A study investigated factors in design of oral language comprehension tests and their relevance in determining actual language proficiency, focusing on the tests' decoding requirements. Task-based, interactive oral language tests were designed: (1) to elicit information about integration of first (L1) and second languages (L2) in the learner's language processing; and (2) to avoid assessment of language skills in isolation. The resulting tests were piloted, and information was gathered on overall task fulfillment and on different aspects of the tasks. One test module, in which the task involved researching an issue and reporting results, illustrates the technique. The examinee's report and interview concerning the research were recorded, and the examinee then wrote a report in L1. The report contains five sections each graded separately on the extent to which the candidate was able to identify and integrate information needed to fulfill the task. Analysis of test performance suggests that conventional oral scales concerned with surface features of language are not adequate predictors of ability to comprehend in oral interactions. In addition, in this test it appeared that comprehension skills were closely related to other complex and integrative skills such as ability to write a coherent report in L1. (MSE)