This article uses a large sample of completed student evaluation forms of university teaching to quantify the impacts of student evaluation of teaching (SET) attributes on teaching effectiveness (TEVAL) scores. Despite much criticism of and support for TEVAL scores measuring teaching effectiveness, detailed quantitative studies of the relationship between instructional attributes and TEVAL scores are lacking. This study helps to fill this gap. Results suggest that the relative influence of teaching attributes on TEVAL scores varies with the level of the course. While students’ perceptions of how well the coursework is organized, explained and presented have large positive impacts on TEVAL scores at all levels, their relative importance varies with the level of the class. Furthermore, the SET attribute “emphasis on thinking rather than memorizing” has little or no substantive impact on TEVAL scores. An implication is that a lecturer stressing this aspect does little to increase her/his TEVAL score. Furthermore, lecturers wishing to raise their TEVAL scores should vary their relative emphasis on different teaching attributes according to the class level. A feature of this study is its use of individual student responses rather than class averages. Therefore, it accounts for all the information provided by the data.