How Do Sexual Assault Characteristics Vary as a Function of Perpetrators’ Level of Intoxication?

Abstract

Although alcohol’s effects are strongly related to the quantity consumed, most studies that have examined the effects of alcohol on sexual assault perpetrators’ behavior have simply assessed if alcohol was consumed, not the amount of alcohol consumed. This study addressed this gap in the literature with a sample of 107 Caucasian and African American men who reported perpetrating some type of sexual assault since the age of 14. The characteristics of the sexual assaults described by men who drank heavily during the incident significantly differed from those described by light drinkers and nondrinkers on a variety of measures including their use of physical force and perceptions of the seriousness of the incident. In contrast, there were few significant differences between light drinkers and nondrinkers. This pattern of results suggests that the amount of alcohol consumed is an important factor in the characteristics and consequences of sexual assault incidents. These findings highlight the importance of sexual assault prevention programs that target men’s heavy drinking.

1. Introduction

Sexual assault occurs at alarmingly high rates and approximately 50% of sexual assault perpetrators consume alcohol prior to or during the assault (for a review see Testa, 2002). Many studies have compared assaults in which perpetrators consumed alcohol to assaults in which they did not, using perpetrators’ self-reports (Zawacki et al., 2003) and victims’ reports of perpetrators’ alcohol consumption (Brecklin & Ullman, 2001). Although these studies have provided valuable information, most of them have not taken into consideration the quantity of alcohol consumed and instead have contrasted drinking perpetrators with nondrinking perpetrators. Combining data from perpetrators who consumed only one or two drinks with those who consumed large quantities of alcohol obscures the effects of high doses. Individuals who consume only a few drinks may not achieve a blood alcohol concentration (BAC) that produces behaviorally significant pharmacological effects (Chermack & Giancola, 1997), although their alcohol expectancies may be activated and even a small dose of alcohol may be sufficient to provide an external justification for inappropriate behavior (George & Stoner, 2000).

Many perpetrators commit multiple sexual assaults and there is no evidence that their drinking is consistent across assaults (Zawacki et al., 2003). Thus, this study extends past research by comparing the characteristics of sexual assaults based on the amount of alcohol consumed by the perpetrator.

2. Method

2.1. Participants

A representative community sample of 163 men were recruited for a study of dating experiences (see Abbey et al., 2006 for a more detailed description of data collection procedures). Only men who indicated they had used a coercive strategy to force some type of sex on a woman since the age of 14 were included in these analyses, reducing the sample size to 107.¹ Perpetrators’ ages ranged from 18 to 49 (M = 31.6, SD = 9.8); 57% were African American and 43% were Caucasian.

2.2. Measures

Perpetrator’s alcohol consumption was assessed at three timepoints: 1) in the 12 hours leading up to the interaction; 2) during the initial stage of the interaction; and 3) during the unwanted sex stage of the interaction. Responses to the three questions were summed and participants were then assigned to one of three groups: no alcohol use, light alcohol use (indicated by 1 – 4 drinks), or heavy alcohol use (indicated by 5 or more drinks).

Sexual assault perpetration was measured using a modified 17-item version of the Sexual Experiences Survey (SES; α = .88; Koss, 1988) that assessed sexually aggressive behavior since the age of 14. Sexual assault outcome severity was coded on a 1 to 4 scale: forced contact, sexual coercion, attempted rape, or completed rape.

After completing the SES, participants were asked to describe one sexual assault in detail. For men who had perpetrated more than one sexual assault, the computer program used an algorithm to select one incident based on simultaneous consideration of multiple factors including outcome severity, his perception of how negative it was, and its recency. Single item questions assessed how well the participant knew the woman, the amount of time he misperceived the woman’s sexual intent, and the amount of physical force used in the sexual assault. Men indicated the amount of consensual sexual activity with the women by selecting from a list of 12 sexual behaviors at two points prior to the forced sexual activity (α = .97). Participants’ isolating and controlling behaviors assessed their level of control over the interaction (α = .83; Norris et al., 1998).

Lastly, participants were asked to reflect on their perceptions of the assault. Single item questions assessed how serious participants felt the experience was, the extent to which he labeled the incident as a sexual offense, how much he was responsible for what happened, and the extent to which he learned something as a result of the incident.

Four covariates were included. Participants’ frequency of heavy drinking in the month prior to the sexual assault was included to control for the effects of their usual consumption level. The victim’s alcohol consumption was also included as a covariate. It was assessed similarly to perpetrator’s alcohol consumption, with the same questions and recoding described above. Social desirability (α = .67; Ballard, 1992) was included to control for willingness to disclose sensitive personal information. Finally, length of time since the sexual assault was included to control for the effects of time on recall.

3. Results

Forty percent of participants (n = 43) were classified as heavy drinkers (5 or more drinks consumed throughout the interaction), 22% (n = 23) as light drinkers (1–4 drinks consumed throughout the interaction), and the remaining 38% (n = 41) as nondrinkers.

Using simultaneous entry discriminant function analysis (DFA), one significant discriminant function emerged which discriminated heavy drinkers from light drinkers and nondrinkers, Wilks’ λ = .34, χ² (28, N = 107) = 103.99, p < .001. Correlations between the predictor variables and the function can be seen in Table 1. Classification rates were examined to assess how well this set of predictors discriminated between groups. Overall, 80.4% of participants were correctly classified, significantly improving upon a chance rate of 35.5%, z (107) = 9.72, p < .001. This set of predictor variables was most successful in identifying heavy drinkers (86.0% correctly identified) and nondrinkers (92.7% correctly identified). Among the light drinkers, 47.8% were correctly identified, 43.5% were incorrectly assigned to the nondrinking group, and 8.7% were incorrectly assigned to the heavy drinking group.

Table 1.

Results from Multivariate Analyses of Covariance, Analyses of Covariance, and Discriminant Function Analysis Examining the Relationships Between Perpetrator’s Level of Alcohol Consumption and Sexual Assault Characteristics

	Perpetrator’s Level of Drinking during the Assault
	None (n = 41)	Light (n = 23)	Heavy (n = 43)	F	Correlation with function
Aspects of the Sexual Assaulta
How well he knew woman (M)	4.90	5.08	4.70	0.25	.03
(SD)	2.22	1.92	2.26
Consensual sexual activity (M)	13.74	15.02	14.09	0.28	.19
(SD)	7.64	6.63	7.77
How long he misperceived woman (M)	0.99a	0.88a	2.12b	5.45**	.25
(SD)	1.68	1.45	1.71
Number of isolating and controlling behaviors (M)	1.34a	1.42a	3.27b	6.94**	.34
(SD)	2.36	2.04	2.39
Amount of physical force (M)	1.04a	1.32a	2.46b	9.59***	.28
(SD)	1.37	1.19	1.40
Severity of assault (M)	2.51a	2.84a	3.39b	4.97**	.30
(SD)	1.13	0.98	1.15
Perceptions of the Sexual Assaultb
Seriousness of incident (M)	2.64a	2.57a	3.99b	6.02**	.22
(SD)	1.86	1.61	1.89
Labeled it a sexual offense (M)	1.37a	1.60a	2.59b	5.72**	.22
(SD)	1.53	1.33	1.56
Attributed responsibility to self (M)	2.43a	2.51a	3.38b	4.47*	.24
(SD)	1.41	1.22	1.44
Amount learned from incident (M)	2.83a	2.70a	3.52b	3.18*	.21
(SD)	1.40	1.21	1.42

Note. Multivariate F ratios were generated from Pillai’s statistic. Means in a row with different subscripts are significantly different from each other using Least Significance Difference (LSD) post hoc test, p’s < .05. Univariate df = 2, 104.

^aF(12,192) = 3.03**;

^bF(8, 196) = 2.47*.

^*p < .05.

^**p < .01.

^***p < .001.

Multivariate analyses of covariance (MANCOVA) were conducted to aid in interpretation of the findings from the DFA to determine the effects of participants’ drinking during the incident on characteristics of the sexual assault divided into two domains: aspects of the sexual assault and perceptions of it. The covariates described above were included. As can be seen in Table 1, the MANCOVAs were significant, thus univariate ANCOVAs were conducted as follow-ups using Least Significant Differences (LSD) for comparisons across groups.

Only 4% of perpetrators were completely unacquainted with the victim. Perpetrators’ alcohol consumption was unrelated to how well they knew the woman and the number of consensual sexual activities they engaged in prior to the assault (see Table 1 for means). Compared with other perpetrators, perpetrators who drank heavily during the incident misperceived her sexual intentions for a longer period of time, employed more isolating and controlling behaviors during the interaction, were more physically forceful, and perpetrated assaults that were more severe. Further, perpetrators who drank heavily viewed what happened as more serious, were more likely to label it as a sexual offense, attributed more responsibility for what happened to themselves, and reported learning more from the incident than other perpetrators.

4. Discussion

As hypothesized, there were numerous differences between sexual assaults committed by men who drank heavily throughout the incident, compared to men who drank lightly and men who did not drink at all. These differences were found after controlling for men’s usual heavy drinking at that point in their lives and victims’ drinking during the interaction, thus they demonstrate the unique effects of perpetrators’ heavy drinking during the sexual assault. Heavy drinking men may be so focused on their own sexual arousal and feelings of entitlement that they miss or ignore messages intended to convey the woman’s lack of interest (Steele & Josephs, 1990). Alcohol administration studies demonstrate that intoxicated men are more aggressive than sober men, particularly when they feel provoked (Chermack & Giancola, 1997). Intoxicated perpetrators may view any form of consensual sexual activity as permission to engage in intercourse, thus feeling wronged and provoked when a woman stops their sexual advances. Their victims may have decided to stop resisting to minimize physical injuries, leading to more completed rapes.

Heavy drinking perpetrators were also most likely to perceive what happened as serious, label it as a sexual offense, attribute responsibility to themselves, and learn something from the incident. Alcohol often provides an excuse for inappropriate behavior, thus these men may have found it easier to acknowledge they did something wrong because they could protect their ego by viewing alcohol as the trigger for their inappropriate actions. The current study’s findings suggest that intoxicated perpetrators may learn from their mistakes. However, it is also possible that alcohol provides an excuse that some perpetrators rely on repeatedly to avoid accepting full responsibility for their behavior. This pattern of results suggests that most of alcohol’s effects on perpetrators’ behavior occur at high doses, demonstrating the importance of quantifying the amount of alcohol consumed during sexual assaults and supporting pharmacological explanations of alcohol’s effects on behavior. This does not mean that societal and individual alcohol expectancies are unimportant. As George and Stoner (2000) noted, alcohol consumption and expectancies work in tandem, each reinforcing the effects of the other.

These findings suggest that sexual assault prevention programs should focus on the cognitive distortions that alcohol produces. Men need to recognize how alcohol affects their perceptions of women and their willingness to use aggression. By focusing on these distortions and responsible drinking behaviors, men can learn to recognize when a woman is expressing discomfort regarding their use of sexual pressure. Men also need to learn that alcohol does not provide an excuse for forced sex and to take responsibility for their behavior when intoxicated. Although prevention programs that target alcohol use are important, this sample included a large number of nonalcohol-involved sexual assaults. Prevention programs that target nondrinking men are also needed. In this study, sober perpetrators did not use as much force and their assaults were less severe, yet research with victims demonstrates that verbally coercive assaults are disturbing and also have negative consequences (Koss, 1988). Sexual assault prevention programs must strive to help men fully understand the seriousness of verbally coercive behavior, even in the absence of physical force.