Give the recent increase in AIDS and HIV infection, there has been a renewed interest in measuring aspects of sexual behavior that are directly relevant to the possible spread of sexually transmitted diseases. Most research relies on self reports of contraceptive and sexual behavior. Despite the primary importance of these measures, there have been relatively few studies on their validity. Self reports of single events are the result of complex psychological processes that may produce biased judgements. The present research will evaluate the accuracy of self report measures of risk behavior and will identify time intervals that are necessary for isolating stable individual differences in risk behavior. The research will analyze accuracy as a function of different question formats and will identify circumstances under which one question format may be superior or inferior to another. On a more theoretical level, the present research will explore an explanatory model of risk behavior that is distinct from traditional models currently receiving attention. Most of these models focus on individual difference variables in the context of cross sectional, correlational designs. The present research will yield data that are distinct from traditional data collection efforts and will permit a careful analysis of more transitory, situational-based influences on risk behavior and how these may lead an individual to avoid risk behavior one week but perform risk behavior another week. Such weekly fluctuations cannot reflect the influence of stable individual difference variables, because the fluctuations occur for a single individual and the influence of the """"""""stable"""""""" variable is the same from week to week. The present study will interview approximately 480 individuals about their risk behavior every week for one year. At 1, 3, 6, 12 months into the project, individuals will recall selected aspects of risk behavior since study inception. Self reports of behavioral patterns will be compared with aggregates of the weekly scores to gain insights into recall accuracy. Numerous control conditions are included to evaluate monitoring and testing effects.
