so much emphasis on arts education as a powerful source of broad innovation. The strongest evidence presented is the somewhat peripheral account of the percentages of tertiary graduates from specific fields holding a highly innovative job, mentioned above.

There is a useful discussion of the requirements for innovation. They include three sets of skills: technical, thinking and creativity, and behavioural and social. The study sees the issue as going beyond research into creativity per se. We have reviewed what is known about the possible impact of the main forms of arts education on the three categories of skills for innovation presented above. We examine verbal, mathematical and spatial skills; creativity; academic motivation; and social skills including self-confidence, empathy, perspective taking, emotion regulation. Neuroscientific literature relating to arts education was also examined.

The chapter then summarises the previous chapters on the effects of arts education on academic skills etc. This could be read in lieu of the detail of those chapters. This summary covers only the effects of music education, treated below.

There are useful observations about arts education and creativity: creativity will not result if the arts are taught in a routine, deadening way; and creativity may be very domain-specific and not spill over into other domains. There is a need for much more research.

…The report shows that learning certain forms of arts instruction does indeed have an impact on the development of very specific skills, as summarised above. The body of empirical research does not cover all skills of interest though – far from it. The kinds of learning that occur in particular art forms shape the kinds of skills that spill over into other areas. Thus, music learning involves auditory training, and music learning “spills over” into skill in speech perception;… theatre involves character analysis and spills over into skill in understanding the perspectives of others…

It should be noted that in a number of areas, the survey found that the effects of arts education had not been demonstrated, not because research had been inconclusive, but because there had not been any, or sufficient, research.

The amount of research discovered is small, considering the scope of the review and represents only a tiny share of educational research. The survey suggests some research priorities for the coming decade. One priority is to use and to develop better methodologies for impact studies. The second priority, even more important, is to develop sound and testable theories about why and how arts education would have an impact on various outcomes of interest.

Methodological: we need more experimental studies to show causation. Possibilities include correlational studies with rigorous control of “possible confounding variables”. BOX 11.1 has a great set of possibilities for research and it is strongly suggested that researchers review them.

Theoretical. Researchers need to build stronger theoretical frameworks on why and how arts education can be hypothesised to develop certain skills which then transfer to other academic subjects…. Any study of transfer should first analyse the kinds of habits of mind taught in the art domain and then develop a plausible hypothesis about the kinds of transfer outcomes one might expect. Thus it is not sufficient to test the hypothesis that infusing many kinds of arts into the academic curriculum will lead to higher test scores. What is needed is a theory of what infusing the arts will do to learning and why that kind of learning should be reflected in the kinds of test scores examined…. Logically, if there is to be transfer from arts learning to some non-arts kind of learning, there must first be arts learning. As Bransford and Schwartz (1999) point out, many educational studies reporting failure of transfer can be traced to limited learning in the original domain…

This latter point is extremely important for advocacy. Broader benefits can flow from music education only if it is continuous and of high quality. To tick the music education box, we need more than some unskilled, tokenistic sing-along moments.

Understanding the impact of arts education on skills for innovation can help education decision makers to design or give incentives for the design of appropriate curricula. What should be the place of the arts in school curricula?

We argue that the main justification for arts education is clearly the acquisition of artistic skills – the current priority objective of arts education in the curricula of OECD countries. By artistic skills, we mean not only the technical skills developed in different arts forms (playing an instrument, composing a piece, dancing, choreographing, painting and drawing, acting, etc.) but also the habits of mind and behaviour that are developed in the arts. Arts education matters because people trained in the arts play a significant role in the innovation process in OECD countries: the arts should undoubtedly be one dimension of a country’s innovation strategy. Ultimately, however, the arts are an essential part of human heritage and of what makes us human, and it is difficult to imagine an education for better lives without arts education.

The main reasons for arts education are intrinsic. In any case, “transfer” is difficult to demonstrate. Research shows that transfer is rare and that its likelihood of occurrence is directly related to the similarity between two situations.

Arts education is important for vocational reasons [are they really intrinsic?] and creative industries are increasingly important to the economy. Arts education produces more consumers of arts. Third, arts graduates seem to have skills useful in highly innovative occupations. 54% have jobs dealing with some sort of innovation.



This chapter reviews the research on the effects of music learning on cognitive outcomes: … Research shows that music lessons improve children’s academic performance, IQ, phonological awareness, and word decoding. We can understand the relationship between music training and phonological awareness since both involve listening skills. Since phonological awareness is related to word decoding, we can also understand why music training might facilitate word decoding skills in young children. But how can we understand the effect of music lessons on IQ and academic performance?

There is research seeking to demonstrate the impact of music education on
• General academic achievement
• IQ
• Reading and the reading-relevant skill of phonological awareness
• Non-native language learning
• Maths
• Visual-spatial skills
• Attention
• Memory

In all cases there are studies showing a positive correlation and often also a causal relationship. But also in all cases there are studies showing no positive correlation or effect. I do not recall any showing a negative correlation or effect ie that music study makes you dumber. Often, the writers say the evidence is positive but not conclusive and more work is needed if causality is to be confirmed.

Music education and general academic achievement
REAP (meta-analysis of studies up to 1998): all artforms showed benefit for non-arts outcomes. But often no conclusions were really possible in the correlational studies because variables such as IQ and SES were not held constant.

Post-REAP. However, each of 3 studies showed positive correlations between instrumental music participation and academic achievement. Two quasi-experimental studies found no effect but they did demonstrate that diverting time to music was not associated with a decline in academic scores – so it is possible to have academic AND music instruction without damage to academic outcomes.

As an observation, readers might stay alert for whether in a study the music learning involves playing an instrument, performing, or other broader music education. The provision of eg education in playing an instrument is more expensive than eg classroom music and there are implications therefore for the objective of delivering an effective music education to all children.

Music education and IQ
While Intelligence Quotient (IQ) is not a usual measure of academic achievement, IQ is an accepted measure of academic intelligence and a good predictor of academic achievement (and other measures of social success). Does music education improve children’s IQ?

Post REAP. In conclusion, the results of studies examining whether music education enhances children’s IQs are positive. Children who take music lessons on top of other forms of schooling at least in Western society have higher IQs than those who do not. However, this advantage is not seen at the adult level.

There were 12 studies of 3 types, and one in each type was not positive. So overall, there was evidence of a positive correlation strong enough to pass the demands of this survey for evidence of causality.

Music education, reading, and the reading-relevant skill of phonological awareness
Some quotations.

Why should music education have an impact on reading and the reading-relevant skill of phonological awareness (awareness of the phonemic components of words)?
Music education develops auditory (listening) skills that could strengthen the auditory perception of speech sounds and thereby stimulate early reading progress.
Another possible connection between music training and reading is that music education involves reading musical notations and symbols. This might reinforce the understanding of reading text or develop skills that are involved in both types of reading. Reading and music both have to do with chronological sequences or phrases, so that music training might again help develop a better awareness of time that transfers to text comprehension.

The auditory cortex does not fully mature until fully into adolescence (Ponton et al., 2000; Shahin et al., 2004). Musical training in early childhood speeds this trajectory: 4-5 year olds studying music have neural responses (called Event Related Potentials, or ERP) that are two to three years more advanced than those of children not studying music when exposed to the timbre of the instrument they are learning (Shahin et al., 2004). It is now established that shared brain areas are involved in the processing of music and speech (Jäncke, 2008; Patel, 2008). Thus researchers have begun to examine the relationship between music education and the kind of auditory perception and processing skills important for both speech perception and reading.

We conclude that a growing body of research demonstrates a causal relationship between musical training and word decoding skills.

Music education and non-native language learning

If music instruction improves auditory perception of speech, as reviewed above, then it seems reasonable to suppose that music instruction makes easier the learning of a non-native language. This was not an outcome investigated by REAP. Adult studies have revealed that musicians are superior to non-musicians at discriminating pitch and detecting pitch violations in a foreign language (Marques, Moreno, Castro and Besson, 2007; Moreno et al., 2009; Schön, Magne and Besson, 2004), and that musical ability correlates with phonological skill in second language learning (Slevc and Miyake, 2006) and with learning of tones in a tonal language (Delogu et al., 2006; Wong and Perrachione, 2007).

Music education and maths
There are many assumptions of strong connections between musical and mathematical ability and so it is not surprising that there are many studies. What is surprising is that connections are not as strong as might be expected.

The REAP meta-analyses of music and maths suggest that there may indeed be a causal link between some forms of music instruction and some forms of mathematics outcomes. Six recent, post-REAP correlational studies show a positive relationship between music training and some form of mathematical reasoning. But more research on this question is needed before we can be sure about the results: from the correlational studies, we cannot conclude that the associations found are an effect of music training rather than evidence of self-selection into music by individuals with math skills; while positive, the existing experimental studies were not numerous enough to allow firm conclusions. More experimental studies are called for. In addition, we need studies that continue to assess separately the effects of music on geometry, which is spatial, and other non-spatial forms of mathematics.

Indeed, some recent research makes us hypothesise that music education is unlikely to have an impact on some maths domains such as arithmetic. Should this be proven, it could still have an impact on geometry, so different types of mathematic outcomes need to be distinguished.

Music education and visual-spatial skills

There are many correlational studies. In my past reading, it seems to me that correlation between music education and acquisition of spatio-temporal skills was one of the most common outcomes. However, that is not the focus here unless the two terms are interchangeable, and in any case, there does not seem to be strong research evidence.

Visual-spatial skills refer to the ability to manipulate mentally figures in 2 and 3 dimensions, as in mental rotation. This is an important skill in mathematics, but also, more broadly, in professions such as engineering, surgery and archaeology.

Hetland (2000) distinguishes two non-mutually exclusive kinds of reasons why music education could enhance visual-spatial skills: neural connection theories and near transfer theories. Evidence from the REAP meta-analyses shows a positive effect of music training on visual-spatial reasoning. However, the one long-term longitudinal study included in REAP failed to show any advantage in visual-spatial skills from music training after three years. This study should cause us to be cautious about assuming that music training has long term positive visual-spatial outcomes. Since we found only one experimental study since REAP on this question, more research is called for.

Music education and attention

Learning a musical instrument requires concentration. Thus we might ask whether learning in music improves the ability to focus one’s attention and concentrate, a benefit that might then account for any improvements in school performance. Music activity involves memorising patterns of tones, notations, and motor sequences. And it requires careful listening and long periods of attention. Thus it is possible that music trains general memory and attentional skills.

A few empirical studies suggest that music training might be associated with improved attention or related brain outcomes, but there is as yet no strong experimental research allowing the conclusion that music education enhances attentional skills. Children with better attentional skills might as well choose to study more music. In the last sentence, I think “as” probably should be deleted.

Music education and memory
A correlational study and a quasi-experimental study showed music education associated with improved memory. There were no experimental studies so causation is not proven.

The study weighs up the evidence and concludes “Music lessons improve children’s academic performance and their IQ, and they improve phonological awareness and word decoding.”

We were not able to find any studies examining specifically whether music education improves children’s domain-generic creativity. In a quasi-experimental study about the effects of integrated mathematics and music activities in kindergarten, Lee and Kim (2006) find no effect of their pedagogy on musical creativity (measured by the Recording Skill Development in Music test), but this may be due to the maths focus of the pedagogy (see Chapter 3, Box 3.5). We did also identify one study comparing musicians to non-musicians on divergent thinking: it demonstrates enhanced divergent thinking in musicians (Gibson, Foley and Park, 2009).

There is not strong evidence of transfer effects from creativity in any of the four art forms. Given the importance we place on arts-induced creativity in our advocacy, the lack of research is a serious deficiency. It is worth noting the paper’s conclusions.

The claim that arts education nurtures children’s creativity seems self-evident. After all, the arts are inherently creative activities. Surprisingly, however, we found little evidence for this hypothesis in the area of multi-arts and visual arts education, though we found studies supporting this hypothesis in the area of theatre and dance. One explanation for the lack of overwhelmingly clear findings that arts education boosts creativity is that the measures used are typically paper and pencil tests of creativity. Perhaps these are poor measures. In addition, there is no reason to think that arts education will make children more creative unless the arts are taught in a way that really pushes children to explore and invent. [Music education, please note!] It is likely that many arts classes ask children to do rather routine things – sing in a group, make Christmas decorations for the school hallway, etc. It is also possible that, like in other disciplines, one needs to reach a certain level of proficiency or mastery before being able to have a more inventive approach to the practiced art, and even more so before such creativity can transfer to other disciplines or practices. However, creativity may be highly domain specific, in which case we would not see transfer of creativity from an art form even to another art form, much less to an academic subject.

There were very few studies and the outcome is inconclusive. I find this strange since informally, I am always coming across reports of children gaining self-confidence through music education and/or performance and I had the impression that there is strong research. See this letter from a mother: [Robina State]. Not of course a study, but a typical anecdote. In a study I did with principals, a number noted that students gained self confidence through participation in the music program. Ditto a study of successful contemporary musicians.

Quote from a principal: “Music, and in particular collective practice in orchestra or choir, is often seen as a means to develop social skills: this would be derived from being part of a band or collective group, communicating musical emotions, or just playing in public and having opportunities to see one’s work applauded and valued. On a more personal level, students could also gain confidence and better master their emotions as they go through the process of learning a musical piece or overcoming stage fright.”

We found one correlational study examining the relationship between self-esteem and music education (Table 9.2). Linch (1994) reported no differences in self-esteem between high school students participating in instrumental music programme and those not participating.

There is also one experimental study showing a positive relationship. However, the study says there is too little evidence overall to draw a conclusion.

Two quasi-experimental studies show a positive relationship. There is insufficient evidence to support the conclusion that music education increases empathy. This is still a question fully open for research.

Since all learning changes the brain, the important question to ask about arts education and the brain is not whether art education changes the brain. Of course it does. The question to ask, if we are interested in the question of transfer, is whether arts education alters the brain in a way that makes learning of another non-arts kind of skill more possible.

The music-brain studies described suggest that instrumental training affects areas of the brain involved in speech perception, auditory working memory, executive functioning, and attention. Most of these studies are correlational, however; experimental studies need to be conducted to determine whether these children and adults had atypical brains to begin with, or whether, as is more likely, their brains were shaped by music training.

The survey concludes with a strong argument for the intrinsic value of arts education.


Copyright The Music Trust © 2024