Health Journal of Research ©2009 by the ICHPER•SD Volume IV, No. 2, pp. 6-9 Predicting HIV/AIDS-Related Risk Behavior Among Men Who Have Sex With Men: An Examination of Psychosocial Perspectives by Todd M. Sabato Men who have sex with men (MSM) continue to be disproportionately affected by the HIV/AIDS epidemic. They represent more than half of all persons living with HIV and over 60% of HIV-infected men. Although as a group MSM no longer account for the majority of new HIV infections, they are estimated to account for 43% of all new infections — more than any other group. Using Social Cognitive Theory as a conceptual framework, this study examined the influence of both psychosocial and demographic features on the sexual risk- taking behaviors of self-identified MSM. AIDS has taken a tremendous toll in the United States. Since in men, 58% were estimated to be in men who have sex with men, the first case was identified in 1981, 956,019 cases of AIDS and an additional 7% were estimated to be among MSM/injection have been reported (Centers for Disease Control and Prevention drug users (CDC, 2006a). Although rates of new HIV infection [CDC], 2006a), and approximately 550,000 Americans have died declined among MSM in the United States between the mid-1980s (CDC, 2006a) from the disease. The disease continues to have and the mid-1990s, (Institute of Medicine, 2000; Quan et al., a tragic impact, not only on those who are living with Human 2000), data suggests that rates of new infections may be increasing Immunodeficiency Virus (HIV) infection, but also on the many (McFarland et al., 2000; San Francisco Department of Public friends, families, and entire communities that have been forever Health and AIDS Research Institute, 2000). Similarly, studies changed by the epidemic. An estimated 1.7 million people in conducted by the Centers for Disease Control, which examined the United States have been infected with HIV (Henry J. Kaiser anonymous blood samples of over 40,000 high-risk patients of all Family Foundation, 2006), and another 426,000 people are living ages, confirms this trend (Linley, 2002). The study found the rate with AIDS (CDC, 2006a). of new infections for MSM to be nine times higher than for women The epidemic continues to disproportionately affect men who and heterosexual men. New AIDS cases related to MSM have also have sex with men (MSM). MSM have taken great strides to increased each year since 2000, rising a total of 10.4% in that time protect themselves from HIV infection; however, as is the case (Health Resources and Services Administration, 2005). in many areas of health related behavior change, MSM have been Reducing the risk of HIV/AIDS among MSM requires unsuccessful in maintaining protected sexual behaviors over long identifying the psychosocial and behavioral factors, such as periods of time (Becker & Joseph, 1988; Ekstrand & Coates, 1990; alcohol and drug use, that are associated with HIV (Avins et Kelly et al., 1990b; Prieur, 1990; Roffman, Gillmore, Gilchrist, al., 1994) and HIV-related sexual risk-taking (Wingood & Mathias, & Krueger, 1990; Stall, Ekstrand, Pollack, McKusick, DiClemente, 1998). Further review of the HIV/AIDS literature & Coates, 1990; Valdiserri et al., 1988). An estimated 365,000 to has identified psychosocial factors that have been shown to 535,000 MSM in the United States are infected with HIV (Karon differentiate between gay and bisexual men who report lower HIV et al., 1996), representing nearly half of all persons living with risk behaviors and those who do not. These factors include age, HIV and approximately 61% of HIV-infected men (CDC, 2006a). substance use, depression, internalized homophobia, and exposure given that only an estimated 5% to 7% of American men have to the HIV/AIDS culture. More specifically, evidence suggests had sex with another man during adulthood (Binson et al., 1995; that, similar to other groups, greater use of alcohol and illicit Rogers & Turner, 1991), these figures are overwhelming. As a substances (Kalichman, Heckman, & Kelly, 1996; Robins, Dew, group, MSM account for 45% of all new infections °V more than Kingsley, & Becker, 1997) are key social factors associated with any other group (CDC, 2006b). Of AIDS cases diagnosed in 2005 high-risk sexual behavior among gay and bisexual men. There is also evidence that internalized homophobia is an important psychological predictor of unprotected anal intercourse among gay Submitted: 04/19/2009 and bisexual men (Folkman, Chesney, Pollack, & Phillips, 1992; Accepted: 06/15/2009 Meyer, 1995; Meyer & Dean, 1995). Few studies, however, have examined the impact of control, perception, and attribution on HIV Todd M. Sabato, Ph.D. risk-related behaviors. It was therefore essential to address these characteristics, and their potential individual and combined impact 0 Journal of Research Sabato, Todd M. upon HIV/AIDS-related risk sexual behavior. used helplessness theory to explain a variety of human difficulties, including academic failure and premature death. Perhaps the best Locus of Control. Previous studies indicate that external known of these studies is Seligman's work (1972, 1975) and his locus of control predicts engagement in high-risk sexual behavior suggestion that learned helplessness may model depression with (Aspinwall, Kemeny, Taylor, Schneider, & Dudley, 1991; Kelly, respect to symptoms, causes, preventions, and cures. Symptoms St. Lawrence, & Brasfield, 1991; Kelly et al., 1990a). Using of helplessness include passivity, cognitive deficits, emotional Bandura's (1977) self-efficacy framework as a predictor of AIDS deficits (including sadness, anxiety, and hostility), a lowering of risk reduction behavior, Aspinwall et al. (1991) found that self- aggression, a lowering of appetite, neurochemical deficits, and an reported self-efficacy regarding consistent condom use, perceived increase in susceptibility to disease (Peterson & Seligman, 1984). risk, response efficacy, and prior sexual behavior accounted for Two problems emerged with the original helplessness model 70% of the variance in the total number of sexual partners and the when applied to human helplessness in the laboratory and to number of anonymous partners over a 6-month interval. Kelly human depression (Peterson & Seligman, 1984). First, the model et al. (1991) conducted a 16-month longitudinal study to assess failed to differentiate boundary conditions for the generalizability change in high-risk sexual behavior after providing subjects of helplessness. Sometimes helplessness generalized to novel with educational interventions. Resumption of high-risk sexual situations (Hiroto & Seligman, 1975), and sometimes it was behavior was significantly associated with the belief that HIV circumscribed or limited to the laboratory setting (Cole & Coyne, infection was determined by external factors, such as chance and 1977). Additionally, aversive events inconsistently precipitated luck (Kelly et al., 1991). Such results support the findings of St. depressive reactions (Lloyd, 1980). Lawrence, Jefferson, Alleyne, and Brasfield (1995) that decreases Second, the original model of helplessness did not explain in high-risk sexual activity were associated with an internal locus the loss of self-esteem among depressives, frequently observed of control. by Beck (1967). Abramson and Sackeim (1977) found that In a second study, Kelly et al., (1990a) found that high-risk depressed individuals blamed themselves for events even when takers were more likely to view their health as being attributable they perceived no control over events. This finding replicated to chance or luck, and less likely to perceive themselves as having individuals' perceptions of uncontrollability previously observed control over their well-being. Participants were more likely to by Maier et al. (1969) and Seligman et al. (1971). However, adopt a passive stance toward taking care of themselves, thereby learned helplessness theory did not adequately address how causal demonstrating an external locus of control. Similarly, Powell, explanations for aversive events influenced self-esteem. Dolan, and Wessely (1990) found that, among subjects with chronic To address the above mentioned shortcomings, Abramson, fatigue syndrome, external locus of control was significantly Seligman, and Teasdale (1978) revised helplessness theory to correlated with pessimistic attributional style, lower self-esteem, include causal explanations for bad events. According to this and greater guilt. revision, when people face uncontrollable bad events, they ask why. The answers that they tell themselves affect how they Learned Helplessness and Attributional Style. Although react to the events. Abramson et al. (1978) argued that three attributional style has not been directly linked to high-risk sex, explanatory dimensions are relevant. First, the cause may be the present research focused on whether a pessimistic attributional explained at something about the person (internal explanation) style would predict engagement in risky behavior. The construct of or it may be something about the situation or circumstance learned helplessness is prerequisite to understanding attributional (external explanation). Second, the cause may be a factor that style. persists across time (stable explanation) or it may be transient Learned helplessness theory was first described by animal (unstable explanation). Third, the cause may affect a variety of researchers (Overmeir & Seligman, 1967; Seligman & Maier, outcomes (global explanation) or it may be limited to the event of 1967). Dogs who received inescapable shock showed striking concern (specific explanation) (Abramson et al., 1978; Peterson & motivational, cognitive, and emotional deficits and appeared Seligman, 1984). helpless when given future opportunities to escape shock in the Whatever particular explanation is made, humans tend to laboratory (Peterson & Seligman, 1984). Such deficits typically habitually choose certain kinds of explanation for both good and are interpreted in cognitive terms (Maier, Seligman, & Soloman, bad events (Peterson & Seligman, 1984), a phenomenon termed 1969; Peterson & Bossio, 1989; Seligman, Maier, & Soloman, attributional style (Abramson et al., 1978). An internal, stable, 1971). That is, during exposure to the shocks, the dogs learned that and global attributional style (i.e., pessimistic attributional style) the shocks were independent of their responses and that, regardless has been empirically demonstrated in depressed individuals of what the dogs did or didn't do, the shocks occurred. This (Eaves & Rush, 1984; Persons & Rao, 1981; Raps, Peterson, learning has been explained as an expectation of future response- Reinhard, Abramson, & Seligman, 1982). Furthermore, causal outcome independence (i.e., the perception of uncontrollability) explanations for an event and expectations about the consequences that generalized to new situations to produce the above-mentioned of an event have similar properties (that is, internal, stable, and deficits (Peterson & Seligman, 1984). global) (Peterson & Seligman, 1984). Knowing an individual's Psychologists have linked the learned helplessness construct attributional style does, therefore, assist clinicians in predicting to a variety of failures in human action. Wortman and Brehm helplessness deficits, much the same as other known risk factors (1975) first investigated the basic helplessness phenomenon in for psychological and physical illnesses (Peterson, 1988; Peterson, the laboratory with human subjects. Garber and Seligman (1980) Maier, & Seligman, 1993; Peterson & Seligman, 1984, 1987). volume 4, issue 2  Sabato, Todd M. Peterson (1982) suggested that perceived control was central to and bad events (Peterson & Seligman, 1987). An internal, stable, health and wellness, and Peterson and Seligman (1987) suggested and global causal explanation is the most debilitating account of six possible pathways by which attributional style may mediate bad events (Abramson et al., 1978) and is important because it illness. First, individuals who offer internal, stable, and global affects an individual's expectations about control and self-efficacy explanations may become passive in the face of disease. Failure (Peterson & Seligman, 1987). to seek or follow medical advice are considered two helpless responses that may exacerbate illness (Seligman, 1975). Second, Self-Efficacy and Regulatory Control individuals with a pessimistic attributional style may neglect the Self-regulation is extremely important for directed behavior basics of health care, because they see no connection between change in that it allows the gradual substitution of internal controls anything they might do and the onset or progression of illness for external controls of behavior. This substitution is gradual, (Becker, 1974). however, and is not guaranteed. Several factors may hinder an Third, people who offer internal, stable, and global explanations individual's ability to exhibit regulatory control, including a lack for bad events may not be good problem-solvers following of self-efficacy, self-control, and one's emotional state. Such uncontrollability (Alloy, Peterson, Abramson, & Seligman, 1984). hindrances often lead to risk behaviors. Fourth, loneliness may be a possible factor linking attributional style Loss of self-regulatory control, characterized by a lack of self- to illness (Anderson, Horowitz, & French, 1983). The individual efficacy and self-control or locus of control, has been found to who makes internal, stable, and global explanations for bad events be a key indicator of HIV risk behaviors such as substance use is socially withdrawn. Since supportive contacts with others may and high-risk sex. Previous studies indicate that external locus buffer illness (Cobb, 1976), the path from attributional style to of control predicts engagement in high-risk sexual behavior illness may lead through social withdrawal and loneliness. (Aspinwall et al., 1991; Kelly et al., 1991; Kelly et al., 1990a). Fifth, depression may be a link between attributional style and High-risk takers have been found to be more likely to view their illness. Schleifer, Keller, Sirus, Davis, and Stein (1985) found health as being attributable to chance or luck, and less likely to that depressed individuals were at increased risk for disease. perceive themselves as having control over their well-being. Finally, pessimistic attributional style may affect an individual's Further, decreases in high-risk sexual activity have been associated physiology. Rodin and Langer (1977) found that the strongest with an internal locus of control (St. Lawrence et al., 1995). The correlates of immunosuppression among the institutionalized results of such research lend credence to the foundation of both aged were stressful life events and the individual's sense of no Learned Helplessness Theory and Social cognitive Theory — that control over these events. More recently, negative attributional expectations of behavior, and subsequent action, can be shaped by style predicted faster immune decline in HIV-positive gay men perception, as well as by causal explanation. In short, perception (Segerstrom, Taylor, Kemeny, Reed, & Visscher, 1996). Internal, and causality impact an individual's expectations about control and stable, and global attributions for negative events accounted for self-efficacy. 8% of the variance in T-cell decline after controlling for all other An existing link between self-regulatory control, self-efficacy, variables. and depression may lead one to engage in unhealthy behaviors. Self-regulatory systems mediate external influences and provide a A Model of High-Risk Sexual Behavior in HIV-Negative Gay basis for purposeful action, which allows for personal control over Men thoughts, feelings, motivations, and actions (Bandura, 1989). A lack The basics of learned helplessness theory (Seligman, 1975) of self-regulatory control is indicated by decreased self-motivation refer to several related cognitive and behavioral aspects of and subsequent self-efficacy, both of which are symptoms of clinical personality which lead to inappropriate and maladaptive passive depression (American Psychiatric Association, 1994). Adults who behavior. People or animals are considered helpless if, through suffer from anxiety and depression are much more likely to engage their inactivity, they fail to control outcomes that are objectively in high-risk activities such as prostitution, both injection and non- responsive to their actions (Hiroto & Seligman, 1975; Overmier & injection drug use, and choosing high-risk partners (Stiffman et al., Seligman, 1967; Seligman & Maier, 1967). Learned helplessness 1995). In addition, depression has been significantly correlated also refers to the role of noncontingent events in producing with risky behavior among both men and women at risk for HIV passivity and to the cognitive mediation of passivity (Overmier & infection (Valente et al., 1993). Seligman, 1967; Peterson & Seligman, 1987). During exposure to noncontingent events, helpless people learn that outcomes occur Social cognitive Theory and High-Risk Behaviors independently of their behavior. Regardless of what they do or While the model of learned helplessness contributes greatly to the do not do, the events take place. They come to expect that future understanding of human behavior in response to uncontrollability, events also will be uncontrollable and this explanation leads to it is essential to examine the impact of numerous concepts over helpless behavior (Peterson & Seligman, 1987). which an individual may exercise personal control. Both personal The reformulated model of learned helplessness (Abramson and environmental factors may provide cues to behavior. The et al., 1978) explains observed human variation in response Social Cognitive Theory addresses such psychosocial dynamics to uncontrollability. This model suggests that individual's influencing health behaviors. expectations of future helplessness are shaped first by their The Social Cognitive Theory defines human behavior as a triadic, perception that events are noncontingent and second by their dynamic, and reciprocal interaction of personal factors, behavior, causal explanation, or habitual attributional style, for both good and the environment (Bandura, 1977, 1986, 1989). According 2 Journal of Research Sabato, Todd M. to the theory, an individual's behavior is uniquely determined by response to a demographic item addressing the respondent's each of these three factors. While the Social Cognitive Theory sexual preference. This definition is consistent with the Centers upholds the notion that response consequences mediate behavior, for Disease Control's surveillance and data reporting techniques it contends that behavior is largely regulated antecedently through (CDC, 1999). While all men were encouraged to complete the cognitive processes. Therefore, response consequences of a questionnaires, only those who identified as MSM were included as behavior are used to form expectations of behavioral outcomes. part of the data analysis. Responses of men who self-identified as It is the ability to form these expectations that gives humans the only having sex with women were excluded from data analysis. capability to predict the outcomes of their behavior before the behavior is performed. In addition, the Social Cognitive Theory Recruitment posits that most behavior is learned vicariously. Participants were recruited in collaboration with a regional not- Based upon the premise that unhealthy behaviors are maintained for-profit organization whose purpose is to provide that affected through periodic social reinforcement, environmental cures, and by HIV with information and support needed to make well- in some cases psychological and physiological reinforcement, informed choices about their lives through advocacy, education, numerous explanations exist for the continued HIV-related risk peer support, and activism. Data collection took place at an annual behaviors that occur among men who have sex with men. Figure fundraising event that promotes unity, visibility, and self-esteem 1 briefly outlines these interactions and their resulting health among the sexual minority community. Trained data collectors consequences. were stationed at data collection points located near the entrance to the festival, as well as high-traffic areas throughout the venue. Instrumentation Demographic Questionnaire. The 10-item demographic questionnaire solicited information on age, race, educational level, employment and socioeconomic status, experience with counseling/ psychotherapy, and participant's history of HIV testing. Attributional Style. The Attributional Style Questionnaire (ASQ) (Peterson et al., 1982) was administered to all participants. The ASQ yield scores for the exploration of events for internal versus external, stable versus unstable, and global versus specific causes, and has been found to correlate positively with self- esteem and the symptoms of learned helplessness (Peterson & Seligman, 1984). The instrument described four hypothetical events and asked participants to generate their own cause for each event, and to rate themselves along a 7-point scale. The three Several studies among gay men have reported significant attributional dimensions (internality, stability, and globality) are associations between high-risk sexual behavior and external locus associated with each event, thereby producing a total of 12 items. of control (Aspinwall et al., 1991; Kelly et al., 1990a; Kelly et al., Participants ranked themselves on each item in the direction of 1991) and internalized homophobia (Folkman et al., 1992; Meyer, increasing internality, stability, and globality (i.e., 1=lowest and 1995; Meyer & Dean, 1995). As well, both alcohol and drug use 7=highest). The sum of the items represented each person's score have predicted high-risk sexual behavior in the general population on this scale. In previous research (Peterson & Seligman, 1984), (Shillington, Cottler, Compton, & Spitznagel, 1995) and among Cronbach's alpha ranged from .72 to to .75. Construct validity has gay men (Mulry, Kalichman, & Kelly, 1994). Depression has also ranged from .71 to .79 (Peterson et al., 1982). In the present study, been found to be predictive of engagement in high-risk sexual reliability for the ASQ was .86 behavior in HIV-negative men (Perkins, Leserman, Murphy, & Sexual Regulation / Locus of Control. The Dyadic Sexual Evans, 1993). Although the psychosocial factors discussed above Regulation scale (DSR) items were developed from open-ended have been examined separately as predictors of mood state, self- interviews with heterosexual and homosexual couples (Catania, esteem, and high-risk sexual behavior, there has been no integrated McDermott, & Wood, 1984). Each of the scale's 11 items were examination of these factors in a single study. The present study assessed on a 7-point Likert-type scale. Concurrent validity of the integrated those factors, along with conceptually-based predictors DSR is evidenced by results showing that the DSR is significantly (attributional style) to examine their respective influences on high- related to dyadic sexual behaviors, cognitions, and affect. risk sexual behavior. Increasing perceptions of internal control, as measured by the DSR, have been found to be associated with increasing frequencies of Methodology intercourse, oral sex from partners, sexual relations (an aggregate Sample of sexual behaviors), affectionate behaviors, increasing levels of The sample consisted of self-identified men who have sex with dyadic sexual satisfaction, and decreasing levels of dyadic sexual men. For research purposes, "men who have sex with men" was anxiety. The consistency of these findings indicates that the DSR operationalized as men who either (a) have sex only with men, is reliably sensitive to the reward properties of dyadic sexual or (b) have sex with both men and women, as indicated by their situations and also indicates that perceptions of control may play a volume 4, issue 2  Sabato, Todd M. role in attaining dyadic sexual satisfaction. Reliability coefficients as unprotected anal intercourse. As such, only participants who have previously ranged from .74 to .83 (Catania, McDermott, & engaged in unprotected anal sex would be deemed to be high-risk Wood, 1984), and was .78 in the present study. takers. High-Risk Sexual Behavior. High-risk sexual behavior was Multi-Indexed Definition of Risk. To broaden the definition of assessed using the UCSF Center for AIDS Prevention Studies high-risk sexual behavior in accordance with the continuum of Sexual Behaviors Questionnaire (Chesney, 1997). This 20-item behavior articulated earlier, a second scoring system was developed scale asked participants to report the number of men they had by the author and used with participants' responses. In this system, had sex with over the preceding three months. Participants were participants were scored dichotomously on all items involving also asked to respond yes or no to questions about condom usage, unprotected oral and anal sex. The sum of the 18 items constituted insertive anal / receptive anal sex, and insertive/receptive oral sex. each participant's total score for high-risk sexual behavior, ranging As described below, two different scoring systems were used to from 0 to 18. The scoring system preserved all of the information measure sexual behavior. obtainable from the 20-item scale. UCSF-Defined Risk. Following the scoring system developed In addition to the constructs addressed above, a series of by the UCSF Center for AIDS Prevention Studies, participants questionnaires were also utilized to gain a proxy measure of were ranked on a 4 point scale based on their responses to 12 items substance use (Brief Michigan Alcohol Screening Test, α = .72) regarding anal intercourse only. Scores could range from 0 to 32. (Pokorny, Miller, & Kaplan, 1972), HIV risk knowledge (Brief This scoring system measures high-risk behavior defined strictly HIv Knowledge Questionnaire, α = .88) (Carey & Schroder, 2002), depression (Center for Epidemiologic Studies Depression Scale, α = .90) (Radloff, 1977), and internalized homophobia (Internalized Table 1. Demographic Information of Self-Identified Men Who Have Sex With Men (n=223) Homophobia Scale, α = .94) (Martin & Deam, 1988). Frequency %* Results Age (range = 18 – 62 years) M = 32.8yrs SD=10.7yrs Descriptive Statistics Race/Ethnicity A total of 225 men completed the paper-and-pencil survey. Two Caucasian 121 55.8 subjects were excluded from the analyses because of a failure to African American 71 32.7 meet one of the study's inclusion criteria of sexual preference (male Other 25 11.5 or both male and female). The final sample therefore consisted of Level of Education 223 self-identified men who have sex with men. Table 1 describes the demographic information provided by these men. High school graduate or less 46 21.1 Some college 58 26.6 College degree 114 52.3 Sexual Regulation. Dyadic Sexual Regulation scores were computed by summing across items so that higher scores indicated Employment Status a greater degree of external control. The range of possible Not employed 36 16.5 scores was between 11 (internal responses for each item) and 55 Employed 182 83.5 (external responses for each item). The mean score was 29.5, an Previous Year’s Income indication that participants were neither internally nor externally $15,000 or less 52 24.0 regulated, but rather a combination of both factors. Two-thirds $15,001 - $34,999 74 34.1 (67.6%) of the respondents "agreed" or "strongly agreed" with the $35,000 or greater 91 41.9 statement "During some sexual encounters I find it pleasurable to Sexual Preference be passive while my partner is the active person." Greater than Sexual activity with men only 177 81.2 half of the participants responded "agree" or "strongly agree" Sexual activity with men & women 42 18.8 with the statements "When I am not interested in sexual activity History of Counseling I feel free to reject sexual advances by my partner." (52.5%), "I Has previously received counseling 94 43.3 often take the initiative in beginning sexual activity." (55.7%), Has not previously received counseling 123 56.7 and "I find it pleasurable at times to be the active member during Length of Time Since Previous HIV Test sexual relations while my partner takes a passive role." (59.4%). Has never been tested for HIV 54 25.0 Respondents were most opposed to the statement "If my sexual Has been tested for HIV in past relations are not satisfying there is little I can do to improve the 9 months 84 38.9 situation." More than six out of every ten participants (61.2%) Has been tested for HIV more than "disagreed" or "strongly disagreed" with this statement. 9 months ago 78 36.1 Attributional Style. The Attributional Style Questionnaire Number of HIV Tests in Last Five Years yielded scores for the exploration of events for internal versus Has not been tested in last five years 61 28.6 external, stable versus unstable, and global versus specific causes. 1 – 5 times in last five years 117 54.9 Participants ranked themselves on each item in the direction of More than five times in last five years 35 16.4 increasing internality, stability, and globality, with possible scores ranging from 12 to 84. The mean score on the questionnaire was * Percentages may not add up to 100 due to missing data 49.3, indicating a trend toward internality, stability, and globality, 4 Journal of Research Sabato, Todd M. the three elements that contribute to pessimistic attributional examine the hypothesis that levels of HIV/AIDS-related sexual style. risk taking could be explained by any of the independent variables Sexual Behavior. As discussed previously, two different (attributional style, depression, sexual regulation, alcohol use, scoring systems were used to measure sexual behaviors. Following internalized homophobia, drug use, and HIV transmission the scoring system developed by the UCSF Center for AIDS knowledge). Participants were classified as practicing low- Prevention Studies, participants were ranked on a 4 point scale or high-risk sexual behaviors based on their Sexual Behavior based on their responses to 12 items regarding anal intercourse Questionnaire scores. Participants whose scores were one only. Responses to each of the items were summed to create a standard deviation above the sample mean were defined, a priori, composite score for sexual risk. Scores could range from 0 to 32. as engaging in high-risk sex. On both measures, twelve to thirteen Again, this scoring system measured high-risk behavior defined percent of the sample indicated that they had engaged in high-risk strictly as unprotected anal intercourse. As such, only participants behaviors over the past three months. who engaged in unprotected anal sex were deemed to be high risk Two separate t-tests were calculated for each independent takers. The mean score was 7.1 variable, in order to address both the UCSF and Multi-Indexed A second system of scoring broadened the definition of high-risk definitions of risk. Heavy alcohol use was consistently associated behavior. The Multi-Indexed scoring system scored participants with high-risk sexual behavior (p<.02 for both the UCSF and dichotomously (0 = no, 1 = yes) on all items involving unprotected Multi-Indexed definitions). Depression was also associated with oral and anal sex. That is, participants who indicated engaging high-risk behavior. Individuals who were classified as engaging in either behavior without a condom received one point for each in high-risk sex reported significantly higher levels of depression behavior. The mean score was 4.6. when compared to their lower-risk counterparts. (p<0.05 for the UCSF measure). Associations Among Independent Study Variables Similar to the findings of the Pearson's correlation, persons Pearson's correlations were utilized as a means to examine the engaging in higher-risk sexual activities indicated more pessimistic relationship between each of the study variables. Many of the attributional tyle (p<0.001 for both measures of risk). Although correlations proved to be statistically significant, with the vast attributional style has yet to be directly linked to high-risk sex, the majority being moderately positive. present results suggest that a pessimistic attributional style would Attributional Style. As seen in previous research, internal, predict engagement in risky behavior. stable, and global attributional style (i.e., pessimistic attributional style) was positively correlated with many psychosocial factors Modeling High-Risk Sexual Behavior related to increased HIV/AIDS risk. Most notably, a pessimistic Using correlation coefficients, relationships were examined attributional style was associated with increased levels of between (a) depression (b) internalized homophobia (c) internality/ internalized homophobia (r=.22, p<0.01) and depressive symptoms externality (d) substance use (e) HIV transmission knowledge, and (r=.27, p<0.01). Further, pessimistic attributional style was (f) an observed, dependent variable (high-risk sexual behavior). positively associated with both alcohol and illicit drug use (r=.21, The hypothesized model, presented in Figure 2, illustrates the p<0.01 and r= .20, p<0.01, respectively). hypothesis that depression, internalized homophobia, internality/ Depression. Despite the relatively low levels of depression externality, substance use, and HIV transmission knowledge exhibited by respondents (M=7.1, SD=5.7), many associations directly affect high-risk sexual behavior. were found between depression and other psychosocial variables. Levels of depression were positively correlated with internalized homophobia (r=.39, p<0.01), suggesting that respondents with higher levels of depression also experienced a greater degree of internalized homophobia. Further, levels of depression were positively associated with alcohol use (r=.43, p<0.01), implying that persons with more depressive symptoms may consume greater quantities of alcohol. A very strong correlation was found between depression and levels of sexual regulation, as measured by the Dyadic Sexual Regulation Scale (r=.50, p<0.01), suggesting that depressed persons are less willing exert personal control when engaging in sexual behaviors. More specifically, the finding draws a relationship between depression and external control, similar to the previous finding between depression and pessimistic attributional style. A relationship also existed between depression and HIV/AIDS- A hierarchal multiple linear regression was conducted in which related sexual behavior, measured by the Multi-Indexed definition all variables were simultaneously entered as factors predicting the of risk (r=.16, p<.05). dependent variable, over and above the effect of the demographic variables. Two separate analyses were conducted given the two High-Risk Sex as a Function of Psychosocial Variables scoring methods used to compute the dependent variable. For A series of independent sample t-tests were conducted to each dependent variable, the hypothesis was tested that men who volume 4, issue 2  Sabato, Todd M. exhibited greater levels of depression, internalized homophobia, attributional style would predict engagement in risky sexual substance use, externality (i.e., pessimistic attributional style and behavior. A latent variable, internality/externality, was composed external sexual control), and lower levels of HIV transmission of two indicators (DSR and ASQ). While attributional style knowledge would report higher engagement in high-risk sexual was found to be a stronger indicator of internality / externality, behavior. the combination of the two indicators nonetheless proved to be a UCSF Defined Risk. First, the model was tested using the scoring statistically significant predictor of high-risk sex (p<.001 for both system developed by UCSF to define high-risk sexual behavior. measures of high-risk sex). That pessimistic attributional style and As shown in Figure 2, a multiple regression, simultaneously external locus of control predict high-risk sexual behaviors is not entering all indicators and variables, revealed that 31% of the unfounded. Peterson (1982) has suggested that perceived control variance in the dependent variable was accounted for by the is central to health and illness, and Peterson and Seligman (1987) predictors (p,0.001). Variables were then examined for individual propose many possible pathways by which attributional style may significance. As hypothesized, participants who exhibited greater mediate behavior and illness. First, individuals with a pessimistic degrees of externality engaged in riskier sexual behaviors (ß = attributional style may become passive in the face of disease. .60, p,0.001), as did persons engaging in substance abuse (ß = Failure to seek or follow medical advice, such as utilizing HIV .15, p<0.03). Internalized homophobia also predicted a trend in risk-reduction strategies, can be considered a helpless response the same direction (ß = .13, p<0.06). Contrary to prediction, the that can invite or exacerbate illness (Seligman, 1975). Further, remaining variables (depression and HIV transmission knowledge) individuals with a pessimistic attributional style may neglect the did not emerge as significant predictors of high-risk sex. basics of care (such as, for example, using condoms during sexual Multi-Indexed Definition of Risk. As shown in Figure 3, , intercourse), because they see no connection between anything thirty-four percent of the variance in the dependent variable was they might do and the onset of an illness (Becker, 1974). accounted for by the predictors (p<0.001). Internality/externality Loneliness may also be a possible factor linking attributional (ß = .62, p<0.001) and drug abuse (ß = .20, p<0.004) contributed style to behavior or illness (Anderson, Horowitz, & French, 1983). significantly to the proportion of variance in high-risk sexual The individual who makes pessimistic explanations for bad events behavior. Contrary to prediction, the remaining variables did not may be socially withdrawn. Since supportive contacts with others emerge as significant predictors of high-risk sex. may buffer negative behavior and subsequent illness (Cobb, 1976), the path from attributional style to behavior and illness may lead through social withdrawal and loneliness. Finally, depression may provide a link between attributional style, behavior, and illness. More depressed individuals, who may subsequently experience lowered self-esteem and self-efficacy, may not waver in the face of uncontrollability. The significant correlation between depression and pessimistic attributional style (p<.001) in the present study supports such a premise. Schleifer, Keller, Sirus, Davis, and Stein (1985) found that depressed individuals were at increased risk for disease. Further, Kelly, St. Lawrence, and Brasfield (1991) found that high-risk sexual behavior was significantly associated with the belief that HIV infection was determined by external factors, such as chance and luck. Kelly et al. (1990a) also found that high-risk takers were more likely to view health as being attributable to chance or luck, both characteristics of pessimistic attributional style. Continued Discussion research that addresses the collective impact of depression and Sexual behavior, including avoidance of anal intercourse, attributional style on high-risk behaviors may provide greater condom usage, and number of sexual partners may not be stable clarity regarding these possible mediating factors. over time (Abid, Joseph, Ostrow, Tal, & Schwartz, 1991; Joseph, Abid, Loopman, & Ostrow, 1990; Stall et al., 1990). Studies have Recommendations for Practice revealed that MSM engage in different levels of risky sexual activity Many psychosocial variables (e.g., attributional style, control at different times in their lives and also have shown that there are factors, and substance abuse) are of importance in understanding fluctuations with regard to the number of sexual partners MSM the continued practice of HIV/-related high-risk sexual behavior. may have over time. To the extent that high-risk sexual behavior Historically, HIV prevention efforts have operated on the is not stable over time, it may reflect a "state" phenomenon. assumption that HIV/AIDS is the preeminent health problem Inconsistent risk-taking may be related to complex strategies of facing at-risk populations, and have generally offered limited risk management based on several characteristics of the individual regard to the other health problems facing such communities. If it and the particular sexual encounter, and may involve both rational is indeed true that the additive effect of interrelated psychosocial and irrational risk calculation (Meyer & Dean, 1995). health conditions increases vulnerability to HIV infection, it may be Although attributional style has not been directly linked to possible to enhance the effectiveness of HIV prevention efforts by high-risk sex, the present study focused on whether a pessimistic working to support a broader health movement within vulnerable  Journal of Research Sabato, Todd M. communities. That is, by working in tandem with organizations Cole, C.S., & Coyne, J.C. (1977). Situational specificity of laboratory- addressing mental health and substance abuse problems within induced learned helplessness. Journal of Abnormal Psychology, 86, 615-623. a community, it may be possible to increase the efficacy of HIV Eaves, G., & Rush, A.J. (1984). Cognitive patterns in symptomatic and prevention efforts and related health promotion efforts. remitted unipolar major depression. Journal of Abnormal Psychology, 93, 31-40. References Folkman, S., Chesney, M.A., Pollack, L., & Phillips, C. (1992). Stress, Abid, S.M., Joseph, J.G., Ostrow, D.G., Tal, M., & Schwartz, S.A. (1991). coping and high-risk sexual behavior. Health Psychology, 11, 218- Relapse in sexual behavior among homosexual men: A 2-year follow- 222. up from the Chicago MACS/CCS. AIDS, 5, 757-760. Garber, J., & Seligman, M.E.P. (Eds.). (1980). Human.Helplessness. New Abramson, L.Y., Seligman, M.E.P., & Teasdale, J.D. (1978). Learned York: Academic Press. helplessness in humans: Critique and reformulation. Journal of Health Resources and Services Administration (2005). Men.who.have.sex. Abnormal Psychology, 87, 49-74. with men and HIV/AIDS in the United States. Rockville, MD: U.S. Alloy, L.B., Peterson, C., Abramson, L.Y., & Seligman, M.E.P. (1984). Department of Health and Human Services. Attributional style and generality of learned helplessness. Journal of Henry J. Kaiser Family Foundation. (2005, September). HIV/AIDS policy Personality and Social Psychology, 46, 681-687. fact sheet. Washington, DC: Henry J. Kaiser Family Foundation. American Psychiatric Association. (1994). Diagnostic.and.statistical. Hiroto, D.S., & Seligman, M.E.P. (1975). Generality of learned manual of mental disorders (4th ed.). Washington, DC: American helplessness in man. Journal of Personality and Social Psychology, Psychiatric Association. 31, 311-327. Anderson, C.A., Horowitz, L.M., & French, R. deS. (1983). Attributional Institute of Medicine. (2000). No time to lose: Getting more from HIV style of lonely and depressed people. Journal of Personality and Social prevention. Washington, DC: National Academy Press. Psychology, 45, 127-136. Joseph, J.G., Abid, S.M. Koopman, J.S., & Ostrow, D.G. (1990). Aspinwall, L.G., Kemeny, M.E., Taylor, S.E., Schneider, S.G., & Dudley, Behavioral change in longitudinal studies: Adoption of condom use J.P. (1991). Psychosocial predictors of gay men's AIDS risk-reduction by homosexual/bisexual men. American Journal of Public health, 80, behavior. Health Psychology, 10, 432-444. 1513-1514. Avins, A.L., Woods, W.J., Lindan, C.P., Hudes, E.S., Clark, W., & Hulley, Kalichman, S.C., Heckman, T., & Kelly, J.A. (1996). Sensation seeking S.B. (1994). HIV infection and risk behaviors among heterosexuals as an explanation for the association between substance use and HIV- in alcohol treatment programs. Journal of the American Medical related risky sexual behavior. Archives of Sexual Behavior, 25, 141- Association,.271, 515-518. 154. Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral Karon, J.M., Rosenberg, P.S., McQuillan, G., Khare, M., Gwinn, M., change. Psychological Review, 84, 191-215. & Peterson, L.R. (1996). Prevalence of HIV-infection in the United Bandura, A. (1986). Social foundations of thought and action: A social States, 1984 to 1992. Journal of the American Medical Association, cognitive.theory. Englewood Cliffs, NJ: Prentice Hall. 276, 126-131. Bandura, A. (1989). Social cognitive theory. In R. Vasta (Ed.), Annals of Kelly, J.A., St. Lawrence, J.S., & Brasfield, T.L. (1991). Predictors of child.development (pp. 1-60). Greenwich, CT: Jai Press. vulnerability to AIDS risk behavior relapse. Journal of Consulting and Beck, A.T., (1967). Depression: Clinical,.experimental,.and.theoretical. Clinical Psychology, 59, 163-166. aspects. New York: Hoeber. Kelly, J.A., St. Lawrence, J.S., Brasfield, T.L., Lemke, A., Amidei, T., Becker, M.H. (1974). The health belief model and personal health Roffman, R.E., et al. (1990a). Psychological factors that predict AIDS behavior. Thorofare, NJ: Slack. high-risk versus AIDS precautionary behavior. Journal of Consulting Becker, M.H., & Joseph, J.G. (1988). AIDS and behavioral change to and Clinical Psychology, 58, 117-120. reduce risk: A review. American Journal of Public Health, 78, 394- Kelly, J.A., St. Lawrence, J.S., Brasfield, T.L., Stevenson, L.Y., Diaz, 410. Y.E., & Hauth, A.C. (1990b). AIDS risk behavior patterns among gay Binson, D., Michaels, S., Stall, R., Coates, T.J., Gagnon, J.H., & Catania, men in small southern cities. American Journal of Public Health, 80, J.A. (1995). Prevalence and social distribution of men who have sex 416-418. with men: United States and its urban centers. Journal of Sex Research, Linley, L. (2002, July). Using STARHS to estimate HIV-1 incidence among 32, 245-254. patients attending STD clinics in selected U.S. cities. Poster presented Carey, M.P., & Schroder, K.E. (2002). Development and psychometric at the XIV International Conference on AIDS, Barcelona, Spain. evaluation of the brief HIV knowledge questionnaire (HIV-KQ-18). Lloyd, C. (1980). Life events and depressive disorder reviewed: I. Events AIDS Education and Prevention, 14, 174-184. as predisposing factors. II. Events as precipitating factors. Archives of Catania, J.A., McDermott, L.J., & Wood, J.A. (1984). Assessment of locus General Psychiatry, 37, 527-548. of control: Situational specificity in the sexual context. Journal of Sex Maier, S.F., Seligman, M.E.P., & Soloman, R.L. (1969). Pavlovian fear Research, 20, 310-324. conditioning and learned helplessness. In B.A. Campbell, & R.M. Centers for Disease Control and Prevention (2006a). HIV/AIDS Church (Eds.), Punishment. New York: Appleton-Century-Crofts. Surveillance Report: Cases of HIV Infection and AIDS in the United Martin, J.L., & Dean, L. (1988). The impact of AIDS on gay men: States.and.Dependent.Areas,.2005. Washington, DC: U.S. Public A.research.instrument. Unpublished technical report, New York: Health Service. Columbia University. Centers for Disease Control and Prevention (2006b). Methamphetamine McFarland, W., Schwarcz, S., Kellogg, T.A., Hsu, L., Kim, A., & Katz, use and HIV risk behaviors among heterosexual men °V preliminary M.H. (2000, June). Implications of highly active retroviral treatment results from five northern California counties. Morbidity.and.Mortality. for HIV prevention: the case of men who have sex with men in San Weekly Report, 55, 273-277. Francisco. Paper presented at the XII International Conference on Centers for Disease Control and Prevention. (1999). Need for sustained AIDS, Durban, South Africa. HIV prevention among men who have sex with men. Retrieved January Meyer, I. (1995). Minority stress and mental health in gay men. Journal of 15, 2002, from http://www.thebody.com/cdc/msm899.html Health.and.Sexual.Behavior,.36, 38-56. Chesney, M.A. (1997). Clinical trial of coping effectiveness training for Meyer, I.H., & Dean, L. (1995). Patterns of sexual behavior and risk taking HIV-positive men. Unpublished manuscript, University of California, among New York City gay men. AIDS Education and Prevention, 7, San Francisco Center for AIDS Prevention Studies. Supplement, 13-23. Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Overmier, J.B., & Seligman, M.E.P. (1967). Effects of inescapable Medicine,.38, 300-314. shock upon subsequent escape and avoidance learning. Journal of volume 4, issue 2  Sabato, Todd M. Comparative and Physiological Psychology, 63, 28-33. Rogers, S.M., & Turner, C.F. (1991). Male-male sexual contact in the Peterson, C. (1982). Learned helplessness and health psychology. Health. USA: findings from five sample surveys, 1970-1990. Journal of Sex Psychology, 1, 153-168. Research, 28, 491-519. Peterson, C. (1988). Explanatory style as a risk factor for illness. Cognitive. San Francisco Department of Public Health and AIDS Research Institute. Therapy and Research, 12, 119-132. (2000). Response to the updated estimates of HIV infection in San Peterson, C. & Bossio, L.M. (1989). Learned helplessness. In R.C. Curtis, Francisco, 2000. Retrieved January 19, 2005, from http://hivinsite. (Ed.), Self-defeating behaviors: Experimental research, clinical ucsf.edu/ari/HIVEstimatesReport8900.pdf impressions,.and.practical.implications. New York: Plenum Press. Schleifer, S., Keller, S., Siris, S., Davis, K., & Stein, M. (1985). Depression Peterson, C., Maier, S.F., & Seligman, M.E.P. (1993). Helplessness: A and immunity. Archives of General Psychiatry, 42, 129-133. theory for the age of personal control. New York: Oxford University Segerstron, S.C., Taylor, S.E., Kemeny, M.E., Reed, G.M., & Visscher, Press, Incorporated. B.R. (1996). Causal attributions predict rate of immune decline in Peterson, C., & Seligman, M.E.P. (1984). Causal explanations as a risk HIV-seropositive gay men. Health Psychology, 15, 485-493. factor for depression: Theory and evidence. Psychological Review, 91, Seligman, M.E.P. (1972). Learned helplessness. Annual Review of 347-374. Medicine,.23, 407-412. Peterson, C., & Seligman, M.E.P. (1987). Explanatory style and illness. Seligman, M.E.P. (1975). Helplessness: On depression, development, and Journal of Personality, 55, 237-265. death. San Francisco: Freeman. Peterson, C., Semmel, A., von Baeyer, C., Abramson, L.Y., Metalsky, Seligman, M.E.P., & Maier, S.F. (1967). Failure to escape traumatic shock. G.I., & Seligman, M.E.P. (1982). The attributional style questionnaire. Journal of Experimental Psychology, 74, 1-9. Cognitive Therapy and Research, 6, 287-299. Seligman, M.E.P., Maier, S.F., & Soloman, R.L. (1971). Unpredictable Pokorny, A.D., Miller, B.A., & Kaplan, H.B. (1972). The brief Michigan and uncontrollable aversive events. In F.R. Brush (Ed.), Aversive. alcoholism screening test. American Journal of Psychiatry, 129, 342- conditioning.and.Learning. New York: Academic Press. 354. St. Lawrence, J.S., Jefferson, K.W., Alleyne, E., & Brasfield, T.L. (1995). Powell, R., Dolan, R., & Wessely, S. (1990). Attributions and self-esteem Comparison of education versus behavioral skills training interventions in depression and chronic fatigue syndromes. Journal of Psychosomatic in lowering sexual HIV-risk behavior of substance-dependent Research, 34, 665-673. adolescents. Journal of Consulting and Clinical Psychology, 63, 154- Prieur, A. (1990). Norwegian gay men: Reasons for continued practice of 157. unsafe sex. AIDS Education and Prevention, 2, 109-115. Stall, R., Ekstrand, M., Pollack, L., McKusick, L., & Coates, T.J. (1990). Quan, V.M., Steketee, R.W., Valleroy, L., Weinstock, H., Karon, J. Relapse from Safer sex: the next challenge for AIDS prevention efforts. & Janssen, R. (2000, June). HIV incidence patterns, trends, and Journal of Acquired Immune Deficiency Syndrome, 3, 1181-1187. association with HIV prevalence in the United States, 1978-1999. Stiffman, A.R., Dore, P., & Cunningham, R.M. (1995). Person and Paper presented at the XIII International Conference on AIDS, Durban, environment in HIV risk behavior change between adolescence and South Africa. young adulthood. Health Education Quarterly, 22, 211-226. Radloff, L.S. (1977). The CES-D Scale: A self-report depression scale Valdiserri, R.O., Lyter, D., Levinton, L.C., Callahan, C.M., Kingsley, for research in the general population. Applied Psychological L.A., & Rinaldo, C.R. (1988). Variables influencing condom use in a Measurement,.1, 385-401. cohort of gay and bisexual men. American Journal of Public Health, Raps, C.S., Peterson, C., Reinhard, K.E., Abramson, L.Y., & Seligman, 78, 801-805. M.E.P. (1982). Attributional style among depressed patients. Journal. Valente, S.M., Saunders, J.M., & Uman, G. (1993). Self-care, psychological of Abnormal Psychology, 91, 102-108. distress, and HIV disease. Journal of the Association of Nurses in AIDS Robins, A.G., Dew, M.A., Kingsley, L.A., & Becker, J.T. (1997). Do Care,.4, 15-25. homosexual and bisexual men who place others at potential risk Wingood, G.M., & DiClemente, R.J. (1998). Partner influences and for HIV have unique psychosocial profiles? AIDS Education and gender-related factors associated with non-condom use among young Prevention, 9, 239-251. adult African-American women. American Journal of Community Rodin, J., & Langer, E.J. (1977). Long-term effects of a control-relevant Psychology, 26, 29-51. intervention with the institutionalized aged. Journal of Personality Wortman, C.B., & Brehm, J.W. (1975). Responses to uncontrollable and Social Psychology, 35, 897-902. outcomes: An integration of reactance theory and the learned helplessness model. In L. Berkowitz (Ed.),. New York: Academic Press. ■  Journal of Research