Research shows that people engage in "risky" sex with "safe" partners and in "safer" sex with "riskier" partners. How is the determination of "risky" or "safe" status made? Factorial survey methodology was used to randomly construct descriptions of romantic partners based on attractive and/or risky characteristics. Respondents evaluated 20 descriptions for attractiveness, health risk, likelihood of going on a date, likelihood of unprotected sex, and likelihood of STD/HIV infection. Respondents were most attracted to and perceived the least risk from attractive descriptions and were least attracted to and perceived the most risk from the risky descriptions. The differences between the "conflicting information" descriptions are attributable to a primacy effect: descriptions that began with attractiveness information but end with risk information were evaluated more positively than those that began with risk and ended with attractive information.