A Review of Data Mining Techniques in Analyzing Faculty Evaluation Feedback

Student feedback on college faculty evaluations often includes personal comments crucial for assessing teaching and making academic decisions. However, these remarks are usually disorganized, making it hard to analyze them using standard methods. In this systematic literature review, we take a closer look at data mining techniques like sentiment analysis, topic modeling, and text classification to understand better what the feedback tells us. We reviewed studies that are relevant or related to analyzing qualitative feedback using sentiment analysis, topic modeling, or text classification, highlighting the problem/objective, the algorithm/method, and the dataset/data source. We aim to determine which method best draws valuable insights from unstructured comments. We compare these techniques based on different metrics. After carefully searching and filtering through a pool of 21,917 studies published between 2015 and 2024, we narrowed it down to 43 articles for an in-depth metaanalysis. Of the 43 articles reviewed, 74 % (32 out of 43) focus on faculty evaluation feedback analysis. Notably, all of the selected studies are journal articles. Among them, 72 % explore sentiment analysis, 19 % examine topic modeling, and 9 % investigate text classification.