This chapter used data from the Young Lives study.  The author's abstract reads:

This chapter investigates the role of children’s time use to produce one cognitive skill (i.e. a verbal score) and two psychosocial skills (i.e. a Self-Efficacy index and a Self-Esteem index). Following a dynamic human capital accumulation approach (Cunha & Heckman, 2008), I estimate linear production functions for both types of skills. Under this framework, I combine time inputs, current and past, and lagged outcomes to examine the relevance of time investments made at younger ages relative to present time investments to produce three different outcomes by the time children reach 15 years old. I also examine the trade-offs of child work among each alternative time input activity. Findings indicate time inputs effects are small for both types of skills, although daily time spent in educational activities is crucial for verbal development, specifically time spent studying and at school, leading to an increase of up to 0.077 standard deviations by age 15. For the Self-Esteem Index, current time (at age 15) spent in leisure and past (at age 8) and current time spent in child work is detrimental for this skill at age 15, decreasing this outcome between 0.057 and 0.63 standard deviations, respectively. I highlight concerns on measurement error for the Self-Efficacy Index, excluding the results in the discussion. On the trade-off analysis of child work, I only find small detrimental effects for the verbal score of current time spent in paid work (at age 15), only when crowding-out time spent in educational activities; and no effects for the Self-Esteem Index.