Current research in the English-as-a-Second-Language (ESL) field has not assessed the comparative ability of indirect and direct test scores to predict academic success. Work of this kind would inform practice in the area of testing and placement. This study investigates the practical problem of making appropriate placement decisions for students whose test results show wide enough discrepancies to indicate placement in different academic programs, or at different levels within programs. The question of whether scores derived from indirect measures (language proficiency tests) or direct measures (writing samples) are better indicators of academic language proficiency is addressed. This study also explores the usefulness of grade point average (GPA) as a measure of academic success, and proposes the use of average accumulated credit per semester (AACPS) as an additional measure. Several researchers have questioned the use of GPA as the sole measure of academic success for ESL students, and this study adds to existing research. The ability of two types of placement tests to predict academic success for ESL students in secondary and university programs is evaluated: one indirect (the Michigan Test of English Language Proficiency); the other direct (a combination of GPA and AACPS measures). Results are mixed. Nine data-rich appendices and 48 references are included. (KFT)