Tim van der Zee
In many, if not most educational institutes, students are asked to evaluate their teachers. This gives the students the ability to give feedback to the teacher on how they perceived the (quality of the) course and instruction. Often, these student evaluations of teachers will also be used as a formal evaluation of the teacher's quality of teaching, and/or of how much the students learned. This begs the question:
Do students learn more from teachers which are more positively evaluated?
The answer is no; the relationship between student learning and student evaluations of teacher is zero. This might be counterintuitive to many, given that these evaluations are often used formally and there have been many meta analyses which do show a relationship. A particularly popular one is a meta analysis by Cohen (1981) which showed a correlation of r = 0.43 between student ratings and learning, which is a rather substantial correlation for educational research.
However, the majority of these previous meta analyses did not account for important factors like publication bias and how small studies often tend to overestimate correlations. A recent meta analysis (Uttl, White, & Gonzalez, 2016) used a wide variety of much more advanced statistical techniques to control for these confounds, and concluded two things:
- Previous meta-analyses (such as Cohen, 1981) used poor meta-analytical methods. Findings showing a strong, positive relationship between student evaluations and student learning are an artifact of these low quality methods.
- When using high quality methods, we can conclude that students do not learn more from teachers which they evaluate more positively. The combined literature of research on this topic is best explained by random variation around a true mean of 0.
References