While the goal of language teaching has been the development of practical communication, most language tests have stressed mastery of discrete linguistic skills. Oral language performance-based testing requires test-takers to use and perform in the language in actual communicative situations. The variations in communication created by the interaction between language and social setting have direct implications for testing oral communication. The extent to which communication variations manifest themselves in testing situations and the extent to which the assessment of oral proficiency is affected by these variations is unclear. A study was undertaken to determine the stability of the assessment of oral proficiency through four oral interview tests. Each oral interview was administered on different occasions by different interviewers, using different speech styles and topics. There was little effect on communication scores due to the variance of occasion or interviewer. However, the assessment of oral proficiency did change significantly from an "interview" style to a "reporting" style, when subjects spoke on different topics. Although there is a lack of stability and low reliability and validity present in the FSI Oral Interview, the test does provide an excellent setting for a variety of tests for different communicative variations. (Author/JK)