Objective: The current study explores non-linear effects of absolute risk and effects of comparative risk information about skin cancer on individuals' safety ratings, affective responses and behavioral intentions. Method: An experimental survey was conducted among college students (N = 563) to test the effects of absolute and comparative risk information across a range of risk levels from 3.7% to 60%. Results: The non-linear model of absolute risk effect was confirmed by the data. Absolute risk information had detectable effects at low levels of risk, but the effects did not change appreciably at the highest absolute risk levels. Regarding comparative risk, the data did not support its effects on individuals' safety ratings, affective responses and behavioral intentions. Conclusion: The study found the non-linear model fits well to the skin cancer related feelings and behavioral intentions. It suggests future research apply non-linear models to a variety of health communication areas.