Wide differences in labor productivity are observed between agriculture and industry in most developing countries. Research suggests that these differences - often denoted a "dual economy" effect - can explain a significant portion of low output per capita in these countries. A central input to the labor productivity calculations is the aggregate labor effort in the agricultural sector. Using findings from the Rural Income Generating Activities (RIGA) database, I reconsider the measure of labor productivity in agriculture and industry. A simple behavioral model of agricultural workers is used to extract information on labor effort from the income data in RIGA, and this is used to estimate the aggregate labor effort in the agricultural sector. With these new estimates, dual economy effects are found to be less severe for most of the RIGA countries, although the effects are more pronounced for several. Using these estimates to adjust a wider sample of country-level data shows that the share of variation in output per capita explained by dual economy effects is around half of previous estimates.