Altering sexual prototypes via prevalence information: an experimental analogue to a sexual intervention program

Abstract

iv INTRODUCTION 1 METHOD 29 RESULTS 38 DISCUSSION 57 NOTES 72 APPENDDCA: ROSENBERG SELF-ESTEEM INVENTORY 73 APPENDIX B: PROTOTYPES 74 APPENDIX C: BACKGROUND INFORMATION 76 APPENDIX D: PREVALENCE INFORMATION 78 APPENDIX E: INTENTIONS AND WILLINGNESS 79 APPENDIX F: AUDIO TAPE TRANSCRIPTS 81 APPENDIX G: TAPE EVALUATIONS 87 APPENDIX H: RECALL TESTS 88 APPENDIX I: RESEARCH SCRIPT 90 REFERENCES CITED 93 Past research has established that the images or prototypes that people hold of the type of person who engages in risk behaviors influence then: actions (Gibbons & Gerrard, 1995; in press). However, to date, no experimental studies have been conducted to determine if these images can be altered, and, if so, what effect this has on individuals' behavioral intentions (BI) and behavioral willingness (BW) to engage in risk behaviors. The current study was designed to attempt to modify images of the type of person who uses a condom (condom user [CU] prototype) and of the person who has multiple sexual partaers (casual sex [CS] prototype) by providing information about prevalence rates of these behaviors. Participants first completed measures of their sexual cognitions at the beginning of the semester, and were then recruited to participate in an experimental session. Specifically, 230 female undergraduates were assigned to one of three prevalence information manipulations: an anecdotal audio tape condition (a fictitious tape consisting of five female undergraduates discussing their sexual attitudes and behaviors), a statistical audio tape condition (an announcer reading summary statistics of campus sexual attitudes and behaviors), or a control condition (no audio tape). It was predicted that anecdotal information would have a greater impact on prototypes, BW, and BI than would statistical prevalence information. Moreover, it was predicted ±at self-esteem and sexual risk behaviors (e.g., lack of condom use, multiple partners) would moderate change in prototypes, BI, and BW. Specifically, it was proposed that high-risk, low self-esteem (LSE) individuals would be the most persuaded by the information and would decrease the favorability of the CS prototype, increase the favorability of the CU prototype, and decrease their risky, sexual BW and BI (in the current study this change was labeled the response). It was also hypothesized that LSE individuals who engage in lowrisk sexual behaviors would respond by either not altering their cognitions or perhaps slightly modifying their cognitions in a direction opposite of the appropriate response. Conversely, high self-esteem (HSE) individuals who are not engaging in risky sexual behaviors will significantiy alter their cognitions in a direction congruent with the appropriate response. It was further predicted that HSE individuals who engaged in risky behaviors would react against the prevalence information which suggested that their behavior was inappropriate (Gerrard, Kurylo, & Reis, 1991; Gibbons, Eggleston, & Benthin, in press). Specifically, it was expected that high-risk HSE individuals would respond to the information by not altering their prototypes, BW, or BI. Results indicated that the prototypes were malleable. The CU prototype tended to become more favorable over time and the CS prototype showed a decline in favorability from the beginning of the semester to the experimental session. No main effect for type of prevalence information (anecdotal versus statistical) was detected. Selfesteem and sexual risk levels did moderate the change in prototypes and BW as expected. Future directions, theoretical connections, and implications for sexual interventions are discussed.