MSc Seminar: "How students describe and rate professors" by Nirmal Narayanan

ABSTRACT:

Student evaluation is widely being recognized across universities and colleges for professor
evaluation, but there are many underlying trends or bias occurring that can influence the
interpretation. RateMyProfessor.com is becoming an increasingly popular publicly accessible
web-based evaluation, due to widespread usage by students for course selection and high quantity of data available for analysis. This seminar will discuss the correlation between review comments of economic professors in Ontario and their numerical rating attributes for identifying dominant aspect on which students comment upon by applying Latent Dirichlet
Allocation (LDA).

Using LDA method, the corpus entries of comments are represented as low dimensional vectors which led to better results. We use this large corpus to answer the following questions: 1) What words do students use to describe their professors? 2) Is there any correlation between the words used and the rating given? And, are there significant differences between the groups analyzed? We discuss the different correlation obtained and the potential implications.