Impact of a College Sexual Violence Prevention Program on Condom Use: A Mediation Analysis

Abstract

College-age students are disproportionately impacted by sexually transmitted infections. Campus programs that reduce sexual violence have received recent investment, are increasingly common, and may offer a platform to increase condom use, but this has not yet been investigated. We explore this novel question through a secondary analysis of a randomized control trial of RealConsent, a web-based, sexual assault program for college women, on three college campuses. By estimating single and multiple-mediator models we examine the relationships between study assignment, the hypothesized mediators: self-efficacy to discuss safer sex, and clarity and assertiveness in sexual communication, and consistent condom use at follow-up. In the single mediator models, self-efficacy for safer sex communication (aOR: 1.11, 95% CI: 1.03–1.19, p = .004), assertiveness in sexual communication (aOR: 1.06, 95% CI: 1.02–1.11, p =.004), and clarity in sexual communication (aOR: 1.03, 95% CI: 1.00–1.05, p = .026) demonstrated significant direct effects on condom use. No statistically significant relationships between RealConsent and the mediators, nor indirect effects were found. In the multimediator model, there were no statistically significant associations identified. Self-efficacy, assertiveness, and clarity in communication about sex may have a positive impact on condom use but we did not find evidence that RealConsent impacted these mediators and thus no mediated effect was identified. Additional research is needed to develop and assess college-based sexual violence prevention programs that include an additional focus on skills specifically related to condom negotiation and use to understand if these widespread programs offer an efficient and effective platform to reduce the impact of sexually transmitted infections (STIs) among this high-risk population.

Keywords

college health self-efficacy sexual communication condom STI prevention

Impact Statement

Despite the fact that college-based sexual violence prevention programs target the same population that is at heightened risk for sexually transmitted infections, the impact of such interventions on condom use has not been assessed. In this secondary analysis of a randomized controlled trial, evidence that a sexual assault, alcohol misuse, and bystander intervention program positively impacted consistent condom use was not found. Additional research to better understand the mechanisms through which these types of programs might impact condom use is needed so that program developers and colleges can leverage these investments in sexual violence prevention to also improve condom use and reduce students’ risk of sexually transmitted infections.

Recent estimates from national surveillance data suggest that rates of some of the most common sexually transmitted infections (STIs)—chlamydia, gonorrhea, and syphilis—have increased substantially in recent years (Centers for Disease Control and Prevention, 2023b). While the impact of the COVID-19 pandemic on diagnosis and reporting requires these estimates to be interpreted with caution, most experts agree that it is probable that the pandemic resulted in underreporting of infections, and so even these troubling trends likely underestimate the need for STI prevention and control (Centers for Disease Control and Prevention, 2023a; Pagaoa et al., 2021; Wright et al., 2022).

While increases in STIs have been observed across subpopulations, youth remain at disproportionate risk. Nearly half of all new STIs are among youth aged 15–24 years of age (Bowen et al., 2018; Sieving et al., 2019). Undergraduate college students, most typically between the ages of 18 and 22 years old, contribute substantially to this high-risk population. Just one-third of all young people in the United States report having had sex before finishing high school, a rate that has been decreasing over the last several years (Centers for Disease Control and Prevention & National Center for HIV, Viral Hepatitis, STD, and TB Prevention, 2023). With college increasingly becoming a period of sexual initiation, it is often characterized by additional sexual exploration which can include more frequent sex with multiple and concurrent partners, condomless sex, and other behaviors associated with higher-risk sex such as drug and alcohol use (Johnson et al., 2010; Whiting et al., 2019). This makes college an increasingly important time to establish foundational healthy sexual behaviors. A 2018 survey conducted by the American College Health Association found that half of all sexually active undergraduate students used condoms sometimes, rarely, or not at all (American College Health Association, 2018). What is more, as the popularity and use of long-acting reversible methods of contraception have increased, there is some evidence that consistent condom use may continue to decline as young people are more concerned with pregnancy than STI protection and so forgo condoms when using these more highly effective contraceptive methods (Steiner et al., 2021). These factors make college students an important population to reach with primary and secondary prevention strategies that are critical to reversing the trend of increasing STI rates.

College, in addition to being a place where many students first experience sex and begin to explore their sexuality, is also a high-risk environment for sexual violence. One study found that 12% of female clients at college health centers had experienced sexual violence in the preceding semester (Fantasia et al., 2018) and other studies estimate that about 20% of female college students will experience sexual assault while in school (Centers for Disease Control and Prevention, National Center for Injury Prevention and Control [U.S.]. Division of Violence Prevention, 2016). In response to these unacceptably high rates of sexual violence, colleges and universities are increasingly investing in programs designed to build risk-reducing behaviors among students, prevent sexual violence perpetration, and promote bystander action. Indeed, the Campus Sexual Violence Elimination Act, passed in 2013, mandates that all institutes of higher education provide campus-wide sexual violence prevention programming (CampusClarity, 2023).

The primary outcomes assessed when testing or evaluating these types of interventions most commonly include reports of sexual violence behaviors such as victimization and perpetration, attitudes surrounding sexual violence and gender norms and roles, knowledge of the risk and risk factors of violence, intentions to and experiences with bystander action, and the needed skills for and the application of risk-reducing behaviors (DeGue et al., 2014; Vladutiu et al., 2010). While often not a primary focus of these interventions, we hypothesize that their impact may extend beyond behaviors most directly related to sexual violence risk and perpetration to others that impact sexual health such as condom use. If supported, this would create an opportunity to leverage investment in campus sexual violence prevention to also address the growing rates of STIs among this same population. A descriptive, cross-sectional, analysis among U.S. college students found a positive association between students’ self-efficacy to obtain sexual consent, a hallmark focus of sexual violence prevention programs, and condom use communication, and consistent use (Edison et al., 2022). We aim to advance this work by conducting a secondary analysis that tests the impact of a web-based, sexual assault, alcohol misuse, and bystander intervention program, RealConsent (women’s version), on condom use to understand if self-efficacy for safer sex communication, and assertiveness and clarity in sexual communication mediated this relationship. Informed by the relationships between self-efficacy, skills, and individual behavior outlined in social cognitive theory (Bandura, 1986), we hypothesized that the women’s version of RealConsent which included a focus on building students’ agency, self-efficacy to discuss sex more generally, and ability to communicate about sex with their partners as a method of sexual violence risk reduction, also positively impacted condom use among this population (Figure 1).

Figure 1.

Hypothesized model of relationship between RealConsent and condom use and proposed mediators, self-efficacy for safer sex communication, assertiveness in sexual communication, and clarity in sexual communication.

Method

Study Design and Recruitment

We conducted a secondary analysis of data collected through a randomized controlled trial of RealConsent (women’s version), conducted from October 2018 to August 2019 at three universities in the Southeastern United States. While the details of the study are described elsewhere (Salazar et al., 2023), briefly the study enrolled college students who identified as women, between 18 and 20 years of age, who were entering their first year of college at one of the three participating institutions. Email contact lists provided by the Universities’ Registrar Office were used to send invitations to potential participants. From there, participants were prompted to complete an eligibility screener and, if eligible, proceeded to complete an informed consent form, registration, and the baseline survey assessment. Participants were then randomized to either the RealConsent intervention or an attention-matched placebo group and invited and reminded to complete the relevant online modules based on this condition assignment. A follow-up survey was also administered 6 months following the completion of the baseline survey. All data were collected through the online platform, Qualtrics. This study was approved by Georgia State University’s Institutional Review Board (H19033) and registered as a clinical trial (ClinicalTrials.gov NCT0372643).

Intervention

RealConsent draws from social cognitive theory by building participants’ confidence in their ability to employ protective tactics that reduce their, and their peers,’ risk of sexual violence and by building self-regulation skills so that participants are better able to set goals for themselves and make plans that can reduce their risk. Informed by extensive formative research, the program consists of four 45-minute modules, accessible through an online portal, that each contains interactive, didactic, and entertainment elements. Among other objectives, the modules were designed to promote positive outcome expectations around sexual communication and increase social learning strategies for assertive communication about sex. Participants assigned to the attention-placebo condition were invited and prompted to participate in Stress and Mood Management, a web-based program that is similar in length and format to RealConsent, but that promotes healthy behaviors and practices to better manage stress, common mood disorders, and support students in identifying and accessing early treatment for depression and anxiety if needed.

Measures

Outcome Variable

Frequency of condom use was assessed at follow-up by asking participants to report how often, since they viewed the program—either RealConsent or the attention-placebo, they used a condom when engaging in a sexual act. Response options were provided on a 5-point scale from never (1) to always (5). For the purposes of this analysis, we collapsed this variable and considered condom use consistent among all those participants who reported always using a condom and inconsistent for any other response.

Mediating Variables

Self-Efficacy for Safer Sex Communication at Follow-Up

Self-efficacy to discuss safer sex was assessed at follow-up with six items. Four items were derived from a previously developed subscale of the Safe Sex Behavior Questionnaire that assessed self-efficacy of multiple different sex communication behaviors (Dilorio et al., 2000) and an additional two items were added by the study team for a total of six items. Participants rated each item on a 5-point scale from 1 “not at all confident” to 5 “very confident.” Responses across all items were summed for a total score ranging from 6 to 30, with higher scores corresponding to higher levels of self-efficacy to discuss sex with a partner. The scale demonstrated acceptable reliability in our analytic sample (α = .89) and construct validity in previous studies among a similar population (Dilorio et al., 1992, 2000) (Table 1).

Table 1.

Measures Used to Assess Theoretical Mediators of RealConsent and Condom Use Outcome.

Hypothesized mediators	M (SD)	Range	Standardized Cronbach’s α	No. of items	Response options	Sample item
Self-efficacy for safer sex communication	26.95 (4.34)	6–30	0.91	6	1 = not at all confident, 2 = somewhat confident, 3 = neither confident or unconfident, 4 = somewhat confident, 5 = very confident	“I can convince a sex partner to use a condom with me”
Assertiveness in sexual communication	23.86 (6.26)	7–35	0.74	7	1= Never to 5 = Always	“If I disagree with information that my partner presents on safer sex practices, I state my point of view”
Clarity in sexual communication	85.97 (12.35)	21–105	0.90	21	1= Never to 5 = Always	Do you ever end up having vaginal intercourse with a guy that you go out with when you don’t really want to, not because you feel forced or coerced, but because of some other concern (such as wanting him to like you or being too embarrassed to talk about it)?” (reverse coded)

Note. M = mean; SD = standard deviation.

Assertiveness in Sexual Communication

Assertiveness in sexual communication was assessed at follow-up using a 7-item subscale of the Safe Sex Behavior Questionnaire (Dilorio et al., 1992). Participants rated each item, which asked about the frequency with which they practice a number of sexual communication behaviors, on a 5-point scale from 1 “never” to 5 “always.” Responses were summed across items for total scores ranging from 7 to 35, with higher scores corresponding to higher levels of assertiveness in sexual communication. The scale demonstrated acceptable reliability in our analytic sample (α = .77) and construct validity in previous studies among a similar population (Dilorio et al., 1992) (Table 1).

Clarity of Sexual Communication

Clarity of sexual communication in dating situations was assessed at follow-up using a 21-item scale adapted from a measure developed by Hanson and Gidycz (1993). Participants responded to each item indicating the frequency with which they practiced the described behavior on a 5-point scale from 1 “never” to 5 “always.” Responses were summed across items for a total possible score ranging from 21 to 105, with higher scores corresponding to higher levels of assertiveness in sexual behaviors. The scale demonstrated acceptable reliability in our analytic sample (α = .90) and in the previous (Table 1).

Control Variables

Condom use at baseline was defined identically to condom use at follow-up, which is described above. Type of university was defined as either public urban, private urban, or public suburban based on the profiles of the three participating schools. For the purposes of this analysis, race was defined as White, Black, Asian, Biracial or Multiracial, and Other. Other races included respondents that marked American Indian, Native Alaskan, Native Hawaiian or other Pacific Islander, or other (specify) in their survey response. Respondents were asked to indicate their ethnicity as Hispanic or Latina or other. Place of residence was categorized as campus dorm or other. Other included on-campus apartment or house, Greek housing, off-campus apartment or house, or at home with parents or guardians. Relationship status was defined as either single or in a committed dating relationship. Respondents were offered choices to indicate married, separated, divorced, widowed, or engaged as well, but no respondents in this analysis indicated these statuses. Due to the limited representation of participants who identified as lesbian (n = 5), queer (n = 6), or other (n = 8) in the analytic sample, sexual orientation was dichotomized as heterosexual or not heterosexual, which included all those respondents who identified as lesbian (n = 5) bisexual (n = 49), queer (n = 6), or other (n = 8). Students were asked if they were members of any athletic sports teams or programs, if they had a job where they received income, if they had ever consumed alcohol, and if they were full or part-time students. All these were coded as binary yes/no variables.

Data Analysis

While participants in the broader study were not necessarily sexually active, for the purposes of this analysis we restricted our analytic sample to those who reported having had vaginal or oral sex since participating in the program at follow-up. Data were analyzed according to an intent-to-treat protocol; those that were assigned to the RealConsent condition were analyzed as such regardless of exposure to or participation in the intervention itself. All analyses were performed with SAS 9.4 (Carry, NC, USA), using the PROCESS macro developed by Hayes (2013). SAS uses listwise deletion for incomplete observations.

We estimated two types of mediation models. In lieu of the four-step approach (Baron & Kenny, 1986), we used the methodology pioneered by Hayes which has no such requirement (Hayes, 2009). First, simple mediation models were estimated to test for relevant mediators independently, controlling for baseline measures of study site, race, ethnicity, place of living, relationship status, sexual orientation, engagement in athletics, job status, having ever drank alcohol, and condom use. Second, a parallel multiple-mediation model with all hypothesized mediators was estimated with the same control variables included. Results estimated the effect of RealConsent on the hypothesized mediators (a-path), the effect of the mediator on the outcome of interest, condom use, at follow-up (b-path), and the direct, unmediated, effect of RealConsent on condom use (c’-path). Together these estimates were also used to calculate the indirect effect of RealConsent on condom use, through each of the hypothesized mediators (ab-path). Bias-corrected bootstrapped 95% confidence intervals around the indirect effects were estimated, and intervals that did not contain the null value were considered significant mediation effects. Because we modeled condom use as a dichotomous variable, regression coefficients for b, c,’ and ab paths are modeled as log-odds metrics and were exponentiated to produce effects on an odds ratio metric.

Results

Characteristics of the Sample

A total of 341 study participants reported having had sex at follow-up and were thus included in this analysis. These participants were distributed nearly evenly between conditions, with 167 (48.97%) having been randomized to the control condition and 174 (51.03%) to RealConsent (Table 2). The sample was made up almost entirely of full-time students (99.71%) who described themselves as single (93.26%). Most students had drunk alcohol at baseline (85.34%), were living in campus dorms (80.29%), and were identified as heterosexual (80.06%). The majority of the analytic sample was White (51.18%), followed by Black (25.29%), Asian or Pacific Islander (12.94%), Biracial or Multiracial (8.24%), and other races (2.35%). Just 13.78% identified as Hispanic, Latinx, or Latina. At baseline, a minority of the sample were engaged in athletics (12.32%) and had a job (36.07%). Participants were largely from a public urban university in the Southeast (49.56%), with (35.48%) attending a suburban public university, and (14.96%) at a private university in the region. No significant differences in demographic characteristics between study conditions, within this subsample, were observed at baseline. At baseline, respondents had moderate scores for self-efficacy, assertiveness, and clarity in sexual communication measures. Respondents assigned to the RealConsent condition had slightly higher mean scores on the clarity of sexual communication scale than those in the control condition (85.46 vs. 87.49, t = 21.94, p = .05). At baseline, a minority of respondents in both intervention (28.74%) and control (38.92%) indicated always using a condom during sex and there was no statistically significant difference identified between study condition assignment.

Table 2.

Baseline Characteristics Among Sexually Active Study Participants Stratified by RealConsent and Control Group.

	Full sample(n = 341)	Control condition (n = 167)	RealConsent (n = 174)	Comparison between control and RealConsent
Socio-demographics
	n (%)	n (%)	n (%)	Chi-square (p)
Engages in Athletics	42 (12.32%)	19 (11.38%)	23 (13.22%)	27 (.61)
Fulltime Student	340 (99.71%)	167 (100%)	173(99.43%)	0.96 (.33)
Ever drank alcohol	291 (85.34%)	145 (86.83%)	146 (83.91%)	0.58 (.45)
Has a job	123 (36.07%)	54 (32.34%)	69 (39.66%)	1.98 (.16)
Race				4.21 (.52)
White	174 (51.18%)	81 (48.50%)	93 (53.45%)
Black	86 (25.29%)	46 (27.54%)	40 (22.99%)
Asian or Pacific Islander	44 (12.94%)	25 (14.97%)	19 (10.92%)
Other	8 (2.35%)	4 (2.4%)	4 (2.3%)
Biracial or Multiracial	28(8.24%)	11 (6.59%)	17 (9.77%)
Hispanic, Latinx, or Latina	47 (13.78%)	25 (14.97%)	22 (12.64%)	0.39 (.53)
Place of living				2.89 (.24)
Campus dorm	273 (80.29%)	129 (77.25%)	1144 (82.76%)
Other	67 (19.71%)	38 (22.75%)	329(16.67%)
Sexual orientation				1.01 (.32)
Heterosexual	273 (80.06%)	130 (77.84%)	143 (82.18%)
Other	68 (19.94%)	37 (22.16%)	31 (17.82%)
Relationship status				0.30 (.58)
Single	318 (93.26%)	157 (94.01%)	161 (92.53%)
In a relationship	23 (6.74%)	10 (5.99%)	13 (7.47%)
University				3.58 (.17)
S.E. private	51 (14.96%)	26 (1115.57%)	25 (14.37%)
S.E. public urban	169 (49.56%)	90 (54.89%)	79 (45.40%)
S.E. public suburban	121 (35.48%)	51 (30.54%)	70 (40.23%)
Hypothesized mediators
	M (SD)	M (SD)	M (SD)	t-value (p)
Baseline self-efficacy for safer sex communication	26.95 (4.34)	26.87 (3.94)	26.52 (5.05)	0.70 (.49)
Baseline assertiveness in sexual communication	23.86 (6.26)	24.01 (6.33)	2372 (6.22)	0.41 (.68)
Baseline clarity in sexual communication	85.97 (12.35)	85.46 (13.24)	87.49 (11.25)	–1.94 (.05)
Outcome of interest
	n (%)	n (%)	n (%)	Chi-square (p)
Frequency of condom use at baseline				8.67 (.12)
Never	63 (19.21%)	32 (19.16%)	31 (17.82%)
Rarely	34 (10.37%)	13 (7.78%)	21 (12.07%)
Sometimes	42 (12.80%)	23 (13.77%)	19(10.92%)
Most of the time	74 (22.56%)	30 (17.96%)	44 (25.29%)
Always	115 (35.06%)	65 (38.92%)	50 (28.74%)

Sexually active = indicated that they had ever had oral or vaginal sex at follow-up.

Theoretical Mediators of Condom Use

We estimated three simple mediation models for condom use, controlling for all covariates and baseline levels of condom use (Table 3). All hypothesized mediators displayed statistically significant, positive associations with condom use (b-paths) such that a one-point increase in scores for self-efficacy for safer sex communication, assertiveness in sexual communication, and clarity in sexual communication increased the odds of consistent condom use by 1.11, 1.06, and 1.03, respectively. Contrary to our hypothesis, however, we did not identify any statistically significant relationships between RealConsent and these hypothesized mediators (a-paths). Likewise, we did not observe any statistically significant indirect effects of the intervention on condom use via these mediators (ab-paths). No statistically significant unmediated relationships between RealConsent and condom use (c’-path) were observed. The multiple-mediation model simultaneously estimated the effect of RealConsent on condom use with all three mediators included (Table 4). In this model, the effects of the hypothesized mediators on condom use (b-paths) were no longer statistically significant. Just as in the single mediator models, there were no statistically significant associations between RealConsent and the hypothesized mediators (a-paths) and none of the hypothesized indirect effects of RealConsent on condom use via the mediators were statistically significant.

Table 3.

Estimates From the Single-Mediator Models on Dichotomous Condom Use at Follow-Up.^a

Hypothesized mediator	a (SE)	b (SE)	c’ (SE)	ab (SE)	95% CI^b
Hypothesized mediator	a (SE)	b (SE)	c’ (SE)	ab (SE)	Lower	Upper
Self-efficacy for safer sex communication (n = 328)	0.1056 (0.4872)	0.1036* (0.0364)	0.2185 (0.2584)	0.0109 (0.0610)	–0.0947	0.1514
Assertiveness in sexual communication (n = 327)	–0.6963 (0.7015)	0.0607** (0.0208)	0.2280 (0.2601)	–0.0423 (0.0486)	–0.1486	0.0469
Clarity in sexual communication (n = 244)	1.4373 (1.8171)	0.0260* (0.0117)	0.4106 (0.3062)	0.0374 (0.0605)	–0.0607	0.1814

95% biased-corrected confidence intervals.

Analysis adjusted for baseline condom use, study site, race, ethnicity, place of living, relationship status, sexual orientation, engagement in athletics, job status, and having ever drank alcohol.

p < .05. **p < .001.

Table 4.

Estimates From the Multiple-Mediator Model on Dichotomous Condom Use at Follow-Up (n = 244).^a

Hypothesized mediator	a (SE)	b (SE)	c’ (SE)	ab (SE)	95% CI^b
Hypothesized mediator	a (SE)	b (SE)	c’ (SE)	ab (SE)	Lower	Upper
Self-efficacy for safer sex communication	–0.0957 (0.5777)	0.0487 (0.0455)	0.5075 (0.3154)	–0.0047 (0.0497)	–0.1110	0.0980
Assertiveness in sexual communication	–1.3542 (0.8043)	0.0430 (0.0288)		–0.0582 (0.0623)	–0.2086	0.0320
Clarity in sexual communication	1.4373 (1.8171)	0.0164 (0.0122)		0.0235 (0.0480)	–0.0459	0.1462

95% biased-corrected confidence intervals.

Analysis adjusted for study site, race, ethnicity, place of living, relationship status, sexual orientation, engagement in athletics, job status, and having ever drank alcohol.

Discussion

RealConsent has already been demonstrated to prevent sexual violence and improve prosocial bystander behavior among college men (Salazar et al., 2014) and to reduce the risk of sexual violence victimization among college women who had previously been victims of violence (Salazar et al., 2023). The analysis presented here aimed to understand if the RealConsent women’s program, which was primarily designed to reduce risky drinking behaviors and protect against sexual violence, might also impact condom use behaviors. We explored the relationship between RealConsent and condom use through three hypothesized mediators: self-efficacy for safer sex communication, assertiveness in sexual community, and clarity in sexual communication. In terms of the outcome of interest, in the single mediator models, all three hypothesized mediators demonstrated small statistically significant and positive direct effects on condom use. Yet, there was no evidence of a statistically significant relationship between RealConsent and these mediators, and no indirect effect was observed. In the multimediator model, which estimated relationships simultaneously, there were no statistically significant relationships observed. In addition to the intent to treat analysis, we also conducted a per-protocol analysis, limiting the analytic sample to those who had completed all RealConsent modules. The results were similar to those presented here with statistically significant b-paths in the single mediator models, but no statistically significant relationships were observed in the full model.

The lack of relationship between RealConsent and the hypothesized mediators was unexpected. While RealConsent did not focus primarily on enhancing self-efficacy to discuss sex, nor assertiveness or clarity in sexual communication more broadly, it did build knowledge and skills around obtaining and giving effective consent for sex which we hypothesized would, in turn, lead to increased communication ability and self-efficacy.

Recent research which found that only communication about condoms specifically was associated with increased condom use (Edison et al., 2022) may explain why, in the multiple-mediator model, the positive relationships between the hypothesized mediators and condom use (the b-paths) are no longer statistically significant. These mediators encompassed measures of communication about a host of sex-related topics, not just or specifically condom use. Research conducted in South Africa has also found a lack of association between women’s self-efficacy and condom use (Closson et al., 2018) which the authors explained by the fact that condoms are male-controlled and so, in addition to self-efficacy, require gender norms and power dynamics that are supportive of negotiation and use. This study and analysis did not seek to address or evaluate these environmental components of social cognitive theory but this may explain the lack of association between the intervention, participants’ self-efficacy, and condom use in our model.

Taken together, these results suggest that to improve condom use, programs likely need to include a specific focus on building skills and strategies for directly discussing condoms and include male partners as allies in shifting norms and power dynamics that currently act as barriers to condom use even among those with high self-efficacy. Efforts to improve discussions around pregnancy prevention or STI prevention, which theoretically could include condoms, are unlikely to impact condom use on their own. Prior research supports this conclusion, as increases in women’s self-efficacy surrounding condom use negotiation, more so than condom use itself, positively impacted condom use intention and behaviors (Guan et al., 2016).

Limitations and Strengths

The primary limitation of this analysis was that it was exploratory in nature. The study was not powered to examine these effects and so may have suffered from insufficient statistical power to detect the relationships of interest, particularly in the full where some missingness further reduced the analytic sample. In addition, RealConsent was not designed specifically to increase condom use behavior; thus, we were examining new hypotheses. A second important limitation is the cross-sectional nature of the data collection procedures used to measure both the hypothesized mediators and the outcome of interest, condom use, which prevents any causal associations as we cannot know the directionality of these relationships. While not a limitation per se, this analysis was intention-to-treat and so the observed effects may be underestimated given some participants in the treatment group did not complete all four modules. Nevertheless, we did perform a per-protocol analysis by limiting the sample further to only those who completed all four RealConsent modules and the findings remained unchanged, although this smaller sample (n = 174) may have suffered more acutely from insufficient statistical power than in the intent-to-treat approach. Finally, while the sample was largely representative of the student populations at each of the participating institutions, our results may not generalize to populations of women college students in other geographic regions of the United States. Even so, the strengths are considerable. The study occurred in three distinct university settings, which created a generally diverse study population. Methodologically, the use of the randomized controlled design and attention-matched placebo produces results robust to internal threats to validity.

Implications on Future Research and Practice

Although this analysis did not find evidence of RealConsent’s effect on self-efficacy for safe sex communication, assertiveness, or clarity in sexual communication nor of a mediated relationship between those variables and RealConsent’s effect on condom use, it does support the independent importance of self-efficacy for safer sex communication, assertiveness in sexual communication, and clarity in sexual communication on condom use. Developers and researchers of college-based sexual assault prevention programs research should focus on identifying the mechanisms through which sexual violence prevention programs on college campuses could improve condom use and other sexual health-related behaviors as a primary aim given the considerable risk college students face for both sexual violence and sexually transmitted infections and the efficiencies potentially gained by simultaneously reaching this population with effective strategies to reduce sexual violence and increase condom use.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The last author is a developer of RealConsent but does not derive financial income from the web-based program. The other authors have no conflicts of interest to declare.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Data used in this analysis were collected with the support of the National Institute on Alcohol Abuse and Alcoholism of the National Institutes of Health under award number R42AA025817.

Clinical Trials Registry

The data analyzed for this project were collected as part of a trial funded by the National Institute on Alcohol Abuse and Alcoholism of the National Institutes of Health under award number R42AA025817.

ORCID iD

Anne Laterra

References

American College Health Association. (2018). American College Health Association-National College Health Assessment II: Undergraduate student reference group data report Fall 2018. https://www.acha.org/documents/ncha/NCHA-II_Fall_2018_Undergraduate_Reference_Group_Data_Report.pdf

Bandura

(1986). Social foundations of thought and action: A social cognitive theory. Prentice Hall.

Baron

R. M.

Kenny

D. A.

(1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182. https://doi.org/10.1037/0022-3514.51.6.1173

Bowen

V. B.

Braxton

Davis

D. W.

Flagg

E. W.

Grey

Grier

Harvey

Kidd

Kreisel

Llata

Mauk

Pagaoa

Picchetti

Presley

R. J.

Ricks

Shapiro

St Cyr

Stenger

Torrone

. . .Wingard

(2018). Sexually transmitted disease surveillance 2018 [Report]. https://doi.org/10.15620/cdc.79370

CampusClarity. (2023). Campus Save Act. http://campussaveact.org/

Centers for Disease Control and Prevention. (2023a). Impact of COVID-19 on STDs. https://www.cdc.gov/std/statistics/2021/impact.htm

Centers for Disease Control and Prevention. (2023b). Sexually transmitted disease surveillance 2021. https://www.cdc.gov/std/statistics/2021/default.htm

Centers for Disease Control and Prevention & National Center for HIV, Viral Hepatitis, STD, and TB Prevention. (2023). Youth risk behavior survey data summary & trends report 2011-2021. https://www.cdc.gov/healthyyouth/data/yrbs/pdf/yrbs_data-summary-trends_report2023_508.pdf

Centers for Disease Control and Prevention, National Center for Injury Prevention and Control (U.S.). Division of Violence Prevention. (2016). Preventing sexual violence on college and university campuses: A summary of CDC activities. https://www.cdc.gov/violenceprevention/pdf/campusvsummary.pdf

10.

Closson

Dietrich

J. J.

Lachowsky

N. J.

Nkala

Palmer

Cui

Chia

Hogg

R. S.

Gray

Miller

C. L.

Kaida

(2018). Gender, sexual self-efficacy and consistent condom use among adolescents living in the HIV hyper-endemic setting of Soweto, South Africa. AIDS and Behavior, 22(2), 671–680. https://doi.org/10.1007/s10461-017-1950-z

11.

DeGue

Valle

L. A.

Holt

M. K.

Massetti

G. M.

Matjasko

J. L.

Tharp

A. T.

(2014). A systematic review of primary prevention strategies for sexual violence perpetration. Aggression and Violent Behavior, 19(4), 346–362. https://doi.org/10.1016/j.avb.2014.05.004

12.

Dilorio

Dudley

Lehr

Soet

J. E.

(2000). Correlates of safer sex communication among college students. Journal of Advanced Nursing, 32(3), 658–665.

13.

Dilorio

Parsons

Lehr

Adame

Carlone

(1992). Measurement of safe sex behavior in adolescents and young adults. Nursing Research, 41(4), 203–208.

14.

Edison

Coulter

R. W. S.

Miller

Stokes

L. R.

Hill

A. V.

(2022). Sexual communication and sexual consent self-efficacy among college students: Implications for sexually transmitted infection prevention. Journal of Adolescent Health, 70(2), 282–289. https://doi.org/10.1016/j.jadohealth.2021.08.012

15.

Fantasia

H. C.

Sutherland

M. A.

Hutchinson

M. K.

(2018). Lifetime and recent experiences of violence among college women. Journal of Forensic Nursing, 14(4), 190–197. https://doi.org/10.1097/jfn.0000000000000211

16.

Guan

Coles

V. B.

Samp

J. A.

Sales

J. M.

DiClemente

R. J.

Monahan

J. L.

(2016). Incorporating communication into the theory of planned behavior to predict condom use among African American women. Journal of Health Communication, 21(9), 1046–1054. https://doi.org/10.1080/10810730.2016.1204383

17.

Hanson

K. A.

Gidycz

C. A.

(1993). Evaluation of a sexual assault prevention program. Journal of Consulting and Clinical Psychology, 61(6), 1046–1052. https://doi.org/10.1037//0022-006x.61.6.1046

18.

Hayes

A. F.

(2009). Beyond Baron and Kenny: Statistical mediation analysis in the new millennium. Communication Monographs, 76(4), 408–420. https://doi.org/10.1080/03637750903310360

19.

Hayes

A. F.

(2013). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford Press.

20.

Johnson

Nshom

Nye

A. M.

Cohall

A. T.

(2010). There’s always Plan B: Adolescent knowledge, attitudes and intention to use emergency contraception. Contraception, 81(2), 128–132. https://doi.org/10.1016/j.contraception.2009.08.005

21.

Pagaoa

Grey

Torrone

Kreisel

Stenger

Weinstock

(2021). Trends in nationally notifiable sexually transmitted disease case reports during the US COVID-19 pandemic, January to December 2020. Sexually Transmitted Diseases, 48(10), 798–804. https://doi.org/10.1097/olq.0000000000001506

22.

Salazar

L. F.

Schipani-McLaughlin

A. M.

Sebeh

Nizam

Hayat

(2023). A web-based sexual violence, alcohol misuse, and bystander intervention program for college women (RealConsent): Randomized controlled trial. Journal of Medical Internet Research, 25, Article e43740. https://doi.org/10.2196/43740

23.

Salazar

L. F.

Vivolo-Kantor

Hardin

Berkowitz

(2014). A web-based sexual violence bystander intervention for male college students: Randomized controlled trial. J Med Internet Res, 16(9), Article e203. https://doi.org/10.2196/jmir.3426

24.

Sieving

R. E.

Gewirtz O’Brien

J. R.

Saftner

M. A.

Argo

T. A.

(2019). Sexually transmitted diseases among US adolescents and young adults: Patterns, clinical considerations, and prevention. Nursing Clinics of North America, 54(2), 207–225. https://doi.org/10.1016/j.cnur.2019.02.002

25.

Steiner

R. J.

Pampati

Kortsmit

K. M.

Liddon

Swartzendruber

Pazol

(2021). Long-acting reversible contraception, condom use, and sexually transmitted infections: A systematic review and meta-analysis. American Journal of Preventive Medicine, 61(5), 750–760. https://doi.org/10.1016/j.amepre.2021.04.032

26.

Vladutiu

C. J.

Martin

S. L.

Macy

R. J.

(2010). College- or university-based sexual assault prevention programs: A review of program outcomes, characteristics, and recommendations. Trauma, Violence, & Abuse, 12(2), 67–86. https://doi.org/10.1177/1524838010390708

27.

Whiting

Pharr

J. R.

Buttner

M. P.

Lough

N. L.

(2019). Behavioral interventions to increase condom use among college students in the United States: A systematic review. Health Education & Behavior, 46(5), 877–888. https://doi.org/10.1177/1090198119853008

28.

Wright

S. S.

Kreisel

K. M.

Hitt

J. C.

Pagaoa

M. A.

Weinstock

H. S.

Thorpe

P. G.

(2022). Impact of the COVID-19 pandemic on Centers for Disease Control and Prevention-funded sexually transmitted disease programs. Sexually Transmitted Diseases, 49(4), e61–e63. https://doi.org/10.1097/olq.0000000000001566