This article tries to answer two related questions: (i) what do children hear while listening to and making sense of music? and (ii) what kind of representational tools can be used to assess this sense-making? To answer these questions, we set up two empirical studies in which 89 children – 8–9-year-olds and 11–12-year-olds (first study) – and 331 children – 8–10-year-olds and 11–13-year-olds, with and without extra music training (second study) – were exposed to a music listening task. The aim of the studies was to get an overall picture of the variety of children's musical representations by means of their graphical notations, and to investigate the impact of age, formal musical training and the characteristics of the musical fragment on these notations. A major finding of the first study was the emergence of two main categories of notations, namely ‘global’ and ‘differentiated’ notations, with a very strong dominance of global over differentiated ones and a negligible impact of subject and task variables. The second study, in which we presented researcher-generated instead of existing musical fragments, yielded a larger number of differentiated notations, and a considerable impact of age and formal musical education as well as of the musical characteristics on these notations. Both studies were ascertaining studies with the aim to describe and analyse the development of children's graphical notations under given instructional conditions. To account for some of the limitations of these studies, some additional design-based research is suggested. Extensive findings/exemplars of both studies can be found on the Cambridge University Press website.