The use of risk ratios, odds ratios and prevalence ratios is limited within dental epidemiological research. These measures of association have been applied to relate the risk indicators' geographical location and sex to the caries prevalence in the maxilla, the mandible and the total dentition of 7-11-yr-old Tanzanian schoolchildren examined in 1986. The risk indicators were studied by means of odds ratio, using logistic regression. There were no associations found between the risk indicators studied and caries prevalence in the maxilla and the total dentition: a significant association was found only in the mandible. This resulted in girls having a greater (PR = 1.4) chance of developing caries than boys, and children living in naturally fluoridated (greater than or equal to 0.5 ppm) rural areas having a lower (PR = 0.5) chance of developing caries than those living in urban and rural areas.