Master's lecture in Applied Statistics - Ólafur Jón Jónsson
The lecture will be streamed: https://eu01web.zoom.us/my/annahelga
Master's student: Ólafur Jón Jónsson
Title: Analysis of results from student evaluation of teaching surveys in the University of Iceland 2013 - 2017
Faculty: Faculty of Physical Sciences
Advisor: Anna Helga Jónsdóttir, Associate Professor at the Faculty of Physical Sciences
Other members of the masters committee: Margrét Sigrún Sigurðardóttir, Assistant Professor at the Faculty of Business Administration and Daði Már Kristófersson, Professor and the Dean of the School of Social Sciences
Examiner: Gunnar Stefánsson, Professor at the Faculty of Physical Sciences
The objective of this study was to evaluate some aspects of the results of the SET (student evaluation of teaching) in the University of Iceland. The difference in how various types of students rate courses and teachers was analyzed, with special emphasis on the gender of the teachers involved. Systematic variability of students, courses and teachers was also analyzed and that made it possible to determine the percentage of the overall variability explained by those factors.
The study consists of SET data obtained from the fall semester of 2013 to the fall semester of 2017. The data was divided into two parts; a course section and a teacher section. The number of observations in the course section was 129,683 in total and in the teacher section there were 183,815 observations in total. The course section of the SET was then divided into five parts of focus; student contribution, course gain, course organization, workload and overall score. The teacher section was divided into three parts of focus; teaching, theoretical motivation and overall score. The overall score parts of focus consisted of a single answer on the scale of 1-10 while the other parts of focus consisted of statements that were answered on a five point Likert scale in which students indicated how much they agreed or disagreed.
Linear mixed models were used to analyze each part of focus. The results showed that the dominant part of the overall variability was unexplained in each model. The variability due to courses or teachers was not as high as expected, which raises concerns about the current state of the SET. Variability explained by students was relatively low indicating that students can be trusted when responding to surveys like these. Due to the large number of observations in the dataset, a slight difference between various groups of students results in a significant statistical difference. Although the worst results in SET is always obtained by the combination that consists of female teachers and male students, the actual difference is very small. The only remarkable difference of groups across almost all models was between the students who failed a course and those who passed the course.