This study compared structured observation with a 24 h pictorial recall of household activities ('sticker diary') to measure the prevalence of handwashing with soap (HWWS) in the community. The study was done within a cluster-randomised trial evaluating a handwashing promotion programme in Bihar, India. HWWS at key occasions in mothers and school children was measured by structured observation in 299 households from 32 villages. Sticker diaries recalling common activities, including personal hygiene, were used to measure HWWS in 299 households from a further 20 villages. Sticker diary HWWS prevalence estimates were about 13% points higher than structured observation estimates, but the differences varied by the type of handwashing occasion. This study confirms structured observation as the method of choice for the study of handwashing behaviours. The sticker diary method may be useful in large-scale surveys. Sticker diaries may overestimate HWWS at important occasions, but probably less so than conventional questionnaire tools.