Measuring child labor: Comparisons between hours data and subjective measures

"This paper examines a subjective measure of child labor as an alternative to hours data for eliciting the distribution of children's time between work, school, and leisure. The subjective child labor questions that were developed have two primary advantages. First, the subjective measures avoid proxy respondent bias in child labor reports made by parents in a standard hours module. Second, the subjective child labor module scales responses to elicit the relative distribution of the shares of children's time without relying on hours data which are prone to severe outlier problems. Adult, proxy respondents are found to produce uniformly lower reports of children's time allocated to work and school than the child's own subjective responses. Conditional labor supply functions are also estimated to examine the marginal effects of child, parent, household and school characteristics between the two types of data. Children's subjective responses are found to increase the magnitude of the marginal effects for child's age, parental education, and school availability with limited differences between household composition and asset variables." from authors' abstract