Response Bias in the Gender Gap in Non-Market Work

Despite increasing female labor force participation, gender differences in non-market work, such as housework and childcare, persist. By exploiting exogenous variation in the questionnaire of the German Socio-Economic Panel, I analyze how response behavior affects estimates of time use in a longitudinal setting. I find a causal impact of questionnaire design on the reported hours of non-market work within individuals over time. Both women and men report more hours of non-market work on weekdays when they are not additionally asked about their time use on weekends. This response bias is larger for men, both in absolute and relative terms, with men overreporting by 10 percent of their average weekday non-market work compared to 2 percent for women. As a result, gender gap estimates are biased downward. The findings imply the broader conclusion that the common practice of irregularity in questionnaire designs has important impacts even on questions asked regularly.