The present study uses panel data models to control unobserved characteristics and to investigate how the presence of spouses in interviews influences reports regarding housework and earnings contributions. Both husbands and wives relatively overreport their housework contributions but do not overreport their earnings contributions. The amount of time spent doing housework lacks a precise measure and involves more subjective estimates than earnings reports. It is argued that the ambiguity of the housework contribution mitigates the guilt felt by overreporting the housework contribution. In addition, without controlling for unobserved characteristics, OLS models overstate the influence of the presence of spouses in the interviews.