Analyzing Gender Bias in Student Evaluations

Sentiment analysis is a computational linguistics task that characterizes positivenegative tone expressed in language data. We apply sentiment analysis to analyze student comments spanning over eight years of College biology, chemistry, physics and math and explore differences in sentiment conveyed toward male and female faculty. We study differences in perceptions found in categories pertaining to instructor competence, organization/presentation, personality/helpfulness and overall satisfaction. We also use term ranking to identify words that are strongly associated with either gender.