<p>Collecting accurate data on children’s time allocation is critical for monitoring progress on the eradication of child labour. However, current practices often rely on proxy respondents reporting on behalf of children. This raises concerns about the accuracy of their responses on how children allocate their time. At the same time, children’s self-report may also be prone to biases. In this study, we use data from Ghana and Burkina Faso to document the differences between children’s self-reported data and proxy respondents' data on children’s participation in household chores, economic activities, and child labour. We also explore the difference in educational indicators by respondent type. Our findings highlight significant discrepancies, indicating that children generally reported higher participation and hours in household chores and economic activities than proxy respondents. We find no discrepancies in education-related indicators in Ghana, but proxies report higher education outcomes in Burkina Faso compared to children. Moreover, we find that the type of respondent affects the estimated impacts of cash transfers on child labour outcomes, as statistically significant impacts are more often observed with proxy-reported data than with child-reported data. Overall, our study suggests the need for greater attention to the source of the data when evaluating statistics and programme impacts on children’s time use and employment.</p>