BACKGROUND: There is a need for valid approaches to measure sexual interactions to assess the impact of behavioural interventions and to predict the impact of behaviour changes. Different methods of asking about sexual behaviour often yield conflicting answers and men often report higher levels of heterosexual activity than women. To better understand self-reported sexual behaviour data and how best to collect it, we analyzed data collected as part of a larger project (ST IMPACTS) on the social and behavioural impact of introducing community-level HIV self-testing (HIVST) with counseling (semi-supervised with pre- and generic post-test counseling provided on delivery or collection of test kits) in an urban Malawian setting. METHODS: Information on sexual behaviour was collected from HIV self-testers over a three-month period. Three different methods were used: retrospective face-to-face interviews (FTFI); audio computer assisted self-interviews (ACASI) and a prospective coital diary. Both retrospective instruments were used before and after the three-month study period. Frequency and cross-tabulation, as well as scatterplots, were used for exploratory analyses. Chi-square tests were used to test for differences in proportions. Spearman's correlation coefficient was used to explore associations between both continuous and ordinal variables and Wilcoxon's paired sample and Mann-Whitney test was used to test for differences in such variables or between variables. RESULTS: There was reasonable agreement between the two retrospective methods although both yielded inconsistent answers e.g. with lower reported numbers of life-time sexual partners at the end than at the beginning of the study period. The diary method elicited higher reported levels of sex with multiple partners than both retrospective instruments which may be due to inadequate recall. Over the study period 37.4% of men and 19.7% of women reported multiple sexual partners using the diary. There was no clear relationship between reported sexual behaviour and HIV status (prevalence 9.6%). CONCLUSIONS: Diaries may therefore have higher validity for sensitive behaviour reporting and thus be the preferred method in similar African contexts in measuring sexual behaviours.