A study was designed to develop a number of tests of oral proficiency and to compare those tests with the existing, highly subjective testing method used at the end of secondary school in Israel. The study used four experimental tests: the oral interview, role playing, reporting task, and group discussion. The experimental tests were analyzed for the extent to which they discriminate among levels of oral proficiency, interrater reliability, and student attitudes. Comparison with the traditional test was based on mean scores, standard deviation, distribution of scores, and correlations. In addition, the relationships among the four experimental tests regarding their representation of different speech styles were examined. It was found that the experimental tests were appropriately discriminatory as a group, that they had content validity, and that student attitudes toward them were favorable. They were found to be preferable to the existing oral test for providing meaningful information; some evidence that they address different constructs of oral proficiency was also found. It is recommended that new tests be introduced into the national educational system. (MSE)