Music training has been found to produce a range of cognitive benefits for young children, although well-controlled evaluation of the effects on psychosocial functioning has been limited. In this study participants were recruited from two grade levels (prep/grade 1, "N" = 210; grade 3, "N" = 149), and were allocated to a music education condition, or a control condition which continued to receive their usual curriculum. In the treatment condition, Kodaly music classes were introduced into the younger cohort's curriculum, while the older cohort received additional instrumental classes which were predominantly strings-based. Results indicated that school-based music classes prevented a decline in global self-esteem measures experienced by the control group in both the younger and older cohorts, and in general and academic self-esteem for the older cohort. Preliminary data also suggest that a similarly engaging, arts-based activity (juggling) may confer similar benefits for older children's self-esteem. While difficulties inherent in performing experimental research within schools prevent strong conclusions, the data imply that increasing the frequency and quality of arts-based activities can be beneficial for the self-esteem of primary school-aged children.