To measure the impact of participating in the RSE program, Gibson and McKenzie (2014a) used propensity score matching to identify a group of households with RSE workers and a group of otherwise similar households without—in terms of characteristics that communities and employers likely paid attention to when selecting workers and characteristics that may affect the decision of whether to apply. Using either difference-in-differences estimation on prescreened samples (based on having propensity scores in the 0.1–0.9 range) or panel data estimation with fixed effects for each household gave similar results. The results in Table 2 show that, over a 2-year period, households participating in the RSE program typically earn 35% more income than other observationally similar households who did not supply RSE workers. Some of this extra income went to an increase in expenditure, especially in Vanuatu, while household savings, dwelling improvements, and bank account ownership also increased by significant margins. The subjective standard of living (based on a randomly selected adult) was found to increase by approximately the same amount (in terms of standard deviations) as the increase in income.¹¹