A study examined why college students take sexual risks and documents the existence of optimistic bias, which suggests that individuals underestimate their personal risk to health hazards in relation to their peers. Subjects, 318 students enrolled in an introductory course in mass communication at a large east coast university, completed an anonymous survey instrument. Results indicated that (1) college students underestimate their personal risk compare to their peers; (2) this optimistic bias is especially high for sexual risks; (3) there was no relationship between optimistic bias and age or grade point average; and (4) males exhibited more bias than females. Findings suggest that the practice of safer sex is better predicted by perception than by actual risk, and that university health centers should consider all students "at-risk" of unplanned pregnancies, sexually transmitted diseases, and HIV/AIDS. Additional research is needed to clarify the influence of other variables on optimistic bias and the extent to which it can be modified by personal risk messages. Contains 43 references. (RS)