This paper investigates whether policies that encourage recipients to exit welfare for full-time employment influence participation in educational activity. The Self-Sufficiency Project ('SSP') was a demonstration project where long-term welfare recipients randomly assigned to the treatment group were offered a generous earnings supplement if they exited welfare for full-time employment. We find that treatment group members were less likely to upgrade their education along all dimensions: high-school completion, enrolling in a community college or trade school, and enrolling in university. Thus, 'work-first'; policies that encourage full-time employment may reduce educational activity and may have adverse consequences on the long-run earnings capacity of welfare recipients. We also find that there was a substantial amount of educational upgrading in this population. For instance, among high-school dropouts at the baseline, 19% completed their diploma by the end of the demonstration. Finally, we simulate the consequences of the earnings supplement in the absence of adverse effects on educational upgrading. Doing so alters the interpretation of the lessons from the SSP demonstration.