This study investigates whether and how (1) cumulative overall exposure to television and (2) cumulative selective exposure to specific television content are related to both estimates of and opinions about people who have mental illnesses. Two hundred fifty-two Belgian high school students completed self-report questionnaires. Measures included time spent watching television, exposure to specific content, perceived realism of television images, direct experience with a person who had mental illness, and estimates of and judgements about the dangerousness of people suffering from a mental illness and their ability to lead a socially integrated and "normal" life. Results suggest exposure to television itself makes but a small contribution to explaining these estimates and attitudes; related concepts such as perceived realism appear to be more important. (Contains 5 tables and 4 footnotes.)