Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2008 Jul 30.
Published in final edited form as: Psychol Women Q. 2008;32(2):145–158. doi: 10.1111/j.1471-6402.2008.00419.x

WOMEN’S SOCIAL BEHAVIOR WHEN MEETING NEW MEN: THE INFLUENCE OF ALCOHOL AND CHILDHOOD SEXUAL ABUSE

Kathleen A Parks 1, Amy L Hequembourg 1, Ronda L Dearing 1
PMCID: PMC2491329  NIHMSID: NIHMS56158  PMID: 18668186

Abstract

Heavy alcohol consumption (Testa & Parks, 1996) and childhood sexual abuse (CSA; Messman-Moore & Long, 2003) have been associated with adult sexual victimization. We examined the social behavior of 42 women under two alcohol conditions (high dose and low dose) in a bar laboratory. Women were videotaped interacting with a man they had just met. Women in the higher dose condition engaged in more open body position and talked, stood, and walked more than women in the lower dose condition. These behaviors are consistent with signs of intoxication or romantic interest. The women in the high-dose condition also frowned more than women in the low-dose condition. An increase in frowning could indicate less comfort or may be considered consistent with an increase in animation during the social interaction given the concomitant increase in other behaviors. Thus, the nonverbal behavior of women in the high-dose condition could be interpreted as mixed signals. CSA victims exhibited fewer head movements (e.g., nods), were less animated, and frowned more than non-CSA victims. These behaviors convey reticence or possibly even anxiety or discomfort during the social interaction. Thus, the nonverbal behavior of women with a history of CSA may convey an unease that could be viewed by a potential perpetrator as vulnerability. Our findings suggest that both acute alcohol consumption and history of CSA may influence nonverbal social behavior and may influence risk for sexual assault by sending mixed cues of romantic interest or signs of vulnerability to potential perpetrators.

Researchers have noted how alcohol use, childhood sexual abuse (CSA) history, and nonverbal behaviors are independently related to victimization, but their complex interrelationships are not well understood. In the following paragraphs, we provide a brief overview of our current understanding of these independent relationships. We then propose an exploratory model to conceptualize the possible linkages among alcohol consumption, history of CSA, risk perception, nonverbal behavior, and risk for (re)victimization.

Alcohol and Victimization

Alcohol has been associated with increased likelihood of sexual victimization (for a review, see Testa & Parks, 1996). Risk for sexual victimization in social situations involving alcohol may be due to alcohol’s pharmacological effects on cognitive abilities. Misperceptions and behaviors that can increase vulnerability to sexual victimization or lead to sexual assault perpetration can be exaggerated under conditions of alcohol intoxication.

Alcohol has been found to increase men’s misperceptions of women’s friendly behavior as sexual intent. Abbey, McAuslan, and Thomson Ross (1998) found that greater frequency of alcohol consumption among college men was related to more frequent misperceptions about women’s sexual intentions, with the frequency of misperceptions positively related to the number of sexual assaults they had committed. Abbey, Zawacki, and McAuslan (2000) found that alcohol consumption influenced both men’s and women’s sexual judgments. After consuming alcohol, men perceived women more sexually than women perceived themselves, and men perceived themselves more sexually than women perceived them. Furthermore, men misperceived women as more sexual and less inhibited when women utilized high levels of nonverbal attention cues (e.g., leaning forward, stretching one’s arms toward one’s partner, using animated speech).

In addition, alcohol decreases women’s ability to effectively evaluate the consequences of their actions and distorts their ability to predict how others may perceive them (Abbey, Thomson Ross, & McDuffie, 1994; Testa, Livingston, & Collins, 2000). Men may perceive a woman as more sexually available based on the woman’s consumption of alcohol or her presence in an environment where alcohol is being served, such as a party or bar (George, Gournic, & McAfee, 1988; Parks & Scheidt, 2000). Not surprisingly, therefore, alcohol has been consistently linked with sexual victimization when the victim, perpetrator, or both have consumed alcohol (for a review, see Abbey, Zawacki, Buck, Clinton, & McAuslan, 2004). However, findings are inconsistent regarding the role of alcohol as a predictor of revictimization (see Breitenbecher, 2001). In contrast, a history of CSA has been identified as a powerful predictor of revictimization, with consistent associations found between CSA history and alcohol use and abuse.

CSA and Revictimization

CSA has been consistently identified as one of the strongest predictors of adult sexual assault among women (for reviews, see Classen, Palesh, & Aggarwal, 2005; Messman-Moore & Long, 2003; Rich, Combs-Lane, Resnick, & Kilpatrick, 2004). In a meta-analysis of 19 empirical studies of sexual revictimization, Roodman and Clum (2001) found an overall moderate effect size (d = 0.59) for revictimization, indicating a “definite relationship” between CSA and adult sexual victimization. With 20% of adult women in the United States reporting a history of CSA and clear documentation of elevated risks for revictimization among these women (Finkelhor, 1994), researchers have devoted increasing attention to better understanding risk factors for sexual revictimization among adult CSA survivors.

Studies have identified a number of consequences of CSA that might contribute to women’s increased risk for revictimization, including deficits in risk perception and interpersonal skills. Cloitre (1998) suggests that an early abuse history disrupts the development of an “appropriate flight-or-fight response” (p. 281), which decreases the likelihood that a woman will respond either effectively or at all in a potential assault situation during adulthood. An ineffective or failed response when faced with threat of assault may result from the impaired cognitive and behavioral skills needed to effectively recognize and extricate oneself from a dangerous situation. Impaired risk perceptions may be attributable to low self-esteem, assertiveness deficits, or reduced avoidance or response skills (Messman-Moore & Long, 2003).

A variety of interpersonal factors may mediate the association between CSA and revictimization, including attachment style, the psychological impact of CSA, attribution and coping related to CSA experiences, and risky adult behaviors (e.g., hyperfemininity, number of sex partners, delinquency/drug use; Gold, Sinclair, & Balge, 1999). Classen, Field, Koopman, Nevill-Manning, and Spiegel (2001), for example, found significant differences in interpersonal problems related to revictimization status among a sample of 52 adult CSA survivors in treatment for post-traumatic stress disorder. Based on a self-report measure of interpersonal problems, non-revictimized women, compared to revictimized women, were significantly more likely to be socially avoidant, have difficulties being assertive, and put the needs of others before their own (i.e., “overly nurturant,” “too responsible”). Although these findings are suggestive of factors mediating CSA and revictimization, insufficient evidence exists to evaluate the role that interpersonal skill deficits play in the relationship between CSA and revictimization (for a review, see DiLillo, 2001).

An additional risk factor that may differentiate CSA victims and nonvictims is rate of substance abuse. Researchers have found higher rates of alcohol and other drug use and abuse among CSA victims compared to nonvictims (Epstein, Saunders, Kilpatrick, & Resnick, 1998; Langeland & Hartgers, 1998; Martin et al., 2005; Miller, Downs, & Testa, 1993; Simpson & Miller, 2002). It is possible that CSA victims may utilize alcohol or illicit drugs as a coping mechanism to manage negative consequences associated with CSA (Simpson, 2003). As previously noted, alcohol use places women at increased risk for sexual assault.

Nonverbal Behavior and Victimization

Although the research outlined above identifies a number of factors that may be associated with revictimization, there are many questions unanswered regarding the mechanisms that increase vulnerability to revictimization among female CSA survivors. One possibility is that women’s non-verbal behavior during social situations with a potential perpetrator interacts with alcohol consumption, CSA history, and/or risk perception to convey vulnerability for revictimization.

Researchers have identified gender differences in non-verbal communication during social and romantic interactions between men and women (e.g., Fichten, Tagalakis, Judd, Wright, & Amsel, 1992; Kowalski, 1992; Moore & Butler, 1989). In the absence of alcohol, Kowalski (1992) found that men often interpret women’s behaviors as sexually responsive when women do not intend them to be perceived that way, thereby increasing women’s vulnerability to unwanted sexual advances or sexual assault. Women exhibit a variety of observable nonverbal cues that influence men’s responses in predictable ways. Moore and Butler (1989) found that women who gazed and smiled frequently, or “high display” women, were more likely to be approached by men than were “low display” women. This was true even when low-display women were more physically attractive than high-display women. They further noted that high-display women, after being approached by a man, engaged in more nodding, leaned closer to the man, and were more likely to cock their head to one side while listening than low-display women. These nonverbal behaviors suggest interest in the interaction and are likely to maintain the interaction for a longer duration.

Potential perpetrators may look for certain nonverbal body movements as indicators of women’s vulnerability to assault. An early study by Grayson and Stein (1981) asked male prison inmates with a history of felony assault to watch videotapes of women walking down a street and to characterize their vulnerability for assault. Grayson and Stein (1981) found that prisoners differentiated potential victims and nonvictims through differences in the women’s body movements. Individuals who were not seen as potential victims had body movements that appeared to flow from the center of their bodies with their body parts appearing more interconnected, whereas potential victims had movements that were less fluid and more sporadic. Interestingly, this description of the potential victim is not dissimilar to descriptions of visibly intoxicated individuals as awkward, stumbling, and uncoordinated. Visibly intoxicated individuals are often targeted by perpetrators of violent crime (Homel, Tomsen, & Thommeny, 1992; Pernanen, 1991).

Using a similar approach, Murzynski and Degelman (1996) conducted an experiment to assess the effects of body movement on judgments of vulnerability to sexual assault by strangers. Based on Grayson and Stein’s (1981) findings, they trained female models to walk mimicking the victim and nonvictim movement. Videotapes of the models then were shown to college students and police officers who consistently identified women displaying the victim movement profiles as more vulnerable and less confident than women displaying the nonvictim movement profile. These two studies suggest that vulnerability for personal victimization may be manifest and observable in an individual’s posture and gait. Thus, distinct patterns of nonverbal behavior may provide specific signals during social interactions with a potential perpetrator that could increase a woman’s vulnerability for sexual victimization. This vulnerability could occur when would-be perpetrators interpret a victim’s nonverbal behavior as either a sign of sexual interest or as a sign of weakness (Sakaguchi & Hasegawa, 2006).

Alcohol Use, CSA history, Risk Perception, and Nonverbal Behavior

In summary, researchers have found evidence to suggest that alcohol intoxication may increase women’s vulnerability to sexual victimization by reducing their ability to perceive cues indicative of risk prior to the assault. Furthermore, women with a history of CSA are at increased risk for revictimization in adulthood. A number of factors have been examined as possible mediating factors between CSA and revictimization, including alcohol consumption. Gender differences in nonverbal behavior during social interactions suggest that misperceptions between men and women also may contribute to women’s risk for victimization. Men and women display and interpret nonverbal cues differently, with certain nonverbal behaviors potentially increasing women’s risks for sexual victimization. These risks are amplified by alcohol use. Although research has identified independent links among alcohol use, CSA history, risk perception, and nonverbal behavior, little is known about their complex interrelationships. These findings should not be construed as placing blame on women for their victimization, but rather should be seen as potential areas through which we can develop strategies to empower women to better protect themselves against would-be perpetrators who choose to exploit what should, in an ideal world, be safe, nonrisky behavior.

The Current Study

As a framework for conceptualizing the complex interrelationships among alcohol consumption, CSA history, risk perception, and nonverbal behavior, we developed a preliminary conceptual model of risk for sexual victimization and revictimization (Figure 1). The solid lines in the model represent relationships that we are able to test in the current study, while dashed lines are relationships that will need to be tested in subsequent studies. In the current study, we tested the main effects of alcohol and CSA on nonverbal behavior and risk perception, as well as the moderating effect of CSA on the relationships between alcohol consumption and nonverbal behavior and alcohol consumption and risk perception. In the model, we suggest several mediated relationships as well; however, these proposed relationships will need to be tested using longitudinal data from future studies. For example, we propose nonverbal behavior as a mediator between history of CSA and risk for revictimization. We should note that not all variables that can impact victimization (e.g., number of sexual partners) are represented in our model nor are they tested in the current exploratory study.

Fig. 1.

Fig. 1

Open in a new tab

Conceptual model of relationships among childhood sexual abuse history, alcohol consumption, nonverbal behavior, risk perception, and risk for sexual victimization or revictimization. Paths represented by dashed lines are not tested in the current analyses.

Our model accommodates the fact that not all women who experience victimization as an adult have a history of CSA. It further addresses the role that alcohol can play in increasing all women’s risk for victimization (Testa & Parks, 1996). Alcohol’s impact on risk for victimization is not direct, but rather indirect, through its pharmacological effects on nonverbal behavior and risk perception. Outward changes in nonverbal social behavior could convey vulnerability to a potential perpetrator, thereby making women more susceptible to sexual assault. Changes in risk perception could prevent women from recognizing a dangerous situation and removing herself prior to a sexual assault.

Our study used a unique paradigm whereby female participants were paired with trained male confederates for a 1.5-hour social interaction in a bar laboratory and provided with the equivalent of either one (.02 blood alcohol level [BAL], low-dose condition) or four (.08 BAL, high-dose condition) standard drinks. Preliminary findings from this study are reported in an earlier publication that summarizes four presentations from a symposium held at the annual meeting of the Research Society on Alcoholism (see Abbey et al., 2002). The current article represents an in-depth analysis of the observed changes in nonverbal behavior of the female participants, whereas the earlier publication focused on the subjective ratings from participants and confederates on their perceptions of the participants’ changes in behavior during the social interactions.

The alcohol conditions were chosen to best represent a realistic evening out in a bar. Based on our previous studies of women bar drinkers, the majority of women (99.5%) report consuming one or more drinks per occasion in a bar, with an average of 6.2 (SD = 2.8) drinks over a 4-hour period (Parks, Miller, Collins, & Zetes-Zanatta, 1998; Parks & Zetes-Zanatta, 1999). Previous studies of social perceptions under alcohol conditions have used auditory, written, and videotaped scenarios of social interactions between a man and a woman (e.g., Abbey, Buck, Zawacki, & Saenz, 2003; Marx & Gross, 1995; Soler-Baillo, Marx, & Sloan, 2005; Testa, Livingston, & Collins, 2000; Testa, VanZile-Tamsen, Livingston, & Buddie, 2006). These scenarios have varied alcohol, risk, and sexual assault cues and asked female participants to imagine themselves as the woman depicted in the story. The current paradigm was unique because rather than envisioning a hypothetical scenario, participants were immersed in a social interaction with a man they had just met, whom they believed to be a participant in the study, under varying alcohol conditions. The man (i.e., the study confederate) was trained to engage in several discrete behaviors (i.e., probe behaviors) that could be perceived positively (e.g., as flirtatious) if the woman was interested in him romantically or negatively (e.g., sexually forward or intrusive) if she was not interested in him. We are aware of only one other study that has used a complementary paradigm that involved female confederates and male participants to assess sexual assault perpetration history and alcohol consumption influences on recognition of female mixed cues of sexual interest (Abbey, Zawacki, & Buck, 2005).

Hypotheses

Based on our model, we hypothesized that differences in nonverbal behavior would be observable among women depending on alcohol dose and history of CSA. More specifically, we hypothesized that alcohol would have a disinhibiting effect on nonverbal social behavior. For example, we expected women in the high-dose alcohol condition to be more behaviorally engaged in the interaction and more receptive toward the man they had just met than women in the low-dose alcohol condition. Thus, we expected that women in the high-dose alcohol condition would exhibit more high-display behaviors (e.g., leaning forward, nodding, gesturing, smiling, etc.) than women in the low-dose condition. In addition, we expected women in the high-dose alcohol condition to have lower risk perception and thus be less likely to notice the probe behaviors (e.g., touching her arm, whispering in her ear) of the male confederates during the social interactions. Given the lack of previous research on changes in nonverbal social behavior based on victimization history, we made tentative hypotheses about the type of differences in nonverbal social behaviors we anticipated between women with and without CSA histories. We anticipated that women with CSA histories would be less engaged or more hesitant during social interactions with a man they just met. Thus, we expected that women with histories of CSA would exhibit more closed body position, lean away, smile and gesture less, and generally exhibit fewer signs of interest in the interaction than women without a history of CSA. In addition, we anticipated that reduced ability to perceive risk for potential assault would manifest as decreased attention to probe behaviors, and thus, CSA-positive participants would be less likely to report noticing probe behaviors. Tentative hypotheses about the interaction between the alcohol condition and CSA were developed. We hypothesized that the effects of alcohol in the high-dose condition would suppress the inhibiting effects of CSA history. Thus, we would expect that the behavior of CSA-positive women in the high-dose alcohol condition would be most like women without a CSA history in the high-dose condition. We further hypothesized that CSA women in the high-dose alcohol condition would be least likely to recognize the probe behaviors.

METHOD

Participants

Selection criteria

Women were recruited for participation in the study through paid advertisements in the primary local, entertainment, and college newspapers in a large city in Western New York State, as well as through posters hung in the women’s restrooms of several popular bars. Advertisements read: “Women who drink in bars needed for a study of alcohol use and social interactions. Must be at least 21 years old to participate. All participants paid.” The majority of women called the study after seeing a newspaper advertisement (90.8%), and a small number of women learned about the study through word of mouth (6.1%) or posters (2.6%). Women were eligible to participate in the study if they were between the ages of 21 and 35 years, reported drinking in bars regularly (i.e., at least once each month over the past year), and were not currently in a romantic relationship lasting more than 6 months. A number of additional criteria were required for eligibility given that the experimental protocol required random assignment to either a low- or high-dose alcohol condition. Women were screened to determine that they were not currently under the care of a physician for any physical or mental condition that required medication or was contraindicative for alcohol use. Women were ineligible if they reported three or more current alcohol-related problems on the short form of the Michigan Alcoholism Screening Test (Selzer, Vinokur, & Rooijen, 1975) or endorsed one or more diagnostic criteria for alcohol or drug abuse as outlined in the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (American Psychiatric Association, 1994). In addition, they were ineligible if they had ever experienced negative physical or psychological consequences as a result of drinking (i.e., blackouts, fainting, flushing). They were screened to ensure that they had consumed a dose of alcohol in the previous month equivalent to the highest experimental dose (i.e., four standard drinks) that they could be given during the study. Thus, our sample consisted of nonproblematic, moderate drinkers. Our final criterion was based on height and weight. Women were required to be within 10 pounds of the ideal weight for their height in order to maximize the likelihood that they would reach the targeted BAL for the experimental protocol (Breslin & Sobell, 1999).

A total of 423 women called the project. Women were provided with a description of the study and were asked whether they would be interested in participating. Thirty (7.1%) women declined to participate. Of the 393 women who were screened for eligibility, 121 (30.8%) were eligible. The most common reasons for ineligibility were: being in a committed relationship for more than 6 months (37.1%), being overweight (27.2%), and reporting problems with alcohol or drugs (24.3%). Of those eligible women, 11 (9.1%) women were used to pilot the procedures for the study and 46 (38.0%) women completed the actual study protocol. Of the remaining 64 eligible women, one tested positive for pregnancy and was not administered alcohol. The other women either did not attend a scheduled session or were canceled because the other woman needed for the procedure did not attend.1

Sample characteristics

Women were randomly assigned to either the low (n = 22) or high (n = 24) BAL condition. Data from one woman assigned to the high-BAL condition were not included in our analyses due to poor videotaping quality during the social interaction. An additional 3 participants (2 from the high- and 1 from the low-BAL condition) were removed from data analyses because they reported suspicion on their final feedback questionnaire about the male confederate’s authenticity as an actual participant in the study. Therefore, we present data from 42 participants, 21 in each alcohol condition. The average age was 23.2 years (SD = 2.0). All of the women were single2, and most were European American (85.7%). Approximately half of the women (45.2%) were college graduates, with most (92.8%) having taken some college classes. Median annual household income was between $20,000 and $30,000, and average individual income was $15,600.

Participants reported regular, moderate to heavy episodic drinking in bars. The majority (64.3%) reported drinking in bars two to three times each week with average consumption of 5.1 (SD = 2.0) drinks per occasion. Most (90.5%) reported feeling moderately or slightly intoxicated after consuming this amount of alcohol.

Procedure

Women were told that the purpose of the study was to assess women’s perceptions and behavior in bars. They were told that their participation would include attending a cocktail hour at the Research Institute during which they would be asked to interact with men and women they did not know while consuming the equivalent of either one or four mixed, alcoholic drinks. Women participated in the protocol in pairs and were paid $10 per hour. A female research assistant acted as the bartender. The male confederates3 arrived at the Research Institute at the same time as the female participants and were introduced as participants in the study. Although some of the social interaction involved the 4 individuals (both the confederates and the participants), most of the interactions occurred in pairs, with each male confederate interacting one on one with the same female participant for the duration of a given session.4

Upon initially entering the bar laboratory, the women gave informed consent. The participants were taken to a separate room and weighed to determine the amount of alcohol required to achieve the desired BAL (.02 or .08). In addition, women were required to take a urine-based pregnancy test in private to ensure that they were not pregnant prior to alcohol administration. Upon returning to the bar laboratory women were allowed to choose their seat at the bar with one of the male confederates. Participants and confederates were told that their beverage conditions were not necessarily the same and we asked them not to share their breathalyzer readings during the test phase of the protocol. The participants were blind to their own target BAL prior to being breathalyzed during the final 20-minute test phase of the laboratory protocol. The male confederates were provided with nonalcoholic beverages and were blind to the target BAL assigned to the women throughout the protocol. The social interactions lasted approximately 1.5 hours. At the end of the alcohol administration and social interactions in the bar laboratory, the women and men were taken to separate recovery rooms. Women in the high-dose condition were provided with food and completed the final questionnaires upon reaching a BAL of .04. Women in the low-dose condition began their questionnaires immediately. All women were provided with a ride home via taxi paid for by the project. Women were not debriefed or released until they had returned to a BAL of .02.

Alcohol administration

Women were given the choice of three alcoholic beverages (rum, vodka, or gin) and a carbonated mixer of their choice (e.g., Coke, ginger ale, etc.). Alcoholic beverages were mixed after the women had been weighed, but before they returned to the bar laboratory after pregnancy testing. Thus, women did not see the actual mixing of their alcoholic beverages. Women in the low-dose condition received 0.77 ml alcohol/kg body weight of 40 proof alcohol mixed at a ratio of 1 ml alcohol to 12.5 ml of carbonated beverage to achieve a target BAL of .02. Women in the high-dose condition received 2.3 ml alcohol/kg body weight of 40-proof alcohol mixed at a ratio of 1 ml alcohol to 3.5 ml of carbonated beverage to achieve a target BAL of .08. For both alcohol conditions the beverage was divided into four or five glasses for consumption over 20 minutes. All participants consumed the required beverage for their alcohol condition.

Social interactions

The 1.5 hours of social interaction comprised four phases: a neutral initial getting-to-know-you phase, drinking phase, absorption period, and test phase. The first phase consisted of a 10-minute conversation on a neutral topic of the participants’ choosing. Each pair chose from a list of possible topics that included favorite foods, restaurants, movies, or leisure activities. Both participants and confederates then completed a brief questionnaire about initial perceptions of the person with whom they interacted.

Participants and confederates continued to interact while seated at the bar and consuming their beverages during the 20-minute drinking phase. Following alcohol consumption, the two participants and confederates engaged in a 20-minute group activity (e.g., darts, cards) to allow for alcohol absorption. This activity occurred in the barroom, but away from the serving bar. This was the only point during which the four individuals were encouraged to interact as a group.

After the alcohol absorption period, participants and confederates returned to the serving bar as pairs. Each individual was breathalyzed and asked to sit facing their original partner for a final 20 minutes. This part of the social interaction was considered the test phase, during which the male confederates enacted the five prescripted probe behaviors toward the female participant with whom he was interacting. The bartender used subtle manipulations of the candles on the bar (at 5 and 15 minutes) and a second breath analysis at 10 minutes to guide the confederates’ timing for the probe behaviors. The entire social interaction was videotaped and nonverbal behaviors that took place during the drinking and test phases of the interaction were later coded using a behavioral coding system designed for this study.

Probe behaviors

The male confederates were trained to engage in five overt, sexually suggestive probe behaviors: complimenting her appearance, moving closer, touching her arm, touching her hair, and whispering in her ear. These behaviors occurred in the order presented above, with each probe designed to be more intrusive than the previous probe. These behaviors were noted by women in previous studies as potentially aggressive sexual or physical behaviors by men (Cue, George, & Norris, 1996). It should be noted that all of these behaviors can be perceived positively (e.g., as flirtatious or as romantic/sexual interest) by a woman if she is attracted to a man.

The deception

During the laboratory procedure, the female participants were told that the men were participants in the study who would be interacting with them in the bar while everyone (2 female participants and 2 male confederates) consumed a sufficient amount of alcohol to achieve the target BAL (.02 or .08). These deceptive measures were necessary in order to create the illusion of social interaction and drinking in a naturalistic bar setting. Female participants were led to believe that the men also were consuming alcohol, even though they were not. The behavioral probes were designed to appear to be genuine, unscripted behaviors initiated by a male participant. The probes were intended to signal to the woman that the male confederate was attracted to her, regardless of whether this was a welcome attraction or not. The probes simulated common behaviors that previous study participants had described as occurring between men and women in bars (Parks et al., 1998).

The debriefing

Following the social interactions, completion of the questionnaires, and return to a BAL of .02 for high-dose participants, all participants were individually debriefed about the study. The debriefing included a more detailed description of the purpose of the study and an explanation that the men with whom they had interacted were not participants but project staff. We explained that the men had been trained to engage in five specific, common, flirtatious/forward behaviors during the final 20 minutes of the social interaction. We explained that these behaviors can indicate that the interaction or person is risky and that, by learning more about how women respond to and perceive these behaviors, particularly under different alcohol conditions, we hope to be able to better understand and prevent situations that put women at risk for aggression in bars. We asked women whether they were upset or embarrassed and explained that most women did not suspect that the men were project staff. We gave women an opportunity to express concerns about the study and to ask any questions. Women then completed a feedback questionnaire. None of the participants showed signs of embarrassment or distress during the paradigm or debriefing.

Measures

Questionnaires

All female participants and male confederates completed initial and final perception measures about the social interactions. In addition, the women completed a series of questionnaires about demographic characteristics; alcohol and other drug use histories; childhood, adolescent, and adult physical and sexual abuse/assault histories; and several individual characteristic measures (e.g., physical health, medication). Within this large questionnaire, CSA was measured with a 5-item measure that assessed experiences ranging from sexual exposure to vaginal penetration and intercourse prior to age 14 (Miller et al., 1993). Participants were asked to indicate the individual with whom the experience had occurred. Any nonconsensual experiences, as well as experiences with a boyfriend, relative, parent, friend, or other adult more than 5 years older were considered CSA. This method of assessing CSA is widely used and follows the convention established by Finkelhor and Browne (1985). Participants were coded dichotomously as either positive or negative for a history of CSA.

Initial perception measure

Participants and confederates were asked to fill out a brief measure of their initial perceptions of the other individual immediately following the first 10-minute getting-to-know-you interaction. Individuals were asked to rate their comfort, enjoyment, and nervousness/anxiousness during the interaction, as well as their interest in getting to know the other person better on a 1 (not at all) to 7 (very) scale. Due to the high correlations among these items, we summed them to create a composite “enthusiasm” score for the participant’s initial interaction with the male confederate.

Final perception measure

Following the test phase, women were taken to a separate room and asked to complete a measure of their final perceptions of the male confederate with whom they had interacted. The male confederates filled out a similar measure of their impression of the female participant. The first several items were the same as the initial perception measure and asked the participant about her comfort, enjoyment, nervousness/anxiousness, and interest during the interaction. These items were again summed to form a general enthusiasm score for interactions with the confederate. The next two items asked the women how much the confederate’s behavior changed after they had both been drinking and whether he had engaged in any behaviors that had made her feel nervous, anxious, or uncomfortable while in the bar setting. These six items were measured on a 7-point scale (1 = not at all to 7 = very). Participants then were asked to indicate the relationship potential (i.e., go on a date, start a relationship, become intimate, get serious) with the male confederate on a 5-point scale (1 = very unlikely to 5 = very likely). A series of items assessed whether the participant had noticed the five probe behaviors and how flattered or insulted she felt by the behaviors.

Behavioral coding system

The Bar Observation Social Interaction Measure (BOSIM) was developed for coding social behavior during naturalistic observations in bars (Parks & Kennedy, 2001; Abbey et al., 2002). It includes ratings of a target subject’s gross motor movements (standing, sitting, walking), body posture and orientation (forward lean, open or closed), eye gaze, facial expressions (smile, frown), and other discrete behaviors commonly observed during social interactions (self grooming/touching, talking, laughing, face covering). Independent observers were asked to rate the frequency with which the target participant engaged in each of the discrete behaviors during the social interaction on a rating scale from 1 (none of the time) to 5 (continuous, more than 75% of the time).

Data Analyses

The videotaped 20-minute drinking and 20-minute test phases for all participants were coded by one of two research technicians who were trained in the use of the BOSIM. These research technicians had not participated in the laboratory protocol and were blind to the alcohol condition of the participants being observed. Both research technicians independently coded 16% of the same drinking and test phases to establish inter-rater reliability. Percent agreement was calculated for all discrete behavioral codes for both the drinking and test phases of the social interactions separately (Uebersax, 2007). Behaviors for which inter-rater reliability was less than 50% were considered unreliable and were not included in further analyses. Reliability was less than 50% for only one discrete behavior, laughter (33.3%), coded during the drinking phase. Inter-rater reliability ranged from 55.6% to 100% for coding of the remaining discrete behaviors during the drinking phase, with an overall reliability of 91.0%. Eye contact (44.4%) and laughter (33.3%) were the only two discrete behaviors with reliability less than 50% coded during the test phase. Inter-rater reliability ranged from 55.6% to 100% for coding of the remaining discrete behaviors during the test phase, with an overall reliability of 80.0%. Coding discrepancies for the 16% of interactions that were mutually coded were resolved through observation of the video and discussion by both coders with the first author.

RESULTS

Prior to all analyses, the behavioral data were checked for inconsistencies and outliers. No inconsistencies or outliers were found that would have affected the analyses. Sixteen women were categorized as positive for a history of CSA, and 29 women were categorized as negative for a history of CSA, based on responses to the CSA questionnaire measure. This measure was administered after the alcohol consumption portion of the study was completed, and CSA history had no bearing on assignment to alcohol condition. The number of CSA-positive women in the low-dose (n = 10) and high-dose (n = 6) alcohol conditions did not differ significantly, X2(1) = 1.84, ns.

Manipulation Check

Peak BAL for women in the low-dose alcohol condition was 0.012 (SD = 0.007) and was reached near the beginning of the test phase of the social interaction. Peak BAL for women in the high-dose condition was 0.085 (SD = 0.014) and was reached during the later part of the test phase. Women in the high-dose condition (M = 2.52, SD = 0.60) rated themselves as significantly more intoxicated on a 1 to 4 scale than women in the low-dose condition (M = 1.33, SD = 0.48), F(1, 41) = 50.00, p < .001, partial η2 = .56.

Power Analysis

We conducted a power analysis to determine our ability to detect main and interaction effects for alcohol condition and CSA history. For the 2 × 3 fixed-effects analyses of variance, the sample size provided roughly 10 cases per cell (in the interaction conditions, as few as 6). We estimated power setting alpha = .05. Our power to detect a main effect of alcohol condition was 70% and power to detect a main effect of CSA history was 60–70%, while power to detect the interaction of CSA and alcohol condition was 45–70% for a large effect size (f = .40).

Initial Perception of the Man Being Met in the Bar

The two male confederates, Mark and Tom (pseudonyms), were both Caucasian and in their mid 20s. Women’s positive rating, or enthusiasm (i.e., comfort, enjoyment, interest, and lack of nervousness), for the confederates after the initial getting-to-know-you phase was generally high for both men; however, ratings were significantly higher from women who interacted with Mark (M = 6.00, SD = 0.60) compared with ratings from women who interacted with Tom (M = 5.15, SD = 0.83), F(1, 41) = 14.25, p < .01, partial η2 = .26. Given this difference, enthusiasm ratings were included as covariates in all analyses of the women’s behavior during the drinking and test phases. There were no differences in the rating of enthusiasm for the confederates during the initial phase based on history of CSA. Women had not yet consumed any alcohol at the time of the initial getting-to-know-you phase of the social interaction, and were not aware of their alcohol assignment (.02 or .08 BAL); therefore, no analyses based on alcohol condition were conducted.

Behavior During Drinking Phase

There were no significant effects of alcohol on behavior during the drinking phase. However, women with a history of CSA moved their heads less frequently, were less animated, and gestured somewhat less than women without a history of CSA (Table 1). There were no significant main effects of the covariate (enthusiasm for the confederate). There was a significant interaction for body lean, and the interaction for face covering approached significance. Women with a history of CSA in the high-dose alcohol condition leaned forward and closer, while women without a history of CSA in the low-dose alcohol condition sat more upright, F(1, 37)=4.36, p < .05, partial η2 =.11. Face covering was generally an infrequent behavior, but women with a history of CSA had the lowest rates of face covering in the low-dose alcohol condition and the highest rates of face covering in the high-dose alcohol condition, F(1, 37) = 3.22, p = .08, partial η2 = .08.

Table 1.

Differences in Behavior During Drinking Phase by History of Child Sexual Abuse (CSA)

CSA History
No CSA
CSA
Behavior M SE M SE F partial η2
Body orientation 4.47 0.18 4.16 0.24 1.10 .03
Body lean 3.72 0.11 3.67 0.15 0.10 .00
Receptivity 4.51 0.13 4.43 0.17 0.73 .00
Face pleasant 4.65 0.11 4.38 0.15 2.03 .05
Smiling 3.21 0.11 2.99 0.14 1.54 .04
Frowning 1.21 0.10 1.31 0.13 0.39 .01
Head movement 2.93 0.09 2.38 0.13 12.17** .25
Covering face 1.85 0.11 1.77 0.14 0.21 .01
Talking 3.12 0.07 3.01 0.09 1.08 .03
Stand/walk 1.03 0.02 1.02 0.03 0.02 .00
Loudness 1.14 0.05 1.03 0.07 1.72 .05
Animated 3.90 0.10 3.53 0.13 4.62* .11
Gesturing 2.98 0.09 2.70 0.12 3.63† .09
Fidgeting 2.27 0.10 2.18 0.13 0.33 .01
Self grooming 1.84 0.53 1.86 0.17 0.01 .00
Open in a new tab
*

p < .05.

**

p < .01.

†

p < .08.

Behavior During Test Phase

During the test phase of the social interaction, during which the participants achieved peak BAL and the confederates engaged in the planned probe behaviors, there were a number of differences in behavior based on alcohol condition (Table 2). Women in the high-dose alcohol condition exhibited more open or receptive body postures, and they talked, stood, and walked more when interacting with the male confederate than did women in the low-dose alcohol condition. In addition, they frowned more than the women in the low-dose alcohol condition. There also was a trend toward women in the high-dose alcohol condition being louder than women in the low-dose alcohol condition. There were no significant findings based on history of CSA. There were two behaviors that had significant effects for the covariate, enthusiasm for the confederate: body orientation, F(1, 37)=4.77, p < .05, partial η2 =.12, and gesturing, F(1, 37)=6.61, p <.05, partial η2 = .15. The only significant interaction was for frowning, F(1, 44) = 4.47, p < .05, partial η2 = .12. Women with a history of CSA in the high-dose alcohol condition had the highest rate of frowning. Frowning was generally a very low base rate behavior so it was notable when it occurred.

Table 2.

Differences in Behavior During Test Phase by Alcohol Condition

Alcohol Condition
Low dose
High dose
Behavior M SE M SE F partial η2
Body orientation 3.86 0.24 4.20 0.27 0.89 .02
Body lean 3.47 0.20 3.49 0.22 0.01 .00
Receptivity 3.94 0.19 4.65 0.21 6.24* .14
Face pleasant 4.72 0.08 4.85 0.09 1.19 .03
Smiling 3.24 0.11 3.26 0.13 0.01 .00
Frowning 1.17 0.07 1.37 0.08 4.12* .10
Head movement 2.75 0.08 2.65 0.09 0.83 .02
Covering face 1.97 0.14 1.84 0.16 0.40 .01
Talking 3.14 0.09 3.43 0.10 4.94* .12
Stand/walk 1.00 0.06 1.31 0.06 12.96** .26
Loudness 1.14 0.09 1.38 0.10 3.46† .09
Animated 3.88 0.12 4.11 0.13 1.71 .04
Gesturing 2.92 0.09 3.06 0.10 1.22 .03
Fidgeting 2.59 0.15 2.67 0.16 0.12 .00
Self grooming 2.30 0.15 2.32 0.16 0.01 .00
Open in a new tab
*

p < .05.

**

p < .01.

†

p < .08.

Final Perception of the Confederate and Probe Behaviors

Although women reported positive social interactions with both Mark and Tom, significant differences in ratings for the two confederates occurred on the final perception measure. As reported in their initial perceptions, women who interacted with Mark rated their enthusiasm (M = 6.29, SD = 0.41) for the social interactions higher than women who interacted with Tom (M = 5.40, SD = 0.96), F(1, 41)=15.01, p < .01, partial η2 = .27. Also, women who were paired with Tom (M = 3.90, SD = 0.89) were significantly more likely to notice the probe behaviors than women who were paired with Mark (M = 2.90, SD = 1.34), F(1, 37) 5.08, p < .05, partial η2 = .17. We further assessed the interactions between confederate identity and alcohol group, as well as confederate identity and CSA history, while including enthusiasm for the social interaction with the confederate as a covariate. The interaction between alcohol group and confederate identity approached significance, F(1, 37) 3.91, p < .06, partial η2 = .10. Women who interacted with Tom noticed the probes at nearly the same rates in both the high-and low-dose alcohol conditions (M = 3.84, SE = 0.36 and M = 3.87, SE = 0.33, respectively), while the women who interacted with Mark were much less likely to notice the probes when they were in the high-dose compared with the low-dose alcohol condition (M = 2.32, SE = 0.34 and M = 3.65, SE = 0.35, respectively). The interaction for CSA history and confederate identity was not significant. In addition, there were no overall differences in enthusiasm for the social interactions between the women and the confederates based on CSA or alcohol condition.

Given the differences in ratings for the two confederates on the final perception measure, the enthusiasm rating was included as a covariate in all analyses of the women’s perceptions of the probe behaviors. Women in the high-dose alcohol condition were less likely to notice the probe behaviors (M = 2.80, SE = 0.27) than women in the low-dose condition (M = 3.76, SE = 0.25), F(1, 37) = 6.87, p < .05, partial η2 = .16. The difference in how women felt about the probe behaviors approached significance. Women in the high-dose alcohol condition reported being more flattered by the probe behaviors (M = 3.72, SE = 0.15) than women in the low-dose condition (M = 3.29, SE = 0.17), F(1, 37) = 3.69, p = .06, partial η2 = .09. The effect of CSA history on whether participants noticed the probe behaviors approached significance. Women with a history of CSA were less likely to notice the probe behaviors (M = 2.92, SE = 0.30) than women without a history of CSA (M = 3.65, SE = 0.23), F(1, 37) = 3.66, p = .06, partial η2 = .09. There was no interaction between CSA and alcohol condition on whether women noticed the probe behaviors. There was no difference in how women felt about the probe behaviors based on CSA history and no interaction between CSA and alcohol condition.

DISCUSSION

Our study provides the first known attempt to examine observable, discrete differences in women’s social behaviors during heterosexual interactions as a function of alcohol consumption under controlled administration conditions in a bar laboratory. Furthermore, we are unaware of any other studies that have attempted to detect differences in discrete social behaviors as a function of CSA history. Thus, our study provides a preliminary assessment of a unique paradigm for evaluating the effects of alcohol and CSA history on nonverbal social behavior. We assessed these relationships as a partial test of a conceptual model of the role of CSA history, alcohol consumption, risk perception, and nonverbal social behavior on women’s risk for adult sexual (re)victimization.

Influence of Alcohol on Behavior

After drinking, during the test phase of the social interaction, differences were found in behavior as a function of alcohol condition. At this point in the paradigm, women were beginning to reach peak BAL (.01 or .08). At a BAL less than .02, women would be unlikely to experience any behavioral effects of alcohol except perhaps an increase in talkativeness and an increase in feelings of well-being; however, at a BAL of .08, women would be likely to be experiencing increased talkativeness as well as slurred speech, increased boldness, decreased sensory perception, and some loss of coordination (Walls & Brownlie, 1985, p. 42). The constellation of changes in behavior we observed (open body position, increased talking and movement) in the women who received the high dose is consistent with greater openness and comfort, decreased inhibition, and increased animation. In a social drinking setting, these behaviors could be interpreted as signs of intoxication. However, the combined effect of behaviors that convey intimacy or welcoming (e.g., open body posture; Mehrabian, 1969) with behaviors that may convey interest in the interaction (e.g., talking) also could be (mis)interpreted as romantic or sexual interest by a potential romantic partner, particularly in a drinking context.

The differences we observed between the high- and low-dose alcohol conditions provide tentative support to our original hypotheses; in the high-dose condition we saw greater behavioral change and women appeared to be somewhat more disinhibited (i.e., talked more, stood more) and more receptive (open body position) toward the male confederate. However, we did not see an increase in the high-display behaviors (e.g., leaning forward, smiling, gesturing) as we originally hypothesized and, counter to our hypotheses, women in the high-dose condition frowned more than women in the low-dose alcohol condition. Frowning is a difficult behavior to interpret. It can have multiple meanings, including discomfort or disagreement during a social interaction, or it can occur as part of a constellation of behaviors within a more animated, expressive conversation (i.e., varying extremes of expression from smiling to frowning). Thus, more complex analyses that assess patterns of nonverbal behaviors that occur during social interactions, as well as additional, larger studies are needed to clearly determine the meaning of these types of behavioral changes in social contexts under varying levels of alcohol consumption.

We originally hypothesized that women in the high-dose alcohol condition would be less likely to notice the probe behaviors than women in the low-dose alcohol condition. Our findings support this hypothesis and suggest that women’s cognitive ability to perceive risk cues and possibly avoid dangerous situations is likely to be impaired with greater alcohol consumption. The paradigm we designed for the current study created a situation in which women were exposed to positive or disinhibiting cues, as well as conflicting or inhibiting cues (i.e., inhibitory conflict). Previous research indicates that a primary reason women report for drinking in bars is to meet members of the opposite sex for romantic purposes (Parks et al., 1998). In addition, both women and men indicate that women who drink in bars appear more sexually available and are at higher risk for victimization (Parks et al., 1998; Parks & Scheidt, 2000). Thus, when drinking in a bar and meeting an attractive man, a woman is presented with a cue (attractive man/relationship potential) that is most likely socially disinhibiting, yet her knowledge of dating risks should make her somewhat cautious of the situation and thus be socially inhibiting. The result is inhibitory conflict. Alcohol myopia (Steele & Josephs, 1990; Steele & Southwick, 1985) suggests that, under conditions of inhibitory conflict, alcohol causes an individual to focus on the most salient cues (e.g., attractiveness, relationship potential). Indeed, women in the high-dose alcohol condition were less likely to notice the probes when they were paired with Mark, the confederate whom they rated with more enthusiasm.

Influence of CSA History on Behavior

We hypothesized that women with a history of CSA would be less likely to notice the probe behaviors than women without a history of CSA. Our results are consistent with this hypothesis and provide tentative support to the idea that women with a CSA history may be less likely to perceive risk cues than women without a CSA history. This finding provides some support for Cloitre’s (1998) theory that a history of CSA causes a disruption in the development of appropriate cognitive abilities needed to recognize and avoid dangerous situations (i.e., appropriate fight-or-flight response).

We found differences in women’s social behavior during and after drinking, based on history of CSA. During the drinking phase, but before achieving peak BAL, women with a history of CSA moved their heads less frequently and were less animated than women without a history of CSA. Greater head movements and animation (i.e., laughter, facial pleasantness, engagement, gestures) during social interactions help maintain interpersonal engagement and convey self-confidence and interest in the other person (Moore, 1985; Moore & Butler, 1989). Our findings suggest that a history of CSA may inhibit or subdue women’s responses when interacting with a man for the first time. Classen et al. (2001) found that women with a history of CSA who were more socially avoidant were more likely to be sexually revictimized than those who were not socially avoidant. They further suggest that this might be due to a lack of social skills needed to protect themselves in risky situations or that these women somehow appear vulnerable to would-be perpetrators.

The Combined Effect of Alcohol and CSA

Women with a history of CSA in the high-dose alcohol condition leaned forward and had the highest rates of face covering during the drinking phase. Leaning forward indicates interest in the interaction and face covering can be seen as a coy or flirtatious behavior during interactions with members of the opposite sex. These two behaviors could be interpreted as counter to the pattern of behavior suggesting social avoidance seen for the main effect of CSA history on behavior. However, face covering also may convey a reticent communication style or discomfort during a social interaction (Burgoon & Koper, 1984). It may be that women with a history of CSA were somewhat anxious during the drinking phase of the social interaction with the male confederate compared with non-CSA women; however, the early effects of alcohol consumption in the high-dose condition may have reduced this anxiety creating a somewhat more ambivalent constellation of behaviors. Alternatively, Sayette (1993) proposed an appraisal-disruption model that would suggest women with a history of CSA who consumed the higher dose of alcohol (as compared to those who consumed the low dose of alcohol) failed to appraise the interaction with a just-met male as stressful (i.e., potentially threatening), due to a disruption in their ability to access information in long-term memory as a result of the pharmacological effects of the higher alcohol dose. This has been termed the Stress Response Dampening effect of alcohol and may help explain why women with a history of CSA and a high dose of alcohol are at elevated risk for revictimization.

During the test phase, women with a history of CSA in the high-dose alcohol condition frowned most. Frowning conveys greater anxiety or less comfort during a social interaction. During the test phase of the social interaction, the men engaged in probe behaviors that were designed to be increasingly more intrusive, yet interpretable as positive if a woman found him attractive; one might anticipate that these behaviors could make a woman with a victimization history anxious and uncomfortable. Thus, it appears that during both the drinking and test phases of the social interaction, women with a history of CSA were more likely to exhibit behaviors indicative of social discomfort or anxiety; however, their behavior also included some signs of interest in the interaction, specifically in the early stages of the interactions (i.e., prior to peak BAL) under the high-alcohol condition. It is conceivable that, with increased alcohol consumption, women’s nonverbal behavior could convey unclear or mixed messages about their interest in the interaction, putting women at greater risk for victimization. Such mixed messages may further increase gender differences in perceptions of romantic or sexual interest that occur during heterosexual interactions. These differences appear to be magnified during interactions that involve alcohol (Abbey et al., 1998; Abbey et al., 1994) and thus could increase the likelihood of sexual assault.

Studies with children indicate that manifest anxiety increases risk for victimization (e.g., Egan & Perry, 1998). In a study of undergraduate women, Sacks and Bugental (1987) found that, in a stressful social interaction, women with helpless attribution styles (i.e., potentially similar to a severe CSA victim; Herman, 1992) became more hostile and were less nonverbally pleasant. Thus, women with a history of CSA who are socially avoidant, uncomfortable, or anxious during social interactions, particularly when drinking, may be at heightened risk for revictimization.

Support for the Proposed Model

The observed changes in nonverbal social behavior as a function of both alcohol consumption and CSA history provide some support for the relationships that we were able to test within our proposed model. We found support for alcohol and CSA effects on risk perception, through differences in women’s ability to notice the probe behaviors. We found direct effects of alcohol on nonverbal social behavior. We found direct effects of CSA on some nonverbal social behavior and limited evidence for CSA as a moderator of alcohol’s effect on nonverbal behavior. It is likely that we did not have sufficient power to detect CSA as a moderator of alcohol’s effects on other social behaviors. These preliminary findings suggest that the relationships described within our model are worth pursuing with larger samples and longitudinal designs in future studies that would allow for complete testing of the model.

Limitations & Future Directions

Although this study provided a unique opportunity to examine behavioral differences in heterosexual social interactions in a bar laboratory setting, several limitations should be noted. This was a preliminary study, thus our sample size was small and nonrandomly selected. The small sample size limited our ability to detect small- to medium-size effects. Retrospective power analyses suggest that our power to detect main and interactive effects was between 50 and 75%, suggesting that our Type II error rate was potentially high. Clearly larger studies are needed to replicate and move beyond the findings of this preliminary study.

The women in our sample were predominantly nonminority and well educated. In addition, they were moderate drinkers without alcohol problems. We did not sample the full range of female drinkers or bar drinkers. The sample was further limited to women who were not over-or underweight. According to data from the U.S. National Health and Nutrition Examination survey data for 1999–2002, 52.8% of women between the ages of 20 and 34 are overweight (Centers for Disease Control, n.d.). It should be noted that it is unclear whether our model is appropriate for explaining sexual aggression among same-sex individuals as it is conceptualized based on male-initiated sexual aggression toward a female victim. We assessed women’s perceptions of a man they did not know; other research has found that women’s appraisals are more positive and that intended responses are less resistant to sexually threatening behaviors when they know their assailant (VanZile-Tamsen, Testa, & Livingston, 2005). We assessed women’s behavior in a relatively safe environment, a bar laboratory at a Research Institute, under drinking conditions designed to achieve a maximum BAL of .08. Thus, our findings from the controlled laboratory bar may not reflect women’s behavior under less controlled conditions when interacting with a man they are meeting for the first time and drinking at a bar in a natural setting. We found differences in women’s risk perception based on the confederate’s identity; thus, in future studies it will be important to obtain confederates who are more equivalent in attractiveness to female participants. Finally, we cannot eliminate alcohol expectancies as a potential influence on the differences in behavior seen for the two alcohol conditions. The use of a more complex alcohol administration design would be necessary to parcel out this effect and can be employed in future studies. Given the initial, limited scope of this study, we did not use a balanced placebo design. In addition, it is likely that both expectancy and pharmacological effects of alcohol will influence behavior in real-world drinking situations where women are usually cognizant of the amount of alcohol they have consumed; therefore, we felt it was not critical to distinguish these effects in this initial laboratory study. Future studies should examine the role of alcohol expectancies on behavior with either a balanced placebo design or, at minimum, a placebo condition and a no-alcohol control along with the higher-dose alcohol condition. In addition, future studies are needed using larger sample sizes to further assess differences between women with and without histories of CSA, as well as other variables that might influence nonverbal behavior (e.g., socio-demographics, psychological health, childhood physical abuse, relationship to perpetrator).

Conclusion

More than a decade ago, Richards (1991) suggested that sexual assault prevention efforts should involve social skills training to make women aware of their outward appearance and nonverbal behavior as well as to teach women to consciously control the nonverbal messages they send. Although this approach might be criticized as placing the blame for sexual assault on the victim, we would argue that providing women with the ability to change their own behavior and perceptions is empowering. Women are faced with interactions with men on a daily basis; it is difficult for women to recognize a perpetrator prior to interacting with him and more difficult to influence his thoughts, intentions, and behaviors. We believe that it is important for women to be able to anticipate potentially dangerous situations, deter a potential perpetrator, and reduce their own risk for sexual assault. Women will be better able to do so by sending clear nonverbal and verbal messages indicating that they are not an easy target and by recognizing potentially dangerous or risky situations with sufficient time to extricate themselves and employ effective resistance.

Acknowledgments

This research and manuscript preparation was supported in part by National Institute on Alcohol Abuse and Alcoholism (NIAAA) grant K01-AA00233 awarded to Kathleen Parks, NIAAA grant K01-AA016105 awarded to Amy Hequembourg, and NIAAA grant K01-AA014685 awarded to Ronda Dearing.

We thank Dr. Maria Testa for her helpful comments on the manuscript.

Footnotes

1

Due to the high no-show rate for the study, we began to send taxis to pick the women up for their scheduled sessions as well as take them home. This improved our completion rates somewhat. Often women were unable or unwilling to reschedule after cancellation of a session.

2

Women were not asked directly about their sexual orientation, however, in our current study of women bar drinkers from this same age group a small percentage (2.7%) self-identify as lesbian. Extrapolation to the sample presented in this manuscript would suggest that possibly one woman in our sample was a lesbian.

3

Male confederates were chosen for the position based on several criteria. We selected males who were moderately attractive, were in their mid to late 20s, were articulate, and understood the ethics of psychological research. In choosing men based on these criteria we hoped the confederates would appeal to, or be least likely to be unappealing to, the majority of women who met criteria for our study.

4

Although it might appear more naturalistic for women to interact with multiple men in a bar, we had participants remain paired with the same confederate throughout the entire experimental paradigm in an effort to control for any variability in their behavior that may have resulted from differences between confederates.

References

  1. Abbey A, Buck PO, Zawacki T, Saenz C. Alcohol’s effects on perceptions of a potential date rape. Journal of Studies on Alcohol. 2003;64:669–677. doi: 10.15288/jsa.2003.64.669. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Abbey A, McAuslan P, Thomson Ross L. Sexual assault perpetration by college men: The role of alcohol, misperception of sexual intent, and sexual beliefs and experiences. Journal of Social and Clinical Psychology. 1998;17:167–195. [Google Scholar]
  3. Abbey A, Thomson Ross L, McDuffie D. Alcohol’s role in sexual assault. In: Watson R, editor. Addictive behaviors in women. Totawa, NJ: Humana Press; 1994. pp. 97–123. [Google Scholar]
  4. Abbey A, Zawacki T, Buck PO. The effects of past sexual assault perpetration and alcohol consumption on men’s reactions to women’s mixed signals. Journal of Social & Clinical Psychology. 2005;24:129–155. doi: 10.1521/jscp.24.2.129.62273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Abbey A, Zawacki T, Buck P, Clinton AM, McAuslan P. Sexual assault and alcohol consumption: What do we know about their relationship and what types of research are still needed? Aggression and Violent Behavior. 2004;9:271–303. doi: 10.1016/S1359-1789(03)00011-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Abbey A, Zawacki T, Buck PO, Testa M, Parks K, Norris J, et al. How does alcohol contribute to sexual assault? Explanations from laboratory and survey data. Alcoholism: Clinical & Experimental Research. 2002;26:575–581. [PMC free article] [PubMed] [Google Scholar]
  7. Abbey A, Zawacki T, McAuslan P. Alcohol’s effects on sexual perception. Journal of Studies on Alcohol. 2000;61:688–697. doi: 10.15288/jsa.2000.61.688. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4. Washington, DC: Author; 1994. [Google Scholar]
  9. Breitenbecher KH. Sexual revictimization among women: A review of the literature focusing on empirical investigations. Aggression and Violent Behavior. 2001;6:415–432. [Google Scholar]
  10. Breslin FC, Sobell SL. Alcohol administration methodology 1994–1995: What researchers do and do not report about subjects and dosing procedures. Addictive Behaviors. 1999;24:509–520. doi: 10.1016/s0306-4603(99)00002-7. [DOI] [PubMed] [Google Scholar]
  11. Burgoon JK, Koper RJ. Nonverbal and relational communication associated with reticence. Human Communication Research. 1984;10:601–626. [Google Scholar]
  12. Centers for Disease Control. Third National Health and Nutrition Examination Survey (NHANES III) Public Use Data Files. nd Retrieved April 12, 2007, from http://www.cdc.gov/nchs/about/major/nhanes/datalink.htm.
  13. Classen C, Field NP, Koopman C, Nevill-Manning K, Spiegel D. Interpersonal problems and their relationship to sexual revictimization among women sexually abused in childhood. Journal of Interpersonal Violence. 2001;16:495–509. [Google Scholar]
  14. Classen C, Palesh OG, Aggarwal R. Sexual revictimization: A review of the empirical literature. Trauma, Violence, & Abuse. 2005;6:103–129. doi: 10.1177/1524838005275087. [DOI] [PubMed] [Google Scholar]
  15. Cloitre M. Sexual revictimization: Risk factors and prevention. In: Follette M, Ruzek JI, editors. Cognitive behavioral therapies for trauma. New York: Guilford; 1998. pp. 278–304. [Google Scholar]
  16. Cue KL, George WH, Norris J. Women’s appraisals of sexual assault risk in dating situations. Psychology of Women Quarterly. 1996;20:487–504. [Google Scholar]
  17. DiLillo D. Interpersonal functioning among women reporting a history of childhood sexual abuse: Empirical findings and methodological issues. Clinical Psychology Review. 2001;21:553–576. doi: 10.1016/s0272-7358(99)00072-0. [DOI] [PubMed] [Google Scholar]
  18. Egan SK, Perry DG. Does low self-regard invite victimization? Developmental Psychology. 1998;34:299–309. doi: 10.1037//0012-1649.34.2.299. [DOI] [PubMed] [Google Scholar]
  19. Epstein J, Saunders B, Kilpatrick D, Resnick H. PTSD as a mediator between childhood rape and alcohol use in adult women. Child Abuse & Neglect. 1998;22(1):223–234. doi: 10.1016/s0145-2134(97)00133-6. [DOI] [PubMed] [Google Scholar]
  20. Fichten C, Tagalakis V, Judd D, Wright J, Amsel R. Verbal and nonverbal communication cues in daily conversations and dating. The Journal of Social Psychology. 1992;132:751–769. [Google Scholar]
  21. Finkelhor D. Current information on the scope and nature of child sexual abuse. Future of Children. 1994;4(2):31–53. [PubMed] [Google Scholar]
  22. Finkelhor D, Browne A. The traumatic impact of child sexual abuse: A conceptualization. American Journal of Orthopsychiatry. 1985;55:530–541. doi: 10.1111/j.1939-0025.1985.tb02703.x. [DOI] [PubMed] [Google Scholar]
  23. George WH, Gournic SJ, McAfee MP. Perceptions of postdrinking female sexuality: Effects of gender, beverage choice, and drink payment. Journal of Applied Social Psychology. 1988;18:1295–1317. [Google Scholar]
  24. Gold S, Sinclair B, Balge K. Risk of sexual victimization: A theoretical model. Aggression & Violent Behavior. 1999;4:457–470. [Google Scholar]
  25. Grayson B, Stein M. Attracting assault: Victims’ non-verbal cues. Journal of Communication. 1981;31:68–75. doi: 10.1111/j.1460-2466.1981.tb01206.x. [DOI] [PubMed] [Google Scholar]
  26. Herman J. Trauma and recovery. New York: Basic Books; 1992. [Google Scholar]
  27. Homel R, Tomsen S, Thommeny J. Public drinking and violence: Not just an alcohol problem. Journal of Drug Issues. 1992;22:679–697. [Google Scholar]
  28. Kowalski R. Nonverbal behaviors and perceptions of sexual intentions: Effects of sexual connotativeness, verbal response, and rape outcome. Basic and Applied Social Psychology. 1992;13:427–445. [Google Scholar]
  29. Langeland W, Hartgers C. Child sexual and physical abuse and alcoholism: A review. Journal of Studies on Alcohol. 1998;59:336–348. doi: 10.15288/jsa.1998.59.336. [DOI] [PubMed] [Google Scholar]
  30. Martin Q, Peters R, Amos C, Yacoubian G, Jones Johnson R, Meschack A, et al. The relationship between sexual abuse and drug use: A view of African American college students in Texas. Journal of Ethnicity in Substance Abuse. 2005;4:23–33. doi: 10.1300/J233v04n01_03. [DOI] [PubMed] [Google Scholar]
  31. Marx B, Gross A. Date rape: An analysis of two contextual variables. Behavior Modification. 1995;19:451–463. [Google Scholar]
  32. Mehrabian A. Significance of posture and position in the communication of attitude and status relationships. Psychological Bulletin. 1969;71:359–372. doi: 10.1037/h0027349. [DOI] [PubMed] [Google Scholar]
  33. Messman-Moore TL, Long P. The role of childhood sexual abuse sequelae in the sexual revictimization of women: An empirical review and theoretical reformulation. Clinical Psychology Review. 2003;23:537–571. doi: 10.1016/s0272-7358(02)00203-9. [DOI] [PubMed] [Google Scholar]
  34. Miller B, Downs W, Testa M. Interrelationships between victimization experiences and women’s alcohol use. Journal of Studies on Alcohol. 1993;11:109–117. doi: 10.15288/jsas.1993.s11.109. [DOI] [PubMed] [Google Scholar]
  35. Moore MM. Nonverbal courtship patterns in women: Context and consequences. Ethology and Sociobiology. 1985;6:237–247. [Google Scholar]
  36. Moore MM, Butler DL. Predictive aspects of non-verbal courtship behavior in women. Semiotica. 1989;76:205–215. [Google Scholar]
  37. Murzynski J, Degelman D. Body language of women and judgments of vulnerability to sexual assault. Journal of Applied Social Psychology. 1996;26:1617–1626. [Google Scholar]
  38. Parks KA, Kennedy CL. How accurate are women’s perceptions of changes in their own behavior after drinking?; Paper presented at the meeting of the Research Society on Alcoholism; Montreal, Quebec, Canada. 2001. Jun, [Google Scholar]
  39. Parks KA, Miller BA, Collins RL, Zetes-Zanatta LM. Women’s descriptions of drinking in bars: Reasons and risks. Sex Roles. 1998;38:701–717. [Google Scholar]
  40. Parks KA, Scheidt DM. Male bar drinkers’ perspective on female bar drinkers. Sex Roles. 2000;43:165–179. [Google Scholar]
  41. Parks KA, Zetes-Zanatta LM. Women’s bar-related victimization: Refining and testing a conceptual model. Aggressive Behavior. 1999;25:349–364. [Google Scholar]
  42. Pernanen K. Alcohol in human violence. New York: Guilford; 1991. [Google Scholar]
  43. Rich C, Combs-Lane A, Resnick H, Kilpatrick D. Child sexual abuse and adult sexual revictimization. In: Koenig L, Doll L, O’Leary A, Pequegnat W, editors. From child sexual abuse to adult sexual risk: Trauma, revictimization, and intervention. Washington, DC: American Psychological Association; 2004. pp. 49–68. [Google Scholar]
  44. Richards L. A theoretical analysis of nonverbal communication and victim selection for sexual assault. Clothing and Textiles Research Journal. 1991;9(4):55–64. [Google Scholar]
  45. Roodman A, Clum G. Revictimization rates and method variance: A meta-analysis. Clinical Psychology Review. 2001;21:183–204. doi: 10.1016/s0272-7358(99)00045-8. [DOI] [PubMed] [Google Scholar]
  46. Sacks CH, Bugental DB. Attributions as moderators of affective and behavioral responses to social failure. Journal of Personality and Social Psychology. 1987;53:939–947. [Google Scholar]
  47. Sakaguchi K, Hasegawa T. Person perception through gait information and target choice for sexual advances: Comparison of likely targets of experiments and real life. Journal of Nonverbal Behavior. 2006;30:63–85. [Google Scholar]
  48. Sayette M. An appraisal-disruption model of alcohol’s effects on stress responses in social drinkers. Psychological Bulletin. 1993;114:459–476. doi: 10.1037/0033-2909.114.3.459. [DOI] [PubMed] [Google Scholar]
  49. Selzer ML, Vinokur A, Rooijen L. A self-administered short Michigan alcoholism screening test (SMAST) Journal of Studies on Alcohol. 1975;36:117–126. doi: 10.15288/jsa.1975.36.117. [DOI] [PubMed] [Google Scholar]
  50. Simpson T. Childhoood sexual abuse, PTSD, and the functional roles of alcohol use among women drinkers. Substance Use & Misuse. 2003;38:249–270. doi: 10.1081/ja-120017248. [DOI] [PubMed] [Google Scholar]
  51. Simpson T, Miller W. Concomitance between childhood sexual and physical abuse and substance use problems: A review. Clinical Psychology Review. 2002;22:27–77. doi: 10.1016/s0272-7358(00)00088-x. [DOI] [PubMed] [Google Scholar]
  52. Soler-Baillo J, Marx B, Sloan D. The psychophysiological correlates of risk recognition among victims and non-victims of sexual assault. Behaviour Research and Therapy. 2005;43:169–181. doi: 10.1016/j.brat.2004.01.004. [DOI] [PubMed] [Google Scholar]
  53. Steele CM, Josephs RA. Alcohol myopia: Its prized and dangerous effects. American Psychologist. 1990;45:921–933. doi: 10.1037//0003-066x.45.8.921. [DOI] [PubMed] [Google Scholar]
  54. Steele CM, Southwick L. Alcohol and social behavior: The psychology of drunken excess. Journal of Personality & Social Psychology. 1985;48:18–34. doi: 10.1037//0022-3514.48.1.18. [DOI] [PubMed] [Google Scholar]
  55. Testa M, Livingston J, Collins R. The role of women’s alcohol consumption in evaluation of vulnerability to sexual aggression. Experimental and Clinical Psychopharmacology. 2000;8:185–191. doi: 10.1037//1064-1297.8.2.185. [DOI] [PubMed] [Google Scholar]
  56. Testa M, Parks K. The role of women’s alcohol consumption in sexual victimization. Aggression and Violent Behavior. 1996;1:217–234. [Google Scholar]
  57. Testa M, VanZile-Tamsen C, Livingston J, Buddie A. The role of women’s alcohol consumption in managing sexual intimacy and sexual safety motives. Journal of Studies on Alcohol. 2006;67:665–674. doi: 10.15288/jsa.2006.67.665. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Uebersax J. Statistical methods for rater agreement. 2007 Retrieved Jan. 10, 2008, from http://ourworld.Compuserv. com/homepages/jsuebersax/agree.htm.
  59. VanZile-Tamsen C, Testa M, Livingston J. The impact of sexual assault history and relationship context on appraisal of and responses to acquaintance sexual assault risk. Journal of Interpersonal Violence. 2005;20:813–832. doi: 10.1177/0886260505276071. [DOI] [PubMed] [Google Scholar]
  60. Walls HJ, Brownlie AR. Drink, drugs, and driving. 2. London: Sweet & Maxwell; 1985. [Google Scholar]

RESOURCES