In population-based surveys on sexual behaviour, men consistently report higher numbers of sexual partners than women, which may be associated with male exaggeration or female under-reporting or with issues related to sampling, such as exclusion of female sex workers. This paper presents an analysis of data collected in the context of a longitudinal study in rural Tanzania, where a sexual partnership module was applied to all participating men and women in the study population. Since the study design included all men and women of reproductive ages and did not involve sampling, these data provide a unique opportunity to compare the consistency of aggregate measures of sexual behaviour between men and women living in the same villages. The analysis shows that non-marital partnerships were common amongst single people of both sexes--around 70% of unmarried men and women report at least one sexual partner in the last year. However, 40% of married men also report having non-marital partners, but only 3% of married women did so. Single women reported about half as many multiple partnerships in the last year as men. Under-reporting of non-marital partnerships was much more common among single women than among married women and men. Furthermore, women were more likely to report longer duration partnerships and partnership with urban men or more educated men than with others. If a woman reports multiple partners, biological data indicate that she is at high risk of contracting HIV. For men, however, there is only a weak association between number of partnerships and the risk of HIV, and it cannot be excluded that men, especially single men, exaggerate the number of sexual partners.