Sexual assault and alcohol consumption: what do we know about their relationship and what types of research are still needed?

Abstract

Approximately half of all sexual assaults are associated with either the perpetrator’s alcohol consumption, the victim’s alcohol consumption, or both. Although the emphasis of this review is on alcohol-involved sexual assaults, their unique aspects can only be evaluated by comparing them to other types of sexual assault. Theoretical perspectives on sexual assault that focus on characteristics of the perpetrator, the victim, and the situation are described. A number of personality traits, attitudes, and past experiences have been systematically linked to sexual assault perpetration, including beliefs about alcohol and heavy drinking. In contrast, only a few experiences have been significantly related to sexual assault victimization, including childhood sexual abuse and heavy drinking. There is support for both psychological and pharmacological mechanisms linking alcohol and sexual assault. Beliefs about alcohol’s effects reinforce stereotypes about gender roles and can exacerbate their influence on perpetrators’ actions. Alcohol’s effects on cognitive and motor skills also contribute to sexual assault through their effects on perpetrators’ and victims’ ability to process and react to each other’s verbal and nonverbal behavior. Limitations with existing research and methodological challenges associated with conducting research on this topic are described. Suggestions are made for future research which can inform prevention and treatment programs.

1. Introduction

Since the 1970s, American feminist scholars have been concerned about the high rates of sexual assault among adolescent and adult women. During the last 30 years, many advances have been made in the measurement of sexual assault, but not in its prevention. This paper describes what is known about alcohol’s role in sexual assault, and what types of research are still needed to guide prevention efforts. Although the focus of this paper is on alcohol-involved sexual assault, it is described in the context of general information about sexual assault. The unique characteristics of alcohol-involved sexual assaults cannot be determined without contrasting them with other sexual assaults. After describing definitions and prevalence rates, complementary models of alcohol’s role in sexual assault are reviewed. Very few methodologically rigorous studies have been conducted that evaluate these models. Thus, this paper ends with suggestions for future research.

2. Sexual assault incidence and prevalence rates

2.1. Definitions

Rape is usually defined as vaginal, anal, or oral sexual intercourse obtained through force or threat of force; a lack of consent; or inability to give consent due to age, intoxication, or mental status. Sexual assault is a more inclusive term, used to describe the full range of forced sexual acts including physically forced kissing or touching, verbally coerced sexual intercourse, and physically forced vaginal, oral, and anal penetration. Each state uses its own definition of criminal sexual behavior; however, these definitions conform to federal law and most state laws (Bureau of Justice Statistics, 1995; Koss, 1996). This paper reviews theories of adolescent and adult sexual assault. Childhood sexual abuse is defined somewhat differently (e.g., can include exhibitionism), has different characteristics, and some different consequences (see Davis & Petretic-Jackson, 2000; Finkelhor, 1990 for reviews).

This paper focuses on female victims and male perpetrators. Research from national studies indicate that less than 5% of adolescent and adult sexual assault victims are male (Bureau of Justice Statistics, 1995; Tjaden & Thoennes, 2000). Perpetrators of these assaults are also usually male (Bureau of Justice Statistics, 1995).

Police reports, national random samples of crime victims, interviews with incarcerated rapists, interviews with victims who seek hospital treatment, general population surveys of women, and surveys of men and women college students have all been used to provide estimates of sexual assault incidence and prevalence (see reviews by Crowell & Burgess, 1996; Spitzberg, 1999). When women are asked to report their experiences of rape and sexual assault, the precise phrasing of the questions strongly affects their responses. Studies that describe the behaviors that constitute sexual assault in simple, nonlegal language provide the best estimates because they include victims and perpetrators who do not realize that their experience fits the legal definition of sexual assault (Koss, 1992). Four different sources of data are included below: community samples of women, college samples of women, community samples of men, and college samples of men. There have been many more studies assessing female victims’ experiences than male perpetrators’ experiences, however, it is important to consider both sources of data. Our goal is not to provide an exhaustive review, but to highlight some of the most influential and/or recent studies.

2.2. Community samples of women

Kilpatrick et al. (National Victim Center, 1992) conducted a national probability survey of 4008 women to determine the lifetime prevalence of rape. Their introduction specifically noted that “Women do not always report such experiences to police” and four behaviorally specific items assessed rape victimization. They found that 13% of American women had been the victim of a completed rape. Seventy-eight percent of the women knew their assailant; 19% of assailants were boyfriends or spouses.

The National Institute of Justice and the Centers for Disease Control and Prevention recently co-sponsored a study of violence against women (Tjaden & Thoennes, 1998). Telephone interviews were conducted with a nationally representative sample of 8000 women and used five behaviorally specific questions to measure attempted and completed rape victimization. Eighteen percent of the women reported that they had been the victims of rape. Only 14% of these rapes were committed by a stranger; 76% were committed by a spouse or date.

Several recent studies have documented extremely high sexual assault rates among women in the military. Merrill et al. (1998) examined prevalence of sexual assault victimization in a sample of 1891 women Navy recruits. Forty-five percent of these women reported being the victim of attempted or completed rape before joining the military. In a representative survey of more than 3000 women veterans, 23% had been sexually assaulted while they were in the military (Skinner et al., 2000).

2.3. Samples of college women

Many researchers have examined the prevalence of sexual assault among college students. There are both scientific and pragmatic reasons for this approach. Although sexual assault occurs to women throughout the lifespan, it is most common in late adolescence and early adulthood (Bureau of Justice Statistics, 1995; National Victim Center, 1992). Approximately one-quarter of this age group is in college (U.S. Bureau of the Census, 1996).

The first of these studies was conducted by Kirkpatrick and Kanin (1957), who surveyed 291 college women from 22 different university classes. Twenty-eight percent of these women had experienced attempts of “offensive” intercourse on a date during the previous year. The specific question used by Kirkpatrick and Kanin was phrased somewhat vaguely, both in terms of whether or not the act was completed and what form of pressure or force was used. This was because of their concern about embarrassing participants to the point that they would not complete the questionnaire. Thus, it is difficult to know exactly what types of events were included or how to compare these data to that gathered in later studies. Nonetheless, 6% of these incidents involved “menacing threats or coercive infliction of physical pain” (p. 53). This early study is noteworthy because it demonstrated that sexual assault was an issue on college campuses long before this problem was openly discussed.

Koss, Gidycz, and Wisniewski (1987) surveyed 3187 women and 2972 men students from 32 colleges selected to represent the higher education enrollment in the United States. Ten behaviorally specific questions were used to assess women’s experiences with forced sexual contact (not involving penetration), verbally coerced sexual intercourse, attempted rape, and rape. Since the age of 14, 15% of the women had experienced an act that met the standard legal definition of completed rape; an additional 12% had experienced attempted rape. Fifty-four percent of the women had experienced some form of sexual assault. Eighty-four percent of the women knew their assailant; 57% were dates. These findings have been replicated by other researchers at various universities (Abbey, Ross, McDuffie, & McAuslan, 1996a; Kalof, 1993; Mills & Granoff, 1992; Muehlenhard & Linton, 1987; Neal & Mangis, 1995). Recently, Brener, McMahon, Warren, and Douglas (1999) examined the prevalence of rape among a representative sample of 4609 female college students. Although they used only a single question to assess rape, they found that 15% of these students had been the victim of a completed rape since the age of 15. This is identical to the completed rape rate reported 12 years earlier by Koss et al. (1987).

2.4. Prospective studies of sexual assault victimization

Almost all of the sexual assault prevalence studies have been cross-sectional. A few authors have conducted prospective studies of sexual assault victimization during college (Gidycz, Coble, Latham, & Layman, 1993: Greene & Navarro, 1998; Himelein, 1995). For example, Himelein (1995) surveyed 100 women at freshmen orientation and again 32 months later. She focused specifically on sexual assaults that occurred on dates. Eight percent of these women were raped on a date since being in college; 29% experienced some level of sexual assault on a date.

Two prospective community surveys found dramatically different sexual assault incidence rates. In a 2-year follow-up of their telephone survey of a national probability sample of 3006 American women, Kilpatrick, Acierno, Resnick, Saunders, and Best (1997) found that 1.6% experienced a new rape since the initial interview. Testa and Livingston (2000) conducted in-depth, in-person interviews with 93 single women who drank at bars and parties and were sexually active. Thirty-one percent of these women had been sexually assaulted between the initial and 1-year follow-up interview; 14% had been raped. Kilpatrick et al.’s (1997) study has the advantage of being a large, representative study. However, reported rape rates may have been low because participants were reluctant to describe their experiences on the telephone. Testa and Livingston’s results are clearly not generalizable to the entire population; however, they document the high prevalence rate among an important subgroup: young, single, women who are actively dating and who drink alcohol.

2.5. Community samples of men

Very few community surveys of men’s sexual assault perpetration have been conducted. Calhoun, Bernat, Clum, and Frame (1997) surveyed 65 young men (average age of 19) in rural Georgia who had previously participated in a longitudinal study of youth. Using Koss et al.’s (1987) measure, 22% of these men reported committing a sexual assault; 6% reported committing an act that met the standard definition of completed rape. Senn, Desmarais, Verberg, and Wood (2000) randomly sampled 195 men from one small Canadian city. Twenty-seven percent of these men perpetrated some form of sexual assault during their lifetime; 8% reported committing an attempted or completed rape. Merrill et al. (1998) surveyed 1754 adult, male U.S. Navy recruits. Fifteen percent of these men reported that they had perpetrated rape or attempted rape before entering the service. Sexual assault prevalence rates were similar for recruits regardless of ethnicity, parents’ income, education, or geographic region. Ageton (1983) conducted interviews with a national probability sample of 1494 youth aged 11–17. Perpetration rates varied based on the year the data was collected and the participant’s age. The highest rate of sexual assault was reported by 17-year-olds interviewed in 1978; 8% of these young men indicated that they had committed one or more sexual assaults. Across ages and cohorts, approximately 3% of adolescent men reported perpetrating a sexual assault. More than 85% of the perpetrators knew the victim; the vast majority of these sexual assaults were committed on dates.

2.6. Samples of college men

Most data about men’s self-reported perpetration rates come from studies of college students. In Koss et al.’s (1987) national study of college students, 7.7% of the men reported that they had committed an act that met the standard legal definition of attempted or completed rape since the age of 14. Twenty-five percent of these men reported that they had committed some form of sexual assault since the age of 14. In studies at individual universities and colleges, up to 15% of the men surveyed have indicated that they perpetrated rape and up to 57% indicated that they perpetrated some form of sexual assault (Abbey, McAuslan, & Ross, 1998; Craig, Kalichman, & Follingstad, 1989; Kanin, 1985; Mills & Granoff, 1992; Muehlenhard & Linton, 1987; Rapaport & Burkhart, 1984). More than 80% of these sexual assaults occurred with women these college perpetrators knew; about half occurred on dates.

3. The prevalence of alcohol-involved sexual assault

Researchers consistently find that approximately half of all sexual assaults are committed by a man who has been drinking alcohol. On average, 50% of convicted rapists consumed alcohol at the time of the assault; a rate that is comparable to that for other violent crimes (Coid, 1986; Collins & Messerschmidt, 1993; Johnson, Gibson, & Linden, 1978; Rada, 1975; Scully, 1991; Vinogradov, Dishotsky, Doty, & Tinklenberg, 1988). Similarly, in Ageton’s (1983) survey of adolescent community sexual assault perpetrators, approximately half of the offenders were drinking prior to the assault. College studies provide similar estimates (Abbey et al., 1998; Koss, 1988; Muehlenhard & Linton, 1987). For example, Abbey et al. (1998) found that 47% of the college men in their sample who acknowledged committing sexual assault were drinking alcohol at the time of the assault.

Similarly, about half of all sexual assault victims report that they were drinking alcohol at the time of the assault (Greene & Navarro, 1998; Harrington & Leitenberg, 1994; Koss, 1988; Miller & Marshall, 1987; Muehlenhard & Linton, 1987; Presley, Meilman, Cashin, & Leich-liter, 1997; Seifert, 1999; Stermac, Du Mont, & Dunn, 1998; Tyler, Hoyt, & Whitbeck, 1998). Many of the community studies described above did not ask sexual assault victims if they had been drinking alcohol, thus much of these data come from college samples. For example, Harrington and Leitenberg (1994) found that 55% of the sexually assaulted college women in their study were at least somewhat drunk at the time of the incident. Stermac et al. (1998) interviewed 1162 women at a sexual assault crisis center. They found that 38% of the women reported being under the influence of alcohol during their assault. In the vast majority of sexual assaults, if either the perpetrator or the victim had consumed alcohol, then both had (Harrington & Leitenberg, 1994; Ullman, Karabatsos, & Koss, 1999a).

As will be described in more detail later, sexual assault surveys do not usually include detailed assessment of the quantity of alcohol consumed or level of intoxication at the time of the assault. Thus, all that is known is who was drinking near the time of the sexual assault, not how much they drank or how intoxicated they were. In one exception, Muehlenhard and Linton (1987) found that dates in which college men and women felt moderately or extremely intoxicated were more likely to involve sexual assault than dates that involved mild intoxication or no alcohol consumption. In another study that used a similar methodology, men drank more on dates that included sexual assault than on self-defined “worst” dates (Abbey, McAuslan, Zawacki, Clinton, & Buck, 2001). Additionally, the strong correlation between perpetrators’ and victims’ alcohol consumption makes it difficult to examine the independent effects of each individual’s intoxication.

4. Theoretical explanations for the relationship between alcohol and sexual assault

A causal relationship between alcohol consumption and sexual assault perpetration is not demonstrated by the fact that approximately half of all men who committed sexual assault were drinking alcohol at the time of the incident. Some men may consciously or unconsciously drink alcohol prior to committing sexual assault in order to justify their behavior. In these cases, the desire to commit sexual assault would be the cause of the perpetrator’s alcohol consumption. Alternatively, personality characteristics, such as narcissism, or life experiences, such as witnessing parental violence in childhood, may lead some men both to drink heavily and to commit sexual assault. In these cases, both alcohol consumption and sexual assault are motivated by other factors.

Each of these explanations has merit and is likely true for some perpetrators. It is our premise that alcohol often acts in a synergistic manner with other variables. As described in more detail below, a man who believes that his date has led him on and acted like she wanted to have sex sometimes feels justified in committing rape. If this man has been drinking alcohol, his likelihood of committing rape is increased.

Social scientists have distinguished between causality and responsibility (Shaver, 1987). Knowing that alcohol contributes to sexual assault does not diminish a perpetrator’s responsibility. Men are morally and legally responsible for sexual assaults that they commit, regardless of whether or not they were intoxicated. Although women’s alcohol consumption may increase their likelihood of experiencing sexual assault, this does not make them responsible for the man’s behavior. Women can empower themselves by identifying actions which may put them at risk, without feeling responsible.

The following sections of the paper consider a variety of explanations for alcohol’s role in sexual assault. As noted above, general explanations of sexual assault are also described because alcohol can exacerbate processes that can occur without alcohol. First, characteristics of perpetrators are considered: personality, attitudes and beliefs, and past experiences. Second, characteristics of victims are considered. Both of these literatures have considered beliefs about alcohol and alcohol consumption as predictors of sexual assault. Then situational characteristics of sexual assault situations are examined. Many of these studies have contrasted alcohol-involved sexual assaults with sexual assaults that do not involve alcohol. This is a vast literature to review in one paper, thus, the description of some issues is relatively brief (review papers are cited for the interested reader). The quality of existing research varies dramatically, and throughout the paper methodological limitations and gaps in the literature are described. In the final section of the paper, concrete suggestions are made for future research to address these shortcomings.

5. Perpetrators’ personality characteristics, attitudes, and experiences

There are many reasons that men commit sexual assault, and different perpetrators can be motivated by different factors (Malamuth, Sockloskie, Koss, & Tanaka, 1991; O’Neil & Harway, 1999; Prentky & Knight, 1991; Seto & Barbaree, 1997). As Barbaree and Marshall (1991) observed, sexual assault always involves an integration of sexual and aggressive motives, although their relative strength may vary among different types of perpetrators. This section of the review separately describes personality, attitudinal, and experiential factors that have been associated with sexual assault perpetration and are summarized in Table 1. For any one perpetrator, these factors will interact. There are two primary sources of data: incarcerated rapists and college students. Very few sexual assault perpetrators are ever convicted; thus incarcerated rapists represent a very small subgroup of all perpetrators. Incarcerated samples overrepresent stranger rapists, assaults that involve the most physical violence and injuries, and low-income, minority perpetrators (Seto & Barbaree, 1997). College student perpetrators represent a more diverse group because most have not been arrested or convicted, knew the victim well, did not use extreme force, and are middle class (Koss, 1988). The extent to which these two groups of perpetrators represent the entire population of sexual assault perpetrators is unknown until etiological research is conducted with nationally representative samples of men.

Table 1.

Perpetrator characteristics that have been linked to sexual assault perpetration

Personality characteristics

Antisocial

Impulsive

High in sexual arousability

Low in empathy

Poor social skills

Attitudes and beliefs

Traditional gender role beliefs

Acceptance of rape myths

Hostility toward women

Acceptance of force in interpersonal relationships

Alcohol expectancies regarding sexuality, aggression, and disinhibition

Stereotypes about drinking women

Alcohol as a cue for consensual sex

Life experiences

Childhood sexual and physical abuse

Delinquency in adolescence

Peers with supportive norms

Early and frequent sexual activity

Heavy alcohol consumption

5.1. Perpetrators’ personality characteristics

Several personality characteristics have distinguished sexual assaulters from nonassaulters, including antisocial personality traits, impulsivity, sexual arousability, and empathy. In their review of the literature, Prentky and Knight (1991) argued that many rapists have antisocial personality disorders that are reflected in a general lifestyle of impulsivity. Antisocial personality disorders are also frequently diagnosed in incarcerated rapists (Brown & Forth, 1997; Prentky & Knight, 1991). College perpetrators are not usually asked to complete measures that assess psychiatric disorders. There is evidence, however, that they also have a general disregard for social norms and other people, as well as a tendency toward aggression. For example, Kosson, Kelly, and White (1997) found that sexual assault perpetration by college men was negatively related to socialization and positively related to narcissism. Rapaport and Burkhart (1984) found that college men’s coercive sex scores were negatively correlated with measures of socialization and responsibility. Impulsivity and aggressiveness are positively related to sexual assault perpetration among college men (Greendlinger & Byrne, 1987; Lisak & Roth, 1988; Petty & Dawson, 1989).

Many sex offenders are more sexually aroused by nonconsenting sexual stimuli than are nonrapist offenders (see review by Lalumiere & Qunisey, 1994). Similarly, college men who perpetrated sexual assault or who reported that they were very likely to rape if they could be assured they would not be punished are more sexually aroused by depictions of rape than are other college men (Malamuth, 1989; Mosher & Anderson, 1986; Porter & Critelli, 1994). Porter and Critelli (1994) asked college men who had previously completed a questionnaire that assessed past sexual assault perpetration to listen to an audiotape of a date rape. Sexual assaulters reported that they were more sexually aroused by the tape and engaged in disinhibitory self-talk (e.g., commented on how aroused the man was or how likely they were to have sex). In contrast, nonassaulters reported that they were not as sexually aroused and used inhibitory self-talk (e.g., criticized the man for what he was doing).

A number of studies have demonstrated that incarcerated rapists have a reduced capacity for empathy (see reviews by Geer, Estupinan, & Manguno-Mire, 2000; Ward, Keenan, & Hudson, 2000). Many rapists are focused on their own feelings and needs and see themselves as entitled to having sex with their victim. Attachment problems in childhood may produce intimacy and empathy deficits in sexual assaulters (Barbaree, Marshall, & McCormick, 1998). In a small sample of community men, Senn et al. (2000) found that men who were unable to feel or express their full range of emotions were more likely to have committed sexual assault. Dean and Malamuth (1997) examined the role of empathy as a protective factor for college men. Among men who were at high risk of committing sexual assault (e.g., history of family violence, frequent sexual experiences), empathy did not influence their fantasies about sexual assault. However, men who were at high risk of committing sexual assault and who were low in empathy were much more likely to perpetrate sexual assault than were men who were high in empathy. These results demonstrate how empathy can keep men from acting on the desire to force sex.

Results regarding social skills deficits in incarcerated rapists have been mixed (see reviews by Geer et al., 2000). This may be due to differences in rapists’ motives, as well as differences in the aspects of social skills which were assessed (Prentky & Knight, 1991). Social skills deficits may need to interact with other characteristics, such as low empathy or hostility toward women in order to motivate sexual assault perpetration.

5.1.1. Implications for alcohol-involved sexual assaults

Sexual assault researchers have not tried to determine if the personality traits of men who commit alcohol-involved sexual assault differ from men who commit sexual assaults that do not involve alcohol. In many instances, alcohol may interact with personality characteristics to make sexual assault perpetration more likely. Many men who are impulsive or hostile have not sexually assaulted a woman; it requires a combination of the man’s predisposition to commit sexual assault and circumstances in which he feels it is acceptable to act on these predispositions. Alcohol consumption disrupts higher-order cognitive processes including abstraction, conceptualization, planning, and problem solving, making it difficult to interpret complex stimuli (Chermack & Giancola, 1997; Leonard, 1989; Peterson, Rothfleisch, Zelazo, & Pihl, 1990). Thus, under the influence of alcohol, people have a narrower perceptual field and are only able to attend to the most salient cues in the situation (Steele & Josephs, 1990; Taylor & Leonard, 1983). The cues that usually inhibit sexually aggressive behavior, such as a concern about future consequences, sense of morality, or empathy for the victim, are likely to be less salient than feelings of anger, frustration, sexual arousal, and entitlement, especially among men who are predisposed to being sexually aggressive. Thus, when intoxicated, men who are predisposed to being sexually aggressive may be more likely to act on their desires.

Rapists often engage in cognitive distortions that allow them to believe that their victims wanted sex or deserved to be forced into having sex because of how they behaved (Scully, 1991; Ward et al., 2000). Sexual assault researchers’ descriptions of rapists’ cognitive distortions are very similar to alcohol researchers’ descriptions of how alcohol affects cognitive processing; both include a decreased ability to notice and evaluate emotions and intentions. Experimental research is needed that directly examines how alcohol consumption interacts with sexual assaulters’ personality characteristics and cognitive styles. It seems reasonable to hypothesize that intoxicated perpetrators are particularly likely to derogate their victims and assume either that they deserved to be hurt or enjoy being hurt, and to use these beliefs to justify their behavior.

Several studies with college men have examined the effects of alcohol on responses to sexual stimuli. Men who thought they were drinking alcohol were more sexually aroused by depictions of forced and consensual sex than were men who did not think they had consumed alcohol (Briddell et al., 1978; George & Marlatt, 1986; George & Norris, 1991; George, Stoner, Norris, Lopez, & Lehman, 2000). In these studies, it was the belief that one had consumed alcohol, rather than alcohol consumption, which affected men’s sexual arousal. George and Marlatt (1986) explained these findings by arguing that the belief that one has consumed alcohol provides men with justification for engaging in socially inappropriate behavior. Both incarcerated and college rapists often use their intoxication as an excuse for their actions (Kanin, 1985; Scully, 1991). George et al. (2000) describe the synergy between alcohol expectancies regarding sexual disinhibition and alcohol myopia. Some men may drink before a party or date because they want to feel sexually disinhibited. However, “as intoxication rises, physiological impairment may lock the drinker on this course by blocking awareness of inhibitory restraints” (pp. 174–175).

5.2. Perpetrators’ attitudes about women and sexual relationships

Sexual assault perpetrators’ attitudes and beliefs about gender roles and sexual relationships have been frequently studied. Men who have committed sexual assault are more likely than other men to endorse traditional stereotypes about male dominance and gender roles (e.g., men being responsible for initiating sex and women for setting the limits), rape myths (e.g., women say no when they mean yes and enjoy forced sex), adversarial beliefs about relationships between men and women (e.g., “all’s fair in love and war”), hostile attitudes toward women, and the acceptability of using force in interpersonal relationships (Byers & Eno, 1991; Craig et al., 1989; Dean & Malamuth, 1997; Koss & Dinero, 1988; Malamuth, Linz, Heavey, Barnes, & Acker, 1995; Malamuth et al., 1991; Mosher & Anderson, 1986).

Many adolescents and young adults believe that there are circumstances that make forced sex acceptable. Forcing sex on a date when the woman had “led the man on,” changed her mind, or sexually aroused the man was endorsed by more than half of the high school males asked these questions (Goodchilds & Zellman, 1984). College students have lower rates of agreement, although forcing sex is perceived as most justifiable in circumstances similar to those endorsed by high school students (Cook, 1995). College student perpetrators of sexual assault believe that using verbal pressure to obtain sex is more acceptable than do non-perpetrators (Abbey et al., 2001). As noted below, these attitudes about forced sex may be activated in drinking situations.

5.3. Perpetrators’ attitudes about alcohol

Fewer researchers have examined the relationship between attitudes about alcohol and sexual assault, although alcohol is often considered an aphrodisiac that increases men’s sexual desire and abilities (see review by Crowe & George, 1989). Additionally, many men expect to feel more disinhibited and aggressive after drinking alcohol (Brown, Goldman, Inn, & Anderson, 1980). For example, when sober college men read a story about a man forcing sex on a date, they reported that they were more likely to behave like the man in the story when he was drunk rather than when he was sober (Norris & Kerr, 1993). This finding indicates that men can envision forcing sex when intoxicated.

Abbey, Ross, McDuffie, and McAuslan (1996b) described the results of a study in which college men completed a number of questionnaires, including a measure of alcohol expectancies and a measure of past sexual assault perpetration. As compared to sober perpetrators, those who had been drinking at the time of the assault were more likely to believe that alcohol increased men’s and women’s sexuality. Although these are correlational data, the findings suggest that men who believe that alcohol increases sexual arousal may feel more comfortable forcing sex when drinking because they can tell themselves that it was the alcohol that made them act that way and that the woman was also sexually aroused.

Women’s drunkenness and sexual behavior are censured more than men’s (Norris, 1994). Sexual assault perpetrators are quick to derogate drinking women (Kanin, 1984; Scully, 1991). In particular, women who drink alone in bars have been characterized as deserving targets because they are presumed to be sexually promiscuous. Even a small amount of alcohol consumption can affect men’s perceptions of a woman’s sexual availability. For example, George, Cue, Lopez, Crowe, and Norris (1995) and George, Gournic, and McAfee (1988) have demonstrated that college students evaluate a woman who has consumed only a few alcoholic drinks as being more promiscuous and more eager to have sex with her date than a woman who did not drink alcohol.

Other studies demonstrate that the mere presence of alcohol leads many students to assume a drinking woman wants sex even when force is used. For example, Norris and Cubbins (1992) had college students read a story about a man and a woman on a date in which the man clearly used physical force against the woman’s wishes to have sex with her. They found that nondrinking college women and men were most likely to view this depiction of acquaintance rape as consensual when both members of the couple had been drinking alcohol. Bernat, Calhoun, and Stolp (1998) compared the responses of college men who had perpetrated sexual assault to those who had not when listening to an audiotape of a date rape. Perpetrators who thought the couple had been drinking alcohol were the slowest to decide the man’s behavior was inappropriate. These studies suggest that after a couple has been drinking together, men (and sometimes women) are much less likely to realize that when a man physically forces a woman to have sex he is committing sexual assault. Participants were sober when they made these judgments about others who were drinking. Thus, it was their beliefs about alcohol, not its pharmacological effects, that influenced their perceptions of whether or not these situations involved forced sex. These studies suggest how strongly alcohol consumption and consensual sex are associated in many individuals’ minds.

5.4. Perpetrators’ life experiences

Sexual assault perpetration has been linked to several types of childhood and adolescent experiences. Men who have committed sexual assault are more likely than other men to have experienced abuse or violence as a child, to have engaged in delinquency in adolescence, to have peers who view forced sex as acceptable, to have early and frequent dating and sexual experiences, and to be heavy drinkers (Ageton, 1983; Boeringer, Shehan, & Akers, 1991; Calhoun et al., 1997; Koss & Dinero, 1988; Malamuth et al., 1991).

5.4.1. Childhood abuse experiences

Childhood sexual abuse frequently has long-lasting deleterious effects on its victims, both male and female (Davis & Petretic-Jackson, 2000; Finkelhor, 1990; Kendall-Tackett, Williams, & Finkelhor, 1993). Common effects observed in sexually abused boys include preoccupation with their sexual identity and inappropriate attempts to reassert their masculinity through aggressive and antisocial behavior (Watkins & Bentovim, 1992; Wellman, 1993). Many researchers have found that male victims of childhood sexual abuse are more likely then other men to become sexual assault perpetrators in adolescence and adulthood (Freeman-Longo, 1986; Longo & Groth, 1983; Seghorn, Prentky, & Boucher, 1987; Senn et al., 2000). For example, being the victim of childhood sexual or physical abuse was significantly related to sexual assault perpetration in a community sample of men (Senn et al., 2000). Childhood sexual abuse has also been linked to sexual assault perpetration among college men (Koss & Dinero, 1988; Malamuth et al., 1991; Stevenson & Gajarsky, 1992). There are many explanations for this relationship including modeling, identification with the aggressor, an attempt to regain a sense of power over one’s life, and a desire to restore one’s sense of masculinity. This is a complex literature that is only briefly reviewed here, and many victims of childhood sexual abuse manage to cope in ways that do not produce long-term negative consequences (see Romano & De Luca, 2000; Watkins & Bentovim, 1992 for thorough reviews).

5.4.2. Delinquency

Delinquency during adolescence is associated with adolescent and adult sexual assault perpetration. For some men, sexual assault may be just one manifestation of their antisocial tendencies (Seto & Barbaree, 1997). For example, Calhoun et al. (1997) measured sexual assault perpetration in a small sample of community men. Using multiple regression, the only significant multivariate predictor was delinquency; the more antisocial and aggressive behavior committed during adolescence, the more likely these men were to have committed sexual assault. Malamuth et al. (1991) found that delinquency was related to sexual assault perpetration among college men indirectly, through its relationship to sexual promiscuity.

5.4.3. Peer environments and norms

Ageton (1983) conducted a longitudinal study in which youths’ attitudes and experiences in early adolescence were used to predict sexual assault perpetration in later adolescence. The only variable that significantly discriminated between perpetrators and nonperpetrators was involvement with delinquent peers. Sexual assault perpetrators were involved in a social network of others who supported and committed unconventional, delinquent acts. Koss and Dinero (1988) found that college men who had committed sexual assault were more likely than other college men to be part of a peer group that perceived women as sexual objects.

College researchers have focused on Greek organizations as peer environments that encourage forced sex. Although not all fraternities condone forced sex, rates of sexual assault perpetration and approval of sexual assault are high among fraternity members (Boeringer et al., 1991; Martin & Hummer, 1989).

5.4.4. Sexual experiences

Men who have sex at an early age and who have many consensual sexual partners are more likely to commit sexual assault than are less sexually active men (Abbey et al., 1998, 2001; Kanin, 1985; Koss & Dinero, 1988; Malamuth et al., 1991, 1995). Early and frequent sexual activity may reflect an individual difference in sexual interest and motivation. It also provides these men with increased opportunities for committing sexual assault (Kanin, 1985; Malamuth et al., 1991). Senn et al. (2000) found that having sexual experiences at a young age and having many sex partners in adolescence were positively related to sexual assault perpetration in a community sample of men. Similarly, Malamuth et al. (1995) found that sexual promiscuity predicted sexual assault perpetration in cross-sectional and longitudinal analyses with college men.

5.4.5. Perpetrators’ usual alcohol consumption

Heavy drinking has also been linked to sexual assault perpetration. Several studies of incarcerated rapists have found that they have high rates of alcoholism (Rada, 1975; Seto & Barbaree, 1997). In college samples, men who report that they drink heavily are more likely than other men to report committing sexual assault and more extreme forms of sexual assault (Koss & Dinero, 1988; Ullman et al., 1999a).

There are many possible explanations of this relationship. First, heavy drinkers are likely to drink heavily in many types of situations, especially at parties and bars which are the types of situations in which sexual assault frequently occurs. Abbey et al. (2001) found that college men who had perpetrated sexual assault tended to consume more alcohol in sexual situations than did college men who had not committed sexual assault. Second, heavy drinkers may simply use intoxication as an excuse for engaging in sexual assault (Scully, 1991). Third, other factors such as impulsivity and delinquency may underlie both the desire to drink heavily and the desire to commit sexual assault (Seto & Barbaree, 1997).

6. Victims’ personality characteristics, attitudes, and experiences

Many researchers have examined the personality characteristics, attitudes, and life experiences of sexual assault victims. As can be seen in Table 2, and in contrast to the studies of perpetrators described above, few significant differences have been found between victims and nonvictims and they explain only small amounts of variance (Koss & Dinero, 1989). Personality traits have not been linked to victimization. These results suggest that women’s personal characteristics are not strong predictors of their likelihood of being sexually assaulted.

Table 2.

Victim characteristics that have been linked to sexual assault victimization

Personality characteristics

None consistently found

Attitudes and beliefs

None consistently found

Life experiences

Childhood sexual abuse

Early and frequent sexual activity

Heavy alcohol consumption

6.1. Childhood sexual abuse

The most consistent predictor of adolescent and adult sexual assault victimization is childhood sexual abuse. For example, Wyatt, Guthrie, and Notgrass (1992) found that women in a community sample who experienced sexual abuse in childhood were more than twice as likely as other women to be revictimized as adults. In another community sample, Russell (1986) reported that among childhood incest victims, 65% had experienced rape or attempted rape after the age of 14. Gidycz et al. (1993) found that college women who had experienced childhood sexual abuse were more likely than other women to be sexually assaulted in college.

Several authors have found that in addition to experiencing childhood sexual abuse, women who have been sexually assaulted in adolescence or adulthood are more likely than other women to have frequent sexual relationships and to be heavy drinkers. For example, in a sample of Navy recruits, women who had been raped were more likely than those who had not to be victims of childhood sexual abuse, to have many sexual partners, and to have drinking problems (Merrill et al., 1999). Similarly, Koss and Dinero (1989) found that college women who had experienced sexual assault in adolescence or adulthood were more likely than other college women to have experienced childhood sexual abuse, to have a larger number of sexual partners, and to drink more heavily.

These findings are usually described in terms of the long-term negative consequences associated with childhood sexual abuse that may lead some of its victims to become heavy drinkers and to have difficulty establishing long-term romantic relationships. Childhood sexual abuse affects survivors’ beliefs and attitudes regarding sex and intimacy, often leading to problematic interpersonal relationships (Kessler & Bieschke, 1999). Childhood sexual abuse survivors often report engaging in risky sexual behaviors including frequent sexual activity with multiple partners (Beitchman, Zucker, Hood, & Dacosta, 1992). A large number of childhood sexual abuse survivors experience abuse in adult romantic relationships, including physical, verbal, psychological, and sexual abuse (Kluft, 1990; Messman-Moore, Long, & Siegfried, 2000; Polusny & Follette, 1995).

Post-traumatic stress disorder (PTSD) may partially explain the relationship between childhood sexual abuse, alcohol use, and revictimization (Alexander, 1993; Chu, 1992; Rowan, Foy, Rodriguez, & Ryan, 1994). Childhood sexual abuse survivors who experience PTSD often report unpleasant affect and memories. To avoid this psychological pain, some survivors turn to substance use. They may be particularly likely to drink alcohol in potential sexual situations as a means of coping with their ambivalent feelings about sex. Drinking in potential sexual situations increases women’s risk of being sexually assaulted both because sexually assaultive men may view them as easy targets and because they may be less able to respond effectively (Polusny & Follette, 1995).

It is important to remember that many victims of childhood sexual abuse do not experience the problems described above (Saywitz, Mannarino, Berliner, & Cohen, 2000; Wilsnack, Plaud, Wilsnack, & Klassen, 1997). More research is needed to identify the protective factors that help many children maintain positive psychological health despite these experiences. As mentioned earlier, the fact that risk factors, such as having many sexual partners and heavy drinking, have been identified among some victims of sexual assault does not imply that these victims are responsible for what happened to them. Instead, this information can be used to develop more effective programs for childhood victims of sexual abuse.

6.2. Regular alcohol use and sexual assault: which comes first?

Two equally plausible causal directions have been proposed for the relationship between women’s regular heavy drinking and sexual assault. Heavy drinking may increase the risk of sexual assault because women who drink heavily may be targeted by perpetrators. Alternatively, sexual assault may increase the risk of becoming a heavy drinker because alcohol may serve as a coping strategy for some women, which they use to reduce the anxiety they feel with men. In cross-sectional studies, the causal direction of this relationship cannot be assessed. Several recent longitudinal studies have examined the relative strength of these two hypotheses and come to somewhat different conclusions.

In their longitudinal study of approximately 3000 women, Kilpatrick et al. (1997) examined the relative evidence for both of these hypotheses. They asked women at the initial interview and 2 years later about their experiences with rape and physical assault. They found only partial evidence for the first hypothesis; women who abused alcohol were not more likely than other women to be assaulted between the two interviews. However, women who abused alcohol and illicit drugs or only drugs were significantly more likely to be assaulted between interviews. Kilpatrick et al. found strong evidence for the second hypothesis. Women who were assaulted between the two interviews were significantly more likely to abuse alcohol and/ or illicit drugs at the follow-up. This study provides critical evidence regarding the relationship between women’s alcohol consumption and sexual assault. No survey, however, can demonstrate causality; for example, another variable, such as a change in marital status, may have increased some women’s risk of heavy drinking and sexual assault. Additionally, their results do not fully address the hypotheses initially stated because they combined physical and sexual assaults. Perhaps alcohol is more likely to make women targets for sexual, but not physical, abuse. Further, they focused on alcohol abuse that approximated DSM-IV criteria, rather than on heavy drinking. A woman would potentially only need to drink heavily on one occasion to be taken advantage of by a sexually aggressive man.

Testa and Livingston (2000) also compared these two hypotheses in a longitudinal study of about 90 women who drank at least three or four drinks of alcohol at least once a week. Unlike Kilpatrick et al. (1997), they found that women who had alcohol problems at the initial interview were more likely to experience a new sexual assault than were other women (drug use was not assessed). They did not find evidence for the second hypothesis. Women who experienced a new sexual assault were not more likely than other women to drink heavily or to experience alcohol problems at the second interview. These findings may be due to a restriction in range; participants were initially selected for this study because they drank fairly heavily.

Finally, White and Humphrey (1998) examined the relationship between alcohol and sexual assault in a longitudinal study of more than 700 college women. Experiencing sexual assault one year doubled the odds of heavy drinking in the following year. Similarly, heavy alcohol use in one year doubled the odds of experiencing sexual assault in the following year. Taken together, these three longitudinal studies suggest that there is a reciprocal relationship between alcohol consumption and sexual assault victimization.

7. Situational factors

Thus far, this review has considered characteristics of perpetrators and victims that might contribute to sexual assault. Men who have a propensity to commit sexual assault do not act on this propensity on all possible occasions, therefore, situational factors must also be considered. For example, bars and parties are locations which heavy drinking men and women frequent. Thus, these are situations in which perpetrators can easily find an intoxicated woman to target for a sexual assault (Parks & Miller, 1997). Several aspects of the situation are considered below and are summarized in Table 3 including the relationship between the perpetrator and the victim, the location of the assault, the perpetrators’s misperception of the victim’s sexual intentions, and alcohol’s effects on how both the perpetrator and victim process information and react to each other.

Table 3.

Situational factors that have been linked to alcohol-involved sexual assaults

Type of relationship between perpetrator and victim

Locations in which perpetrator and victim spend time together

Perpetrator’s misperception of the woman’s degree of sexual interest

Victim’s risk assessment

Perpetrator’s aggressiveness

Victim’s resistance

Victim’s injuries

7.1. The relationship between the perpetrator and the victim

Several authors have found that the relationship between the perpetrator and the victim is more casual in alcohol-involved sexual assaults as compared to sexual assaults that do not involve alcohol (Abbey et al., 1996a; Harrington & Leitenberg, 1994; Ullman et al., 1999a; Ullman, Karabatsos, & Koss, 1999b). For example, Harrington and Leitenberg (1994) found that alcohol consumption was more likely when the man and woman were casually acquainted rather than steady dates. Norris, Nurius, and Gaylord (1998) also found that sexual assaults that involved alcohol were more likely to be perpetrated by a nonromantic friend or acquaintance; in contrast, sexual assaults that did not involve alcohol were more likely to be perpetrated by a romantic partner. Drinking victims had known the perpetrator for a shorter length of time (about 1 year) than had nondrinking victims (about 2 years).

7.2. The type of setting in which the sexual assault occurred

Not surprisingly, alcohol-involved sexual assaults are more likely than other sexual assaults to involve spontaneous interactions and time spent at a party or bar (Abbey et al., 1996a; Ullman et al., 1999a, 1999b). The perpetrator and victim are less likely to be alone the entire time when either of them has been drinking alcohol (Abbey et al., 2001; Norris et al., 1998). In one study, one-third of the women who regularly drank in bars had been sexually assaulted at a bar or immediately after being at a bar (Parks & Miller, 1997).

7.3. Men’s misperceptions of women’s sexual intentions

Research with college students has consistently found that men perceive women as behaving more sexually and as being more interested in having sex with them than the women actually are (Abbey, 1982; Abbey, Zawacki, & McAuslan, 2000; Harnish, Abbey, & DeBono, 1990). This finding is relevant to situational theories of sexual assault because some perpetrators may at least initially misperceive their female companion’s cues. When sexual assaults occur on dates or in other types of social situations such as parties, the man may expect consensual sex. This expectation does not justify forced sex or mitigate its seriousness. However, when sexual assaults occur under circumstances in which consensual sex is a feasible outcome, it is important to consider how men interpret women’s behavior.

The cues used to convey sexual interest are often ambiguous; thus, it is easy to mistake someone’s friendliness for flirtation. For example, when a man meets a woman at a party and she seems very interested in an anecdote about his trip to Australia, this might be a sign of sexual attraction. It is just as likely that she is being attentive because she is planning to travel there herself. If this man is sexually attracted to this woman, he has to decide how to interpret her response. His initial moves are likely to be subtle, so that he can save face if rejected. If he perceives an encouraging response, then he will make another move; thus his confidence increases as the interaction continues. This indirect form of indicating sexual interest is common, and usually men and women manage to make their intentions clear and avoid embarrassment if their companion is not interested in them (Abbey, 1987). The vagueness of cues, however, frequently leads to miscommunication.

7.3.1. Alcohol enhances men’s misperception of women’s behavior

Expectancies are often self-fulfilling; people tend to process ambiguous information in such a way that it confirms their initial hypothesis (Snyder & Stukas, 1999). If a man who is on his fifth date with a woman is expecting the evening to end with sexual intercourse, then he is likely to interpret any smile or touch as a sexual sign (Roche & Ramsbey, 1993). Alcohol consumption exacerbates this process, thereby increasing the probability that forced sex will occur. As noted earlier, alcohol consumption disrupts higher-order cognitive processes, making it difficult to evaluate complex stimuli (Hindmarch, Kerr, & Sherwood, 1991; Steele & Josephs, 1990; Tranel, Anderson, & Benton, 1994). Intoxication encourages people to ignore peripheral cues and emphasize the most salient cues in the situation. People’s expectations determine which cues are most salient to them. Thus, intoxication makes it very easy for a potential sexual assault perpetrator to interpret friendly cues as a sexual invitation if that is the behavior he is anticipating.

Abbey et al. (2000) examined the role of alcohol consumption in initial sexual mispercep-tions. Using the balanced placebo design (Rohsenow & Marlatt, 1981), unacquainted college men and women talked to each other for 15 min after drinking either an alcoholic or a nonalcoholic beverage. As found in past research, men thought their partners were more sexually attracted to them than their partners indicated that they were. When drinking alcohol, both men and women were perceived as behaving more sexually than when they were sober. In support of the alcohol myopia theory, drinkers and nondrinkers also appeared to evaluate their partners’ cues somewhat differently. Drinking participants perceived both men and women as being higher in sexuality when they made any hints about being interested in seeing their partner again (e.g., “I’d like to see that movie too”; “You go to that club? I’ve been there too”). These types of comments did not affect sober participants’ judgments. This finding suggests that drinkers perceived these ambiguous comments as signs of sexual interest, whereas nondrinkers were more cautious in their interpretation.

7.3.2. Alcohol reduces women’s ability to assess risk

Alcohol produces the same types of cognitive deficits in women as in men. Therefore, if a woman feels that she has made it clear that she is not interested in sex at this point in time, intoxication will make it difficult for her to recognize signs that suggest her companion has misunderstood her. For example, in Harrington and Leitenberg’s (1994) sample of sexually assaulted college women, most had engaged in some consensual sexual activity with the man (e.g., kissing, petting) before the assault occurred. Sober victims, however, were less likely than intoxicated victims to engage in consensual sexual activities. This suggests that intoxicated women may not have discerned the impact that their kisses were having.

A woman’s recognition of the man’s misperception of her sexual intentions does not insure that the misunderstanding is resolved. Many women find it difficult to directly confront a man who has misperceived them; they feel that they must protect their date’s ego and let him down gently. This is particularly true when women want to maintain the relationship (Lewin, 1985; Murnen, Perot, & Byrne, 1989). Norris, Nurius, and Dimeff (1996) interviewed a woman who explained how difficult it is to let a man know that his sexual advances are unwanted, “[you] start to worry … And you’re also thinking, ‘I don’t want to feel dumb. I don’t, you know, want him to tell his friends, or start calling me a prude”’ (p. 135). Indirect or weak refusals, however, are rarely taken seriously. Abbey et al. (2001) compared the characteristics of sexual assaults and worst dates described by college men. As compared to worst dates, sexual assaults involved a longer period of time in which the man misperceived the woman’s sexual intentions. The longer a man thinks that a woman wants to have consensual sex with him, the more likely it is that he will feel justified forcing sex when he finally realizes her “no” is serious because he feels that he has been led on by her.

Testa and Livingston (1999) interviewed sexual assault victims who were single, drank at bars and parties, and were sexually active. These women described how alcohol-involved sexual assaults occurred after they behaved in ways they typically perceived as too risky. For example, they accepted a ride home from a party with a man they did not know well or allowed him into their apartment. These women felt that they missed danger cues that they usually would have noticed because they were intoxicated. Norris et al. (1996) suggested that women must often “walk a cognitive tightrope” (p. 137) when interacting with men. When on a date or interacting at a party, women want to have fun and be liked, although they know that not all men can be trusted and that they must be on the alert. Therefore, women must balance their desire to have fun and their desire to be safe. Alcohol myopia may lead women to take risks they would normally avoid because being intoxicated allows them to ignore danger cues and focus on their enjoyment.

Laboratory research can only ethically examine the early stage of the misperception process. Survey research is needed to examine the role of misperception in sexual assaults. Muehlenhard and Linton (1987) compared the characteristics of college students’ dates that did and did not involve sexual assault. Men believed that dates on whom they had forced sex had “led them oncognitive tightrope” (p. 137) when interacting with men. When on a date or to a greater extent than did dates on whom they had not forced sex. Similarly, women who had experienced forced sex on a date were more likely than those who had not to believe that the man felt “led on, cognitive tightrope” (p. 137) when interacting with men. When on a date or although women reported that this had not been their intention. Using structural equation modeling, Abbey et al. (1998) found that misperception mediated the relationship between alcohol consumption and perpetration. The more alcohol college men consumed in heterosocial situations, the more frequently they had misperceived a woman’s sexual intentions; and frequency of misperception was positively related to the number of sexual assaults they had committed.

7.4. Alcohol’s effects on the man’s use of force and the woman’s response

At the point that the sexual assault occurs, what role does alcohol play? Are sexually assaultive men more aggressive if they have been drinking alcohol? Are their victims more likely to be injured? Are victims who are drinking less likely to fight back? Many studies have examined the role of either the perpetrator’s or the victim’s alcohol consumption on the outcome of the sexual assault. Disentangling the effects of perpetrators’ as compared to victims’ drinking is extremely difficult because they are so highly correlated.

The cognitive deficits associated with alcohol consumption can enhance men’s likelihood of behaving aggressively because intoxicated men may have more difficulty generating non-aggressive solutions to gaining sexual satisfaction (Sayette, Wilson, & Elias, 1993). An intoxicated man is likely to focus on salient cues, such as his sexual arousal, frustration, and sense of entitlement, rather than distal cues such as the potential pain and suffering of his victim, his internalized sense of morality, or the possibility that he will be punished. Thus, an alcohol-induced sense of disinhibition and reduction in anxiety and self-appraisal makes it easier for men to use physical force to obtain sex (Crowe & George, 1989; Ito, Miller, & Pollock, 1996; Pernanen, 1993).

In laboratory studies of aggression, which use a learning or competition paradigm, intoxicated men retaliate strongly if they feel threatened and once they begin behaving aggressively, it is difficult to make them stop (Taylor & Chermack, 1993). Although most of these studies used male aggressors and victims, intoxicated men are also aggressive toward female targets (Giancola & Zeichner, 1995). These findings suggest that sexual assaults involving intoxicated perpetrators are more likely than other sexual assaults to include extreme levels of forced sex, more violent behavior, and more injuries to the victim. Sexual assault research also suggests that prompt action can sometimes deter a rapist (Seigel, Sorenson, Golding, Burnam, & Stein, 1989; Ullman, 1997). Alcohol impairs motor skills, thereby limiting a woman’s ability to effectively resist sexual assault (Koss & Dinero, 1989). This has led researchers to hypothesize that intoxicated victims will be more likely than other victims to experience extreme levels of sexual assault and to engage in less resistance. Less resistance, in turn, is hypothesized to be related to less perpetrator violence and fewer injuries.

Findings regarding these hypotheses are mixed. For example, two recent studies examined the relationship between alcohol and sexual assault outcomes using Koss et al.’s (1987) large sample of college students. One study (Ullman et al., 1999a) examined perpetrators’ reports of their own drinking and their victims’ drinking. In a complementary manner, the second study (Ullman et al., 1999b) examined victims’ reports of their own drinking and their perpetrators’ drinking. Both studies tested path models that included the effects of usual alcohol consumption, alcohol consumption during the assault, the perpetrator’s aggression, the victim’s resistance, and the severity of the assault (ranging from sexual contact to completed rape).

In both data sets, there was a strong relationship between the man’s and woman’s alcohol consumption during the assault; the standardized beta was 0.81 in the perpetrators’ analyses and 0.67 in the victims’ analyses. Quantity of alcohol consumed was not evaluated; only whether or not it was consumed. In the perpetrators’ data set (Ullman et al., 1999a), the woman’s alcohol consumption during the assault, not the man’s, was positively related to how aggressive the man was during the assault. Perpetrators’ aggressiveness, in turn, was positively related to how much victims resisted, which was positively related to severity of the assault. Given how strongly the man’s and woman’s alcohol consumption were correlated, it is not surprising that only one emerged as a significant multivariate predictor (Stevens, 1996). However, based on the research described above, Ullman et al. had expected perpetrators’ alcohol consumption to be positively related to perpetrators’ aggressiveness (alcohol consumption making men more violent) and victims’ alcohol consumption to be negatively related to perpetrators’ aggressiveness (alcohol consumption making the victim unable to resist effectively so less violence is required). Realistically, if perpetrators’ and victims’ alcohol consumption are strongly, positively correlated, they cannot relate to other variables in opposite directions.

Somewhat different results were found with the victims’ data set (Ullman et al., 1999b). The relationship between perpetrators’ and victims’ drinking was somewhat lower (0.67 vs. 0.81). As hypothesized, victims’ alcohol consumption during the assault was negatively related to perpetrators’ aggressiveness, whereas perpetrators’ alcohol consumption was positively related to perpetrators’ aggressiveness. As found in the men’s data set, perpetrators’ aggressiveness was positively related to victims’ resistance, which in turn was positively related to the severity of the assault.

Martin and Bachman (1998) examined similar hypotheses with the National Crime Victimization data set. A strength of this study is that it includes a nationally representative sample of 279 rape victims. There are two major limitations. First, women victims were asked to report on the perpetrators’ alcohol consumption but not their own. Second, only whether or not perpetrators consumed alcohol was assessed, rather than the quantity of alcohol consumed. Contrary to their hypothesis, they found that perpetrators were more likely to drink alcohol in attempted rapes than in completed rapes. However, as expected, victims were more likely to be injured when perpetrators drank alcohol. The relationship between perpetrators’ drinking and victims’ injuries was also examined by Tjaden and Thoennes (2000). They used logistic regression to examine risk factors associated with intimate partner violence in a representative sample of 4896 women. Perpetrators’ use of alcohol or drugs was significantly related to women’s likelihood of being injured during a rape.

Several other researchers have found that completed rapes are more common among intoxicated victims suggesting that sober victims are more likely to find a way to escape or effectively resist (Abbey et al., 1996b; Harrington & Leitenberg, 1994; Ullman & Knight, 1993). For example, Harrington and Leitenberg (1994) found that acquaintance rape victims who reported being at least somewhat drunk were less likely to use physical resistance strategies than were victims who were not drunk.

Many rapists realize that it is harder for women to resist sexual advances when intoxicated, thus they try to get their female companion drunk as a way of obtaining sex (Bowker, 1979; Kanin, 1985; Mosher & Anderson, 1986; Tyler et al., 1998). Kanin (1985) found that 75% of the college date rapists he interviewed purposely got women intoxicated in order to have sexual intercourse with them. Rapists may also encourage women to play drinking games so that they become intoxicated (Johnson, Wendel, & Hamilton, 1998).

A serious methodological limitation of these studies concerns how perpetrators’ alcohol consumption was assessed. Alcohol consumption was simply measured as whether or not any alcohol was consumed prior to the assault, not the quantity consumed or how far in advance of the assault. Thus, it is impossible to determine from these studies how intoxicated the perpetrators were at the time of the assault; if they had only one drink several hours before the assault they may not have been intoxicated at all. Men at extremely high levels of intoxication are likely to experience sexual and motor impairments that make sexual assault completion unlikely.

8. Suggestions for future research

Surveys of victims and perpetrators have been the primary source of information about alcohol-involved sexual assaults. Alcohol administration studies have also provided useful data. These two research methodologies have different strengths and limitations; thus, they complement each other well. This paper has highlighted the best research. However, there are still many gaps in existing knowledge about alcohol’s role in sexual assault. In this final section of the paper, suggestions are made for research that would fill in some of these gaps and better inform prevention and treatment programming.

8.1. Surveys with perpetrators and victims

More research is needed that uses community samples and prospective designs. Almost all of what we know about sexual assault perpetrators comes from imprisoned rapists and college students. The few community studies that have been conducted had very small samples of unknown representativeness (Calhoun et al., 1997; Senn et al., 2000). Large-scale nationally representative surveys of men are needed to document the characteristics of sexual assault perpetrators and to compare the characteristics of intoxicated and sober offenders. These surveys present more challenges than do general population victimization surveys because men are being asked to report on past criminal behavior and may be extremely reluctant to acknowledge engaging in such activities. The careful use of survey research techniques designed to enhance personal disclosures, such as computer-assisted self-interviews, should enhance men’s willingness to disclose past perpetration (Turner et al., 1998). Additionally, in some circumstances, a Certificate of Confidentiality can be obtained from the federal government to protect this type of data from subpoena; knowledge of this should reassure participants that their honesty will not lead to prosecution.

The literature review provided above demonstrates that personality, attitudinal, and experiential factors have been consistently linked to sexual assault perpetration. Most of these studies, however, have not attempted to determine if perpetrators of alcohol-involved assaults systematically differ from other perpetrators. Alcohol researchers have examined individual difference variables related to sexual and aggressive behavior in the laboratory such as hostility and attraction to sexually explicit materials (George et al., 2000; Taylor & Chermack, 1993). Future research is needed to determine the extent to which some of the concepts examined in general aggression and sexuality research help explain alcohol-involved sexual assaults. For instance, aggression researchers have demonstrated that as compared to individuals with normal cognitive functioning, individuals who are low in general cognitive functioning react more aggressively (Hoaken, Giancola, & Pihl, 1998). Further, deficits in executive cognitive functioning have been found in heavy drinkers (Blume, Marlatt, & Schmaling, 2000). These findings suggest a mechanism through which heavy drinking can reduce cognitive functioning, which in turn can increase aggressiveness among intoxicated perpetrators. In the sexuality domain, although results have been mixed, many alcohol administration studies have found expectancy effects (Crowe & George, 1989). These studies lead to the hypothesis that perpetrators with strong alcohol expectancies regarding sexuality will be most likely to misperceive the actions of a drinking woman as implying a higher degree of interest in sex than she intended. More knowledge about the ways in which men who commit alcohol-involved sexual assaults are similar to and different from other perpetrators can lead to the development of prevention and treatment programs targeted for men who have committed sexual assault when intoxicated or who are heavy drinkers and may be at risk of committing sexual assault when intoxicated.

Most surveys of sexual assault have been cross-sectional, making it impossible to determine which beliefs or experiences were predictors of sexual assault and which were consequences of sexual assault. For example, if men who have committed alcohol-involved sexual assaults have stronger alcohol expectancies regarding sex than do other perpetrators (Abbey et al., 1996b), is this a predisposing belief which encouraged perpetration or is it a consequence of committing a sexual assault when intoxicated? With one exception, the few prospective studies that have been conducted have attempted to determine whether there is more support for the hypothesis that heavy drinking leads to sexual assault victimization or that sexual assault victimization leads to heavy drinking. As noted above, these studies have produced mixed results (Kilpatrick et al., 1997; Testa & Livingston, 2000; White & Humphrey, 1998). Not only do they suggest that these effects are reciprocal, they also highlight the fact that patterns of relationships may differ among different subpopulations. Testa and Livingston’s (2000) work with young women who drink heavily and frequent bars demonstrates the importance of studying high-risk populations, as well as general population samples.

An additional limitation of most existing prospective studies is that they have followed adults for only a few years. Ageton (1983) provided an important exception by following youth from early to late adolescence. Her finding that early involvement with delinquent peers predicted later sexual assault perpetration suggests avenues for early intervention and prevention. Studies that interview young adolescents and follow them into adulthood are needed to disentangle the causal direction between variables such as alcohol use, delinquent behavior, attitudes supportive of violence against women, and sexual assault perpetration.

Measures of alcohol consumption in sexual assault surveys are extremely limited. Typically, researchers only assess whether or not alcohol was consumed, not the level of intoxication. Surveys need to include measures of the size and number of drinks consumed, the time period within which they were consumed, and perceived intoxication. With improved measurement, more sophisticated hypotheses can be examined. For example, although small doses of alcohol increase sexual arousal, high doses inhibit performance (Crowe & George, 1989). This suggests a curvilinear relationship between alcohol consumption and outcome severity, because at high levels of intoxication perpetrators are too impaired to continue. Similarly, dose may affect the victim’s response. Several researchers have suggested that when the victim is extremely intoxicated, perpetrators use less force because she is unable to resist (Martin & Bachman, 1998; Ullman et al., 1999a). However, the dichotomous measures of alcohol consumption used in past studies have not allowed this hypothesis to be directly evaluated. Here too, a curvilinear hypothesis is credible. Mildly intoxicated women may be more willing to physically resist because alcohol myopia makes them less concerned about hurting the man’s feelings; however, extremely intoxicated women may experience such extensive motor and cognitive impairments that they are unable to respond. These are just some of the complex issues that could be addressed with better indicators of alcohol consumption. Again, it is difficult to develop targeted prevention or treatment programs without such detailed information. Are women at greater risk of experiencing sexual assault whenever they drink or only when they drink to intoxication or only when they are with a man who is also intoxicated? Do men who have committed alcohol-involved sexual assaults only commit sexual assaults when intoxicated or do they also commit sexual assaults when sober? Knowing this has important implications for perpetrator treatment programs.

Some of the explanations of alcohol’s role in sexual assault focus on the man’s alcohol consumption (e.g., his beliefs about alcohol’s impact on his sexual desire, his cognitive impairments), whereas others focus on the woman’s alcohol consumption (e.g., men’s perceptions of drinking women, her cognitive impairments). It is difficult to evaluate the independent influences of the perpetrators’ and victims’ drinking because they are confounded in most sexual assaults. Researchers need to acknowledge this problem when they are unable to separately consider the effects of each individual’s alcohol consumption (Martin & Bachman, 1998). Research is needed with very large samples so that the unique effects of perpetrators drinking alone, victims drinking alone, and perpetrators and victims drinking together can be determined.

Most sexual assault research has involved large, quantitative surveys. These studies are important, however, qualitative research is also needed with individuals from different subgroups to insure that what we have learned about alcohol’s role in sexual assault in general populations also applies to individuals from different ethnic and cultural groups. Rates of alcohol consumption vary among different ethnic groups, thus rates of alcohol-involved sexual assault are also likely to differ (Office of Applied Statistics, 1997). Abbey et al. (1996a) found that alcohol was involved in fewer of the sexual assaults that African American women experienced as compared to Caucasian women, although the cross-sectional predictors of sexual assault did not vary by ethnicity. This study, however, assessed only a limited number of variables, therefore, research that includes more detailed questions about alcohol’s role in sexual assault may find differences. For example, some religious groups have prohibitions against alcohol consumption. This suggests that if women from these groups experience an alcohol-involved sexual assault, they may be derogated by their support network. Anecdotal reports suggest that perpetrators often feel less responsible for sexual assault if they had been drinking alcohol; whereas women often feel more responsible (Abbey et al., 1996b; Scully, 1991). More research is needed that explores in depth how alcohol affects perceptions of responsibility and whether factors such as subgroup norms about alcohol consumption, the quantity consumed, or the location where it is consumed affect these perceptions.

Qualitative research with fraternity men has demonstrated how peer group norms about women as sexual commodities contribute to sexual assault perpetration (Martin & Hummer, 1989). Similar studies are needed with noncollege populations who participate in other types of all male groups, such as sports teams or fraternal organizations, to determine if there are shared norms that encourage or discourage sexual assault. Several studies have documented that as compared to other sexual assaults, alcohol-involved sexual assaults typically involve a more casual social relationship between perpetrator and victim and are more likely to occur after being in locations such as parties or bars (Abbey et al., 1996a; Ullman et al., 1999b). Research that explores how perpetrators select a potential victim at a party or a bar would help inform prevention programs.

Another issue that could be explored in qualitative research concerns the interrelationships between women’s experience of childhood sexual abuse, heavy drinking, and revictimization in adulthood (Wilsnack et al., 1997; Wyatt et al., 1992). Do some victims of childhood sexual abuse drink more than usual in situations that remind them of the abuse? Do they sometimes drink to make it easier to engage in sexual activities but then are too intoxicated to stop a man who forces more sex than they wanted? More focused treatment programs could be developed for child victims if these long-term effects were better understood.

8.2. Alcohol administration research

Only experimental research can demonstrate causality. Laboratory research in which participants are randomly assigned to drink an alcoholic or a nonalcoholic beverage insures that any differences that are found are due to alcohol rather than other variables such as pre-existing personality differences. The major drawback to experimental research is its external validity. What outcomes adequately simulate sexual assault yet can be ethically manipulated by experimenters? Hall and Hirschman (1994) have argued that men’s willingness to expose a female confederate to sexually violent materials is a reasonable proxy for sexual assault perpetration. In their research, many more sexual assault perpetrators than nonperpetrators were willing to show women sexually violent materials, even though they knew that these materials upset the woman. Noting the limitations of vignette research, Bernat et al. (1998) suggest that responses to an audiotape of a date rape can correspond to the judgments that individuals make in a potential sexual assault situation. Additional methodological research is needed to determine what types of materials provide the most appropriate proxies.

Alcohol administration studies are also needed that focus on the processes that lead up to sexual assault. Laboratory research can examine the effects of alcohol on men’s perceptions of women’s sexual interest in them under different types of circumstances. For example, research on alcohol’s effects on higher-order cognitive processing suggests that intoxicated men focus on the most salient cues in the situation. As described above, in the types of social situations in which sexual assault usually occurs, the most salient cues are likely to encourage men to assume that the woman is interested in engaging in sexual activities with them. What happens when the most salient cues suggest that the woman is not sexually attracted to the man? Alcohol myopia theories suggest that intoxicated men are more likely to attend to these cues than sober men. If such effects could be demonstrated in the laboratory, they would provide important information about strategies that women could use with intoxicated men to reduce their risk of being assaulted. Additionally, alcohol administration studies can be conducted with men whose sexual assault perpetration status is already known. Differences in the behavior of perpetrators and nonperpetrators when drinking alcohol and when sober can be compared. For example, they could roleplay how they would act in certain types of situations with a female confederate. Are intoxicated perpetrators less able to accurately read the cues the confederate sends or to convey their own sexual interest? Do they ignore women’s negative cues and overreact to ambiguous or positive cues as alcohol myopia theories would suggest? Here too, the results could be used to develop more effective prevention and treatment programs.

9. Conclusion

Our understanding of the causes, circumstances, and consequences of sexual assault have increased greatly during the past 30 years. Yet, much is still unknown, particularly about alcohol’s role in sexual assault. Alcohol is described as a risk factor in most college acquaintance rape programs; however, they do not provide concrete, realistic prevention plans based on this information (Bohmer & Parrot, 1993). Sexual assault and alcohol researchers need to collaborate with prevention and treatment practioners to develop well-informed prevention strategies for college and noncollege populations. If substance abuse therapists have ideas about strategies to reduce men’s aggressiveness when drinking, these can be tested in laboratory studies before being implemented in prevention programs.