The development and validation of oral native language tests for Dutch adolescents was conducted for both assessment and curriculum development purposes at the secondary level. The oral language situations tested were the monologue, dialogue, and polylogue or group conversation on a topic. An American inventory of common speech subjects was used to construct test situations, and the test subskills were drawn from the students' textbooks. The situations, subjects, and subskills were combined into three tests, which were administered to 14-year-olds in all areas of the country and all dialect regions. The tests were taped, and teachers were asked to rate the dialogue and monologue tests on content, organization, language, delivery, and communication. The interrater reliability and homogeneity of assessment criteria were analyzed statistically. On the whole, it was found that the two test types were closely correlated and that overall scores would have sufficed for ranking students in the case of almost all raters. However, when only one rater is available, as in a classroom testing situation, the analytic method of scoring is preferred to the overall scoring method. (MSE)