Quantitative assessment of course evaluations

Student evaluation of teaching has been used in educational institutions around the world as a means of providing feedback on the quality of teaching. Nowadays, it is one of the most widespread tools used to inform teachers and administration about the instruction given in an institution.

The goal of the thesis is to develop efficient tools to analyze the data from student evaluations of teaching and courses at the Technical University of Denmark.

The thesis explores both classical and modern methods of multivariate statistical data analysis to address different issues of student evaluation of teaching (SET). In particular, the thesis includes results on the investigation of the association between the student evaluations of the course and the student evaluations of the teachers, the investigation of the effects of the mid-term evaluation on the endof-term evaluations and the investigation of the student non-response on SETs. In order to utilize information from open-ended qualitative student answers, text-mining methods were applied in order to extract points of students praise and complaints.

The methods proposed contribute to the knowledge about student evaluation at the Technical University of Denmark. The results provided some new information that will help teachers and university managers to better understand results of course evaluations.

Mid-term course evaluation was found to be able to capture both types of course issues: issues that can be addressed during the semester and also issues that can only be addressed at the next semester. Therefore, it seems to be preferable to conduct general mid-term evaluations instead of end-of-term evaluation, so the current course students can benefit. Additionally, it might be beneficial to conduct a short end-of-term evaluation with very limited number of questions that focus on general course issues after the final exams in order to obtain student feedback on the entire teaching and learning process, including the alignment of assessment of students’ learning with course objectives and teaching activities.

Student-specific and course-specific characteristics was found to be related with whether students participate in SETs and with how students evaluate courses and teachers. The DTU administrations should be aware that high achievers are more likely to participate in course evaluation survey and are more likely to give higher scores to courses. Students diversity on the course should be taken into account while making comparisons of evaluation results between courses.

In the student written feedback was found be able to provide additional knowledge of student point of satisfaction or dissatisfaction. However, in order to build an automated tool that can help to extract patterns from student comments higher quality of the collected data is needed.