This article compares stylised (questionnaire-based) estimates and diary-based estimates of housework time collected from the same respondents. Data come from the Home On-line Study (1999-2001), a British national household survey that contains both types of estimates (sample size = 632 men and 666 women). It shows that the gap between the two types of estimate is generally smaller in the case of women. But the gap between the estimates in the case of women is associated with the amount of housework performed as secondary activities and the level of irregularity in housework hours. Presence of dependent children, on the other hand, inflates the gap for both men and women. Men holding traditional gender-role attitudes tend to report more housework time in surveys than in diaries, but the tendency is reversed when they undertake long hours of housework. The overall results suggest that there are systematic errors in stylised housework time estimates.