A survey of heterosexually active college students gathered information about condom use, self-efficacy (SE), outcome expectancies, sexual attitudes, peer group influences, acquired immunodeficiency syndrome (AIDS) knowledge, and perceived vulnerability to AIDS. On the basis of A. Bandura's (1986) social-cognitive theory, a structural model with SE as the central mediator was formulated and evaluated with LISREL. This model explained 46% of the variance in condom use from judgments of SE and effects attributable to peers and 53% of the variance in SE from outcome expectancies and peer group influences. Sexual attitudes, AIDS knowledge, and perceived vulnerability did not predict condom use. Most students were well-informed about human immunodeficiency virus (HIV) transmission but reported not feeling at risk, even though many engaged in risky sexual behavior.
PIP: Psychological factors, associated, with sexual risk behavior were investigated guided by A. Bandura's (1986) social-cognitive theory. It was examined whether a self-efficacy (SE) model is capable of predicting condom use from expectancies, social influences, peer group comparison, sexual attitudes, perceived risk for HIV infection, and AIDS-related knowledge. A self-report questionnaire was developed to gather information. A final sample of 212 single heterosexual and currently sexually active undergraduate students (103 men and 109 women) completed an anonymous survey. The ethnic composition was 83% Caucasian, 8% African-American, 3% Hispanic and 6% other. The state religious affiliation was 28% Catholic, 37% Jewish. 24% Protestant, and 11% other. The mean age of the respondents was 21.4 years (men = 21.8 years; women = 21.1 years). The subjects reported a mean number of 1.8 sex partners during the past 6 months and a mean number of 5.4 sex partners during the past 3 years. Only 29% of the subjects reported consistently having condoms during the past 6 months. 73% claimed pregnancy prevention and only 17% mentioned fear of AIDS. 80% of the students perceived themselves only slightly at risk for HIV infection. The fit of the model was evaluated with LISREL VII. The variables in the model accounted for 46% of the measurement error variance in condom use and for 53% of the variance in SE. Consistent with social-cognitive theory, condom use was predicted by SE beliefs, but it was also significantly predicted by peer group comparison. SE was predicted by self-comparisons with one's peer group and by outcome expectancies: positive expectancies about condom use (disease and pregnancy prevention) were directly related to Se, whereas negative expectancies (reduction of pleasure) were inversely associated with SE. Consistent with predictions, AIDS knowledge was unrelated to condom use. A positive correlation between AIDS knowledge scores and perceived vulnerability to HIV infection was found, but perceived vulnerability was not a significant predictor of either SE or condom use.