Social desirability bias, which is the tendency to under-report socially undesirable health behaviours, significantly distorts information on sensitive behaviours gained from self-reports. We designed a list randomisation method to indirectly elicit condom use among female sex workers and tested it among 651 female sex workers in Senegal, a country where sex workers face high social stigma and where the AIDS epidemic is mainly concentrated among this population. Based on our list randomisation, we found that the condom use rate in the last sexual intercourse with a client was 78%, which is significantly lower than the 97% obtained when asked directly in the survey. When estimating condom use among the subgroups, we found that female sex workers who are at a higher risk of infection are less likely to use condoms.