Measuring Housework in Time Use Surveys

A methodology is devised to decompose the error in the measurement of housework time in surveys. Using data from the Panel Study of Income Dynamics (PSID) and the Americans' Use of Time Survey, we estimate that a telephone (mail) survey rather than an interview survey leads to a mean underestimation of 60 (30) min/week for women and 18 (13) min for men. We also find that the errors created by using a narrow definition of housework leads to an underestimation of between 10 and 12 h/week for both men and women in the PSID. An erratum to this article is available at .