This study investigated the relationships among several different indicators of high school performance: test scores, dropout rates, transfer rates, and attrition rates. Hierarchical linear models were used to analyze panel data from a sample of 14,199 students who took part in the National Education Longitudinal Survey of 1988. The results generally support the notion of an alternative as opposed to a common view of school effectiveness: Schools that are effective in promoting student learning (growth in achievement) are not necessarily effective in reducing dropout or transfer rates. In fact, after control for student inputs, high schools exhibit relatively little variability in dropout rates but considerable variation in transfer rates. In addition, characteristics of schools that contributed to performance in one area often did not contribute to performance in another. Given these findings, the authors suggest that, along with test scores, dropout and transfer rates should be used to judge school performance. (Contains 33 footnotes, 5 tables and 2 figures.)