A small study that compared teaching Spanish-speaking children English vocabulary using a song or a spoken poem has found definite and long-term advantages to the song form.
The study involved 38 Spanish-speaking Ecuadorian children (aged 9-13), of whom 22 were randomly assigned to learn a 29-word English text as an oral poem, and 16 learned it as a song. None of the children had had any formal instruction in English; all had some limited music training. The children were given 4 training sessions and 3 testing sessions over two weeks, with a final test for 13 children six months later.
Children in the song condition out-performed those in the spoken condition on every measure: their ability to recall the passage verbatim, pronounce the words, and translate target terms from English to Spanish.
While pronunciation of vowels was notably better, though there was no difference in consonants.
Long-term recall is of course the main question of interest: six months after this little experiment, with no English instruction since, those from the song condition could recall without prompting an average of 8.83 words out of 10 target words, compared with 0.29 words for those from the spoken condition. However, there was no significant difference in translation success, which was extremely low in both cases (2.26 vs 1.07 — this compares with 4.03 vs 2.69 at the end of training).
The song itself, its melody and rhythmic structure, was remembered very well. The children in the song condition also enjoyed the learning sessions much more.
The study is small, and comes with several caveats, but it does provide support for the use of songs as an adjunct to foreign language learning.
(2015). The efficacy of singing in foreign-language learning.
Psychology of Music. 43(5), 627 - 640.