Abstract
Nonconsensual condom removal (NCCR) is a form of sexual violence that a significant minority of men use to obtain unprotected sex. Experiencing NCCR is associated with serious physical and mental health consequences, such as sexually transmitted infections, unplanned pregnancies, anxiety, and depression. Alcohol use is known to be associated with sexual violence in general; however, limited research has been done on the association between alcohol-related factors and NCCR specifically. Thus, the present study examined the associations between event-level drinking, daily drinking, drinking motives, alcohol expectancies, and NCCR. Young, single, heterosexually active men (N = 96) completed cross-sectional measures of their NCCR behavior, event-level drinking, drinking motives, and alcohol expectancies. A total of 19 (19.8%) participants reported successfully engaging in NCCR at least once since age 14. Results showed that consuming more alcohol during the event, having more highly intoxicated partners, drinking to enhance sex, and holding more negative alcohol expectancies were associated with increased likelihood of NCCR. To effectively decrease the rates of NCCR, prevention efforts could aim to decrease event-level alcohol consumption for both men and their partners and to challenge men’s beliefs regarding alcohol’s role in sexual behavior. Given the current study’s limitations, future studies should consider employing an ecological momentary assessment design to avoid recall bias and include more diverse samples to increase the generalizability of findings.
Keywords
Sexually transmitted infections (STIs) and unintended pregnancies are common and costly public health concerns (Centers for Disease Control and Prevention [CDC], 2021a, 2021b). The overturn of Roe vs. Wade sparked political debate and created a divisive political climate in the United States where women’s bodily rights have become contested sites of violence and repression. There are numerous protests and demonstrations around the United States where conservative politicians and activists have been pushing for more restrictive abortion laws while progressive politicians and activists advocate for the preservation of women’s rights and bodily autonomy (Quinn, 2022). Additionally, previous research has estimated high pregnancy-related mortality among Black and Hispanic women, highlighting that consistent and correct condom use is of imperative importance to prevent unintended pregnancy (Stevenson, 2021). However, studies show that condoms are not being used consistently or correctly. In 2011 to 2015, male condoms were only used by 45.2% of men at their last intercourse (Copen, 2017). In fact, some men purposefully resist the use of condoms, via noncoercive and coercive (i.e., aggressive or manipulative) tactics (termed “condom use resistance” [CUR]) to obtain condomless sex with a partner who wishes to use condoms (Davis, Schraufnagel, et al., 2014). Nonconsensual condom removal (NCCR), also known as stealthing, is a specific form of coercive CUR which refers to the removal of condoms before or during sex without a partner’s knowledge or consent (Davis et al., 2019). When someone removes a condom without their partner’s knowledge or consent, they are disregarding their partner’s wishes and putting them at risk for health consequences such as STIs and unintended pregnancies. Because this violation exceeds the scope of initial consent and poses increased risks associated with sexual activity, NCCR can be considered “rape-adjacent” (Brodsky, 2017). Additionally, the secretive nature of NCCR violates recipients’ autonomy and dignity during sexual encounters. Consequently, recipients of NCCR are at higher risk for experiencing negative mental health consequences such as anxiety and depression (Czechowski et al., 2019; Ebrahim, 2019; Latimer et al., 2018; Tarzia et al., 2020). The goal of the present study was to understand the risk factors associated with NCCR perpetration to generate evidence that can guide prevention and intervention efforts to address this concern. We focus on alcohol-related factors, given the known association between alcohol use and sexual violence more generally (Abbey, 2011).
NCCR is a relatively common practice, especially among young men (Davis, 2019). In a study of 626 heterosexual men aged 18 to 30 years, 9.8% had engaged in NCCR since the age of 14 years, and 42.6% of those who had engaged in NCCR reported engaging in it three or more times. Moreover, many men consider NCCR as normative (Davis, Schraufnagel, et al., 2014) and do not perceive it as sexual assault or as emotionally or physically harmful (Ahmad et al., 2020; Davis et al., 2019; Nguyen et al., 2021), despite studies outlining several negative physical and psychological consequences. In addition to greater STI and unintended pregnancy risks (Czechowski et al., 2019; Davis, 2019), recipients of NCCR report feeling bothered, disrespected, upset, and fearful after being victimized (Czechowski et al., 2019; Tarzia et al., 2020). Several countries, including the United Kingdom, Canada, and Germany, consider NCCR to be illegal (Wolters, 2021). Given the detrimental consequences of NCCR, recent legislative efforts in the United States have been devoted to NCCR as well. For example, a bill that allows recipients of NCCR to sue offenders for damages was recently passed in California, making California the first state in the United States that has codified NCCR as illegal (Honderich & Popat, 2021). Similar legislation has been proposed in New York, New Jersey, Massachusetts, and Wisconsin, as well as on the federal level to make NCCR illegal and allow victims to seek civil damages after experiencing NCCR (Stealthing Act of 2022, 2022; Wolters, 2021).
Alcohol and NCCR
As a first step toward effectively addressing and preventing NCCR, research delineating its risk and protective factors is needed. Both theory and research have linked alcohol use to sexual risk behaviors in general and condom nonuse in particular. According to alcohol myopia theory, alcohol intoxication inhibits people’s ability to extract meaning from complex and distal cues (Steele & Josephs, 1990). For example, during sexual encounters, an intoxicated individual may focus on the more salient cues, such as their sexual arousal, while ignoring the more distal cues, including consequences associated with NCCR, such as STI transmission and the violation of consent. Furthermore, intoxication has been found to decrease one’s ability to engage in sexual communication and negotiation (Scott-Sheldon et al., 2016) and lead to more favorable attitudes toward having unprotected sex among those who were more sexually aroused (MacDonald et al., 2000).
Alcohol expectancy theory, which is another theoretical approach that helps explain the association between alcohol use and risky sexual behavior, posits that behaviors exhibited after drinking are likely to be consistent with previously held beliefs about the effects of alcohol (Jones et al., 2001). As a result, intoxicated individuals tend to behave based on what they have learned to expect from alcohol. For example, among intoxicated men, having greater aggression-related alcohol expectancies (i.e., believing that alcohol makes people more aggressive) leads to greater sexual aggression motivation and emotions after reading a hypothetical sexual encounter scenario in which the female partner refused unprotected sex (Davis, 2010). Studies have shown that alcohol expectancies specific to condom use (i.e., the expected effect of alcohol on the ability to use and negotiate condom use) mediate the association between drinking and sexual risk intentions, especially among men (LaBrie et al., 2002). Moreover, condom nonuse has been found to be more likely among individuals with the expectation that alcohol would lead to risky sexual behaviors (Dermen & Cooper, 2000).
Finally, the motivational model of alcohol use purports that people drink for different reasons, such as to enhance positive emotions or to cope with negative emotions, and these reasons for drinking may influence their alcohol consumption and drinking-related problems, such as violent sexual behaviors (Cooper et al., 1995). Sex-related drinking motives include motivations to consume alcohol to enhance sexual experiences or to cope with sex-related distress (Bird et al., 2019). Although there is currently no research that has investigated the relationship between sex-related drinking motives and condom nonuse, there is evidence to support the association between drinking motives and violent behaviors in general. In a sample of male college athletes, drinking to cope significantly predicted higher frequency of violent behaviors (Martens et al., 2003)
These findings support a link between alcohol use and sexual violence and suggest that drinking expectations and motivations influence this association. However, although NCCR is considered a form of sexual violence, the association between alcohol-related factors and NCCR in particular is less understood. Previous research has connected alcohol consumption to CUR more generally (Davis, Schraufnagel, et al., 2014; Davis, Stappenbeck, et al., 2014). Among a sample of women who had experienced NCCR, 68% of them reported that their partner had consumed alcohol at the time of incident (Latimer et al., 2018); yet, the effects of alcohol expectancies and drinking motivations on NCCR perpetration are unknown.
The Present Study
Despite the known relationship between alcohol consumption and sexual violence, very limited research has focused on the relationship between alcohol consumption and NCCR specifically. Such research could provide valuable insights for developing intervention programs aimed at reducing instances of NCCR (Davis et al., 2023). Therefore, this study addresses this gap by exploring the relationship between NCCR and alcohol-related factors, including event-level alcohol consumption, alcohol expectancies, and drinking motives, using a sample of 102 men with a coercive CUR history. We hypothesize that greater individual and partner event-level alcohol consumption, stronger condom nonuse and sexual aggression-related alcohol expectancies, and stronger sexual enhancement and coping drinking motives would be positively correlated with the likelihood that men engage in NCCR. Findings from the current research could provide specific evidence that may help guide the development and implementation of intervention efforts.
Method
Sample Recruitment
This study was approved by the Institutional Review Board before data collection. Participants were recruited from various cities around the United States using online advertisements (e.g., Craigslist) soliciting single males aged 18 to 30 years interested in dating women for an online study on dating experiences. Interested men were directed to an online screening survey on the study website. Eligibility criteria included: Residing in the United States, not currently in a monogamous relationship, and at least one successful lifetime instance of coercive CUR with a woman who wanted to use a condom. Once deemed eligible, participants received a phone call from study staff to verify their age, gender, date of birth, and location. Verified participants were sent a link to the 45 to 60-min online questionnaire and received a $30 Amazon gift card as a compensation upon survey completion. All identifiable data were removed before analysis, and study data were only accessible to research personnel to ensure confidentiality.
Of the 1,316 participants who completed the online screening survey, 199 (15.12%) met the eligibility criteria. Of those, 95 (47.74%) participants were deemed ineligible during telephone screening or study staff were unable to reach them. Thus, 104 participants were sent the questionnaire, with 102 completing it. Six of the participants were removed from analysis because they did not report having engaged in CUR in the primary survey, despite having endorsed these behaviors in the online screener. The final sample consisted of 96 men. A power analysis was conducted in G*Power (Faul et al., 2009) to determine the required sample size needed to detect whether alcohol use is associated with NCCR perpetration. Results showed that a sample size of 82 participants would yield a power of 0.80 to detect medium effects.
Participants
Participants had a mean age of 25.7 years (SD = 3.5). Fifty-six (58.3%) of the participants were White, 14 (14.6%) were Black/African American, 13 (13.5%) were Asian/South Asian, 11 (11.5%) were multiracial, 1 (1%) was American Indian/Alaskan Native, and 1 (1%) participant responded “Other.” Forty-four participants (45.8%) had completed some college, 35 (36.8%) participants were college graduates, 5 (5.2%) participants were high school graduates, 2 participants (2.1%) had completed some high school education, and 4 (4.2%) participants had completed trade or vocational/technical school education. Sixty-five (67.7%) participants were employed full-time, 23 (24%) participants were employed part-time, and 8 (8.3%) participants were not employed at the time of survey completion. Participants had a median household income of $41,000 to $50,999 per year. Twenty-five (26.3%) participants were students.
Measures
Nonconsensual Condom Removal
NCCR was assessed with one item from the CUR Scale (Davis, Stappenbeck, et al., 2014) asking participants if they had “successfully avoided using a condom with a woman who wanted to use one by agreeing to use a condom but removing it before or during sex without telling her” since the age of 14 years. Participants reported on the frequency of successfully engaging in NCCR on a scale from 0, 1, 2 . . . to 20 or more. Responses were dichotomized into 0 (frequency of 0) versus 1 (frequency greater than 0).
Event-Level Drinking
Participants answered questions about alcohol consumption during the event in which they engaged in NCCR. Participants reported on whether they consumed alcohol during the event (Yes [1], No [0]), whether their partners consumed alcohol during the event (Yes [1], No [0]), and the number of drinks they consumed at the event (ranging from 1 [one drink] to 26 [more than 25 drinks]). Participants also reported their highest level of intoxication on the day of the event, their level of intoxication before sex, and their perceptions of their partners’ intoxication level before sex using a seven-point Likert scale ranging from 1 (Not at all intoxicated) to 7 (Extremely intoxicated).
Sex and Aggression-Related Alcohol Expectancies
Participants rated their agreement with 31 statements regarding their expectations of how alcohol influences aggression and sexuality on a five-point Likert scale (Abbey et al., 1999; Davis et al., 2015) ranging from 1 (Not at all) to 5 (Very much). Five separate subscales were examined, and mean scores for each subscale were calculated: aggression (seven items, α = .91, e.g., “When drinking alcohol I am more mean”), sexual affect (six items, α = .91, e.g., “When drinking alcohol I feel more intimate”), sexual drive (six items, α = .95, e.g., “When drinking alcohol I am more likely to initiate sex”), sexual coercion perpetration (six items, α = .89, e.g., “When drinking alcohol I am more sexually coercive”), and women’s vulnerability to sexual coercion (six items, α = .97, e.g., “When drinking alcohol women are more likely to be taken advantage of sexually”).
Alcohol Expectancies for Condom Use
Participants responded to three questions that asked how much they expected alcohol to affect their ability to negotiate and use condoms on a seven-point Likert scale ranging from 1 (Not at all like me) to 7 (Totally like me/Always true) (α = .89; LaBrie et al., 2002). A sample item was “Drinking makes me forget about using condoms.” A mean score was computed.
Sex-Related Drinking Motives
Participants responded to 12 questions that asked about their sex-related reasons for drinking on a five-point Likert scale ranging from 1 (Never/Almost never) to 5 (Always/Almost always; Bird et al., 2019). Mean scores were calculated for sexual enhancement (seven items, α = .94, e.g., “To feel more self-confident and sure of yourself during sexual activity”) and coping with sex-related distress (five items, α = .95, e.g., “To feel less depressed about sexual activity”) subscales.
Social Desirability
Participants responded “True” (1) or “False” (0) to 13 questions on the shortened version of the Marlowe–Crowne Social Desirability Scale (α = .60; Reynolds, 1982) that measured participants’ likelihood of responding in a socially desirable manner. A total sum score was calculated.
Analytic Approach
All analyses were conducted using SPSS version 27.0 (IBM Corporation, 2020). Descriptive statistics were calculated to describe sample characteristics and the prevalence of NCCR. Chi-square tests and t-tests were used to compare differences in demographics and event-level drinking between men who had and had not engaged in NCCR. Due to small sizes of subgroups, race, ethnicity, and relationship status were recoded into dichotomous variables in chi-square tests (e.g., White vs. non-White, Hispanic vs. non-Hispanic, in committed relationship vs. not in committed relationship). Point-biserial correlation analyses were conducted to examine the associations between NCCR, event-level alcohol consumption, drinking motives, and alcohol expectancies. Finally, two binary logistic regression models were used to examine the relationship between sex-related drinking motives (Model 1), sex-related aggression alcohol expectancies (Model 2), and NCCR. Social desirability was added as a control variable in all logistic regression models. Data for this study are available by emailing the corresponding author.
Results
Nineteen (19.8%) participants reported engaging in NCCR during their lifetimes.
Demographic Information and NCCR
Level of education was significantly associated with NCCR: Men who went to vocational school and men with college and graduate degrees were more likely to have engaged in NCCR compared to men who attended some high school, had a high school degree or GED (General Education Development test), and attended some college (X2[5, N = 96] = 11.86, p = .04). No other significant differences were found.
Alcohol-Related Variables and NCCR: Bivariate Associations
There was a significant positive correlation between NCCR and number of drinks consumed at the NCCR event (rpb[85] = 0.26, p = .01), men’s perceptions of their partners’ level of intoxication (rpb[79] = 0.34, p < .01), drinking to enhance sex-related experiences (rpb[95] = 0.34, p < .01), condom use alcohol expectancies (rpb[87] = .31, p < .01), sex drive alcohol expectancies (rpb[90] = .31, p < .01), and sexual coercion perpetration alcohol expectancies (rpb[90] = .45, p < .01). No other statistically significant correlations were found (see Table 1).
Point-Biserial Correlations Between NCCR and Other Variables.
Note. 1. Men’s number of drinks before sex event; 2. Men’s highest level of intoxication during sex event; 3. Men’s highest level of intoxication before sex event; 4. Men’s perceived level of partners’ intoxication during sex event; 5. Men’s perceived level of partners’ intoxication before sex event; 6. SRDM—Enhancement; 7. SRDM—Coping; 8. AECU; 9. SARAE—Aggression; 10. SARAE—Sexual affect; 11. SARAE—Sexual drive; 12. SARAE—Vulnerability to sexual coercion; 13. SARAE—Sexual coercion perpetration. NCCR = nonconsensual condom removal.
Significant at p < .05.
Significant at p < .01.
Results of a chi-square test showed that NCCR was more likely to occur with a partner who consumed alcohol at the event than with a partner who did not consume alcohol, X2(2, N = 95) = 6.13, p = .02. Of the men who had engaged in NCCR, 72.2% (n = 13) were with a partner who had consumed alcohol during the event, and 27.8% (n = 5) men reported that their partner had not consumed alcohol at the time of the NCCR event.
Alcohol-Related Variables and NCCR: Multivariate Associations
In Model 1, controlling for all other variables, the odds of engaging in NCCR increased by a factor of 212% (OR = 3.12 [1.49, 6.56], SE = .38, p = <.01) among men who drank to enhance sex-related experiences (see Table 2). Furthermore, in Model 2, controlling for all other variables, the odds of engaging in NCCR increased by 81% among men with stronger beliefs that alcohol would impair condom use (OR = 1.81 [1.02, 3.20], SE = .29, p = .04). Additionally, controlling for all other variables, the odds of engaging in NCCR increased by 374% among men with stronger sexual coercion-related alcohol expectancies (OR = 4.74 [1.60, 21.64], SE = .66, p = < .01).
Logistic Regression Models.
Note. SRDM = sex-related drinking motives; SARAE = sex and aggression-related alcohol expectancies.
Significant at p < .05.
Significant at p < .01.
Discussion
This cross-sectional quantitative study tested the association of several postulated alcohol-related variables in relation to men’s NCCR behavior. This study adds to the current literature by examining drinking-related correlates of NCCR, including event-level alcohol consumption, sex-related alcohol expectancies, and sex-related drinking motives, among a national sample of 102 young men. NCCR was fairly common among the men in our high-risk sample, with 18.3% of men reporting that they had successfully engaged in NCCR in the past.
Our hypotheses were partially supported. At the event level, bivariate results suggest that men who consumed more alcohol during sex events were more likely to engage in NCCR. This is in line with alcohol myopia theory (Steele & Josephs, 1990) and previous research supporting a relationship between alcohol consumption and sexual violence in general. The myopic effect of alcohol, focusing men’s attention to sexual arousal cues, may make it more difficult for men to consider the negative consequences associated with NCCR. Moreover, our study found that having an intoxicated partner was associated with higher likelihood of NCCR. Interestingly though, men’s alcohol expectancies regarding their partners’ vulnerability to sexual coercion were not significantly associated with their likelihood of engaging in NCCR. Therefore, more research on event-level intoxication and NCCR is needed to investigate the underlying factors that may have contributed to men’s higher likelihood of engaging in NCCR with a drinking partner. In sum, our results suggest that event-specific drinking may be a critical predictor of NCCR behavior and a potentially effective avenue for intervention.
Our study found that drinking to enhance sex-related experiences was associated with a higher likelihood of NCCR. This finding is consistent with previous work showing that college students with strong drinking to enhance sex motives were more likely to engage in risky sexual behaviors during drinking events regardless of their level of intoxication, and that drinking to cope did not have a significant association with risky sexual behavior (Kilwein & Looby, 2018). Thus, drinking to enhance sexual experiences may be associated with higher likelihood of engaging in risky sexual behaviors in general and engaging in NCCR specifically. Although alcohol tends to be perceived as an aphrodisiac, studies suggest that both chronic and acute alcohol intoxication may have inhibitory effects on penile erectile function (George, 2019). Addressing this myth about sex-enhancing effects of alcohol consumption may be a possible area of intervention to decrease rates of NCCR and risky sexual behaviors in general.
Finally, the current results provide support for alcohol expectancy theory in relation to NCCR behavior. Stronger beliefs that alcohol would negatively impair condom use and increase sexual coercion were positively correlated with NCCR. These findings are consistent with prior work showing that alcohol expectancies for aggression increase the likelihood of intoxicated aggression in general and may moderate the association between aggression alcohol expectancies and sexual aggression specifically (Davis, 2010; Quigley & Leonard, 2006). Such associations suggest that prevention efforts could emphasize encouragement of respectful sex practices after alcohol consumption rather than focus on alcohol’s role in increasing sexually coercive and violent behavior as this content could ultimately reinforce alcohol expectancies associated with NCCR perpetration.
Limitations and Future Directions
The current results should be interpreted in light of some limitations. First, the cross-sectional design of the current study does not enable us to assess causal relationships; future studies could address this limitation by employing experimental and longitudinal designs, possibly with larger samples. Relatedly, the use of self-report measures may be subject to recall bias. Participants might not be able to accurately recall details of events that had occurred a long time before data collection. Future research could consider using ecological momentary assessment and daily diary assessments to obtain more accurate event-level data. Second, the sample in our study limits the external validity of our findings. Given the unique health consequences and associated disparities that female victims of NCCR experience (i.e., unintended pregnancy), our study only included young, heterosexually active men who were not involved in long-term monogamous relationships; thus, the findings from our study might not be generalizable to other populations such as female perpetrators, older adults, individuals who do not have sex with women, and individuals in long-term relationships. Notably, men who have sex with men (MSM) also experience NCCR. In a study conducted in Australia, 19% of MSM reported having experienced NCCR (Latimer et al., 2018). Past research has shown that individuals from gender and sexual minority populations tend to be more vulnerable to sexual violence (Martin-Storey et al., 2018). Therefore, more research on different populations, especially gender and sexual minority populations, is needed to test the generalizability of our findings and investigate effective prevention efforts among these groups. Additionally, the majority of our participants identified as White. Therefore, the findings from this study might not accurately reflect the experiences of men from other racial groups. Future research should aim to recruit a more racially diverse sample to increase the external validity of findings. Third, in this study, only sex-related drinking motives were included. Future research could investigate the relationship between NCCR and other drinking motives, such as socialization and conformity (i.e., avoiding social rejection), which are associated with increased alcohol use and impaired self-control (Merrill & Read, 2010). Finally, the quantitative nature of this research does not provide in-depth exploration of the phenomena studied here. Future research employing qualitative methods could be used to gain a deeper understanding of the relationship between alcohol-related factors and NCCR.
Conclusions
The present findings demonstrate that factors related to alcohol consumption—including event-level intoxication, drinking to enhance sex, and specific alcohol expectancies—are associated with men’s NCCR behavior. Our research addresses current gaps in the NCCR literature and provides possible avenues for intervention through investigating the relationship between specific alcohol-related factors and NCCR. Intervention strategies focused on reducing alcohol consumption, for both men and their partners, may be beneficial in decreasing the incidence of NCCR. However, although mass media campaigns that aim to reduce alcohol consumption have been associated with increased knowledge and changes in attitudes and beliefs about alcohol, they tend to have limited success in actually reducing alcohol consumption (Young et al., 2018). Young et al. (2018) recommend that future research should continue to address this challenge and aim to conduct larger campaigns that explore the pathways of behavioral change and the cost-effectiveness of interventions. Furthermore, our results suggest that beliefs that drinking enhances sex-related experiences contribute to an increased likelihood of NCCR. Thus, interventions aimed at challenging such commonly held beliefs may help reduce rates of NCCR. Finally, since holding expectancies that alcohol consumption would impair condom use and increase sexual coercion was associated with increased likelihood of NCCR, public health professionals could incorporate interventions that emphasize how consensual and respectful sexual behaviors are still possible after drinking, instead of promoting messages that link alcohol consumption to sexually violent behaviors.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interests with respect to the authorship and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research and/or authorship of this article: Funding for this study was provided by the National Institute for Alcohol Abuse and Alcoholism (2R37AA025212) and the Institute for Social Sciences Research at Arizona State University (Seed Grant) to the last author.
