A case-control study was conducted to determine the protective effect of measles vaccine under routine vaccination conditions in Dar es Salaam, Tanzania. A total of 172 cases of measles were obtained over two six-week periods from hospitals likely to admit such cases in the city and their mothers interviewed using a questionnaire that had been developed for the purpose. In addition, each case was matched to four controls for age, sex and place of residence and their mothers were also interviewed using the same questionnaire as for the cases. Vaccination status of the cases and the controls was ascertained by asking the mothers as well as obtaining vaccination information from the Mother-and-Child or Road-to-Health (MCH) clinic cards normally issued to mothers at first vaccination of their children. The overall protective effect of measles vaccine in all age groups studied was found to be 54% (95% confidence limits (CL)36%-67%) when mothers' recall for vaccination status was used. This level of protection was considered to be low compared with the efficacy of the vaccine under experimental conditions in controlled trials previously reported. However when MCH Clinic cards were used to ascertain vaccination status the protective effect rose to as high as 96% (95% CL 83%-99%). Potential biases in the study design and their role in masking or exaggerating the vaccine efficacy are discussed. It was therefore concluded that the protective effect of measles vaccine under routine vaccination conditions prevailing in Dar es Salaam region at the time of this study was as high as observed under experimental conditions when ascertainment of vaccination status was record-based. The study therefore recommended that, whenever in doubt about vaccine efficacy, simple study designs like the case-control method should be used to quickly evaluate vaccine efficacy by utilizing record-based vaccination status ascertainment methods.