This working paper investigates whether there is a bias in the measurement of the Self-Sufficiency Project treatment effect because up to 20 per cent of the intended sample did not agree to take part. It does this by comparing the estimated impact of the program that was obtained using only experimental data with the impacts obtained using administrative data on the individuals who did not take part in the experiment.
Published: October 2003
Policy Area: Income Security - Welfare and Employment
Population: EI Recipients - Low-income Populations - Low-skilled Workers - Social Assistance Recipients - Women - Communities and Families