Most research on children's play takes a context-dependent, adult-focused observational approach to the measurement of play. The current two studies present the development and psychometric properties of the Child Self-Report Playfulness (CSRP) scale, which was presented via "puppet-show" to two samples of children. Study 1, across 98 children between 5 and 7 years of age, showed that the items of the CSRP had good internal consistency, were stable across the 6-week test--retest period and, for the most part, showed no differential item functioning across age, gender, and language ability. Study 2 involved a longitudinal sample of children followed from Reception (age 5, N = 244) to Year 1 (age 6). Findings revealed favourable psychometric properties using longitudinal confirmatory factor analysis and measurement invariance. Associations between child playfulness and teacher- and peer-ratings of social behaviour were examined, and showed specific associations between playfulness and teacher-rated play interaction, rather than social maturity, problem behaviour or social preference. Overall, the CSRP appears to be a promising approach to the measurement of dispositional playfulness in young children.