Data on students' performance at the Oslo Dental Faculty for 1977-86 were divided in two 5-year periods and analyzed in accordance with admission levels, largely on the basis of academic performance in junior college. During these years admission level decreased considerably, whereas the frequency of 'not passed', 'dropouts', and candidates using prolonged student time increased. The dental school grade average and the distribution of high- and low-performance candidates varied with the admission level. This trend was visible most clearly in the first 5-year period, when the admission point range included a considerable number of high admission level students. However, the admission level was not a good predictor of students' performance in the large middle or low admission level groups prominent in the second 5-year period. Poor preclinical results were not compensated for by clinical skills. All comparison of students' performance gave results in favor of the female students, but no sex-related differences were statistically significant. Most dropouts left dental school during the 1st year without visible examination difficulties. Students with several 'not passed' tended to repeat examinations, dropping out at a later stage or graduating after prolonged student time with poor results. With low interest in odontology, selection of students on the basis of minute differences in academic performance in the lower admission point scale is of limited value as a predictor of students' performance.