This paper reports a second-cycle validation study of the test of English at Matriculation (TEAM) following revisions made to the test in 1993. Candidates' scores from four academic sessions (1993-1997) were used to assess the relationship between the performance at the beginning of their degree course and their eventual academic outcome. Results suggest that: (1) the predictive capacity of the revised test is similar to that of the other tests of English for academic purposes; and (2) performance in the listening section remains the strongest single linguistic predictor of candidates' success in Master's degrees (even though it might have been expected that writing skills would have been a better predictor considering that most measures of student achievement in higher education is in written form). The overall correlation is slightly lower than for the first version of TEAM. It is concluded that the test as presently constructed is adequate for the purposes for which it is intended. (Contains 13 references.) (Author/KFT)