Skip to main content
NCBI home page
Journal of Studies on Alcohol and Drugs logoLink to Journal of Studies on Alcohol and Drugs
. 2025 Mar 31;86(2):195–205. doi: 10.15288/jsad.24-00019

The Role of Substance Misuse and Sexual Victimization History on Sexual Violence Risk Perception in Men Who Have Sex With Men: A Novel Paradigm

Brooke E Wells a, Damon Mitchell b, D J Angelone c,,*, Dustin Fife c, Megan Korovich c, Elizabeth Pakan c, Justin Sitron a
PMCID: PMC11980405  PMID: 39315593

Abstract

Objective:

Sexual victimization (SV) is common among men who have sex with men (MSM), as is dating and sexual networking (DSN) app use. We developed a novel laboratory paradigm (“G-Date”) of sexual violence risk perception in DSN app environments and explored its validity and the role of substance misuse and SV history on sexual violence risk perception.

Method:

Using convenience sampling, we recruited 145 MSM to use G-Date to interact with two bogus speed dates whose responses were scripted to be risky or nonrisky. Risky dates displayed several cues indicating risk for sexual violence perpetration. Dependent variables included pre/post changes in ratings of their dating partner's appeal, ratings of the presence of each embedded risk cue, and the duration of the speed dates.

Results:

Compared with nonrisky dates, participants terminated risky dates significantly sooner, rated them higher in each risk cue, and gave them pre-to postdate ratings of partner appeal that declined significantly more. Participants' drug misuse was associated with reduced interest in the nonrisky date but not the risky date and predicted shorter speed date length and lower partner appeal ratings across date types. Substance-facilitated SV history interacted with alcohol misuse and date type to predict sexual violence risk perception.

Conclusions:

Results provide evidence for the paradigm's validity and suggest that drug misuse and substance-facilitated SV history shape MSM's risk perception in DSN apps. Efforts to prevent SV among MSM should consider individual characteristics, including substance misuse, in risk perception.

Sexual victimization (sv) is common among men who have sex with men (MSM) (Dame et al., 2020; Freeman et al., 2023; Klein et al., 2023; Siconolfi et al., 2021). Although rates vary, national surveys suggest that nearly 40% of MSM have experienced completed or attempted contact SV at least once (Chen et al., 2020). Beyond surviving the experiences of SV, the effects of SV may include mental health problems (Dame et al., 2020; Klein et al., 2023), sexual risk behavior (Turpin et al., 2021), and substance misuse (Duncan et al., 2018; Freeman et al., 2023).

A key context in which MSM experience risk for SV is dating and sexual networking (DSN) apps. DSN app use is common among MSM (Badal et al., 2018), who report using apps to negotiate sexual health and preferences (Albury & Byron, 2016), including consent (Dietzel, 2024). There is also the risk of in-person sexual violence perpetrated by people one has met via apps (Dietzel, 2021; Filice et al., 2022; Gewirtz-Meydan et al., 2024). In fact, qualitative data identify a range of negative experiences with app-based partners, including sexual violence via force, coercion, and intoxication (Woerner et al., 2023). Reflecting that reality, MSM report concerns in their DSN app encounters, and some engage in risk reduction strategies online and in person (Colliver, 2023; Dietzel, 2024; Wu & Ward, 2018). Learning about the ways MSM navigate potential cues for sexual violence risk in DSN apps could inform prevention efforts.

Experimental manipulations and measurement of sexual violence risk perception

Historically, sexual violence risk research has used written vignette-based paradigms from second- or third-person points of view (Angelone et al., 2022; Messman-Moore & Brown, 2006; Parrott et al., 2012; Rinehart et al., 2018). Researchers have also developed third-person audio or video vignettes of a dating or party situation embedded with risk cues (Marx & Gross, 1995; Parks et al., 2016). These paradigms are thought to measure participants' ability to detect SV risk potential. For example, participants rate their impression of the potential perpetrator or the point at which they feel uncomfortable or would leave the situation. Although useful in understanding risk perception, there is limited ecological validity in that appraisals are hypothetical, and research has exclusively focused on heterosexual encounters (Gidycz et al., 2006).

To address ecological validity in sexual violence risk perception research, researchers have developed laboratory-based paradigms that measure decision-making in actual interactions (Davis et al., 2014). For example, “Edudate” is an online speed dating paradigm designed to study the identification of risk cues among heterosexual college women (Angelone et al., 2009). Findings with Edudate have highlighted a complex interplay of participant and environmental factors that affect risk perception, suggesting that these elements can be parsed and potentially embedded in prevention interventions. To extend research beyond heterosexual populations, and using Edudate as a model, our team developed G-Date, a speed dating paradigm for MSM (Angelone et al., 2024). G-Date differs from Edudate in content (e.g., photos and stimuli designed for MSM vs. heterosexual women) and in appearance and functionality. Whereas Edudate was modeled on computer-based dating platforms, G-Date was designed for smartphones and to mimic contemporary DSN apps (e.g., Grindr). In G-Date, users are led to believe they are beta-testing a new app where they “try out” two speed dates. Unbeknownst to participants, their speed dates' responses are predetermined such that one partner portrays risk cues (risky date) and the other does not (nonrisky date). Guided by existing literature about individual-level predictors of sexual violence perpetration among MSM, and focusing on cues that could be communicated in text, we included three categories of risk cues: excessive substance use (Basile et al., 2021), antisociality (low empathy, dominance, aggressiveness, etc.) (Waldis et al., 2023), and internalized homophobia (Badenes-Ribera et al., 2019). We developed responses from the speed date partner that indicated the presence of one or more of these cues and refined these in pilot testing. Participants' ratings of the partner displaying risk cues (compared with one who does not) serve as an indicator of sexual violence risk perception, and their level of engagement with the risky date serves as an indicator of their behavioral response to the risky situation.

Factors shaping responses to risk cues

Although many factors could alter sexual violence risk perception, we focus on alcohol misuse, drug misuse, and SV history (any history and history of substance-facilitated SV [SFSV]) as they may be particularly relevant for and are underresearched among young MSM. Further, substance misuse and the effects of SV are potentially amenable to intervention. Research across populations finds that SV, particularly childhood sexual abuse, predicts subsequent SV (Blackburn et al., 2023; Caamano-Isorna et al., 2021; Fereidooni et al., 2022), and LGB people may be more likely than heterosexual people to experience revictimization (Daigle & Hawk, 2022). Although any SV history could increase the risk for subsequent victimization, SFSV may be more strongly connected to negative post-assault mental and behavioral health outcomes (Astle et al., 2024) and subsequent SFSV (Anderson et al., 2020). Some laboratory-based analog research depicting sexual violence found that SV history predicts lower risk detection and/or response to risk (Garrido-Macías et al., 2022; Neilson et al., 2018), whereas other research found no effects of SV history (Bridges et al., 2021; de Waal et al., 2022). Notably, Eshelman et al. (2015) found that SFSV history was associated with risk perception and behavioral response, whereas forcible SV history was not. Research also indicates that substance use and misuse increase SV risk (Domingos & de Lira, 2024; Edwards et al., 2021; Parrott et al., 2023; Tubman et al., 2024) and decrease bystander sexual violence risk perception (Marcantonio et al., 2024). Because MSM have high rates of substance use and misuse (Schuler et al., 2022), these may be key factors in relation to sexual violence risk perception.

Present study

MSM are at high risk for SV and many use DSN apps, which could pose risks for in-person SV. Although laboratory paradigms for sexual violence risk perception in women exist, corresponding paradigms geared toward MSM have yet to appear in the literature. G-Date was developed as a laboratory-based tool for studying sexual violence risk perception among MSM in the context of DSN app use. The current study has two aims. First, we examined the paradigm's initial validity and hypothesized that, compared with the nonrisky partner, participants would (a) rate the risky partner less appealing from pre-to postdate (partner appeal), (b) rate the risky partner higher in risk cues (cue detection), and (c) demonstrate less engagement with the risky partner (behavioral response). Second, we explored SV history (general and substance-facilitated) and alcohol and drug misuse as potential predictors of sexual violence risk perception outcomes. Regarding relationships between SV history, substance misuse, and our measures of sexual violence risk perception (partner appeal, cue detection, and behavioral response), no directional hypotheses were developed because of the novelty of the paradigm and lack of empirical work in this area; instead, we used exploratory data analyses to examine these associations.

Method

Before participant recruitment and data collection, the project was approved by the third author's institutional review board; all participants completed informed consent.

Participants

Potential participants were recruited via advertisements on social media platforms and in-person locations (e.g., LGBT centers, college campuses). Participant eligibility included (a) 18–30 years old, (b) identification as a man, (c) sexual contact with a man in the previous 12 months, and (d) DSN app use in the last 12 months. A total of 145 men enrolled between December 2022 and November 2023. The mean age was 24.3 years (SD = 3.4). Almost half of the sample (47%) identified as White, with the remaining identifying as Black/African American (14%), Latinx/e/o (12%), Asian (9%), or of mixed racial/ethnic identity (15%) or another race/ethnicity (3%). Most (81%) identified as cisgender, with 18% identifying as transgender men and one person identifying as nonbinary. Most (61%) identified as gay, queer (14%), or bisexual (14%), with the remainder identifying as pansexual (6%), same-gender-loving (2%), or another label (3%).

Procedure

Participants were told that they would be trying out G-Date, a new speed dating app for MSM that was undergoing beta testing. This software ostensibly permitted them to engage in an interaction limited to simultaneously answering a set of questions purported to predict compatibility. Participants were told that the experimenter was interested in obtaining feedback on the app before public release. G-Date was built on a platform like popular DSN apps; thus, its functionality and appearance would be familiar to those with DSN app experience.

Participants completed the study in person (on an iPad mini) or online via Zoom (using their own devices). After a brief app tutorial and assistance in taking and uploading a profile photo, participants completed the G-Date paradigm without assistance. Although participants were led to believe that they were engaging in real-time speed dates, the dating partners did not exist and all responses from the partner during question-and-answer (Q&A) exchanges were pre-determined. Each participant completed two dates (in random order), one with a partner whose Q&A responses were free of risk cues (nonrisky partner) and one with a partner whose Q&A responses contained risk cues (risky partner). Risk cues indicated a loss of control over alcohol and drug use, antisociality, and internalized homophobia. The research team developed the initial nonrisky and risky Q&A responses, which were then evaluated and refined via qualitative and quantitative pilot studies (Angelone et al., 2024). The G-Date paradigm protocol took about 45 minutes and occurred in two stages: profile setup/review of potential partners, and Q&A speed dates followed by partner ratings.

Profile setup/review of potential partners

Participants logged in and set up their profiles, providing information about themselves and what they want in partners, as is typical on DSN apps. Next, participants viewed photos of prospective partners who presumably met their standards and rated each photo (1 to 5 stars) to which they gave a “thumbs up.” After 60 photos, participants were informed that they had been matched with several partners and would have an opportunity to participate in two speed dates.

Q&A speed dates and partner ratings

Speed dates proceeded in a structured and predetermined manner. The app selected speed date partners whose profile photos participants rated highest (if they rated more than two profile pictures with the highest rating, the app randomly chose two). Before starting each speed date, participants rated their date's physical attractiveness, personality appeal, and their interest in meeting in person. Each speed date was conducted using “G-Date's patent-pending Q&A system,” in which they and their date would respond to questions designed to maximize the assessment of a partner's compatibility. During the Q&A, a question appeared on screen and participants had 90 seconds to respond. After 90 seconds, participants saw their response delivered to their date and received their date's response, which presumably had been typed at the same time. After the first three Q&As, participants could end or continue the date after each of the subsequent responses. In the nonrisky speed date, none of the partner's responses contained risk cues. In the risky speed date, there were no risk cues in the first two Q&A responses, but the third and subsequent responses included risk cues such that every participant was exposed to at least one risky response. For example, in response to the question, “What's one thing important to you in a partner, other than physical appearance?” the risky date indicated antisociality by responding, “IMHO physical appearance is the most important. No fats or femmes. You gotta know what the inside of a gym looks like. I take care of myself and I expect anyone I'm fucking to do the same.” To indicate internalized homophobia, the risky date responds to a question about what an ex would say about them with, “Guys I've been with would say I'm a MAN … masc4masc! whatever you want to call it or label it no rainbow flag shit for me.” As an example of a risk cue for substance misuse, the risky date responds to a question asking for their idea of a good night out with friends by saying, “I've def gotten a little out of control when I've been partying and then you never know what's going to happen. After a certain point, all bets are off. I do not believe half the shit my exes say I've done when I've been like blackout drunk. At least I haven't been arrested yet Inline graphic.” After the 12th Q&A response, the speed date automatically ended. After each speed date, participants rated their date on the same three attractiveness and interest scales as before the speed date. After completing both dates, participants saw side-by-side photos of their dates and compared them on six risk cue characteristics (see Measures).

Survey completion and debriefing

After the G-Date paradigm, participants completed a 20-minute Qualtrics survey, which included demographic questions, the Alcohol Use Disorders Identification Test (AUDIT), Drug Use Disorders Identification Test (DUDIT), Sexual Experiences Survey, and other personality and background measures as well as an open-ended item querying their perception of the purpose of the study (used to check participants' suspiciousness, which resulted in removing data from seven participants). Finally, participants were debriefed and compensated $30.

Measures

Ratings of dating partner. Participants rated (a) their partner's physical attractiveness (1 = extremely unattractive, 7 = extremely attractive), (b) their partner's personality (1 = extremely unappealing, 7 = extremely appealing), and (c) their interest in meeting their partner in person (1 = extremely uninterested, 7 = extremely interested) before and after each speed date. Differences in ratings from pre-to postdate served as measures of change in partner appeal.

Speed date engagement. The number of Q&A text exchanges (3–12) with each date before ending the speed date served as an indicator of behavioral response.

Partner comparisons. At the conclusion of the speed dates, participants compared their dating partners on six characteristics to examine their perceptions of differences between the two dates given the differences in embedded risk cues. Four items targeted antisociality (empathic vs. narcissistic, easygoing/genuine vs. controlling/manipulative, enjoyable vibe vs. creepy vibe, and clearly respectful of sexual boundaries vs. not respectful of sexual boundaries), one targeted internalized homonegativity (very comfortable with sexual identity vs. many issues), and one item targeted alcohol/drugs (clearly does not have issues vs. clearly has issues). To preserve the study's cover story, these were presented as a date comparison and keyed on a 1–7 scale such that higher ratings indicated more of the positively valenced (e.g., empathic) terms. Responses to these items served as our measures of risk cue detection.

Alcohol Use Disorders Identification Test. This 10-item scale measures past-year alcohol use and negative consequences (Babor et al., 2001). Items (scored 0 to 4) are summed for a total score. The AUDIT is reliable and valid across diverse populations (Lange et al., 2019). Cronbach's α in this sample was .82, and scores ranged from 0 to 27 (M = 5.41, SD = 4.81).

Drug Use Disorders Identification Test. Eleven items measure past-year drug use and negative consequences (Berman et al., 2005). Items (scored 0 to 4) are summed for a total score. The DUDIT is reliable and valid across populations (Hildebrand, 2015). Cronbach's α in this sample was .90, and scores ranged from 0 to 35 (M = 5.12, SD = 6.92).

Sexual Experiences Survey–Short Form Victimization (SES-SFV). The SES-SFV measures attempted and completed SV experiences since age 14, which it considers adult SV (Koss et al., 2007). Individual items capture oral, anal, vaginal, and nonpenetrative victimization (e.g., unwanted kissing/touching) through five perpetration tactics (including substance incapacitation). Although primarily used with women, scores are correlated with childhood SV and intimate partner victimization in men (Anderson et al., 2018). We used a version with updated instructions and genderneutral language (Canan et al., 2020). We scored the SES in two ways. First, we created a dichotomous variable (any SV) indicating any history of SV since age 14; participants with any history (including attempted or completed) were coded as 1 (and those without such history coded as 0). Most participants (68.5%) reported a history of adult SV. Second, we created a dichotomous variable (SFSV) indicating any history of SFSV since age 14 with participants coded as 1 if they endorsed someone “taking advantage of me when I was too drunk or out of it to stop what was happening” at least once (and 0 for those without such history), wherein 33.3% reported a history of SFSV.

Data analysis strategy

All analyses were conducted in R (Version 4.2.1), using the flexplot package for visualizations and model comparisons (Fife, 2022), the VGAM package for modeling a truncated Poisson model, and the party package for random forest (RF; Strobl et al., 2008). Assumptions were evaluated visually using diagnostic plots (e.g., residual-dependence plots, scale-location plots). Missing data were handled with listwise deletion, as there were only four missing observations.

To determine whether participants' activity in G-Date reflected differences in response to the risky versus nonrisky candidate, we conducted a series of t tests on the six items in the partner comparisons (risk cue detection), pre/post changes in dating partner ratings (partner appeal), and speed date engagement (behavioral response). We also conducted a series of linear models (i.e., analyses of covariance) to determine whether there was an interaction effect between experiment format (in-person vs. virtual) and date type for each dependent variable. Because none of these interactions had large effect sizes or reached statistical significance, the experiment format was not included in subsequent analyses.

To determine which predictor variables (date type, any SV, SFSV, DUDIT, and AUDIT) best predicted our risk perception variables, and given that our intentions were exploratory, we used a data mining strategy. This procedure was adopted from Fife and D'Onofrio (2022) and uses three sequential analytic strategies. First, RF—a data mining algorithm that aggregates hundreds (or thousands) of decision trees—was used to identify the predictors worthy of further exploration. The advantage of RFs is that they natively detect interactions and nonlinear effects, cross-validate well, and compute a “variable importance” metric for each predictor variable that allows researchers to rank order variables in terms of predictive utility. Second, after the RF algorithm identified the top predictors, we visualized the fits of the RF models for these same predictor variables. The visuals then guided the development of statistical models. When visuals suggested a specific hypothesis worthy of testing (e.g., a model with vs. without an interaction or a model with vs. without a specific main effect), we supplemented the visuals with model comparisons (i.e., hierarchical regressions). These model comparisons allow us to test hypotheses that any main or interaction effects are important above and beyond our covariates. See supplemental materials for a step-by-step description of our strategy. (Supplemental material appears as an online-only addendum to this article on the journal's website.)

Results

Paradigm validation

As summarized in Table 1, participants rated the risky candidate as having more issues with substances and sexual identity and more antisociality across indicators (see also Supplemental Table C and Supplemental Figure C). Regarding partner appeal, participants' pre-to postdate ratings demonstrated greater improvements in attractiveness, personality appeal, and interest in meeting the nonrisky date relative to the risky date. Similarly for behavioral response, participants demonstrated significantly less engagement with the risky date than the nonrisky date. In all cases, effect sizes were medium to large with large Bayes factors (BFs).

Table 1.

Means, 95% confidence intervals, Cohen's d values, Bayes factor (BF), p values, and R2 values for risk perception comparisons

graphic file with name jsad.24-00019tbl1.jpg

Dependent variable Nonrisky M (SD) Risky M (SD) Δ Lower Upper d BFa p R 2
Risk cue detection
 In control of drug/alcohol use 5.04 (1.19) 3.93 (1.44) 1.11 0.65 1.56 0.84 1.05E+08 <.0001 .151
 Comfort with sexual identity 6.36 (1.06) 4.02 (1.99) 2.35 1.79 2.90 1.47 2.03E+23 <.0001 .352
 Respects sexual boundaries 5.54 (1.25) 3.92 (1.53) 1.62 1.13 2.10 1.16 1.52E+15 <.0001 .252
 Empathic (vs. narcissistic) 5.48 (1.21) 3.48 (1.48) 2.00 1.53 2.47 1.48 3.39E+23 <.0001 .355
 Easygoing/genuine (vs. controlling) 5.75 (1.15) 3.82 (1.50) 1.93 1.47 2.39 1.44 4.35E+22 <.0001 .345
 Enjoyable vibe (vs. creepiness) 5.80 (1.10) 3.95 (1.56) 1.85 1.39 2.32 1.37 4.92E+20 <.0001 .322
Behavioral response
 Speed date engagement 7.65 (4.05) 5.1 (2.64) 2.55 1.37 3.73 0.75 1.52E+06 <.0001 .123
Partner appeal Nonrisky MΔ(SD) Risky MΔ(SD) Δ (Δ) Lower Upper d BFa p R 2
Attractiveness 0.04 (0.70) -0.45 (1.11) 0.49 0.17 0.81 0.53 4.78E+02 <.0001 .067
Personality appeal 1.03 (1.51) -1.01 (1.74) 2.04 1.48 2.60 1.25 3.65E+17 <.0001 .283
Interest in meeting 0.32 (1.40) -1.39 (1.75) 1.71 1.16 2.25 1.08 1.99E+13 <.0001 .227
Open in a new tab

Notes: Risk cue detection and partner appeal variables were rated on a 1–7 scale, whereas the behavioral response variable could range from 3 to 12. For all models, date type (risky vs. nonrisky) was the independent variable. For the lower table, the Δs indicate the difference between pre- and post-ratings, whereas Δ(Δ) represents the average difference between risky and nonrisky differences.

a

The BF is a ratio of evidence in favor of one model over the other. In this case, the BF evaluates the evidence in favor of a model with date type as a predictor versus a model without date type as a predictor. Because all BFs are larger than one (and all are significantly larger than one), we can conclude that date type affects all 10 dependent variables.

Predictors of risk perception: A data mining analysis

Table 2 summarizes the findings from the data mining analyses. Date Type and Predate partner appeal ratings served as covariates. Predate ratings were not of interest because we expected them to be correlated with postdate ratings, and Date Type was included since all outcomes were significantly different according to date type. As such, the reduced model always contained Date Type and Predate ratings and represented the “status quo” model (Rodgers, 2010). The remaining variables were our key focus and comprise any SV, SFSV, DUDIT, and AUDIT. We constructed “full” models such that they included both covariates from the reduced models and any variables suggested by the data mining/visuals analysis.

Table 2.

Summary statistics for exploratory models that include predictors of interest

graphic file with name jsad.24-00019tbl2.jpg

Full (Block 2) Model Reduced (Block 1) Model R 2 Semi-partial R2 p BF
Risk cue detection
 Enjoyable Vibe = Date Type + DUDIT + DUDIT × Date Type Enjoyable Vibe = Date Type .358 .017 .036 0.115
 Control of drug/alcohol use = Date Type + Any SFSV + Any SFSV × Date Type Control of Drug/Alcohol Use = Date Type .176 .023 .008 2.32
 Empathy = Date Type + AUDIT + Any SFSV + AUDIT × Any SFSV Empathy Ratings = Date Type .368 .013 .021 0.935
Partner appeal
 Personality Appeal = Predate + Date Type + DUDIT + DUDIT × Date Type Personality Appeal = Predate + Date Type .339 .021 .022 0.187
 Attractiveness = Predate + Date Type + DUDIT Attractiveness = Predate + Date Type .468 .009 .038 0.551
Behavioral response
 Speed date engagement = Date Type + DUDIT Speed Date Engagement = Date Type n.a. n.a. .002 8.04
Open in a new tab

Notes: Semi-partial R2 valuate the added variance explained by the full model relative to the reduced model. (For censored Poisson models [or even Poisson models], it does not make sense to compute R2 values so these metrics are not available.) P values test the null hypothesis that the full and reduced models are identical. Bayes factors (BFs) report the evidence in favor of the full model over the evidence in favor of the reduced model. In other words, values greater than one indicate the full model is preferred, while values less than one indicate the reduced model is preferred. Bolded models show when the p value and BF agree on which model is preferred. All other models show disagreement between the BF and p values, with the p values always favoring the full model and the BF favoring the reduced. In these instances, the choice of which model is preferred is ambiguous. n.a. = not applicable.

As noted above, we used data mining to eliminate variables, visualize predictors to identify plausible hypotheses to test, and build full and reduced models for each dependent variable. The specification of the chosen full model is shown in Column 1 of Table 2, and the specification for the reduced model is shown in Column 2. For each regression, we report the full model's R2 as well as the semi-partial R2. We also report the p value and BFs that compare the full/reduced model; p values less than .05 and BF values greater than 1 indicate that the full model is preferred. Because the “speed date engagement” dependent variable was a count variable (number of questions engaged before ending the date) with only 12 questions possible, we used a “censored” Poisson regression (using the VGAM package in R).

As summarized in Table 2, the p value favors the full model in all cases, whereas the BF favors the full model for the speed date engagement variable and for the alcohol/drug issues variable.1 The DUDIT emerged as a potentially important predictor in three models: personality appeal (explaining 2.1% of the variance, p = .022, BF = 0.187), attractiveness (explaining almost 1% of the variance, p = .038, BF = 0.551), and speed date engagement (p = .002, BF = 8.04). The models for outcomes wherein no predictors beyond the covariates emerged as important in the data mining algorithm are not presented in Table 2 or in the figures. The relationships between the identified predictors and relevant dependent variables are displayed in Figures 1 and 2. Higher DUDIT scores were associated with lower attractiveness ratings and speed date engagement across date conditions. Higher DUDIT scores were also associated with lower personality appeal ratings and less enjoyable/creepier vibe ratings in nonrisky dates, but not in risky dates. Although having a history of any SV did not emerge as an important predictor in any model, SFSV emerged as an important predictor in two models: alcohol/drug use issues (explaining 2.3% of the variance, p = .008, BF = 2.32) and empathy (predicting 1.3% of the variance, p = .021, BF = 0.935). Participants with no SFSV history rated nonrisky dates as having more control over their substance use compared with risky dates, whereas those with SFSV history rated the two dates similarly. AUDIT scores interacted with SFSV history such that among participants with no SFSV history, empathy ratings increased as AUDIT scores increased (controlling for date type), whereas for those with SFSV history, empathy ratings decreased as AUDIT scores increased.

Figure 1.

Figure 1.

Open in a new tab

Effects of Drug Use Disorders Identification Test (DUDIT) on attractiveness, enjoyable vibe, personality appeal, and speed date engagement. Note: In both plots, the black dots/lines represent the scores under the risky condition and the gray triangles/dotted lines represent the scores under the nonrisky condition.

Figure 2.

Figure 2.

Open in a new tab

Effects of substance-facilitated sexual victimization (SFSV) on perceptions of control of alcohol/drug use and empathy. AUDIT = Alcohol Use Disorders Identification Test.

Discussion

Our results provide initial support for the internal and construct validity of this novel paradigm: Compared with nonrisky dates, participants perceived risky dates as having higher levels of each of the risk cues; their ratings of personality appeal, attractiveness, and interest in meeting declined significantly more for risky dates; and participants engaged significantly less with the risky date. In addition, exploratory analyses suggest that substance misuse may shape the ways MSM engage with potential partners in DSN apps. Higher DUDIT scores (reflecting greater consumption and problematic drug use) were associated with lower attractiveness ratings and a shorter date in both conditions. Higher DUDIT scores were also associated with less appealing personality ratings and less enjoyable/creepier ratings for the nonrisky date.

MSM experiencing drug misuse may be less attracted to partners who do not indicate sexual violence risk cues because of higher sensation seeking (Regan et al., 2020). Participants may be less attracted to partners they perceive to be low in sensation seeking or unlikely to engage in activities to fulfill their desire for novelty and sensation. Future research should parse the dimensions of substance misuse (heavy consumption vs. negative consequences of use) and specific substances and use motivations to elucidate the mechanisms of effect on sexual violence risk perception. MSM with high DUDIT scores may have found dates in both conditions less attractive and engaged less because of their impressions from profile pictures. Artificial intelligence–generated images showed men with normative looks, perhaps suggesting less interest in substance use.

Having any history of adult SV was not associated with risk perception or behavioral response. However, SFSV history was associated with perceiving the substance-related issues of risky and nonrisky dates similarly, whereas those with no such history more strongly distinguished substance-related risk when it was present. SFSV history also interacted with AUDIT scores in predicting the empathic/narcissistic risk cue rating, such that higher AUDIT scores were associated with greater perceptions of empathy (controlling for date type) among those without SFSV histories and lower perceptions of empathy among those with SFSV histories. Findings support previous research suggesting unique associations between substance-involved SV and risk perception and suggest individual factors, such as alcohol misuse, that may shape these associations. There could be an indirect path from SV history to risk perceptions via substance use or misuse, as demonstrated in other research (Blackburn et al., 2023; Cusack et al., 2021).

Our results may also suggest strategies that therapists and educators could use to support MSM's safety as they navigate DSN apps and experiences. Sexuality education should include skills training in consent as well as sexual violence risk detection that applies to online and in-person settings and supports navigating online to in-person transitions. As DSN app use is common among adolescents (Macapagal et al., 2020), adolescent sex education should teach culturally appropriate protective behavioral strategies for this context. Further, sexuality and relationship education should address the role of substance use in sexual and romantic relationships to support sexual safety and relationship health even in the context of substance use. Because many young people lack access to school-based sex education, this sex education should happen in a range of settings, such as social service agencies where professionals have rapport and experience working with youth (e.g., Lavie-Ajayi, 2020). Substance use prevention/intervention efforts should also address connections between use and sexual behavior, such as by developing consent models that account for substance use (Smith et al., 2021).

Limitations

Generalizing these findings to sexual violence risk perception among MSM should be done considering several limitations. Participants may have behaved differently in this app than in their regular DSN app use. Although G-Date stimuli demonstrated internal and construct validity, there are limitations in ecological validity: There may have been different contextual factors related to participation in G-Date (e.g., beta testing an app for compensation, experimenter presence, limited profile information) that influenced participants to behave riskier or more cautiously in G-Date than they might in their regular DSN app use. A related limitation concerns the generalizability of the risky behavior in a dating app to decision-making during in-person meetings. The relative anonymity and safety of online interactions could have disinhibited online behavior but not precluded safety concerns regarding in-person meetings. Further, the number of Q&As before date termination could be examined alongside participant responses because more engagement could reflect direct responses to risk cues. We plan to address these considerations as well as predictive validity in follow-up surveys that assess participants' subsequent DSN behavior, SV, and substance use, among other variables.

We note three limitations related to our sample. First, given that nearly 20% of the sample are transgender or nonbinary, their past dating app experiences (e.g., rejection and fetishization) (Griffiths & Armstrong, 2024) and high SV rates (Messinger et al., 2022) are likely relevant to their risk perception and navigation; future research should examine the role of gender modality. Second, the sample's high education level may limit the generalizability of our findings. Finally, about 5% of the sample (n = 7) expressed suspiciousness about the true purpose of the study and were excluded from analyses, suggesting some limitations to the cover story. However, this percentage is similar to manipulation checks in comparable paradigms (Angelone et al., 2009, 2022).

Future directions

Although many factors are potentially related to sexual violence risk in MSM, this initial test of the validity of the paradigm examined only two overarching constructs: SV history and substance misuse. Nonetheless, a more comprehensive evaluation of the myriad factors that likely predict risk perception in MSM should be conducted, including constructs articulated in minority stress theory, such as internalized homonegativity, discrimination, and social support (Parrott et al., 2024; Tubman et al., 2023). Further, although we used self-reported substance misuse as predictors of speed date engagement and perceptions, future research could use G-Date in conjunction with alcohol administration (beyond alcohol misuse), particularly as this study provides evidence of the paradigm's validity. For example, participants could be administered alcohol (or a control beverage) and asked to engage the G-Date paradigm to explore the role of in vivo alcohol intoxication on sexual violence risk perception and its interaction with historical and personality characteristics. Finally, recognizing the various motivations for DSN app use (Choi & Bauermeister, 2022), experimental research could also parse distinctions in sexual violence risk perception according to sexual arousal, perceptions of partner availability (e.g., in rural vs. urban locations), participant gender modality (i.e., transgender vs. cisgender), and sexual position (e.g., insertive vs. receptive).

Acknowledgments

The authors acknowledge the contributions of student research assistants at Rowan University who played a primary role in developing the materials for the app, facilitating focus groups/interviews, as well as assisting with recruitment and data collection, including Corey Doremus, Alexandra Nicoletti, Sarah Sosland, and Ebru Yucel. The authors also acknowledge the contributions of Stan Kurkovsky for leading the team of computer science students at Central Connecticut State University, who developed the app used in the laboratory paradigm, including Rachel Bernard, Chris Conlon, Shad Nadeau, Donna Nguyen, and Antonio Zea. We also thank Bob Freeman at the National Institute on Alcohol Abuse and Alcoholism for his support during the development of the grant that supported this project. Finally, we acknowledge the consultants who supported the development and execution of this project, including William George, Kelly Cue Davis, and David Pantalone.

Footnotes

This research was funded by National Institute on Alcohol Abuse and Alcoholism Grant No. R15AA028637 (principal investigator: D. J.Angelone).

1

It is not surprising or concerning that p values and BFs disagree. They are asking different questions (p value relates to the probability of the full/reduced models being equal in fits, whereas BFs evaluate the evidence in favor of one model over the evidence in favor of the other) and emerge from different philosophical traditions (frequentists vs. Bayesians). When both metrics agree (e.g., for speed date engagement variable), there is greater confidence in conclusions than when the metrics disagree (e.g., for Personality Appeal). When metrics disagree, it is harder to choose a model and signals less confidence in whatever choice one makes.

References

  1. Albury K., Byron P. Safe on my phone? Same-sex attracted young people's negotiations of intimacy, visibility, and risk on digital hook-up apps. Social Media + Society. 2016;2(4) doi: 10.1177/2056305116672887. [DOI] [Google Scholar]
  2. Anderson J. C., Chugani C. D., Jones K. A., Coulter R. W. S., Chung T., Miller E. Characteristics of precollege sexual violence victimization and associations with sexual violence revictimization during college. Journal of American College Health. 2020;68(5):509–517. doi: 10.1080/07448481.2019.1583237. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Anderson R. E., Cahill S. P., Delahanty D. L. The psychometric properties of the Sexual Experiences Survey–Short Form Victimization (SES-SFV) and characteristics of sexual victimization experiences in college men. Psychology of Men & Masculinity. 2018;19(1):25–34. doi: 10.1037/men0000073. [DOI] [Google Scholar]
  4. Angelone D. J., Mitchell D., Carola K. Tolerance of sexual harassment: A laboratory paradigm. Archives of Sexual Behavior. 2009;38(6):949–958. doi: 10.1007/s10508-008-9421-2. [DOI] [PubMed] [Google Scholar]
  5. Angelone D. J., Mitchell D., Wells B., Korovich M., Nicoletti A., Fife D. Assessment of sexual violence risk perception in men who have sex with men: Proposal for the development and validation of “G-Date.”. JMIR Research Protocols. 2024;13:e57600. doi: 10.2196/57600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Angelone D. J., Mitchell D., Yucel E., Davis K. C., George W. H. An evaluation of the methodological underpinnings of a laboratory paradigm for the study of sexual violence. Journal of Interpersonal Violence. 2022;37(23–24):NP22092–NP22113. doi: 10.1177/08862605211068081. [DOI] [PubMed] [Google Scholar]
  7. Astle S., McAllister P., Spencer C., Rivas-Koehl M., Toews M., Anders K. Mental health and substance use factors associated with sexual violence victimization and perpetration in university samples: A meta-analysis. Sexuality Research and Social Policy. 2024;21(1):388–399. doi: 10.1007/s13178-023-00830-2. [DOI] [Google Scholar]
  8. Babor T. F., Higgins-Brindle J. C., Saunders J. B., Monteiro M. G. AUDIT: The Alcohol Use Disorders Identification Test: Guidelines for use in primary health care. 2nd ed. World Health Organization; 2001. https://iris.who.int/handle/10665/67205 [Google Scholar]
  9. Badal H. J., Stryker J. E., DeLuca N., Purcell D. W. Swipe right: Dating website and app use among men who have sex with men. AIDS and Behavior. 2018;22(4):1265–1272. doi: 10.1007/s10461-017-1882-7. [DOI] [PubMed] [Google Scholar]
  10. Badenes-Ribera L., Sánchez-Meca J., Longobardi C. The relationship between internalized homophobia and intimate partner violence in same-sex relationships: A meta-analysis. Trauma, Violence, & Abuse. 2019;20(3):331–343. doi: 10.1177/1524838017708781. [DOI] [PubMed] [Google Scholar]
  11. Basile K. C., Smith S. G., Liu Y., Lowe A., Gilmore A. K., Khatiwada S., Kresnow M.-j. Victim and perpetrator characteristics in alcohol/drug-involved sexual violence victimization in the U.S. Drug and Alcohol Dependence. 2021;226:108839. doi: 10.1016/j.drugalcdep.2021.108839. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Berman A. H., Bergman H., Palmstierna T., Schlyter F. Evaluation of the Drug Use Disorders Identification Test (DUDIT) in criminal justice and detoxification settings and in a Swedish population sample. European Addiction Research. 2005;11(1):22–31. doi: 10.1159/000081413. [DOI] [PubMed] [Google Scholar]
  13. Blackburn A. M., Bystrynski J. B., Rieger A., Garthe R. C., Piasecki M., Allen N. E. Sexual assault revictimization among sexual minority individuals: A systematic review and meta-analysis. Psychology of Violence. 2023;13(4):286–296. doi: 10.1037/vio0000477. [DOI] [Google Scholar]
  14. Bridges A. J., Dueweke A. R., Marcantonio T. L., Ham L. S., Wiersma-Mosley J. D., Jozkowski K. N. Two studies investigating associations between sexual assault victimization history and bystander appraisals of risk. Violence Against Women. 2021;27(10):1736–1757. doi: 10.1177/1077801220940390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Caamano-Isorna F., Adkins A., Moure-Rodríguez L., Conley A. H., Dick D. Alcohol use and sexual and physical assault victimization among university students: Three years of follow-up. Journal of Interpersonal Violence. 2021;36(7–8):NP3574–NP3595. doi: 10.1177/0886260518780413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Canan S. N., Jozkowski K. N., Wiersma-Mosley J., Blunt-Vinti H., Bradley M. Validation of the Sexual Experience Survey–Short Form Revised using lesbian, bisexual, and heterosexual women's narratives of sexual violence. Archives of Sexual Behavior. 2020;49(3):1067–1083. doi: 10.1007/s10508-019-01543-7. [DOI] [PubMed] [Google Scholar]
  17. Chen J., Walters M. L., Gilbert L. K., Patel N. Sexual violence, stalking, and intimate partner violence by sexual orientation, United States. Psychology of Violence. 2020;10(1):110–119. doi: 10.1037/vio0000252. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Choi S. K., Bauermeister J. A latent profile analysis of online dating patterns among single young men who have sex with men. AIDS and Behavior. 2022;26(4):1279–1288. doi: 10.1007/s10461-021-03485-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Colliver B. Understanding perceptions of victimization: A critical analysis of gay and bisexual male Grindr users negotiations of safety and risk. The British Journal of Criminology. 2024;64(2):487–502. doi: 10.1093/bjc/azad043. [DOI] [Google Scholar]
  20. Cusack S. E., Bourdon J. L., Bountress K., Saunders T. R., Kendler K. S., Dick D. M., Amstadter A. B. Prospective predictors of sexual revictimization among college students. Journal of Interpersonal Violence. 2021;36(17–18):8494–8518. doi: 10.1177/0886260519849680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Daigle L. E., Hawk S. R. Sexual orientation, revictimization, and polyvictimization. Sexuality Research and Social Policy. 2022;19(1):308–320. doi: 10.1007/s13178-021-00543-4. [DOI] [Google Scholar]
  22. Dame J., Oliffe J. L., Hill N., Carrier L., Evans-Amalu K. Sexual violence among men who have sex with men and two-spirit peoples: A scoping review. The Canadian Journal of Human Sexuality. 2020;29(2):240–248. doi: 10.3138/cjhs.2020-0014. [DOI] [Google Scholar]
  23. Davis K. C., Parrott D. J., George W. H., Tharp A. T., Hall G. C. N., Stappenbeck C. A. Studying sexual aggression: A review of the evolution and validity of laboratory paradigms. Psychology of Violence. 2014;4(4):462–476. doi: 10.1037/a0037662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. de Waal M. M., Christ C., Messman T. L., Dekker J. J. M. Changes in risk perception after sexual victimization: Are we following the right track? Journal of Interpersonal Violence. 2022;37(13–14):NP11699–NP11719. doi: 10.1177/0886260519848790. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Dietzel C. “That's straight-up rape culture”: Manifestations of rape culture on Grindr. The Emerald international handbook of technology-facilitated violence and abuse. 2021:351–368. doi: 10.1108/978-1-83982-848-520211026. [DOI] [Google Scholar]
  26. Dietzel C. Clickable consent: How men who have sex with men understand and practice sexual consent on dating apps and in person. The Journal of Sex Research. 2024;61(3):481–494. doi: 10.1080/00224499.2023.2235584. [DOI] [PubMed] [Google Scholar]
  27. Domingos V. A. C., de Lira A. N. Risk and protective factors associated with intimate partner violence with gay men: A scoping review. Trauma, Violence, & Abuse. 2024;25(3):2264–2281. doi: 10.1177/15248380231209738. [DOI] [PubMed] [Google Scholar]
  28. Duncan D. T., Goedel W. C., Stults C. B., Brady W. J., Brooks F. A., Blakely J. S., Hagen D. A study of intimate partner violence, substance abuse, and sexual risk behaviors among gay, bisexual, and other men who have sex with men in a sample of geosocial-networking smartphone application users. American Journal of Men's Health. 2018;12(2):292–301. doi: 10.1177/1557988316631964. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Edwards K. M., Siller L., Littleton H., Wheeler L., Chen D., Sall K., Lim S. Minority stress and sexual partner violence victimization and perpetration among LGBQ+ college students: The moderating roles of hazardous drinking and social support. Psychology of Violence. 2021;11(5):445–454. doi: 10.1037/vio0000394. [DOI] [Google Scholar]
  30. Eshelman L. R., Messman-Moore T. L., Sheffer N. The importance of substance-related sexual victimization: Impact on substance use and risk perception in female college students. Journal of Interpersonal Violence. 2015;30(15):2616–2635. doi: 10.1177/0886260514553630. [DOI] [PubMed] [Google Scholar]
  31. Fereidooni F., Daniels J., Lommen M. Predictors of revictimization in online dating. Journal of Interpersonal Violence. 2022;37(23–24):NP23057–NP23074. doi: 10.1177/08862605211073715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Fife D. Flexplot: Graphically-based data analysis. Psychological Methods. 2022;27(4):477–496. doi: 10.1037/met0000424. [DOI] [PubMed] [Google Scholar]
  33. Fife D. A., D'Onofrio J. Common, uncommon, and novel applications of random forest in psychological research. Behavior Research Methods. 2022;55(5):2447–2466. doi: 10.3758/s13428-022-01901-9. [DOI] [PubMed] [Google Scholar]
  34. Filice E., Abeywickrama K. D., Parry D. C., Johnson C. W. Sexual violence and abuse in online dating: A scoping review. Aggression and Violent Behavior. 2022;67:101781. doi: 10.1016/j.avb.2022.101781. [DOI] [Google Scholar]
  35. Freeman J. Q., Cha S., Wejnert C., Baugher A. Physical and sexual violence and sexual behaviors among men who have sex with men in 22 U.S. cities—National HIV Behavioral Surveillance, 2017. Journal of Interpersonal Violence. 2023;38(1–2):37–59. doi: 10.1177/08862605221078821. [DOI] [PubMed] [Google Scholar]
  36. Garrido-Macías M., Valor-Segura I., Expósito F. Women's experience of sexual coercion and reactions to intimate partner sexual violence. Journal of Interpersonal Violence. 2022;37(11–12):NP8965–NP8988. doi: 10.1177/0886260520980394. [DOI] [PubMed] [Google Scholar]
  37. Gewirtz-Meydan A., Volman-Pampanel D., Opuda E., Tarshish N. Dating apps: A new emerging platform for sexual harassment? A scoping review. Trauma, Violence, & Abuse. 2024;25(1):752–763. doi: 10.1177/15248380231162969. [DOI] [PubMed] [Google Scholar]
  38. Gidycz C. A., McNamara J. R., Edwards K. M. Women's risk perception and sexual victimization: A review of the literature. Aggression and Violent Behavior. 2006;11(5):441–456. doi: 10.1016/j.avb.2006.01.004. [DOI] [Google Scholar]
  39. Griffiths D. A., Armstrong H. L. “They were talking to an idea they had about me”: A qualitative analysis of transgender individuals' experiences using dating apps. The Journal of Sex Research. 2024;61(1):119–132. doi: 10.1080/00224499.2023.2176422. [DOI] [PubMed] [Google Scholar]
  40. Hildebrand M. The psychometric properties of the Drug Use Disorders Identification Test (DUDIT): A review of recent research. Journal of Substance Abuse Treatment. 2015;53:52–59. doi: 10.1016/j.jsat.2015.01.008. [DOI] [PubMed] [Google Scholar]
  41. Klein L. B., Dawes H. C., James G., Hall W. J., Rizo C. F., Potter S. J., Martin S. L., Macy R. J. Sexual and relationship violence among LGBTQ+ college students: A scoping review. Trauma, Violence, & Abuse. 2023;24(4):2196–2209. doi: 10.1177/15248380221089981. [DOI] [PubMed] [Google Scholar]
  42. Koss M. P., Abbey A., Campbell R., Cook S., Norris J., Testa M., Ullman S., West C., White J. Revising the SES: A collaborative process to improve assessment of sexual aggression and victimization. Psychology of Women Quarterly. 2007;31(4):357–370. doi: 10.1111/j.1471-6402.2007.00385.x. [DOI] [Google Scholar]
  43. Lange S., Shield K., Monteiro M., Rehm J. Facilitating screening and brief interventions in primary care: A systematic review and meta-analysis of the AUDIT as an indicator of alcohol use disorders. Alcoholism: Clinical and Experimental Research. 2019;43(10):2028–2037. doi: 10.1111/acer.14171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Lavie-Ajayi M. Informal sex education by youth practitioners. Young. 2020;28(5):485–501. doi: 10.1177/1103308819899564. [DOI] [Google Scholar]
  45. Macapagal K., Kraus A., Moskowitz D. A., Birnholtz J. Geosocial networking application use, characteristics of app-met sexual partners, and sexual behavior among sexual and gender minority adolescents assigned male at birth. The Journal of Sex Research. 2020;57(8):1078–1087. doi: 10.1080/00224499.2019.1698004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Marcantonio T. L., Haikalis M., Misquith C., Leone R. M. Alcohol's effects on the bystander decision-making model: A systematic literature review. The Journal of Sex Research. 2024;61(5):783–798. doi: 10.1080/00224499.2023.2267547. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Marx B. P., Gross A. M. Date rape: An analysis of two contextual variables. Behavior Modification. 1995;19(4):451–463. doi: 10.1177/01454455950194003. [DOI] [Google Scholar]
  48. Messinger A. M., Guadalupe-Diaz X. L., Kurdyla V. Transgender polyvictimization in the U.S. Transgender Survey. Journal of Interpersonal Violence. 2022;37(19–20):NP18810–NP18836. doi: 10.1177/08862605211039250. [DOI] [PubMed] [Google Scholar]
  49. Messman-Moore T. L., Brown A. L. Risk perception, rape, and sexual revictimization: A prospective study of college women. Psychology of Women Quarterly. 2006;30(2):159–172. doi: 10.1111/j.1471-6402.2006.00279.x. [DOI] [Google Scholar]
  50. Neilson E. C., Bird E. R., Metzger I. W., George W. H., Norris J., Gilmore A. K. Understanding sexual assault risk perception in college: Associations among sexual assault history, drinking to cope, and alcohol use. Addictive Behaviors. 2018;78:178–186. doi: 10.1016/j.addbeh.2017.11.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Parks K. A., Levonyan-Radloff K., Dearing R. L., Hequembourg A., Testa M. Development and validation of a video measure for assessing women's risk perception for alcohol-related sexual assault. Psychology of Violence. 2016;6(4):573–585. doi: 10.1037/a0039846. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Parrott D., Leone R. M., Hequembourg A., Shorey R. C., Eckhardt C., Stuart G. L. An integrative model of alcohol-facilitated intimate partner aggression perpetration in sexual and gender diverse couples. Journal of Studies on Alcohol and Drugs. 2025;86(2):218–228. doi: 10.15288/jsad.24-00008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Parrott D. J., Bresin K., Hequembourg A., Velia B., Swartout K. M., Stappenbeck C. A., Masyn K. E., Grom J. L. Dyadic effects of minority stress and problematic alcohol use on sexual intimate partner violence in same sex couples. Aggressive Behavior. 2023;49(3):198–208. doi: 10.1002/ab.22072. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Parrott D. J., Tharp A. T., Swartout K. M., Miller C. A., Hall G. C. N., George W. H. Validity for an integrated laboratory analogue of sexual aggression and bystander intervention. Aggressive Behavior. 2012;38(4):309–321. doi: 10.1002/ab.21429. [DOI] [PubMed] [Google Scholar]
  55. Regan T., Thamotharan S., Hahn H., Harris B., Engler S., Schueler J., Fields S. A. Sensation seeking, sexual orientation, and drug abuse symptoms in a community sample of emerging adults. Behavioural Pharmacology. 2020;31(1):102–107. doi: 10.1097/fbp.0000000000000523. [DOI] [PubMed] [Google Scholar]
  56. Rinehart J. K., Yeater E. A., Treat T. A., Viken R. J. Cognitive processes underlying the self-other perspective in women's judgments of sexual victimization risk. Journal of Social and Personal Relationships. 2018;35(10):1381–1399. doi: 10.1177/0265407517713365. [DOI] [Google Scholar]
  57. Rodgers J. L. The epistemology of mathematical and statistical modeling: A quiet methodological revolution. American Psychologist. 2010;65(1):1–12. doi: 10.1037/a0018326. [DOI] [PubMed] [Google Scholar]
  58. Schuler M. S., Collins R. L., Ramchand R. Disparities in use/misuse of specific illicit and prescription drugs among sexual minority adults in a national sample. Substance Use & Misuse. 2022;57(3):461–471. doi: 10.1080/10826084.2021.2019776. [DOI] [PubMed] [Google Scholar]
  59. Siconolfi D., Storholm E. D., Vincent W., Pollack L., Rebchook G. M., Huebner D. M., Peterson J. L., Kegeles S. M. Prevalence and correlates of sexual violence experienced by young adult Black men who have sex with men. Archives of Sexual Behavior. 2021;50(8):3621–3636. doi: 10.1007/s10508-021-02011-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Smith L. A., Kolokotroni K. Z., Turner-Moore R. Making and communicating decisions about sexual consent during drug-involved sex: A thematic synthesis. The Journal of Sex Research. 2021;58(4):469–487. doi: 10.1080/00224499.2019.1706072. [DOI] [PubMed] [Google Scholar]
  61. Strobl C., Boulesteix A.-L., Kneib T., Augustin T., Zeileis A. Conditional variable importance for random forests. BMC Bioinformatics. 2008;9(1) doi: 10.1186/1471-2105-9-307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Tubman J. G., Lee J., Moore C. Factors associated with sexual victimization among transgender emerging adults. Journal of Interpersonal Violence. 2024;39(11–12):2832–2852. doi: 10.1177/08862605231221836. [DOI] [PubMed] [Google Scholar]
  63. Tubman J. G., Moore C., Lee J., Shapiro A. J. Multivariate patterns of substance use, minority stress and environmental violence associated with sexual revictimization of lesbian and bisexual emerging adult women. Journal of Lesbian Studies. 2023;29(1):36–56. doi: 10.1080/10894160.2023.2240552. [DOI] [PubMed] [Google Scholar]
  64. Turpin R. E., Salerno J. P., Rosario A. D., Boekeloo B. Victimization, substance use, depression, and sexual risk in adolescent males who have sex with males: A syndemic latent profile analysis. Archives of Sexual Behavior. 2021;50(3):961–971. doi: 10.1007/s10508-020-01685-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Waldis L., Herzberg P. Y., Herold J., Nothhelfer K., Krebs J., Troche S. Predictors of sexual aggression among gay men and lesbian women: An application of Malamuth's confluence model. Aggressive Behavior. 2023;49(2):154–164. doi: 10.1002/ab.22062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Woerner J., Chadwick S. B., Antebi-Gruszka N., Siegel K., Schrimshaw E. W. Negative sexual experiences among gay, bisexual, and other men who have sex with men using GPS-enabled hook-up apps and websites. The Journal of Sex Research. 2024;61(8):1142–1157. doi: 10.1080/00224499.2023.2269930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Wu S., Ward J. The mediation of gay men's lives: A review on gay dating app studies. Sociology Compass. 2018;12(2) doi: 10.1111/soc4.12560. [DOI] [Google Scholar]

Articles from Journal of Studies on Alcohol and Drugs are provided here courtesy of Rutgers University. Center of Alcohol Studies

RESOURCES