Can children in classrooms learn new languages through songs? A look at what research on this topic can tell us

Language teachers believe that singing songs with young learners is a valuable teaching practice that helps children to learn new languages (see this Deanery Digest for more details). However, there appears to be little research on whether this is true. And, if it is, exactly what is it about songs that helps support children’s foreign language development. To find out if teachers’ beliefs are supported by evidence, we conducted a systematic review of research studies that evaluated using songs, chants or nursery rhymes as whole-class language-learning activities, and which measured specific features of children’s second or foreign language development. For example, do children learn more vocabulary through singing songs or through chanting, or is children’s pronunciation improved by singing songs compared to other ways of teaching?

A systematic review aims to find all the relevant research evidence addressing a particular topic or question, and consider it together, rather than relying on only one or two selected studies. In this way, we can be more confident that we have a complete picture of what the evidence tells us. As a consequence, we can be more confident about the recommendations for classroom practice.

We conducted a thorough and systematic search of the key databases of research in education, psychology, and linguistics for research that aimed to find out what the effects of using songs, chants and nursery rhymes is on second or foreign languge development. To do this, we wrote a detailed search strategy that helped us discover any studies that met our criteria for being included in the review.

To be included, a study needed to have investigated using songs, chants or nursery rhymes as a whole-class activity with learners aged two to eighteen years (in school settings, not at home) and to have measured a feature of the children’s language development. There was no restriction on which aspect of language was being investigated.

After we identified the papers, we extracted all the relevant information, such as who took part in the study, what type of school they were in, which country the study took place in, which languages were being taught, the methods the researchers used, and what their findings were about the aspects of language development that they investigated.

Finally, we assessed how well conducted each study was. For example, were they fair tests? Were differences between groups of children at the start of each study accounted for when analysing the results at the end? Were there any other influences that could have affected the studies’ results? How many children took part in each study? And so on.

We found nearly 3000 potentially relevant studies and whittled these down to a final 60 that met all our criteria. They came from 23 countries and had been conducted from the 1970s onwards, with 47 published since 2009.

34 studies took place in primary schools, with 13 in secondary and 13 in pre-schools. 33 studies (over half) measured vocabulary learning as an outcome, indicating that this was an important area for researchers. This category was followed by speaking (17 studies), then grammar (15), listening (11), reading (8), and writing (3). Studies quite often measured multiple aspects of language development. There was not much coherence in how things were measured: most studies seemed to be reinventing the wheel rather than building on previous work.

However, when we assessed how well conducted each study was, we found that only three studies met the highest standards and provided ‘strong’ or highly trustworthy evidence to contribute to the discussion. 14 studies contributed ‘moderately’ trustworthy evidence. The remaining 43 studies fell into the category of ‘limited trustworthiness’. These studies did not always state their aims clearly, collect data that was useful in addressing those aims, or run appropriate and careful analyses.

This lack of quality research was quite a surprise, given how strongly some practitioners advocate for using songs for teaching languages.

The overall picture painted by the research was quite unclear: we neither know what the potential effects of singing songs might be on young language learners’ various language development outcomes, nor how these effects compare with using other teaching approaches (e.g., telling stories).

On the bright side, there was no evidence of songs being detrimental to language development. It is just unclear, from the research, whether songs are more or less useful than any other teaching approaches we might choose.

The next step is to begin carefully accumulating research from studies where we can draw solid conclusions about how using songs compares to other teaching approaches. And this is what we did in our next study, which you can read about in this Deanery Digest.