Detecting Gender Bias in Course Evaluations

TL;DR

The paper tackles gender bias in university course evaluations by analyzing student comments in English and Swedish with NLP techniques. It employs a bilingual bag-of-words pipeline (CountVectorizer) and two classifiers, logistic regression and random forest, to predict examiner gender from feedback after anonymization. Findings suggest language-specific cues correlate with examiner gender, with Swedish data showing stronger signals; the work demonstrates tangible linguistic differences in evaluations that may reflect bias and lays the groundwork for deeper embedding-based analyses. This approach provides a path toward more nuanced, cross-linguistic assessments of bias in educational feedback and informs future methodological improvements.

Abstract

An outtake from the findnings of a master thesis studying gender bias in course evaluations through the lense of machine learning and nlp. We use different methods to examine and explore the data and find differences in what students write about courses depending on gender of the examiner. Data from English and Swedish courses are evaluated and compared, in order to capture more nuance in the gender bias that might be found. Here we present the results from the work so far, but this is an ongoing project and there is more work to do.

Figures (1)

• Figure 1: Overall impression for courses with male and female examiners, separated by teaching language.