In this paper, we investigate the issue of using protected at-tributes in college dropout prediction in real-world contexts. Protected attributes are traits or characteristics based on which discrimination is prescribed as illegal, such as gender, race, age, religion, and genetic information.
