The predictive utility of an expanded AIDS Risk Reduction Model (ARRM) among adult gay and bisexual men.
ACKNOWLEDGEMENTS: This research was supported by Behavior Sciences Research in HIV infection, grant # NIMH T32-MH19139 (Zena Stein, M.D., Principal Investigator), while the first author was a Postdoctoral Fellow at the HIV Center for Clinical and Behavioral Studies, New York State Psychiatric Institute and Columbia University, New York. We would also like to acknowledge the support by NIMH center grant 2-P50-MH43520 (Anke Ehrhardt, Ph.D., Principal Investigator) for the research, and that of a Mental Health Clinical Research Center, grant NIMH 30906-13, from the NIMH for the computer analyses.
A model of behaviour change specific to HIV/AIDS, the AIDS Risk Reduction Model (ARRM) (Catania, Coates, Kegeles, Ekstrand, Guydish, & Bye, 1989; Catania, Kegeles, & Coates, 1990) has been proposed to address some of the deficiencies in earlier models of health risk and sexual risk behaviour. The ARRM, as Catania et al. (1989) have noted, represents a hybridization of the Health Belief Model (Becker, 1974; Janz & Becker, 1984), the Theory of Reasoned Action (Ajzen & Fishbein, 1980; Fishbein & Middlestadt, 1989), self-efficacy theory (Bandura, 1977, 1989), diffusion theory (Rogers, 1983), help-seeking models (Fisher, Winer, & Abramowitz, 1983), and the Transtheoretical Model of Change (Prochaska & DiClemente, 1983; Prochaska, Redding, Harlow, Rossi, & Velicer, 1994). The ARRM also shares similarities with the Information-Motivation-BehaviouraI Skills Model of AIDS-preventive behaviour (Fisher, 1997; Fisher, Fisher, Williams, & Malloy, 1994).
Within the ARRM, three stages of behaviour change have been identified: (1) Labelling, that is, recognizing and accurately identifying one's sexual behaviour as putting one at risk for contracting HIV; (2) Commitment, which represents a conscious decision to reduce high risk sexual behaviour and substitute lower risk sexual behaviour; and, finally, (3) Enactment, which encompasses the acts of seeking solutions and practicing new behaviours and dispositions with the end result being sexual behaviour at lowered risk for HIV. Catania et al. (1990b) have proposed sets of variables that they believe relate to specific ARRM stages, and they have also identified several general factors (social norms and aversive emotional states) that are I relevant to all ARRM stages. Variables believed to be specifically associated with the ARRM labelling stage are risk relevant health beliefs, prior experience with health-related outcomes, social norms and networks, and knowledge about the routes of transmission of HIV (Catania et al., 1990b; Catania, Coates, & Kegeles, 1994). Perceptions of enjoyment of high risk sexual activity, self-efficacy at performing safer sex behaviours, cost-benefits analysis of condom use, and reference group norms regarding condom use are believed to be related to the ARRM commitment stage (Catania et al., 1990b; 1994). The ARRM enactment stage is believed to be predicted by adequacy of the sexual communication between partners, particularly the negotiation of condom use between partners (Catania et al., 1994), and by help-seeking to gain information about safer sex practices or to remedy deficient behavioural skills for the maintenance of safer sex behaviour (Catania et al., 1990b).
Two initial studies reported by Catania and associates have provided some empirical support for the ARRM. Data gathered from a sample of 564 San Francisco gay men (Catania et al., 1989) demonstrated an association between acts of unprotected anal intercourse and ratings of sexual enjoyment of safer sex practices. In addition, the men in this cohort had a high incidence of help-seeking for safer sex information from both formal and informal sources (Catania et al., 1989). In a second seminal study of the ARRM, Catania et al. (1994) presented data from a sample of 716 unmarried heterosexual adults with at least one risk factor for contracting HIV. The authors examined variables (including new variables not included in the original theoretical formulation) that were correlated with each of the ARRM labelling, commitment and enactment stages, as well as the ability of early ARRM stages to predict later ARRM stages. Analyses for the commitment and enactment stages were conducted separately for primary and secondary sexual partners, and condom use during vaginal intercourse was used as the measure of the enactment stage. The labelling stage indicator, "feeling at risk for HIV infection", was associated with having a history of genital herpes and believing that people other than they were likely to contract HIV. Perceived condom norms and condom enjoyment ratings were associated with the commitment stage, "making a decision to change one's sexual risk behaviour", for both primary and secondary sexual partners, while the labelling stage was predictive of commitment for secondary partners, but not primary sexual partners. In contrast, a history of genital herpes was related to commitment to use condoms with primary, but not secondary, sexual partners. The enactment stage was predicted by commitment, health protective communication, and sexual enjoyment for both primary and secondary sexual partners. However, social norms regarding condom use were related to enactment for secondary, and not primary, sexual partners (Catania et al., 1994).
More recently, the ARRM has been investigated for its usefulness with populations of male heterosexual intravenous and non-intravenous drug users, HIV-positive heterosexual women, and incarcerated adolescents. Malow and Ireland (1996), following the ARRM framework, examined the psychosocial correlates of 111 heterosexual, non-injection, cocaine-dependent men in treatment. They found that a higher number of sexual partners was associated with greater perceived susceptibility of contracting HIV, lower sexual self-efficacy, higher lifetime incidence of sexually transmitted diseases, and using alcohol or drugs during sex. Longshore, Anglin and Hsieh (1997) found that intentions to share injection equipment less often were associated with perceived risk of HIV infection which, in turn, was predicted by participants' knowledge of HIV transmission, perceived susceptibility to HIV, and perceived peer norms. Kowalewski, Longshore and Anglin (1994) examined ARRM stage-specific predictors of intentions to use condoms using a sample of 161 injection drug users. Labelling oneself at risk for HIV infection, ARRM Stage 1, was predicted by knowledge about HIV for condom users, and ARRM Stage 2, commitment to change behaviour, was predicted by self-efficacy, peer norms regarding condom use, and the perceived pleasure associated with condom use. Kline and VanLandingham (1994) examined the use of condoms with primary partners for a sample of 214 HIV-positive women. These authors found that partner-related factors were most important in predicting consistent condom use. Women who had high perceived power to influence their partner's condom use, had partners who were HIV-negative or had partners who did not want more children, were more likely to use condoms consistently with their partner. Lanier and Gates (1996) used ordered probit analysis to examine data collected from 393 adolescents and found that, generally, the ARRM was a useful theoretical framework for examining HIV risk-related behaviour, but that specific forms of enactment were associated with different predictors, and that previous life experiences should be formally incorporated into the model.
The purpose of the current investigation was to attempt a comprehensive evaluation of the predictive ability of the ARRM and to determine the extent to which proposed stage-specific predictor variables are related to ARRM stages. The ARRM has become one of the most popular theoretical frameworks for evaluating the effects of HIV/AIDS risk reduction interventions (Flowers, Sheeran, Beail & Smith, 1997). Because of the popularity of the ARRM, it is important to critically evaluate some of its theoretical tenets, and consider whether the ARRM requires modification to serve as a comprehensive and valid model for use as a framework to construct program evaluations for HIV/AIDS interventions. To this end, two regression models of each of the ARRM stages were evaluated. First, a standard model was evaluated which included those predictor variables which had been specified by Catania et al. (1989; 1990b). Second, an expanded model was evaluated which assessed the contributions of response bias variables, previous sexual risk behaviour, the personality disposition of sensation-seeking, sex guilt, perceived threat of AIDS, barriers to the performance of safer sex, and difficulties with safer sex performance. These additional predictors were chosen by the authors to capture dimensions of sexual risk behaviour not adequately addressed by the ARRM. A response bias measure was included to examine the potential effects of socially desirable responding, which has long been a prominent concern for the validity of questionnaires that measure personality constructs, attitudes, or particularly sensitive subjects such as sexual behaviour (Paulhus, 1991; Jaccard & Wilson, 1991). Previous sexual risk behaviour was considered important because it was a variable that had been shown to be predictive of current sexual risk behaviour (Aspinwall et al., 1991); it also added a longitudinal dimension to the ARRM. A measure of sensation-seeking was included because it has been shown to be related to sexual behaviour (Zuckerman, 1984), and has been proposed as a confounding personality variable for the assessment of the relationship between alcohol and drug use and high risk sexual activity (Jaccard & Wilson, 1991; Leigh & Stall, 1993). Similarly, sex guilt is relevant because it has been shown to be related to sexual behaviour choices, such as the decision to use contraceptives (Gerrard, 1987), and has been proposed to influence sexual risk-taking behaviour (Jaccard & Wilson, 1991). Measures of barriers to safer sex performance and difficulties with safer sex performance were included to assess obstacles to the acquisition and maintenance of safer sex behaviour (Kelly, St. Lawrence, Betts, Brasfield & Hood, 1990). It was hypothesized that the additional predictors would explain significant amounts of unique variance in the analyses of the ARRM stages. Further, it was expected that the expanded prediction model would explain significantly more variance than the standard model for all ARRM stages. Important new variables found to predict sexual risk behaviour will be considered for their applications to safer sex education and safer sex skills negotiation enhancement.
PARTICIPANTS The 119 participants in the current study are a sub-sample of a convenience sample of an original 207 self-identified gay and bisexual men who had recruited during 1987-1988 in New York City for a five year study of the progression of HIV disease (Gorman et al., 1991). The 119 men in this report (74 HIV+ and 45 HIV-) were those participants still enrolled in the longitudinal study at the time of the seventh biannual assessment (three and one-half years after baseline assessment), and who agreed to complete the series of self-report measures. Since the cohort at that point consisted of 131 men, the 119 participants who completed the measures represented a return rate of approximately 91%. Sociodemographic and ethnic information for this sample of 119 gay and bisexual men indicated that the group was overwhelmingly white and non-Hispanic (94%), approaching their fourth decade of life (M = 39.8 years), single and never married (84.6%), middle class (mean of 50.0 on the Hollingshead Two Factor Index (Hollingshead, 1957)), and well-educated (M = 16.3 years). No significant demographic differences were found between HIV+ and HIV-participants.
SEXUAL RISK BEHAVIOUR ASSESSMENT Participants were examined with detailed assessment protocols in the medical, psychiatric, neurological, neuropsychological, and psychosexual domains at regular six month intervals throughout the course of the longitudinal study (Gorman et al., 1990). Sexual risk behaviour was assessed using the Sexual Risk Behavior Assessment Schedule-Adult-Homosexual Men (SERBAS-HOM) (Meyer-Bahlburg, Ehrhardt, Exner, Calderwood & Gruen, 1988). The SERBAS-HOM is an extensive semi-structured interview schedule designed to cover diverse aspects of sexual risk and non-risk behaviour, sexual orientation identification, partner relationships, drug use accompanying sex, and sexual functioning. A more extensive discussion of the conceptual development, reliability, and interviewer training associated with the SERBAS-HOM can be found in both Meyer-Bahlburg et al. (1991) and Dolezal, Meyer-Bahlburg, Remien and Petkova (1997). The complete current-behaviour version of the SERBAS-HOM which covers the six months preceding the evaluation date takes approximately 50 minutes to administer.
MEASURES DERIVED FROM FACE TO FACE INTERVIEWING WITH THE SERBAS-HOM ARRM Enactment Stage Indicator: Current Sexual Risk Behaviour Index and Previous Sexual Risk Behaviour Index (SRBI-7) The measure of the enactment stage was a composite risk index derived from a subset of sexual activity variables obtained with the SERBAS-HOM. It was measured at a point approximately six months after participants completed the questionnaires for the study. A sexual risk index was chosen as a summary measure because it permitted a parsimonious examination of the relationship of predictors to the criterion.
The sexual risk index used is a slightly modified version of the risk index used by Joseph et al. (1989) with the Chicago Multisite AIDS Cohort study. This particular sexual risk index measure was chosen for several reasons. It is a measure that has been used in several publications (Beltran, Ostrow, & Joseph, 1993; Dolezal et al., 1997; Herek & Glunt, 1995; Ostrow, 1989) and, because it has been used with the Chicago Multisite AIDS Cohort study, there is a large body of data associated with the index. The index has been shown to be related to HIV seroconversion events for a homosexual/bisexual male sample (Ostrow, 1989), has been used to examine the pattern of sexual behaviour change over time (Joseph et al., 1989), and has been used to examine the relationship between receiving HIV test results and subsequent risk behaviour. The sexual risk index was modified to include the practice of insertive anal intercourse, in addition to receptive anal intercourse. This modification to the index was necessary because of epidemiological research that has demonstrated that HIV infection can be transmitted to an individual performing insertive anal sex without a condom (Detels et al., 1989). The index is based on sexual contact with men, and assigns participants to one of four risk categories. These categories are ordinal in nature and are defined as: (1) No Risk: celibate; (2) Low Risk: if monogamous, uses condom for insertive and receptive anal sex; if non-monogamous, does not engage in anal sex; (3) Moderate High Risk: if monogamous, practices insertive or receptive anal sex without consistent condom use; if nonmonogamous, condom is used consistently during any anal sex; (4) High Risk: non-monogamous; practices receptive or insertive anal sex without consistent condom use. The Enactment Stage Predictor consisted of the index score derived from sexual risk behaviour data from Time 8 of the cohort while the corresponding risk behaviour from the period of six months earlier, at the time of administration of the initial questionnaires, Time 7, was used to construct the index for previous sexual risk behaviour, SRBI-7.
Occasions That Alcohol Or Marijuana Were Used During Sex Score Within the SERBAS-HOM interview, individuals are questioned regarding the number of occasions that they used any alcohol or specific classes of drugs during sex. The specific classes of drug use that are assessed are hallucinogens, barbiturates, tranquillizers, opiates, amphetamines, psychedelics, "designer drugs" (Extasy, MDA, Special K), and "poppers" (ethyl, amyl, or butyl chlorides). Based on an examination of the frequency of alcohol and drug use during sex, only two substances, alcohol and marijuana, were used with high enough frequency to permit useful analyses of the data. The data for the alcohol and marijuana use during sex scales represent counts of the number of times in the previous six months that either alcohol use or marijuana use occurred during sex.
MEASURES DERIVED FROM QUESTIONNAIRES AND SCALES Self-report scales covered the areas of perceived severity of AIDS, knowledge of HIV transmission, normative values regarding safer sex performance, behavioural intentions and attitudes relating to safer sex, help-seeking behaviour, drug and alcohol use during sex, sex guilt, sensation-seeking, and response biases. From these items, scales were constructed according to standard psychometric procedures. The self-report scales served to operationalize the predictors of the ARRM stages. These scales are described and summarized in Table 1.
Table 1 Description and Psychometric Properties of Self-Report Sales Scale Name Description Reference Labeling Stage Indicator self-estimate of Catania et al. sexual risk behaviour (1988) HIV Transmission Knowledge knowledge of HIV Catania et al. transmission routes (1988) Attitude-Condom Use-Anal endorsement of Gillis & Exner Sex condom use for (1991) anal sex Attitude-Monogamy endorsement of Gillis & Exner monogamy (1991) Peer Norm-Condom Use-Anal friends/partners Gillis & Exner Sex attitudes toward (1991) condom use for anal sex weighted by motivation to comply Peer Norm-Monogamy friends/partners Gillis & Exner attitudes toward (1991) monogamy weighted by motivation to comply Self-Deception component of socially Paulhus (1989) desirable responding Impression Management component of socially Paulhus (1989) desirable responding Commitment Stage Indicator future likelihood of Exner(1989) engaging in unprotected sexual behaviours Pleasure Unprotected Anal enjoyment insertive Catania et al. Sex & receptive anal sex (1988) without condoms compared to using condoms Self-Efficacy-Safer Sex self-rated confidence Catania et al. in negotiating safer (1988) sex Perceived Threat of AIDS threat to health Catania et al. from HIV/AIDS (1988) Help-seeking for Safer Sex frequency of Gillis & Exner Skills utilization of (1991) community and personal resources for developing safer sex skills Difficulties Performing reported difficulties Catania et al. Safer Sex with the initiation (1988) and maintenance of safer sex practices Sex Guilt guilt and discomfort Mosher (1988) relating to sexual activities and fantasies Changes Sexual Risk self-rated concern Exner (1989) Behavior with making changes in sexual risk behaviour Barriers Safer Sex barriers to the Exner (1989) Performance initiation and maintenance of safer sex behaviour Sensation Seeking construct of Zuckerman sensation-seeking & Myers (1983) as defined by Zuckerman (1971) Scale Name # items # points alpha Labeling Stage Indicator 1 4 n/a HIV Transmission Knowledge 7 6 .52 Attitude-Condom Use-Anal Sex 1 5 n/a Attitude-Monogamy 1 5 n/a Peer Norm-Condom Use-Anal Sex 3 5 n/a Peer Norm-Monogamy 3 5 n/a Self-Deception 20 7 .66 Impression Management 20 7 .85 Commitment Stage Indicator 6 13 .62 Pleasure Unprotected Anal Sex 4 6 n/a Self-Efficacy-Safer Sex 9 6 .80 Perceived Threat of AIDS 1 6 n/a Help-seeking for Safer Sex Skills 15 5 .82 Difficulties Performing Safer Sex 7 4 .84 Sex Guilt 30 7 .89 Changes Sexual Risk Behaviour 1 2 n/a Barriers Safer Sex Performance 15 4 .69 Sensation Seeking 39 forced n/a choice
Self-Report Scales of ARRM Labelling and Commitment Stage Indicators
LABELLING STAGE INDICATOR: PERCEIVED SEXUAL RISK OF HIV INFECTION OR RE-INFECTION (CATANIA, KEGELES, & COATES, 1988) The outcome measure of the labelling stage was assessed by a single item, "My sexual behaviour is putting me at risk for HIV infection or re-infection." This question was answered on a four-point scale varying from "strongly agree" to "strongly disagree". Since this measure consisted of only one item, internal reliability estimates could not be computed.
COMMITMENT STAGE INDICATOR (EXNER, 1988) This was a 17-item scale assessing future likelihood of engaging in a variety of protected and unprotected sexual behaviours. Internal reliability (coefficient alpha) based on this sample for the full 17-item scale was estimated as .83. The outcome measure of the commitment stage was a mean of six specific behavioural intentions that were related to sexual risk behaviours measured by the Sexual Risk Behaviour Index. These intentions were rated on a 13-point scale ranging from "extremely unlikely" to "extremely likely". The specific behavioural intentions entered into the analyses were: (1) sexual abstinence; (2) having only one sexual partner; (3) using a condom during receptive anal sex with orgasm; (4) using a condom during receptive anal sex prior to orgasm; (5) using a condom during insertive anal sex with orgasm; and (6) using a condom during insertive anal sex prior to orgasm. Internal reliability (coefficient alpha) for this 6-item commitment indicator based on the current sample was estimated as .62.
PROCEDURE The study was conducted in accordance with the ethical guidelines for informed consent and confidentiality specified by the Institutional Review Board of the New York State Psychiatric Institute and the American Psychological Association. The sexual behaviour interview, SERBAS-HOM, was conducted as regularly scheduled for the seventh follow-up assessment. Two to three weeks prior to their regularly scheduled assessment, participants were mailed a copy of the questionnaire package, the consent form, and a brief explanation of the study. Participants completed the questionnaires at home, and turned them in at the time of their interview. Questionnaires were then checked by the interview staff, and missing responses and questions were resolved at that point. If participants did not remember to bring the completed questionnaire set with them, an additional questionnaire package with a return envelope was given to them to complete. Of the existing cohort of 131 gay and bisexual men in the longitudinal at the seventh assessment, 119 (90.8%) returned completed and usable questionnaires for this study. Sexual risk behaviour reports obtained using the SERBAS-HOM at the subsequent assessment period six months later (four years from baseline assessment) were used as the basis for calculating the dependent measure for the Enactment Stage indicator.
DATA ANALYTIC STRATEGY Multiple regression analysis using block entry was used to assess the predictive ability of the specified variables. Multiple regression is a statistical technique that allows one to determine which subset of a set of variables has the highest association with the criterion to be predicted, in this case, sexual risk behaviour. Multiple regression is a particularly appropriate statistical procedure to use when many variables are correlated with the criterion and are also correlated with each other. The strength of the relationship and the direction can be determined by examining the unstandardized regression coefficient, B. Similar to a correlation coefficient, B has a probability of significance, p, associated with it. Also similar to a correlation coefficient, the magnitude of B represents the strength of the relationship between the variable and the criterion, and the sign of B represents a direct (+ sign) or indirect (- indirect) relationship. Each time a variable or set of variables is entered into a multiple regression analysis in an attempt to predict a criterion, the amount of explained variance associated with those variables is calculated. Amount of explained variance is expressed as the square of a correlation coefficient labelled [R.sup.2]. Since the total variance is equal to 100%, expressed as an [R.sup.2] = 1.0, the importance of a variable in predicting the criterion is judged by the amount of explained variance attributable to the variable relative to total variance explained. For this study, a total of seven separate regression analyses were performed. Two regression analyses were performed for each stage of the ARRM model (labelling, commitment and enactment). First, a standard model regression analysis was conducted using the variables proposed by Catania et al. (1990b). Next, an expanded model regression analysis which included the response bias measures, previous sexual risk behaviour, and additional personality measures, was conducted. Finally, an overall regression analysis was performed which included the outcome measures of the labelling and commitment stages, in addition to the other variables, as a predictor of actual sexual risk behaviour (enactment).
The sequence of entry of variables into the regression analyses was determined on theoretical and logical bases when possible. Being important sociodemograpbic and clinical status measures, age and HIV status of the participants were entered first into the analyses. Since the majority of the subjects in the study was HIV+, HIV status was considered an important participant characteristic to test for differences. It was anticipated that HIV status would be an important stage predictor which would modulate the effect of other predictors. Measures of response bias (Self-Deception and Impression Management) were entered next in an attempt to control for possible biases in the self-report measures used (Paulhus, 1991). Previous sexual risk behaviour was entered next into the analyses as suggested by Weinstein and Nicolich (1993). Finally, the specific predictors hypothesized to be associated with the individual stages were entered last into the regression analyses. Thus, the predictive ability of these stage-specific variables was assessed while controlling for all previous variables entered.
ARRM STAGE 1: LABELLING: "IS MY SEXUAL BEHAVIOUR PUTTING ME AT RISK?"
The measure of the labelling stage was a single item measured on a 4-point Likert-like scale ranging from 1 to 4, The mean score for the sample for labelling was equal to 3.41, a finding which indicates that participants strongly believe that their sexual behaviour is not putting them at risk for HIV infection or HIV re-infection. The distribution of participants' response to this labelling stage indicator was as follows: 2 participants (1.7%) indicated that they "strongly agreed", 21 (18.1%) of the participants indicated that they "agreed somewhat" with the statement, 21 (18.1%) indicated that they "disagreed somewhat" with the statement, while the majority of 72 participants (62.1%) indicated that they "strongly disagreed". A total of 3 participants (2.5%) were unable or failed to complete this item.
ARRM STAGE 2: COMMITMENT: BEHAVIOURAL INTENTIONS To ENGAGE IN SAFER SEX
The mean score for the commitment indicator for the total sample was 2.34, and the corresponding standard deviation was 2.09. This mean score indicated a greater intention among participants than at present to engage in condom use during anal sex and to be monogamous. The commitment indicator had a median of 2.00 and a range of -4.00 to +6.00, thus indicating a large variance in individuals' specific safer sex intentions.
ARRM STAGE 3: ENACTMENT: FOUR POINT SCALE OF SEXUAL RISK BEHAVIOUR
The enactment stage indicator, the Sexual Risk Behaviour Index, was assessed prospectively at a point six months after the self-report measures were completed. Based upon calculations for the sexual risk behaviour index, the participants in the study were ranked as follows: (1) No Risk = 31 (26. 1%); (2) Low Risk = 44 (37.0%); (3) Moderate High Risk = 30 (25.2%); and (4) High Risk = 14 (11.8%). The mean Sexual Risk Index measure for the entire sample based on this ranking was equal to 2.23 (a value between the Low Risk and Moderate High Risk categories), and had a standard deviation equal to 0.97. A Sexual Risk Behaviour Index for the previous assessment period (the period during which the self-reports were completed) was also computed. There was a moderately high correlation between the current Sexual Risk Behaviour Index and the SRBI-7 Behaviour Index (r=.53, p [is less than] .001). This indicated that those individuals in higher sexual risk behaviour categories at the previous assessment were likely to also be in higher sexual risk categories at the subsequent assessment. It should also be noted that the various ARRM stages are moderately to highly inter-correlated.
STAGE-SPECIFIC PREDICTION: LABELLING STAGE Demographic variables (Age and HIV Status) were entered as a block into the regression analyses at step one to predict labelling (self-perception of HIV risk transmission for HIV- participants or of HIV reinfection for HIV+ participants). The combination of Age and HIV status accounted for only 1.1% of the total variance in Labelling (see Table 2A). In step two, five stage-specific predictors were entered as a block. These were: (1) HIV Transmission Knowledge; (2) Attitude Toward Condom Use During Anal Sex; (3) Attitude Toward Monogamy; (4) Peer Norms Regarding Monogamy; and (5) Peer Norms Regarding Condom Use during Anal Sex. This block accounted for an additional 3.2% of explained variance. However, the low variability associated with the HIV Transmission Knowledge scale and the resultant attenuated correlation did not permit an adequate test of the association of HIV transmission knowledge and labelling. The total amount of variance explained by the variables, as represented by [R.sup.2], was equal to only 4.3% (see Table 2A). The overall regression of the predictors on labelling was not significant (F(7,111) [is less than] 1.0). Due to the finding of a non-significant overall regression, interpretation of the individual regression coefficients of the predictors entered is not justified.
Table 2A Standard Model: Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Labelling Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age -.012 .27 -.08 HIV Status -.109 .50 -.07 .011 HIV Transmission Knowledge .006 .72 .02 Attitude-Condom Use-Anal Sex -.133 .19 -.12 Attitude-Monogamy .076 .29 .05 Peer Norm-Condom Use-Anal Sex -.001 .83 -.04 Peer Norm-Monogamy -.008 .36 -.07 .043
For the expanded prediction model of labelling, the additional predictors of Self-Deception and Impression Management were entered into the analysis at step two, SRBI-7 Behaviour was entered at step three, and was followed by the stage-specific predictors of labelling described above (see Table 2B). The overall regression for the expanded model of labelling was significant (F(10,108) = 3.97; p [is less than] .001). SRBI-7 (b = .433) and Attitude Toward Monogamy (b = .131) were significant predictors of labelling. However, SRBI-7 alone accounted for the majority of the 26.9% of total explained variance (see Table 2B).
Table 2B Expanded Model Including Response Bias Measures and Previous Risk Behaviour (SRBI-7): Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Labelling Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age -.006 .55 -.08 HIV Status -.035 .81 .07 .011 Self-Deception -.005 .42 -.18 Impression Management -.001 .77 -.16 .052 SRBI-7 .433 .001 .46 .219 HIV Transmission Knowledge .005 .74 .02 Attitude-Condom Use-Anal Sex -.103 .25 -.12 Attitude-Monogamy .131 .04 .05 Peer Norm-Condom Use-Anal Sex -.004 .40 -.04 Peer Norm-Monogamy .00 .99 .07 .269
STAGE-SPECIFIC PREDICTION: COMMITMENT STAGE The commitment stage indicator was operationalized as the mean score of behavioural intentions to use condoms during anal sex, to be monogamous, and to be sexually abstinent. For the regression analysis, demographic variables (Age and HIV status) were entered as a block in step one to predict commitment. Age and HIV status together accounted for 2.1% of the total variance for commitment (see Table 3A). In step two, six stage-specific predictors were entered into the analysis. These stage-specific predictors were: (1) Pleasure from Unprotected Anal Sex; (2) Safer Sex Efficacy; (3) Attitude-Condom Use-Anal Sex item; (4) an Attitude-Monogamy item; (5) Peer Norm-Condom Use-Anal Sex item; and (6) a Peer Norm-Monogamy item. This block accounted for an additional 8.8% of explained variance. Thus, the final [R.sup.2] for the analysis was equal to .109 which represents 10.9% of explained variance (see Table 3A). Since the overall regression of the predictors on commitment was not significant (F(8,110)=1.68), interpretation of the individual regression coefficients of the predictors will not be attempted.
Table 3A Standard Model: Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Commitment Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age -.013 .59 -.02 HIV Status .612 .11 .14 .021 Pleasure Unprotected Anal Sex -.105 .23 -.18 Self-Efficacy-Safer Sex .013 .58 .10 Attitude-Condom Use-Anal Sex .512 .04 .23 Attitude-Monogamy .086 .59 .07 Peer Norm-Condom Use-Anal Sex -.001 .93 .03 Peer Norm-Monogamy .017 .42 .12 .109
For the expanded prediction model of commitment, Age and HIV Status were entered into the analysis in step one (see Table 3B). In step two, the additional predictors of Self-Deception and Impression Management were entered into the analysis, and resulted in an increase of explained variance of 1.4%. SRBI-7 was entered at step three, and accounted for an additional 16.4% of explained variance. In step four, the seven stage-specific predictors (Perceived Threat of AIDS, Pleasure Unprotected Anal Sex, Safer Sex Efficacy, Attitude-Condom Use-Anal Sex, Attitude-Monogamy, Peer Norm-Condom Use-Anal Sex, and Peer Norm-Monogamy) were entered into the analysis. The addition of these stage-specific variables resulted in a modest increase in explained variance (5.6%). The overall regression for the expanded predictors model on labelling was significant (F(12,106) = 3.31; p [is less than] .01). The two significant predictors of commitment in the expanded prediction model were SRBI-7 which had a value of b (unstandardized regression coefficient) equal to -.940 (p [is less than] .001), and Attitude-Condom Use-Anal Sex (b =.502, p [is less than] .05) (see Table 3B). Again, however, Previous Risk Behaviour accounted for the majority of the explained variance.
Table 3B Expanded Model Including Response Bias Measures and Previous Risk Behavior (SRBI-7): Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Commitment Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age -.023 .34 -.02 HIV Status .433 .25 .14 .021 Self-Deception .008 .62 .13 Impression Management -.011 .26 .03 .035 SRBI-7 -.940 .001 -.40 .199 Perceived Threat of AIDS -.003 .98 .027 Pleasure Unprotected Anal Sex -.101 .22 -.18 Self-Efficacy for Safer Sex -.016 .49 .10 Attitude-Condom Use-Anal Sex .502 .03 .23 Attitude-Monogamy -.017 .91 .07 Peer Non-n-Condom Use-Anal Sex .004 .72 .03 Peer Norm-Monogamy .006 .77 .12 .255
STAGE-SPECIFIC PREDICTION: ENACTMENT STAGE The enactment stage indicator was operationalized by the Sexual Behaviour Risk Index score. The demographic variables (Age and HIV status) were entered as a block into the regression analysis at step one to predict enactment (see Table 4A). Age and HIV status together accounted for only 1.9% of the total variance of enactment (see Table 4A). In step two, ten stage-specific predictors of enactment were entered as a block into the analysis. These were: (1) Help-Seeking for Safer Sex Skills scale; (2) Self-Efficacy for Safer Sex scale; (3) Pleasure Rating of Unprotected Anal Sex score; (4) Difficulties with Performing Safer Sex scale; (5) Alcohol Use During Sex score; (6) Marijuana Use During Sex score; (7) Attitude Toward Condom Use During Anal Sex item; (8) Attitude Toward Monogamy item; (9) Peer Norm Regarding Condom Use during Anal Sex item; and (10) Peer Norm Regarding Monogamy item. This block accounted for an additional 19.5% of explained variance which was a significant increase in explained variance (F(12,106) = 2.88, p [is less than] .01). The value of [R.sup.2] was equal to .214 or 21.4% of total variance at the final step. In contrast to the standard models for the analyses of the labelling and commitment stages, the overall regression of the predictors on enactment for this standard model was highly significant (F(12,106) = 2.63; p [is less than] .01). Of the twelve predictors of enactment entered into the regression analyses of the standard model, the Self-Efficacy for Safer Sex scale (b = -.024), the Difficulties with Performing Safer Sex scale (b = .097), and the Attitude Toward Monogamy item (b = -.143) were significant.
Table 4A Standard Model: Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Enactment Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age .004 .97 -.11 HIV Status -.258 .14 -.09 .019 Help Seeking for Safer Sex Skills .010 .29 .10 Self-Efficacy-Safer Sex -.024 .03 -.24 Pleasure Unprotected Anal Sex -.071 .07 -.08 Difficulties Performing Safer Sex .092 .006 .29 Alcohol Use During Sex -.005 .89 .04 Marijuana Use During Sex .006 .65 .05 Attitude-Condom Use Anal Sex -.156 .16 .19 Attitude-Monogamy -.143 .048 -.18 Peer Norm-Condom Use Anal Sex .003 .66 .04 Peer Norm-Monogamy .003 .71 -.07 .214
For the expanded model of prediction of enactment, Self-Deception and Impression Management accounted for 0.3% of total variance after they were entered into the analyses at step two following Age and HIV Status (see Table 4B). SRBI-7 was then entered into the analysis at step three, and explained 29.2% of total variance. Next, an expanded set of stage-specific predictors was entered into the regression analyses at step four. This expanded set consisted of 14 stage-specific predictor variables, ten of these variables from the standard model (Help Seeking for Safer Sex Skills, Self-Efficacy for Safer Sex, Pleasure Rating of Unprotected Anal Sex, Difficulties with Performing Safer Sex, Alcohol Use During Sex, Marijuana Use During Sex, Attitude Toward Condom Use During Anal Sex, Attitude Toward Monogamy, Peer Norm Regarding Condom Use during Anal Sex, and Peer Norm Regarding Monogamy), and four additional predictors (Concern About Making Changes to Sexual Risk Behaviour item, Barriers to Safer Sex Performance scale, Sensation Seeking, and Sex Guilt). The entry of this block resulted in a increase of explained variance of 10.9%. The overall regression for the expanded predictors model on enactment was significant (F(19, 99) = 4.31; p [is less than] .001). The three significant predictors listed in order of their unstandardized regression coefficients (b) are: (1) SRBI-7 (.553); (2) Self-Efficacy for Safer Sex (-.029); and Self-Deception (.016). The unstandardized regression coefficients for Sex Guilt, Attitude-Condom Use-Anal Sex, and Peer Norm-Monogamy were near the established significance level of p [is less than] .05 with values of p of [is less than] .10.
Table 4B Expanded Model Including Response Bias Measures, SRBI-7, and Supplemental Variables: Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Enactment Stage Indicator
final final first [R.sup.2] Predictors B p order r block Age .009 .41 -.11 HIV Status -.239 .13 -.09 .019 Self-Deception .016 .03 .03 Impression Management .006 .20 -.05 .022 SRBI-7 .553 .0001 .52 .314 Help Seeking for Safer Sex .017 .85 .10 Self-Efficacy-Safer Sex -.029 .013 -.24 Pleasure Unprotected Anal Sex -.053 .13 -.08 Safer Sex Difficulties .041 .19 .29 Alcohol Use During Sex -.001 .72 .04 Marijuana Use During Sex .003 .77 .05 Sex Guilt -.252 .09 -.16 Changes Sexual Risk Behaviour -.078 .69 .20 Barriers Safer Sex Performance -.011 .29 .04 Sensation Seeking -.003 .84 .20 Attitude-Condom Use Anal Sex -.173 .08 -.19 Attitude-Monogamy -.067 .32 -.18 Peer Norm-Anal Sex .002 .67 .04 Peer Norm-Monogamy .014 .09 -.07 .423
OVERALL PREDICTION OF ENACTMENT STAGE FROM PRECEDING STAGES Following the stage model of behaviour change outlined in the ARRM, the prediction of the enactment stage (stage 3 of ARRM) was attempted using the indicators of labelling (stage 1 of ARRM) and commitment (stage 2 of ARRM) in addition to the stage-specific predictors of Enactment (see Table 5). As before, demographic variables (Age and HIV Status) were entered first, followed by Self-Deception and Impression Management in step two. In step three, SRBI-7 was entered. The model at that point was equivalent to the Expanded Enactment Stage model and had the same set of predictors and the same amount of explained variance. In step four, however, the labelling stage indicator was introduced into the regression analysis. The addition of labelling accounted for a non-significant addition of 0.2% of explained variance. The commitment indicator was entered at step five, and accounted for an additional 2.1% of explained variance. In step six, the expanded set of stage-specific predictors of enactment were entered (Help-Seeking for Safer Sex Skills, Self-Efficacy-Safer Sex, Pleasure Unprotected Anal Sex, Safer Sex Difficulties, Alcohol Use During Sex, Marijuana Use During Sex, Sex Guilt,, Changes Sexual Risk Behaviour, Barriers Safer Sex Performance, Sensation Seeking, Attitude-Condom Use-Anal Sex, Attitude-Monogamy, Peer Norm-Condom Use-Anal Sex, and Peer Norm Monogamy). The entry of this block of 14 predictors accounted for an additional 11.5% of explained variance which was a marginally significant change in explained variance (F(19, 99) = 1.61; p = .09) (see Table 5). The final model for the prediction of sexual risk behaviour from the previous ARRM and the stage-specific predictors of the ARRM enactment stage accounted for 45.2% of the total variance, a highly significant result (F(21,97) = 4.21; p [is less than] .001).
Table 5 Unstandardized Regression Coefficients, Significance of Predictors, and Variance Accounted for in the Enactment Stage Indicator for the Final Model
final final first [R.sup.2] Predictors B p order r block Age .005 .63 -.11 HIV Status -.193 .22 -.09 .019 Self-Deception .015 .03 .03 Impression Management .005 .24 -.05 .022 SRBI-7 .470 .0001 .52 .314 Labelling Stage Indicator -.051 .68 .27 .316 Commitment Stage Indicator -.090 .04 -.36 .337 Help Seeking for Safer Sex Skills .048 .59 .10 Self-Efficacy-Safer Sex -.032 .007 -.24 Pleasure Unprotected Anal Sex -.057 .10 -.08 Safer Sex Difficulties .034 .30 .29 Alcohol Use During Sex -.004 .91 .04 Marijuana Use During Sex .003 .75 .05 Sex Guilt -.297 .05 -.16 Changes Sexual Risk Behaviour -.027 .89 .20 Barriers Safer Sex Performance -.010 .36 .04 Sensation Seeking .003 .98 .20 Attitude-Condom for Anal Sex -.116 .26 -.19 Attitude-Monogamy -.053 .43 -.18 Peer Norm-Anal Sex .002 .66 .04 Peer Norm-Monogamy .014 .10 -.07 .452
Of the 21 predictors of enactment entered into the regression analyses in six separate blocks, five predictors remained significant at the .05 level in the final regression equation. These five significant predictors, ranked in terms of their unstandardized regression coefficient, (b), were: (1) SRBI-7 Index (b = .470); (2) Sex Guilt (b = -.297); (3) Commitment Stage Indicator (b = -.090); (4) Self-Efficacy for Safer Sex (b = -.032); and (5) Self-Deception (b = .015) (see Table 5). Thus, in this final model, higher Sexual Risk Behaviour Index scores were associated with higher levels of previous sexual risk behaviour, lower sex guilt, lower intentions to engage in safer sex, lower self-efficacy for safer sex, and higher levels of self-deception.
For all three ARRM stages, the expanded model accounted for substantially more explained variance of the stage indicator than did the standard model, largely due to the predictive power associated with previous sexual risk behaviour. Thus, the results of this investigation support the expansion of ARRM to include additional variables, particularly previous level of sexual risk behaviour, and are consistent with the findings of Aspinwall et al. (1991) and Dolezal et al. (1997). This finding also has important implications for the continued need for HIV prevention programs suggesting that, for white gay and bisexual urban men who are knowledgeable about HIV/AIDS, the level of sexual risk behaviour engaged in by an individual remains relatively constant over a six month period. Further reduction in levels of sexual risk behaviour would possibly be facilitated by HIV prevention efforts focused on such areas as behavioural skills negotiation, sexual assertiveness, self-esteem and other areas delivered in a culturally and community sensitive format as advocated by Kelly (1995).
Participants in the study, similar to those in van der Velde et al. (1991), appeared to accurately perceive and report their level of sexual risk. In addition, it was found that similar to results of van der Velde et al. (1991) and Gerrard, Gibbons and Bushman (1996), individuals who perceive themselves to be at higher risk are less likely to intend to use condoms during anal sex and less likely to be sexually abstinent. This would suggest that HIV/AIDS intervention programs in urban gay male communities need to focus their attention on those individuals who report high perceived risk, rather than on the general community.
The failure of variables to predict the ARRM stages of labelling and commitment for this sample should not necessarily be interpreted as a failure of the model. It is well documented that the AIDS epidemic has affected different segments of the U.S. population at different times and to different degrees. The spread of the epidemic, the awareness of AIDS, and the initiation of successful community-level campaigns to decrease sexual risk behaviour has occurred first in large urban areas with highly politicized and organized gay communities, such as New York City and San Francisco. For these gay and bisexual men, substantial reductions in the number of sexual partners and acts of receptive anal sex had already occurred by the early 1980s (Martin, Garcia, & Beatrice, 1989). Most of them have passed the labelling and commitment stages as hypothesized by ARRM, and are in the enactment phase. Thus, any test of the predictive ability of the ARRM in such a population during 1991, as has been done in this investigation, is done at a time when the level of overall sexual risk behaviour for the group has been substantially reduced. Furthermore, the variables that are associated with continued high risk sexual behaviour or a return to higher levels of sexual risk behaviour are likely to be different from those variables associated with the substantial initial changes in risk behaviour. Thus, demonstrating the validity of the various stages of ARRM under such conditions as these will be more difficult than with a group that is earlier in the behaviour change process. Similarly, the use of a sample that does not perceive itself at risk for HIV infection, and has not made any changes in sexual behaviour to reduce risk, would constrain attempts to establish the validity of ARRM.
Other limitations of the sample used in this investigation must be emphasized. A paramount limitation to the generalizability of the study is the fact that the investigation used only gay- and bisexual-identified men who lived in a large urban epicentre for HIV infection. The gay and bisexual men recruited for the study were of largely European-American heritage, high SES, well-educated and from a highly urbanized and politically active gay community. While this group is still at high risk for HIV infection, they are only one of many communities where HIV risk reduction efforts should occur. The relevance of the current findings to gay and bisexual men of colour is thus unclear. However, the similarity of the findings to the results of Kelly et al. (1995) and Herek and Glunt (1995) with small city and rural gay men would suggest that the results have applicability to those groups.
Predictors of the enactment stage of ARRM are particularly important to consider for their possible relevance to HIV risk reduction interventions. In this analysis, lower levels of sex guilt, lower self-efficacy for safer sex, lower intentions to engage in safer sex, and higher levels of self-deception were prospectively associated with higher reported sexual risk behaviour. The association of lower levels of self-efficacy with higher levels of sexual risk behaviour is emerging as a frequent and consistent finding in studies of the sexual risk behaviour of gay men (Aspinwall et al., 1991; de Wit, van Griensven, Kok, & Sandfort, 1993; Herek & Glunt, 1995; Kalichman, Roffman, Picciano, & Bolan, 1997; Kelly, St. Lawrence, Betts, Brasfield, & Hood, 1990; McKusick, Coates, & Morin, 1990). The overall pattern of significance of predictors is quite similar to that found by Aspinwall et al. (1991), who found similar results for the predictive power of previous sexual risk behaviour and self-efficacy, despite the fact that these investigators used total number of sexual partners and number of anonymous partners as their dependent measures, and did not assess intentions to engage in behaviour as one of their predictors. Given the current findings, and the evidence for an association of self-efficacy and safer sex behaviour among a range of populations (Kalichman et al., 1990; Fisher & Fisher, 1992), it would appear that a consensus regarding the importance of self-efficacy in the performance of safer sex behaviour is emerging. Indeed, research on self-efficacy as a determinant of safer sex behaviour has progressed to the point where investigators have proposed a path analytic model of condom use with self-efficacy as the central mediator (Wulfert & Wan, 1993; Goldman & Harlow, 1993). Based on the current analyses and the predominant view in the intervention literature (Kelly et al., 1990; Fisher & Fisher, 1992), self-efficacy is most probably more strongly associated with the ARRM enactment stage, rather than the commitment stage, as originally proposed by Catania et al. (1990b). Since Catania et al. (1994) only looked at self-efficacy as a predictor of commitment and not of enactment, this may explain why they failed to find a significant association.
The role of sex guilt in relation to sexual risk behaviour for gay men had not been previously addressed, despite a wealth of research on the correlates of sex guilt and suggestions about its applicability as a personality correlate of sexual risk behaviour (Jaccard & Wilson, 1991). The finding that those individuals higher in sex guilt engaged in less sexual risk behaviour is consistent with the findings of Rimberg and Lewis (1994) in heterosexual college students and has several plausible explanations. It is possible that individuals with higher sex guilt are more likely to be concerned with following sexual risk reduction guidelines, and tend to avoid sexual risk activities. Alternatively, individuals with higher sex guilt might tend to engage in less sexual activity and would be more likely to be sexually abstinent, thus lowering their risk for contracting HIV.
The lack of a significant association between sexual risk behaviour and alcohol or marijuana use during sex is of interest. Frequently, though not always (Martin, 1990), increased substance use accompanying sexual activity has been associated with higher sexual risk behaviour in studies of gay and bisexual men (Kalichman, Kelly, & Rompa, 1997; Kelly et al., 1991; Leigh & Stall, 1993; Stall, McKusick, Wiley, Coates, & Ostrow, 1986). In fact, research that was conducted with this same cohort of gay and bisexual men by Remien, Meyer-Bahlburg, Williams, Exner, Rabkin, Gorman and Ehrhardt (1992) and Dolezal et al. (1997) found a relationship between the presence of a history of drug use disorders and engaging in higher sexual risk behaviour. Of particular relevance to alcohol and drug use are studies of a newly devised scale of sexual sensation-seeking (Kalichman, Johnson, Adair, Rompa, Multhauf, & Kelly, 1994) and level of sexual risk behaviour in gay men (Dolezal et al., 1997; Kalichman et al., 1994). Though the current study found that the Zuckerman and Myers (1983) general measure of sensation-seeking was unrelated to sexual risk behaviour, the pattern of significance reported by Dolezal et al. (1997) for sexual sensation-seeking and sexual risk behaviour among gay men was complex, and probably mediated the scales association with alcohol and drug use (Kalichman et al., 1994).
Results of the present investigation and a recent quantitative review of ARRM studies (Flowers et al., 1997) would suggest that the ARRM has good potential as a useful heuristic to guide assessment and intervention efforts for HIV prevention, though the specific correlates of risk for some populations have yet to be assessed. Flowers et al. (1997) make the important observation that relationship status needs to be considered within the ARRM framework, as much research has shown sexual risk behaviour differs as a function of intimacy of relationship. Accepting the precepts of ARRM as a starting point for constructing interventions, future research should focus on testing the ARRM within the context of an intervention, preferably a large scale community-based model, as several authors have recognized the urgency of doing so (Fisher & Fisher, 1992; Kelly, Murphy, Sikkema, & Kalichman, 1993).
Adib, S.M., Joseph, J.G., Ostrow, D.G., & James, S.A. (1991). Predictors of relapse in sexual practices among homosexual men. AIDS Education and Prevention, 3, 293-304.
Adib, S.M., Joseph, J.G., Ostrow, D.G., Tal, M., & Schwartz, S.A. (1991). Relapse in sexual behavior among homosexual men: A 2-year follow-up from the Chicago MACS/CCS. AIDS, 5, 757-760.
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social behavior. Englewood Cliffs, NJ: Prentice-Hall.
Aspinwall, L.G., Kemeny, M.E., Taylor, S.E., Schneider, S.G, & Dudley, J.P. (1991). Psychosocial predictors of gay men's AIDS risk-reduction behavior. Health Psychology, 10, 432-444.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191-215.
Bandura, A. (1989). Perceived self-efficacy in the exercise of control over AIDS infection. In V. M. Mays, G. W. Albee, & S. F. Schneider (Eds.), Primary prevention of AIDS (pp. 128-141). Newbury Park, CA: Sage Publications.
Becker, M. (Ed.) (1974). The health belief model and personal health behavior. Thorofare, NJ: Charles B. Slack.
Beltran, E. D., Ostrow, D. G., & Joseph, J. G. (1993). Predictors of sexual behavior change among men requesting their HIV-1 antibody status: The Chicago MACS/CCS cohort of homosexual/bisexual men, 1985-1986. AIDS Education and Prevention, 5, 185-195.
Catania, J. A., Coates, T. J., & Kegeles, S. (1994). A test of the AIDS risk reduction model: Psychosocial correlates of condom use in the AMEN cohort survey. Health Psychology, 13, 548-555.
Catania, J. A., Coates, T. J., Kegels, S. M., Ekstrand, M., Guydish, J. R., & Bye, L. L. (1989). Implications of the AIDS risk reduction model for the gay community: The importance of perceived sexual enjoyment and help-seeking behaviors. In V.M. Mays, G.W. Albee, & S.F. Schneider (Eds.) Primary prevention of AIDS (pp. 242-261). Newbury Park, CA: Sage Publications.
Catania, J. A., Gibson, D. R., Chitwood, D. D., & Coates, T.J. (1990a). Methodological problems ins AIDS behavioral research: Influences on measurement error and participation bias in studies of sexual behavior. Psychological Bulletin, 108, 339-362.
Catania, J.A., Kegels, S.M., & Coates, T.J. (1988). Psychosocial measures for studies of AIDS Unpublished manuscript from the Psychosocial Research Group and the USCF Center for AIDS Prevention Studies, San Francisco, CA.
Catania, J. A., Kegeles, S. M., & Coates, T. J. (1990b). Toward and understanding of risk behavior: An AIDS risk reduction model (ARRM). Health Education Quarterly, 17, 53-72.
Centers for Disease Control. (1986). Classification system for Human T-Lymphotrophic virus type III/ lymphadenopathy-associated virus infections. Morbidity and Mortality Weekly Report, 35, 334-339.
Detels, R., English, P., Visscher, B.R., Jacobson, L., Kingsley, L.A., Chmiel, J.S., Dudley, J.P., Eldred, L.J., Ginzburg, H.M. (1989). Seroconversion, sexual activity, and condom use among 2,915 seronegative men followed for up to 2 years. Journal of Acquired Immune Deficiency Syndrome, 2, 77-83.
de Wit, J.B.F., van Griensven, G.J.P., Kok, G., & Sandfort, T.G.M. (1993). Why do homosexual men relapse into unsafe sex? Predictors of resumption of unprotected anogenital intercourse with casual partners. AIDS, 7, 1113-1118.
Dolezal, C., Meyer-Bahlburg, H.F.L., Remien, R.H., & Petkova, E. (1997). Substance use during sex and sensation seeking as predictors of sexual risk behavior among HIV+ and HIV- gay men. AIDS and Behavior, 1, 19-28.
Evans, B.A., McLean, K.A., Dawson, S.G., Teece, S.A., Bond, R.A., MacRae, K.D., Thorp, R.W. (1989). Trends in sexual behavior and risk factors for HIV infection among homosexual men, 1984-7. British Medical Journal, 298, 215-218.
Exner, T.M., Meyer-Bahlburg, H.F.L., & Ehrhardt, A.A. (1992). Sexual self control as mediator of high risk sexual behavior in a New York City cohort of HIV+ and HIV- gay men. Journal of Sex Research, 29, 389-406.
Exner, T.M. (1989). Self-report scales for Core II at Round 2. Unpublished manuscript, HIV Center for Clinical and Behavioral Studies, New York State Psychiatric Institute, College of Physicians and Surgeons of Columbia University, New York, NY.
Exner, T.M., & Gillis, J.R. (1991). Self-report scales for Core II at Round 7. Unpublished manuscript, HIV Center for Clinical and Behavioral Studies, New York State Psychiatric Institute, College of Physicians and Surgeons of Columbia University, New York, NY.
Fisher, E., Winer, D., & Abramowitz, S. (1983). Seeking professional help for psychological problems. In A. Nadler, J. Fisher, & B. DePaulo (Eds.), New directions in helping: Vol. 3. Applied perspectives on help-seeking and receiving. New York: Academic Press.
Fisher, J.D., & Fisher, W.A. (1992). Changing AIDS risk behavior. Psychological Bulletin, 111, 3, 455-474.
Fisher, J.D., Fisher, W.A., Williams, S.S., & Malloy, T.E. (1994). Empirical tests of an Information-Motivation Behavioral Skills model of AIDS-preventive behavior with gay men and heterosexual university students. Health Psychology, 13, 238-250.
Fisher, W.A. (1997). A theory-based framework for intervention and evaluation in STD/HIV prevention. Canadian Journal of Human Sexuality, 6, 105-111.
Fishbein, M., & Middlestadt, S.E. (1989). Using the theory of reasoned action as a framework for understanding and changing AIDS-related behaviors. In V.M. Mays, G.W. Albee, & S.F. Schneider (Eds.), Primary prevention of AIDS (pp. 93-110). Newbury Park CA: Sage Publications.
Flowers, P., Sheeran, P., Beail, N., & Smith, J.A. (1997). The role os psychosocial factors in HIV risk-reduction among gay and bisexual men: A quantitative review. Psychology and Health, 12, 197-230.
Folkman, S., Chesney, M.A., Pollack, L., & Phillips, C. (1992). Stress, coping, and high-risk sexual behavior. Health Psychology, 11, 218-222.
Gerrard, M. (1987). Sex, sex guilt, and contraceptive use revisited: The 1980's Journal of Personality and Social Psychology, 52, 975-980.
Gerrard, M., Gibbons, F.X., & Bushman, B.J. (1996). Relation between perceived vulnerability to HIV and precautionary sexual behavior. Psychological Bulletin, 119, 390-409.
Gillis, JR., & Exner, T.M. (1991). Self-report scales for Core II at Round 7. Unpublished manuscript, HIV Center for Clinical and Behavioral Studies, New York State Psychiatric Institute, College of Physicians and Surgeons of Columbia University, New York, NY.
Gillis, J.R., Meyer-Bahlburg, H.F.L., & Exner, T.M. (1992, July). A comparative test of models predicting sexual risk behavior. Paper presented at the VIII International Conference on AIDS/ III STD World Conference, Amsterdam, The Netherlands.
Goldman, J.A., & Harlow, L.L. (1993). Self-perception variables that mediate AIDS-preventive behavior in college students. Health Psychology, 12, 489-498.
Gorman, J.M., Kertzner, R., Todak, G., Goetz, R. R., Williams, J. B., Rabin, J., Meyer-Bahlburg, H.F.L., Mayeux, R., Stem, Y., Lange, M., Dobkin, J., Spitzer, R., & Ehrhardt, A.A. (1991). Multidisciplinary baseline assessment of gay men with and without HIV infection. I: Overview of study design. Archives of General Psychiatric 48, 120-123.
Herek, G.M., & Glunt, E.K. (1995). Identity and community among gay and bisexual men in the AIDS era: Preliminary findings from the Sacramento Men's Health Study. In G.M. Herek, & B. Greene (Eds.), AIDS, identity, and community: The HIV epidemic and lesbians and gay men (pp. 55-84). Thousand Oaks, CA: Sage Publications.
Hollingshead, A. B. (1957). Two factor index of social position. New York: Psychological Corporation.
Jaccard, J., & Wilson, T. (1991). Personality factors influencing risk behaviors. In J.N. Wasserheit, S.O. Aral, & K. Holmes (Eds.), Research issues in human behavior and sexually transmitted diseases in the AIDS era, (pp. 177-197). Washington, DC: American Association for Microbiology.
Janz, N., & Becker, M. (1984). The health belief model: A decade later. Health Education Quarterly, 11, 1-47.
Joseph, J.G., Kessler, R.C., Wortman, C.B., Kirscht, J.P., Tal, M., Caumartin, S., Eshleman, S., & Eller, M. (1989). Are there psychological costs associated with changes in behavior to reduce AIDS risk? In V.M. Mays, G.W. Albee, & S.F. Schneider (Eds.), Primary prevention of AIDS (pp. 209-224). Newbury Park, CA: Sage Publications.
Kalichman, S.C., Johnson, J.R., Adair, V., Rompa, D., Multhauf, K., & Kelly, J. (1994). Sexual sensation seeking: Scale development and predicting AIDS-Risk behavior among homosexually active men. Journal of Personality Assessment, 62, 385-397.
Kalichman, S.C., Kelly, J.A., & St. Lawrence, J.S. (1990). Factors influencing reduction of sexual risk behaviors for human immunodeficiency virus infection: A review. Annals of Sex Research, 3, 129-148.
Kalichman, S.C., Kelly, J.A., Rompa, D. (1997). Continued high-risk sex among HIV seropositive gay and bisexual men. Health Psychology, 16, 369-373.
Kalichman, S.C., Roffman, R.A., Picciano, J.F., & Bolan, M. (1997). Sexual relationships, sexual behavior, and HIV infection: HIV-seropositive gay and bisexual men seeking prevention services. Professional Psychology. Research and Practice, 28, 355-360.
Kelly, J.A. (1995). Changing HIV risk behavior. Practical strategies. New York, NY: The Guilford Press.
Kelly, J.A., Murphy, D.A., Sikkema, K.J., & Kalichman, S.C. (1993). Psychological interventions to prevent HIV infection are urgently need: New priorities for behavioral research in the second decade of AIDS. American Psychologist, 48, 1023-1034.
Kelly, J.A., Sikkema, K.J., Winett, R.A., Solomon, L.J., Roffman, R., Heckman, T.G., Stevenson, L.Y., Perry, M.L., Norman, A.D., & Desiderato, L.J. (1995). Factors predicting continued high-risk behavior among gay men in small cities: Psychological, behavioral, and demographic characteristics related to unsafe sex. Journal of Consulting and Clinical Psychology, 63, 101-107.
Kelly, J.A., & St. Lawrence, J. S. (1988). AIDS prevention and treatment: Psychology's role in the crisis. Clinical Psychology Review, 8, 255-284.
Kelly, J.A., St. Lawrence, J. S., Betts, R., Brasfield, T.L., & Hood, H.V. (1990). A skills-training group intervention model to assist persons in reducing risk behaviors for HIV infection. AIDS Education and Prevention, 2, 24-35.
Kirscht, J.P. & Joseph, J.G. (1989). The health belief model: Some implications for behavior change, with reference to homosexual males. In V.M. Mays, G.W. Albee, & S.F. Schneider (Eds.), Primary prevention of AIDS (pp. 111-127). Newbury Park, CA: Sage Publications.
Kline, A., & VanLandingham, M. (1994). HIV-infected women and sexual risk reduction: The relevance of existing models of behavior change. AIDS Education and Prevention, 6, 390-402.
Kowalewski, M.R., Longshore,D., & Anglin, M.D. (1994). The AIDS Risk Reduction Model: Examining intentions to use condoms among injection drug users. Journal of Applied Social Psychology, 24, 2002-2027.
Lanier, M.M., & Gates, S. (1996). An empirical assessment of the AIDS Risk Reduction Model (ARRM) employing ordered probit analyses. Journal of Criminal Justice, 24, 537-547.
Leigh, B.C., & Stall, R. (1993). Substance use and risky sexual behavior for exposure to HIV: Issues in methodology, interpretation, and prevention. American Psychologist, 48, 1035-1045.
Longshore, D., Anglin, M.D., & Hsieh, S.C. (1997). Intended sex with fewer partners: An empirical test of the AIDS risk reduction model among injection drug users. Journal of Applied Social Psychology, 27, 187-208.
Malow, R.M., & Ireland, S.J. (1996). HIV risk correlates among non-injection cocaine dependent men in treatment. AIDS Education, and Prevention, 8, 226-235.
Martin, J.L. (1990). Drug use and unprotected anal intercourse among gay men. Health Psychology, 9, 450-465.
Martin, J.L., Garcia, M.A., & Beatrice, S.T. (1989). Sexual behavior changes and HIV antibody in a cohort of New York City gay men. American Journal of Public Health, 79, 501-503.
McKusick, L., Coates, T.J., & Morin, S. (1990). Longitudinal predictors of unprotected anal sex among gay and bisexual men in San Francisco: The AIDS behavioral research project. American Journal of Public Health, 80, 978-983.
Meyer-Bahlburg, H.F.L., Ehrhardt, A.A., Exner, T.M., Calderwood, M., & Gruen, R.S. (1988). Sexual Risk Behavior Assessment Schedule-Adult, Baseline Interview for Gay Men (SERBAS-A-HOM-1). Unpublished. College of Physicians and Surgeons of Columbia University, Department of Psychiatry, New York, NY.
Meyer-Bahlburg, H. F. L., Exner, T. M., Lorenz, G., Gruen, R., Gorman, J. M. &Ehrhardt, A. A. (1991). Sexual risk behavior, sexual functioning, and HIV-disease progression in gay men. Journal of Sex Research, 28, 3-27.
Montgomery, S.B., Joseph, J.G., Becker, M., Ostrow, D., Kessler, R.C., Kirscht, J.P. (1989). The health belief model in understanding compliance with preventive recommendations for AIDS: How useful? AIDS Education and Prevention, 1, 303-323.
Mosher, D.L. (1987). Revised Mosher Sex Guilt Inventory (pp. 152-154). In C.M. Davis, W.L. Yarber, & S.L. Davis (Eds.), Sexuality-related measures: A compendium. Lake Mills, 10: Graphic Publishing Company.
Ostrow, D.G. (1989). Risk reduction for transmission of Acquired Immunodeficiency virus in high risk communities. Psychiatric Medicine, 7, 79-96.
Ostrow, D.G., Beltran, E., & Joseph, J. (1994). Sexual behavior research on a cohort of gay men, 1984-1990: Can we predict how men will respond to interventions. Archives of Sexual Behavior, 23, 531-552.
Paulhus, D.L. (1989). Measuring self-deception and impression management in self-reports: The Balanced Inventory of Desirable Responding -- Version 6. Unpublished manual.
Paulhus, D.L. (1991). Measurement and control of response bias. In J.P. Robinson, P. Shaver, & L. Wrightsman (Eds.), Measures of personality and social psychological attitudes (pp. 17-59). New York: Academic Press.
Prochaska, J.O., & DiClemente, C.C. (1983). Stages and processes of self-change of smoking: Toward an integrative model of change. Journal of Consulting and Clinical Psychology, 51, 390-395.
Prochaska, J.O., Redding, C.A., Harlow, L.L., Rossi, J.S., & Velicer, W.F. (1994). The Transtheoretical Model of Change and HIV prevention. A review. Health Education Quarterly, 21, 471-486.
Remien, R.H., Meyer-Bahlburg, H.F.L, Williams, J.B. W., Exner, T.M., Rabkin, J.G., Gorman, J. G., & Ehrhardt, A. A. (1992, July). A history of drug use disorders predicts current high-risk sexual behavior in gay men. Paper presented at the VIII International Congress on AIDS/III World STD Conference, Amsterdam, The Netherlands.
Rimberg, H.M., & Lewis, R.J. (1995). Older adolescents and AIDS: Correlates of self-reported safer sex practices. Journal of Research on Adolescence, 4, 453-464.
Rogers, E.M. (1983). Diffusion of innovation (third edition). New York: The Free Press.
Rosenstock, I.M. (1974). The Health Belief model and preventive health behavior. Health Education Monographs, 2, 354-386.
Smith, K.W., McGraw, S., Costa, L.A., & McKinlay, J.B. (1996). A self-efficacy scale for HIV risk behaviors: Development and evaluation. AIDS Education and Prevention, 8, 97-105.
Stall, R.D., McKusick, L., Wiley, J., Coates, T., & Ostrow, D. (1986). Alcohol and drug use during sexual activity and compliance with safe sex guidelines for AIDS: The AIDS Behavioral Research Project. Health Education Quarterly, 13, 359-371.
Upchurch, D.M., Weisman, C.S., Shepherd, M., Brookmeyer, R., Fox, R., Celentano, D.D., Colletta, L., & Hook, E. W. (1991). Interpartner reliability of reporting of recent sexual behaviors. American Journal of Epidemiology, 134, 1159-1166.
van der Velde, F.W., van der Plight, J., & Hooykaas, C. (1994). Perceiving AIDS-related risk: Accuracy as a function of difference in actual risk. Health Psychology, 13, 25-33.
Weinstein, N.D., & Nicolich, M. (1993). Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psychology, 12, 235-245.
Winslow, R. W., Franzini, L. R., & Hwang, J. (1992). Perceived peer norms, causal sex, and AIDS risk prevention. Journal of Applied Social Psychology, 22, 1809-1827.
Wulfert, E., & Wan, C.K. (1993). Condom use: A self-efficacy model. Health Psychology, 12, 346-353.
Zuckerman, M. (1984). Sensation seeking: A comparative approach to a human trait. The Behavioral and Brain Sciences, 7, 413-471.
Zuckerman, M., & Myers, P. L. (1983). Sensation seeking in homosexual and heterosexual males. Archives of Sexual Behavior, 12, 347-356.
Correspondence regarding this paper should be addressed to Prof. J. Roy Gillis, Department of Adult Education, Community Development, and Counselling Psychology, OISE/UT, 252 Bloor Street West, 7th floor, Toronto, Ontario, Canada M5S 1V6. email: JGILLIS@OISE.UTORONTO.CA.
|Printer friendly Cite/link Email Feedback|
|Author:||Gillis, J. Roy; Meyer-Bahlburg, Heino F.L.; Exner, Theresa; Ehrhardt, Anke A.|
|Publication:||The Canadian Journal of Human Sexuality|
|Date:||Mar 22, 1998|
|Previous Article:||Impact of an AIDS prevention video on AIDS-related perceptions.|
|Next Article:||Learning to strip: the socialization experiences of exotic dancers.|