The Interaction of Rape Myth Acceptance and Alcohol Intoxication on Bystander Intervention

Abstract

Bystander interventions for sexual assault promote third-party interference. People who endorse rape myths blame victims more and perpetrators less; consequently, rape myth acceptance (RMA) can impede helping behaviors toward sexual assault victims. Acute alcohol intoxication may exacerbate the effects of RMA on bystander intervention. In this study, we examined the influence of RMA—and potential moderating effect of acute alcohol intoxication—on predictors of bystander intervention. Young adults (N = 128) completed a survey in a lab setting, then consumed either an alcoholic or control beverage, read and listened to a fictional sexual assault scenario, and finally completed a semi-structured interview and postexperiment survey assessing their perceptions of the scenario. Using multivariate analysis of covariance (MANCOVA), we found people with higher RMA blamed the victim more and perpetrator less; they were also less likely to perceive responsibility to intervene for a sexual assault victim. Alcohol intoxication did not exacerbate these effects. That is, alcohol intoxication was not a context in which RMA was expressed more strongly. We recommend bystander programs continue to address RMA, specifically as a barrier to intervening.

Introduction

Approximately one in five college women experiences sexual assault (Muehlenhard, Peterson, Humphreys, & Jozkowski, 2017). As a result, universities are required to provide sexual assault “primary prevention and awareness programs” that must include education about bystander intervention (Campus SaVE Act, 2011). Bystander intervention programs engage potential third-party observers, who are present at nearly a third of reported sexual assaults (Banyard, Moynihan, & Crossman, 2009). However, endorsement of rape myths and alcohol intoxication may be associated with a reduced willingness to intervene (Orchowski, Berkowitz, Boggis, & Oesterle, 2016). To extend this previous research, we examined the influence of rape myth acceptance (RMA) on blame and perceived responsibility to intervene—and the potential moderating effect of acute alcohol intoxication via a laboratory-based alcohol administration study.

Rape myths are inaccurate beliefs about the etiology of rape that justify men being sexually violent toward women (Bohner, Eyssel, Pina, Siebler, & Viki, 2009). People who endorse rape myths are more likely to blame victims and less likely to think victims are severely affected by sexual assault. People who consume alcohol tend to have higher RMA (Navarro & Tewksbury, 2017). These negative attitudes toward victims can impede people’s willingness to help them (Banyard & Moynihan, 2011; McMahon, 2010). RMA may be associated with potential bystanders’ perceived responsibility—or lack thereof—to intervene for the victim, because those who endorse rape myths blame the victim more and the perpetrator less.

Social settings in which alcohol is consumed are common contexts for sexual assault (e.g., Armstrong, Hamilton, & Sweeney, 2006). Acute alcohol intoxication may influence bystanders’ perceptions of risk for their friends (e.g., Drouin, Jozkowski, Davis, & Newsham, 2018); it may also exacerbate the effects of RMA on bystander behavior. In a cross-sectional study assessing alcohol consumption patterns and bystander intent, Fleming and Wiersma-Mosley (2015) found that men who consume alcohol were less likely to intervene when they knew the perpetrator, whereas women with higher alcohol expectancies were less likely to intervene when the perpetrator was not known—above and beyond their alcohol consumption. The alcohol myopia theory (Steele & Josephs, 1990) posits that alcohol narrows one’s attention to salient cues in the environment, including pre-existing belief systems like RMA (Davis, Hendershot, George, Norris, & Heiman, 2007). Potentially evidencing this, Orchowski and colleagues (2016) found that men who engage in heavy alcohol use were less likely to endorse bystander intentions—partially due to sexist and rape supportive beliefs. Because many potential bystanders consume alcohol when they might witness sexual misconduct or violence (McMahon & Banyard, 2012), we investigated whether acute alcohol intoxication moderated the association between RMA and predictors of bystander intervention.

Our study extends previous research in three important ways. First, using a laboratory-based, alcohol administration experimental design, we were able to examine how alcohol intoxication influences participants’ responses. Second, we were able to conduct an in-the-moment assessment of bystander intention compared with previous research, which has primarily examined participants’ retrospective bystander intention. Third, our sample comprises women and men, extending Orchowski et al.’s (2016) work that cross-sectionally examined these constructs among men only. We proposed three hypotheses.

Hypothesis 1:

RMA will be positively associated with victim blaming and negatively associated with perpetrator blaming.

Hypothesis 2:

RMA will be negatively associated with perceived responsibility to get involved for the victim.

Hypothesis 3:

RMA will be more strongly associated with blame and perceived responsibility to get involved for intoxicated participants compared to sober participants.

Method

Participants

Young adults between the ages of 21 to 29 years were recruited from a college town in the southern United States. The final sample included 64 women and 64 men (M_age = 23.27, SD = 2.41), who primarily identified as non-Hispanic White (n = 100, 78.1%). Most were college students (n = 108, 84.4%). See Table 1 for more information on the demographic characteristics of the sample.

Table 1.

Sociodemographic Characteristics by Condition (N = 128).

	Alcohol (n = 64)		Control (n = 64)
	n	%	n	%	χ2	df	p
Gender					—	1	—
Women	32	50.0	32	50.0
Men	32	50.0	32	50.0
Race/ethnicity					2.732	5	.741
White	48	75.0	52	81.3
Black	2	3.1	2	3.1
Latino	3	4.7	5	7.8
Asian	2	3.1	3	4.7
American Indian	1	1.6	0	0.0
Multiracial	3	4.7	1	1.6
Missing	5	7.8	1	1.6
Relationship status					.725	3	.867
Exclusive	35	54.7	31	48.4
Nonexclusive	3	4.7	3	4.7
Casual	5	7.8	6	9.4
None	18	28.1	22	34.4
Missing	3	4.7	2	3.1
Sexual orientation					2.215	4	.696
Heterosexual	53	82.8	57	89.1
Homosexual	1	1.6	0	0.0
Bisexual	4	6.3	3	4.7
Unsure	1	1.6	1	1.6
Queer	0	0.0	1	1.6
Missing	5	7.8	2	3.1
Education level					1.731	4	.785
Current student	19	29.7	20	31.3
<High school	1	1.6	2	3.1
Some college	18	28.1	16	25.0
Undergraduate degree	16	25.0	20	31.3
Graduate degree	7	10.9	4	6.6
Missing	3	4.9	2	3.1

Procedure

The current study was part of a larger alcohol administration experiment assessing the effects of alcohol intoxication on the stages of Latané and Darley’s (1970) model of bystander intervention. For a full description of the procedure approved by our institutional review board, see Ham, Wiersma-Mosley, Wolkowicz, Jozkowski, Bridges, & Melkonian (2019).

Participants completed a screening interview and baseline surveys; then, they were taken into a bar lab and informed of their randomly assigned beverage condition (i.e., alcohol vs. control). Alcohol condition participants (n = 64) received a dose of 100-proof vodka that was titrated to achieve a peak breath alcohol content (BrAC) of 0.08% (e.g., Davis, 2010). Control participants (n = 64) consumed nonalcoholic soda water.

Next, participants listened via headphones and read along to a fictional sexual assault scenario (see Supplemental Material). The scenario and corresponding follow-up questions were written specifically for this study, informed by previous research (Bennett & Banyard, 2016; Jozkowski, 2015). According to the story, participants introduce two of their friends, a woman and man, to each other at a party. Explicit statements about the woman’s alcohol consumption and behavioral signs of intoxication indicate that she is intoxicated; the man is described as sober. The two characters later begin engaging in seemingly consensual sexual activity that becomes nonconsensual—as evidenced by nonverbal refusals by the woman.

After listening to the scenario, participants completed a semistructured interview and postexperiment survey assessing their perceptions of the sexual assault. Participants were then debriefed; intoxicated participants remained in the lab for detoxification until deemed safe per National Institutes for Alcohol Abuse and Alcoholism (i.e., BrAC < 0.04%).

Measures

RMA.

In the baseline survey, participants completed a 24-item version of the Illinois RMA Scale (Payne, Lonsway, & Fitzgerald, 1999), rating each rape myth from 1 (not at all agree) to 7 (very much agree). Scores are calculated by summing items; higher scores indicate greater endorsement of rape myths. These 24 items were strongly intercorrelated (sample Cronbach’s α = .840).

Bystander attitudes.

Regarding blame attributions for the fictional sexual assault, participants rated how much the victim and perpetrator were each at fault. Response scales ranged from 1 (not at all) to 10 (entirely). Participants also indicated the extent they perceived it was their responsibility to get involved for the victim in the fictional sexual assault. This response scale ranged from 1 (not at all your responsibility) to 10 (entirely your responsibility). These items were specifically written for this study to correspond with the scenario provided to examine participants’ responses to the different stages of the bystander intervention model (Latané & Darley, 1970). They were developed via an iterative process by a team of researchers with expertise in sexual violence and alcohol intoxication and reviewed by an external content expert. For more information on the study’s methodological approach, see Authors (redacted).

Analysis

We tested a model to assess the effects of RMA as a continuous variable and its interaction with alcohol condition on three dependent variables related to bystander behavior: (a) blaming the victim for the sexual assault, (b) blaming the perpetrator, and (c) perceiving responsibility to intervene for the victim. We used multivariate analysis of covariance (MANCOVA) because these dependent variables are related conceptually and statistically (|r|s ≥ .389, ps < .001). We tested effects with Pillai’s trace test statistic, which is as robust as nonparametric parameter estimates (Pituch & Stevens, 2015). Omnibus multivariate tests were calculated at an α-level of .05. We then used a Bonferroni correction for each set of univariate tests; thus, we set our α-level to .017. We report η² as an effect size. All analyses were conducted using SPSS 25.

Results

Our participants varied in RMA based on gender, t(90.47) = −3.94, p < .001. Participants who identified as female endorsed significantly fewer rape myths, M = 26.69, SD = 7.02, than those who identified as male, M = 34.51, SD = 14.15. Descriptive statistics for all dependent variables are reported in Table 2; there were no other gender differences. Supporting Hypothesis 1, higher RMA was associated with increased victim blaming (r = .377, p < .001) and decreased perpetrator blame (r = −.466, p < .001) for the sexual assault. Supporting Hypothesis 2, higher RMA was associated with lower responsibility to intervene in the sexual assault (r = −.265, p < .001).

Table 2.

Descriptive Statistics for Study Variables.

	Female		Male
	M	SD	M	SD	t	df	p value
Rape myth acceptance	26.69	7.02	34.51	14.15	−3.935	90.47	<.001
Victim blaming	2.63	1.87	3.30	2.33	−1.801	118.65	.074
Perpetrator blaming	8.89	1.67	8.22	2.45	1.785	109.20	.077
Responsibility to intervene	9.29	1.19	8.81	1.82	1.754	126	.082

Note. Degrees of freedom that are not whole numbers reflect variables that violated the homogeneity of variance assumption.

Model Testing

The MANCOVA model accounted for 26.8% of the variance in participants’ victim blaming, 17.9% of perpetrator blaming, and 8.2% of perceived responsibility to intervene for the victim. There were significant main effects of RMA, F(3,120) = 13.290, p < .001, η² = .249, and alcohol condition, F(3,120) = 3.544, p = .017, η² = .081. There was also a significant interaction effect between these two variables, F(3,120) = 3.162, p = .027, η² = .073; however, none of the univariate interaction effects were significant. Therefore, we interpreted the significant main effects at the univariate level (see Table 3).

Table 3.

Regression Analysis of Dependent Variables on RMA and Alcohol Condition (N = 126).

	Victim Blaming			Perpetrator Blaming			Responsibility to Intervene
	B	SE	p value	B	SE	p value	B	SE	p value
RMA	.049	.019	.011	−.098	.017	<.001	−.047	.015	.002
Alcohol	−.737	.345	.035	.997	.323	.002	.450	.266	.093
RMA × Alcohol	.051	.030	.092	.042	.028	.137	.033	.023	.162

Note. RMA = rape myth acceptance.

When controlling for alcohol intoxication condition, people with higher RMA blamed the victim more, F(1,125) = 24.263, p < .001, η² = .166, and perpetrator less, F(1,125) = 29.774, p < .001, η² = .196. Controlling for alcohol condition, people with higher RMA were also less likely to think they were responsible to intervene in the sexual assault scenario, F(1,125) = 6.747, p = .011, η² = .052. Controlling for RMA, people who were intoxicated blamed the victim more, F(1,125) = 4.568, p = .035, η² = .036, and perpetrator less, F(1,125) = 9.540, p = .002, η² = .073, compared with sober participants. However, alcohol condition was not associated with perceived responsibility to intervene—above and beyond the effects of RMA.

Discussion

Stronger endorsement of rape myths is associated with blaming victims of sexual violence (Bohner et al., 2009). Extending this, our findings indicate that RMA also impedes blame-related aspects of bystanding. Specifically, we found that people with higher RMA are less likely to perceive responsibility to intervene for a victim of sexual assault. Similar to RMA, acute alcohol intoxication affected blame attribution and perceived responsibility to intervene in harmful ways; however, it did not exacerbate the association between RMA and victim blaming.

Because individual differences in RMA seem to more strongly impede bystanding than acute alcohol intoxication, we echo other researchers by recommending that bystander programs specifically focus on addressing rape myths—especially as a mechanism to reduce victim blaming—and increase perpetrator blaming (Banyard & Moynihan, 2011; McMahon, 2010). Although alcohol intoxication did not potentiate the effects of RMA on bystander perceptions, it still decreased risk detection (Ham et al., 2019), which can negatively influence intervening (Latané & Darley, 1970). Thus, we support recommendations that alcohol intoxication be addressed in bystander programs. In addition, because men tend to report more frequent heavy episodic drinking and adverse consequences of alcohol consumption (e.g., Holmila & Raitasalo, 2005; Wilsnack, Vogeltanz, Wilsnack, & Harris, 2000) and have higher RMA than women (e.g., Canan, Jozkowski, & Crawford, 2016), we recommend that programs specifically focus on addressing these factors with high-risk men. Researchers should continue to investigate other potential individual differences that are barriers to bystander intervention of sexual assaults such as perceptions of sexual consent and refusal communication as well.

That RMA can lead to blaming perpetrators less may be contextualized by previous research on RMA and consent communication. One study found that students who reported using somewhat aggressive strategies to communicate consent were more likely to endorse rape myths (Jozkowski, Sanders, Peterson, Dennis, & Reece, 2014). Similarly, participants in our study who scored higher on RMA may have interpreted the perpetrator’s aggressive strategies as consent indicators rather than coercive tactics. Bystander programs should emphasize the nuances of consent and refusal communication, so people can effectively discern subtle sexual refusals.

Although our study demonstrated methodological rigor, there are important limitations to note. Participants may have answered in a way that is socially acceptable rather than with their actual behavior or feelings. Future research should consider using a similar lab-based controlled study design to assess actual bystander behaviors via the use of confederates. Due to this being a time- and resource-intensive study, our sample may not have been large enough to adequately assess moderation effects. Because we found a significant interaction at the multivariate level, we encourage future researchers to further investigate the possibility that acute alcohol intoxication exacerbates the relationships between RMA and constructs related to bystander intervention for sexual assault (e.g., victim blaming; perceived responsibility to intervene). We were also unable to assess group differences for other demographic characteristics besides gender. As such, we encourage researchers to replicate our study with a larger sample that is powered to make additional comparisons. Researchers may also consider a study design that oversamples minorities regarding demographic variables of interest (e.g., race/ethnicity, sexual orientation) to make these comparisons as well.

Biographies

Author Biographies

Kristen N. Jozkowski, PhD, is the William L. Yarber Endowed professor of Sexual Health in the School of Public Health and a senior research scientist at the Kinsey Institute for Research in Sex, Gender, and Reproduction at Indiana University.

Malachi Willis, MA, is a PhD student in health promotion and health behavior and project coordinator in the Sexuality Education and Consent Studies Research Lab at the University of Arkansas.

Lauren E. Hurd, MA, is a PhD student in clinical psychology and involved in the Laboratory for Anxiety and Substance Abuse Research at the University of Arkansas.

Lindsay S. Ham, PhD, is a professor of Clinical Psychology and Director of the Laboratory for Anxiety and Substance Abuse Research at the University of Arkansas.

Ana J. Bridges, PhD, is a professor of Clinical Psychology and Director of the Diversity Research and Enhanced Access for Minorities Lab at the University of Arkansas.

Jacquelyn D. Wiersma-Mosley, PhD, is an associate professor of Human Development and Family Studies at the University of Arkansas.