Students' evaluations of teacher effectiveness are explored as a multidimensional construct, emphasizing the Students' Evaluations of Educational Quality (SEEQ) instrument developed by H. W. Marsh (1987). An overview is presented of studies in which Marsh evaluated longitudinal data from an archive of responses to nearly 1 million SEEQ instruments representing 50,000 courses collected over 13 years. The focus is on a cohort of nearly 200 teachers evaluated over the period. The analysis of studies considers: (1) the generalizability of the SEEQ factor structure; (2) higher order factor structure; (3) the generalizability of ratings over time; (4) models of covariance stability; and (5) analyses of teacher profiles. Analyses provided clear support for the generalizability of the SEEQ factor structure over time, over courses, and over teaching at different levels. Higher order factor analyses suggest that higher order factors that might underlie SEEQ factors are not particularly useful in providing a smaller number of scores with which to summarize SEEQ responses. Each instructor was found to have a reasonably distinct profile. Mean ratings of the same teachers evaluated consistently showed no systematic increases or decreases. An appendix contains the SEEQ instrument. Four tables and three figures present study data. (SLD)