EFFECTS OF A DATE RAPE INTERVENTION ON
RAPE PROCLIVITY AND ACCEPTANCE OF RAPE-SUPPORTIVE ATTITUDES AMONG MALE COLLEGE STUDENTS:
A SOCIAL LEARNING APPROACH
JULIE M. ABRAMS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
I would like to thank my committee, Dr. Carolyn M. Tucker, Dr. Phyllis Meek, Dr. Greg Neimeyer, Dr. Barbara Probert, and Dr. Robert Ziller, for their scholarly advice and guidance. Special thanks are extended to Dr. Tucker who, as chairperson, mentor, colleague, and friend, has been an admired and respected role model and a source of inspiration and support. Without her scholarly abilities and dedication, this dissertation would not have been possible.
Acknowledgments are due to several professionals and students at the University of Florida (UF) and elsewhere. Russ Sabella is immensely appreciated for conducting several of the workshops in the study, thus making data collection possible. Dr. John Dixon is gratefully acknowledged for his statistical consultation.
Dr. Andrea Parrot and Ms. Janet Salmons-Rue of Cornell University are respectfully acknowledged for their permission to use materials associated with their campus rape prevention program, "Stop Date Rape," as models for developing portions of the workshop interventions. Bill Abrams is warmly appreciated for his time and talent in the videotape production of "Date Rape Prevention: A ii
Demonstration," which was used in some of the workshop presentations. Peggy Moore and Tom Britt are thanked for their performances in this videotape.
The University of Florida's Sexual Assault Recovery
Service is acknowledged for sharing materials on date rape prevention. Marc Spector and Kristin Smith are thanked for their assistance during the planning phases of the project.
I wish to extend special thanks to all of my family and friends whose patience, support, and encouragement made this dissertation possible. My mother and father, Patricia and Robert Abrams, are thanked for their years of encouragement, guidance, hard work, and sacrifice to create the opportunities for me to learn. Warm thanks are also extended to all of my friends, whose support and friendship have given me the inspiration, confidence, and balance necessary to complete this endeavor.
TABLE OF CONTENTS
ACKNOWLEDGMENTS . . . . . . . . . .
LIST OF TABLES . . . . . . . . . vi
ABSTRACT . . . . . . . . . . . vii
I INTRODUCTION . . . . . . . . . I
Limitations of Previous Research . . . . 3 Rationale for the Study . . . . . . 4
Research Questions . . . . . . . . 6
2 REVIEW OF LITERATURE . . . . . . . 8
Rape . . . . . . . . . . . 8
Definition of Rape . * * * 8
Incidence and Prevalenc; *f*Ra*pe . . . . 11 Consequences of Rape . . . . . . . 19
Rape Myths . . . . . . . . . 25
Sex-Role Socialization and Cultural Norms
Regarding Violence . . . . . . 30
Rape Prevention Research and Practice . . 34
Social Learning Theory and Human Aggression . 48
Principles of Social Learning Theory . . . 49
Social Learning Theory and Aggression
Research . . . . . . . . . 59
Social Learning Theory and Rape . . . . 62
Effects of Modeling . . . . . . . 63
Effects of Reinforcement . . . . . . 66
Implications for Date Rape Prevention
Programming . . . . . . . . 68
3 METHOD . . . . . . . . . . . 71
Subjects . . . . . . . . . . 71
Instruments . . . . . . . . . 72
Procedure . . . . . . . . . . 78
Research Design . . . . . . . . 85
Hypotheses . . . . . . . . . . 87
4 RESULTS . . . . . . . . . . 89
Preliminary Analyses . . . . . . . 89
Tests of the Hypotheses . . . . . . 90
Post Hoc Analyses . . . . . . . . 104
5 DISCUSSION . . . . . . . . . . log
Summary of Results . . . . . . . . 109
Interpretation of the Results . . . . . 112 Limitations of the Study . . . . . . 117
Suggestions for Further Research . . . . 118 Conclusion . . . . . . . . . . 120
A DEMOGRAPHIC INFORMATION FORM . . . . . 122
B SEXUAL EXPERIENCES SURVEY . . . . . . 124
C RAPE MYTH ACCEPTANCE SCALE . . . . . . 126
D ACCEPTANCE OF INTERPERSONAL VIOLENCE SCALE . . 129 E MARLOWE-CROWNE SOCIAL DESIRABILITY SCALE (20) 130 F INFORMATION TEST . . . . . . . . 132
G INFORMED CONSENT FORM . . . . . . . 134
H DEBRIEFING FORM . . . . . . . . 137
REFERENCES . . . . . . . . . . . 139
BIOGRAPHICAL SKETCH . . . . . . . . . 148
LIST OF TABLES
4-1 Item Means for the Rape Myth Acceptance Scale
at Baseline, Immediate Posttest, and Delayed
Posttest . . . . . . . . 95
4-2 Item Means for the Acceptance of Interpersonal
Violence Scale at Baseline, Immediate
Posttest, and Delayed Posttest . . . 97
4-3 Pearson Correlation Coefficients Among the
Dependent Measures at Baseline . . . 99
4-4 Frequency and Percentage of "Yes" Responses to
Sexual Experiences Survey Items (N = 186) 100
4-5 Pearson Correlation Coefficients Between
Marlowe-Crowne Social Desirability Scale
(20) Scores and the Dependent Measures . 102
Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
EFFECTS OF A DATE RAPE INTERVENTION ON
RAPE PROCLIVITY AND ACCEPTANCE OF RAPE-SUPPORTIVE
ATTITUDES AMONG MALE COLLEGE STUDENTS:
A SOCIAL LEARNING APPROACH
Julie M. Abrams
Chairperson: Carolyn M. Tucker Major Department: Psychology
The effects of four workshop/presentation
interventions on male college students' acceptance of date rape myths, acceptance of interpersonal violence, and selfreported likelihood of using force in a sexual interaction were investigated. Specifically, the study was designed to test the effects of social learning theory-based interventions on rape-supportive attitudes.
Using a 2 X 3 repeated measures factorial design, 189 male college students were exposed to Intervention I (information about rape myths, rape trauma syndrome, and strategies for reducing the likelihood of rape); Intervention II (information in Intervention I, videotaped modeling of assertive interpersonal dating behaviors, and positive reinforcement of anti-rape behaviors and statements
made in the videotape and by subjects); or a Control Condition (two videotapes that were unrelated to rape). Half of the subjects in each of the three conditions were exposed to a male presenter; half were exposed to a female presenter.
Results revealed less disparity in rape myth acceptance scores between the two presenter conditions during the second semester of data collection than during the first semester. No such differences were obtained with regard to acceptance of violence or likelihood of using physical force or verbal coercion. In addition, there were no overall effects for intervention type, presenter, semester, or any other 2- or 3-way interaction for any of the dependent variables tested.
A moderate correlation was found between the dependent measures. Interitem correlations for the Rape Myth Acceptance Scale and the Acceptance of Interpersonal Violence Scale were low to moderate, suggesting that these measures have low internal consistency and questionable construct validity for the sample tested.
Further research is needed on methods of rape attitude assessment with college populations. Descriptive data on the present sample of college males indicate a continuing need for research to assess and reduce rape proclivity and rape-supportive attitudes.
Rape is the most common yet most underreported violent crime in the United States, according to estimates presented by the Federal Bureau of Investigation (FBI, 1981). It is estimated that one rape occurs approximately every six minutes (FBI, 1986) and that for every rape that is reported, there are ten that remain unreported (Russell, 1984). It is further estimated that one in three women will be raped during her lifetime (FBI, 1986) and in the majority of these cases, the victim will know her assailant (Koss, 1985).
Most unreported rapes are committed by acquaintances (Koss, 1985) and many of these occur on dates (Rapaport & Burkhart, 1984). In one study, 27.5% of college women surveyed reported having experienced acts that met legal definitions of rape or attempted rape (Koss, Gidycz, & Wisniewski, 1987). In another study, 15% of college men sampled admitted to having forced a date to have sexual intercourse (Rapaport & Burkhart, 1984). About 35% of college males admit to at least some possibility of raping a woman in the future if they could be assured anonymity (Malamuth, 1981).
University students are a high risk group for rape because they are similar in age to the majority of rape victims and offenders. Most rape victims are in their late teens or early twenties and almost half of all alleged rapists who are arrested are under age 25 (FBI, 1986). This, coupled with an increase in autonomy, availability of alcohol and drugs, and vulnerability due to uncertain surroundings and peer pressure, increases the likelihood of date/acquaintance rape on college campuses and makes this issue a serious concern nationwide.
Rape has a tremendous impact on victims and significant others. Victims often endure severe emotional, psychological, and physical trauma that can last over a period of months, years, and even a lifetime. Such effects have been referred to as rape trauma syndrome (Burgess & Holmstrum, 1974) and encompass changes in one's sense of self, relationships with others (including sexual relationships), and physical functioning. Family, friends, and significant others are often affected by rape as well. Many may feel shocked, angry, guilty, helpless, and frustrated (Orzek, 1983). Some may question the integrity of the victims, blaming them for their predicament or doubting their honesty. Fueled by a host of culturally entrenched stereotyped, prejudicial, and false beliefs about rape, many significant others become victims themselves. It
is estimated that over half of all intimate partnerships dissolve after a rape has occurred (Crenshaw, 1978).
Many cultural myths about rape, rape victims, and
rapists exist in support of rape. Feminist theory contends that these myths are reflective of the patriarchal, competitive, sex-role stereotyped society in which we live (Brownmiller, 1975; Herman, 1984). Such myths serve to create an unsafe environment for the rape victim (Burt, 1980) and relieve the offender of responsibility for the crime.
Limitations of Previous Research
In recent years, the literature on rape has
proliferated and has addressed such topics as incidence rates (Koss et al., 1987; Russell, 1982), attitudinal correlates (Burt, 1980; Feild, 1978; Rapaport & Burkhart, 1984), counseling/recovery issues (Burgess & Holmstrum, 1974; Halpern, Hicks, & Crenshaw, 1978; Katz, 1984), and prevention efforts (Borden, Karr, & Caldwell-Colbert, 1988; Parrot, 1985). The literature regarding prevention efforts, however, has been primarily descriptive in nature (Roark, 1987; Sandberg, Jackson, & Petretic-Jackson, 1987) and has mainly focused on strategies women can employ to reduce the likelihood of rape (Gray, Lesser, Quinn, & Bounds, 1990; Kidder, Boell, & Moyer, 1983; Krulewitz & Kahn, 1983; Krulewitz & Nash, 1979; Levine-MacCombie & Koss, 1986).
There has been relatively little research that has empirically investigated the effectiveness of date rape prevention activities, particularly efforts focusing on males (e.g., Lee, 1987). In addition, many of the studies that have been conducted have methodological limitations such as the use of informal evaluations to measure outcomes (Pace & Zaugra, 1988) and the omission of a treatment control group (Gilbert, Heesacker, & Gannon, 1991). Only one study was found which used a theoretical framework to guide the rape intervention research (Gilbert et al., 1991).
Rationale for the Study
The current research was proposed for three reasons:
(1) to investigate empirically the effectiveness of a date rape intervention, (2) to perform this investigation within the framework of a social psychological theory of human aggression, and (3) to target the prevention effort to males (because males represent the majority of rape offenders). The study was designed to extend beyond the use of informal evaluations to document the success of the intervention/ presentations. It proposed to assess the effects of a date rape intervention using existing measures that have been found to be valid and reliable.
Because rape is an act of violence and power, and not an act of lust (Brownmiller, 1975), it is helpful to consider the various social psychological theories of human
aggression in understanding the problem of rape. A review of the social psychological literature on human aggression reveals that the theoretical model that has received the most consistent empirical support to date is social learning theory as proposed by Albert Bandura (Bandura, 1973, 1977).
Briefly, social learning theory postulates that
aggression is primarily learned first by observing others behaving aggressively and second by witnessing the consequences of such aggression. Aggressive behavior may also be shaped and maintained through positive reinforcement from external or internal sources. Theoretically, people can act to increase or decrease their aggressive behavior by changing the environmental conditions that may induce aggression, altering cognitions which support aggression, and providing positive or negative reinforcement for such behavior (Bandura, 1973, 1977).
While several theorists and researchers have used
social learning theory principles to guide their thinking and research regarding the causes and correlates of rape (Boeringer, Shehan, & Akers, 1991; Donnerstein, 1980; Malamuth & Check, 1981), no empirical research was found that overtly applied social learning theory to the understanding of the rape prevention. The purpose of the proposed research, therefore, was to use social learning theory to investigate the effectiveness of two date rape prevention interventions (each conducted under two different
conditions) in reducing male college students, (a) selfreported likelihood of using force in a sexual interaction,
(b) acceptance of date rape myths, and (c) acceptance of interpersonal violence.
My questions for research were:
(1) Are rape prevention workshop/presentations that include
information on rape myths, rape trauma, and strategies
for reducing the likelihood of rape effective in
reducing male college students' acceptance of rape
myths, acceptance of interpersonal violence, and selfreported likelihood of using force in a sexual
(2) Are rape prevention workshop/presentations that include
the modeling and reinforcement of assertive
interpersonal behavior in addition to information on rape myths, rape trauma, and strategies for reducing
the likelihood of rape more effective in reducing male
students' acceptance of rape myths, acceptance of
interpersonal violence, and self-reported likelihood of
using force in a sexual interaction than are workshop/
presentations that do not include such modeling and
(3) Are rape prevention workshop/presentations that are
conducted by a male more effective in reducing male
college students' acceptance of rape myths, acceptance of interpersonal violence, and self-reported likelihood of using force in a sexual interaction than the same workshop/presentations conducted by a female?
REVIEW OF LITERATURE
An overview of the problem of rape in the United States is presented in this review. The magnitude and extent of rape, the consequences of rape, cultural norms and values that exist in support of rape, and current rape prevention strategies will be discussed. A social psychological analysis of sexual assault will be given in terms of Albert Bandurals social learning theory. Specific attention will be paid to the effects of modeling, reinforcement, and implications for date rape prevention outreach. Definition of Rape
The definition of rape varies greatly by jurisdiction. The Uniform Crime Report (UCR) Program of the Federal Bureau of Investigation (FBI) defines rape as "carnal knowledge of a female forcibly and against her consent" (FBI, 1986). This definition, thus, limits rape to assaults against females that involve the penetration of the vagina by a penis. Its requirement of force excludes some instances of child sexual abuse. The FBI's Uniform Crime Report Program provides widely cited national statistics regarding the
incidence and prevalence of rape. However, its rather narrow definition of rape may contribute to the underestimation of the magnitude and extent of the problem. In reporting crime to the FBI, some state law enforcement officers believe that the UCR definition forces them to underreport incidences of rape (Martin, DiNitto, Norton, & Maxwell, 1984).
In contrast to the FBI, several states have more
encompassing definitions of rape. Florida, for example, is recognized as having one of the most "comprehensive" or advanced sexual assault statutes in the country. According to the Florida statutes, sexual battery is defined as "oral, anal or vaginal penetration by union with a sexual organ of another or the anal or vaginal penetration by another by any other object" (Florida Statutes, 1983). These state statutes, thus, include not only vaginal penetration by a penis, but nonconsensual sodomy (anal intercourse), fellatio (oral intercourse on a male), cunnilingus (oral intercourse on a female), and penetration by fingers, hands, or foreign objects (Martin et al., 1984). They include instances of unwanted sexual contact in which there is no show of force, assaults against males, and instances of statutory rape. on the following pages, attempts will be made to clarify what definition of rape is used in the referenced research.
In defining rape, it is important to note that rape is an act of power and violence; it is not an act of passion or
lust (Brownmiller, 1975). Viewing rape as a sexual act in which the victim really wants to be dominated and overpowered contributes to the myth that victims ask to be raped. Similarly, it alleviates assailants of the responsibility for any wrongdoing and contributes to the belief that women (the victims, in most cases) are masochistic persons. Unfortunately, such beliefs also support the continued usage of such oxymorons as "rape fantasy."
Much of the early research on rape assumed a
typological approach to the conceptualization of rape. That is, a person was seen as either a rapist or a nonrapist, a victim or a nonvictim. Evidence now exists to support a dimensional view of rape. A dimensional view of rape still regards rape as an extreme behavior, but places rape on a continuum with normal male behavior (Koss & Oros, 1982). The continuum of sexually aggressive behavior is considered to include verbal coercion, threat of force, and actual use of physical force to obtain sexual intercourse. These behaviors, however, are not seen as a series of escalating events where one behavior necessarily leads to another (Koss & Gidycz, 1985).
Different terms have been coined to refer to specific
types of rape. Acquaintance rape refers to rape that occurs between persons who know one another. Date rape is a form of acquaintance rape in which the assailant and the victim
have an agreement for a social engagement. Gang rape refers to rape committed by two or more persons (C.P. Walsh, personal communication, October 12, 1990). This review will first address the general issue of rape, but will focus specific attention on the problem of date and acquaintance rape on college campuses. Because women do indeed represent virtually all of reported rape victims (Law Enforcement Assistance Administration CLEAA), 1975), this manuscript will focus on male offenders and female victims of sexual violence.
Incidence and Prevalence of Rape
Rape is the most frequent yet underreported violent crime in the United States (FBI, 1981). According to the Federal Bureau of Investigation, 87,340 rapes occurred in 1985 with one rape occurring approximately every 6 minutes (FBI, 1986). While these figures may seem shocking, they are considered to be underestimates of the full scope of rape for two reasons: (1) they include only those crimes that fit the FBI's narrow definition of rape, and (2) they are based only on those rapes that are reported to the police (Koss et al., 1987). It is thought that for every rape that is reported, 3-10 rapes are committed but are unreported (LEAA, 1975; Russell, 1984). Estimates of the actual frequency of rape are made more difficult by the fact that only a small proportion of reported rapes actually
result in convictions (Ross et al., 1987; Martin et al., 1984).
one of the main sources of crime estimates including
rape is the Bureau of Justice's National Crime Survey (NCS). Although frequently cited, the NCS has been criticized on conceptual and methodological grounds. For example, some of the NCS questions are considered to be vague and require the respondent to infer what is being asked. The survey assumes that victims of sexual assault use the term "rape" to conceptualize their experiences. In addition, because questions about rape are embedded in a list of questions about violent crime in general, respondents who are rape victims but who do not consider themselves to be victims of crime, may be less likely to give accurate accounts of their experiences. The survey is also criticized in that it assumes a typological view of sexual aggression and victimization and thus ignores the finer gradations of sexual assault (Koss et al., 1987).
Russell (1982) investigated the prevalence and
incidence of rape and attempted rape and found that of a sample of 930 adult women, 24% described experiences that met the criteria for rape. Only 9.5% of these women reported the crime to the police. In another study, about 15% of a sample of college men admitted to having sexual intercourse against their date's consent and only 39% denied
any coercive sexual involvement whatsoever (Rapaport & Burkhart, 1984).
Similar results were obtained by Miller and Marshall (1987) in a study of 795 undergraduate and graduate students. These researchers found that 27% of the women sampled reported psychologically or physically forced sexual intercourse. Interestingly, only 3% conceptualized the experience as "rape." Additionally, 15% of the men sampled indicated that they had coerced a woman into sex, but only it defined this experience as "rape." These findings suggest that large numbers of "hidden" victims and offenders abound.
Koss and colleagues (Koss et al., 1987) conducted the most encompassing research study to date on the incidence and prevalence of rape. Sampling 6,159 students from 32 institutions of higher learning and using an instrument designed to reflect a dimensional view of rape, these researchers asked subjects to indicate their experience with various levels of sexual victimization or aggression. They found that 53.7% of the women sampled revealed some form of sexual victimization and 25.1% of the men reported some involvement with sexual aggression. The most serious form of sexual victimization experienced by the women was rape for 15.4%, attempted rape for 12.1%, and sexual coercion for 11.9%. The most serious level of sexual aggression reported by the men was rape for 4.4%, attempted rape for 3.3%, and
sexual coercion for 7.2%. The institutions at which females reported the highest rape rates were major universities (17%) and private colleges (14%). The region in which the highest proportion of males admitted rape was the Southeast
Prevalence rates for sexual victimization and
aggression were also found to differ by ethnic group. Rape was reported by 16% of White women, 10% of Black women, 12* of Hispanic women, 7% of Asian women, and 40% of Native American women. Among men, rape was reported by 4% of Whites, 10% of Blacks, 7% of Hispanics, 2% of Asians, and 0% of Native Americans (Koss et al., 1987).
Based on this national sample, incidence rates for rape and attempted rape, i.e., the number of women per thousand who experienced rape, were calculated to be 83 per 1,000 women for a six-month period (using state definitions of rape) or 38 per 1,000 women (using the FBI's narrower definitions of rape and attempted rape). Perpetration rates for rape and attempted rape were calculated to be 34 per 1,000 men for a six-month period (using state statutes) or 9 per 1,000 men (using the FBI's definition) (Koss et al., 1987). It is pointed out that the victimization rate using the FBI's definition of rape is 10-15 times greater than statistics reported by the National Crime Survey and that the rape perpetration rate using the FBI's definition is 2-3 times greater than corresponding NCS figures. These data
provide some evidence to suggest that the NCS fails to assess the full scope of rape.
The discrepancy between victimization and perpetration rates is likely to be due, at least in part, to the fact that some of the incidents reported by the women undoubtedly occurred before college by males other than those surveyed. It was suggested, however, that some of the men surveyed may not have accurately perceived the degree of force or coercion involved in their interactions and may have misinterpreted a woman's nonconsent (Koss et al., 1987). If this is so, increased attention to communication and perceptions may be important during rape education programs.
Consistent with the findings of other researchers
(e.g., Miller & Marshall, 1987), results obtained by Koss et al. (1987) indicate that few rapes are acknowledged by the victim (27%). Even fewer rapes are ever reported to the police (5%) or prompt the victim to seek victim assistance services (5%). Many rapes are never revealed to anyone at all (42%). Because much of the existing research on rape is based on samples of women who are self-acknowledged rape victims (recruited through paper ads, police reports, court records, emergency rooms or counseling services), "hidden victims" (those who don't acknowledge, report, or reveal their rapes) are often omitted from study. It is recommended that those "hidden" victims be included in future research (Koss et al., 1987); a similar case may be
made for hidden offenders, i.e., those who may not recognize their aggression and those who are not reported, tried, and convicted.
An examination of the data on the incidence and
prevalence of rape in Florida is sobering. In 1982, Florida had the third highest rape index in the nation (with 53.6 rapes per 100,000 persons), second only to Alaska and Nevada. Among Florida's counties, Alachua had the seventh highest rape index (with 64.65 rapes per 100,000 persons) (Martin et al., 1984). Alarming as these statistics may be, they are not necessarily negative indicators for Florida. The high crime indices may indeed be reflective of Florida's increasing and transient population and concomitant social problems. However, they may also be due, in part, to the existence of rape victim support services, which may foster higher reporting rates.
A 10-month study of the problem of rape in Florida
involving personal interviews with over 200 people from rape crisis centers, law enforcement agencies, hospitals, mental health agencies, citizen action groups, and the state attorney's office, yielded a profile of rape in Florida similar to that of the nation. Most rapes were found to occur among acquaintances and between persons of the same race. The vast majority of victims were female (93-95%) and white (about 64%) and almost all offenders were male (Martin et al., 1984). Further, the vast majority of reported rapes
are assaults committed by one offender, although there were occasional reports of gang rape (Martin et al., 1984).
Although the rape assessment conducted in Florida
looked at rape among the general population, the majority of rape research, including many of the national incidence and prevalence studies, targetscollege students as subjects. Rape research on college students is important for several reasons. First, college students are particularly at risk for rape because of their age. The most common age for rape victims is between 16 and 19. The second highest rate occurs among persons between the ages of 20 and 24 (Bureau of Justice Statistics, 1984). Additionally, almost half of all alleged rapists who are arrested are under age 25 (FBI, 1986).
Second, college students of traditional age are especially vulnerable to victimization in general. Typically, these students are in an unfamiliar setting, for the first time without old support systems and parental supervision. Their identities are generally not yet established, their sexuality may not yet be fully explored, and they may have illusions of invincibility. Their cohorts may be-experimenting with new freedoms and exertfngptVrpressure to conform to the group (Roark, 1987).
Third, the environment in which college students live may also contribute to increased risk of rape. one aspect of the university environment that is often associated with
increased risk of date/acquaintance rape is the widespread .1 ise of alcohol among college students. Research shows that about 50% of women who are victims of rape and about 65% of their assailants were drinking before the rape occurred (Johnson, Gibson, & Linden, 1978). Alcohol use can impair judgment and decrease assertiveness. Its purposeful use to increase a potential victim's vulnerability is illegal under certain circumstances in Florida. Under Florida law, it is illegal to have sexual intercourse with a person who, without their consent or prior knowledge, has been given )intoxicants which mentally or physically incapacitate them, thus disabling them from giving consent. It is also illegal to have sexual intercourse with a person who is physically helpless, i.e., unconscious, asleep, or for any other reason unable to communicate unwillingness to the act (Florida Statutes, 1983).
Another aspect of university life which is said to play a role in some rapes is the existence of closely knit, allmale groups such as fraternities. Fraternities, in general, have been characterized as having a narrow, stereotypic idea of masculinity and heterosexuality, an emphasis on group loyalty, competition, and supremacy, and an acceptance of the use of alcohol as a means to gain sexual access (Martin & Hummer, 1989). Indeed, one study showed that a disproportionate number of alleged campus rape offenders were members of fraternities (O'Shaughnessey & Palmer,
1990). In addition, the majority of reported gang rapes on campus are said to involve fraternity members (Roark, 1987). Fraternity members are also reported to be more likely than "independents" to use verbal coercion, drugs, and alcohol in order to obtain sexual intercourse (Boeringer, Shehan, & Akers, 1991).
Some evidence exists to support the idea that men and women enter college with different ideas regarding sexual relationships. Giarrusso, Johnson, Goodchilds, and Zeliman (1979) found that of a sample of 432 teenagers between the ages of 14 and 18, 76% of the boys and 56% of the girls said there were certain circumstances under which it was acceptable for a male to use force to obtain sexual intercourse. Such circumstances included when a girl gets a boy sexually aroused, and when a girl says she plans to have sex with a boy and then changes her mind. In a similar study of 272 female and 268 male college students, Muehlenhard (1988) found that both men and women believed that date rape was more "justified" when the woman initiated the date, when they went to the man's apartment and when the man paid the expenses. The degree of justification was greater for men compared to women under these conditions. Consequences of Rape
Rape has far-reaching consequences for the victim, significant others, and society. It often has severe
emotional, social, cognitive, and behavioral effects that may last for years and even a lifetime. Below is an overview of the literature on the impact of rape, including some discussion of the effects of rape on significant others and society at large.
Rape trauma syndrome, as identified by Burgess and
Holmstrum, (1974), has been the cornerstone of our current understanding of the impact of rape on victims and survivors. This syndrome has two basic stages: 1) an immediate or acute phase, involving a disruption of life style and 2) a long-term phase involving a reorganization of the self and a resolution of personal feelings about the rape. Recovery from the physical and emotional trauma of rape is thought to proceed through these stages.
The acute phase may last for a few days or a few weeks. During this crisis stage, the rape victim commonly experiences a wide range of emotions, including shock, disbelief, fear, anxiety, tension, hurt, alienation, powerlessness, defenselessness, distrust, depression, vulnerability, guilt, shame, embarrassment, confusion, anger, and loss of control (Burgess & Holmstrum, 1974; Doweiko, 1981; Grossman & Sutherland, 1983; Guest, 1977). Some rape victims are visibly expressive of their feelings; others appear calm and controlled. Mood swings are common.
Physical reactions during the acute phase are typical. some rape victims report feeling a general bodily soreness.
others report having physical symptoms that are specific to the parts of the body that were violated. These may be expressed as an irritation of the mouth and throat, vaginal discharges and/or itching, a burning sensation upon urination, and rectal pain or bleeding. Many rape victims experience disrupted sleep patterns (insomnia, night terror) and disrupted eating patterns (abdominal pain, loss of appetite, nausea) (Burgess & Holmstrum, 1974; Grossman & Sutherland, 1983).
The long-term phase may last for months or even years. During this stage, rape victims deal with the impact the rape has had on their lives. Many make changes in their style of living. Some move, change or get an unlisted phone number, visit relatives, or remain home much of the time. Often victims have difficulty concentrating and report having nightmares. In their dreams, they may either feel like victims of violence or they may feel like victimizing others. They also have to deal with any phobias they may have developed as a result of the rape. These phobias may include a fear of sex, crowds, being alone, or being near those who have similar characteristics as the assailant (Burgess & Holmstrum, 1974; Grossman & Sutherland, 1983).
Physical symptoms during the long-term reorganization period may include chronic gynecologic problems, changes in the menstrual cycle, gastrointestinal problems, and conversion reactions. Emotionally, rape survivors may
experience continued sleeplessness, loss of appetite, anxiety, depression, mood swings, lowered self-esteem, difficulty trusting men, fear of sex, and reluctance to engage in sexual relations (Rosenburg, 1986).
In addition to the two-phase model offered by Burgess and Holmstrum (1974), other theorists have postulated a triphasic response to rape (Doweiko, 1981; Katz, 1984). Doweiko (1981) states that a middle phase exists between the initial crisis and subsequent resolution phases which he terms "outward adjustment." This phase is characterized by denial, suppression, and rationalization and may last for several weeks or months until the victim acknowledges and accepts the rape and begins to work toward resolution. Katz (1984) describes the middle phase as part of the post-crisis adjustment. During this phase before long-term reorganization, the rape victim may continue to experience many emotional and physical symptoms, including feeling disoriented and out of control and having nightmares.
Recovery from rape has also been conceptualized in terms of the many losses associated with sexual assault: loss of self-identity, security, friendships, status within the community, sexual identification (Whiston, 1981), selfrespect, power, control, and possibly virginity, as well as a threat of loss of life (Freiberg & Bridwell, 1976). According to this loss model, survivors of rape may experience many feelings commonly associated with the grief
process, including denial, depression, and anger (Freiberg & Bridwell, 1976). Feelings of loss associated with independence, privacy, and self-esteem may be aggravated when others make decisions for the victim following a rape.
The degree of emotional trauma experienced as a result of rape may vary according to the nature of the crime. It is argued that rape by an acquaintance, friend, or relative is much more psychologically harmful than rape committed by a stranger (Koss, Leonard, Beezley, & Oros, 1985). The violation of trust in nonstranger rape cases makes the process of maintaining current relationships and risking new attachments and emotional intimacy much more difficult for the survivor of nonstranger rape. Because victims of nonstranger rape, compared to victims of stranger rape, are less likely to report the crime committed against them (Russell, 1984), they are more likely to endure their suffering alone.
The impact of rape extends beyond the victim to include family members, friends, and significant others. It is common for loved ones to express (1) feelings of guilt, self-blame, and a desire to overprotect, (2) feelings of frustration, anger, and revenge, and (3) a tendency to blame the victim for his or her predicament (viewing rape as a sexual, rather than a violent act) (Orzek, 1983; Rodkin, Hunt, & Cowan, 1982). Friends and loved ones may be in crisis, may feel victimized, and may feel confused and
misguided by false information (myths) regarding rape (Egidio & Robertson, 1981; Rodkin et al., 1982).
Couples are particularly at risk following a sexual
assault. Statistics show that 50-80% of raped females lose their boyfriends or husbands within one year of the assault (Halpern et al., 1978). Partners may have questions regarding sexually transmitted diseases, pregnancy, and sexual desirability. Sexual behavior may be misinterpreted due to the different emotional stages experienced by each partner following a rape. For example, the sexual assault survivor may avoid sexual contact, which may, in turn, cause the partner to feel unwanted. The partner, on the other hand, may try to force sexual contact, which may result in the victim feeling used and/or robbed of personal autonomy (Orzek, 1983; Rodkin et al., 1982). Couples counseling or sex therapy may be indicated for couples with continued relationship difficulties following a sexual assault (Halpern et al., 1978).
The impact of rape is visible in the demand for victim assistance services. From counseling, law enforcement, medical assistance, and legal affairs, rape survivors are often in need of information, services, support, and advocacy. They often need information regarding medical procedures (including the physical exam and protection against pregnancy and infection), the legal system (including whether or not to prosecute), the impact of the
assault on significant others, and options for psychological counseling (Courtois, 1979; Doweiko, 1981; Guest, 1977). Each of these services are provided at some cost to the public at large.
In addition to straining legal, counseling, and medical services, rape may also present a cost to employers in terms of poor concentration, decreased work performance, and days absent from work. on university campuses, sexual violence may be evidenced by a loss of concentration, impaired academic performance, and lowered self-esteem.
Rape is a crime supported by many cultural myths.
Because these myths exist, many victims are reluctant to file police reports and to seek professional counseling. Rape myths create an unsafe environment for rape victims. They often place the blame for rape on victims and relieve assailants of responsibility for any wrongdoing.
one of the existing myths is that rape is primarily a sexual act caused by sexual frustration or maladjustment. Research has failed to demonstrate support for this contention. There is, however, substantial evidence to support the view that rape is an act of sexual aggression (Briere & Malamuth, 1983; Burt, 1980; Malamuth, 1981). In one study, a number of sexuality variables failed to predict self-reported likelihood of raping or using force in a
sexual interaction. However, a variety of rape-supportive beliefs (such as blaming the victim and viewing sexual violence as arousing to women) did significantly predict likelihood of sexual aggression (Briere & Malamuth, 1983). In another study, men who reported a higher likelihood of raping displayed more aggressive behavior towards women in a laboratory setting (Malamuth, 1981).
It is a myth that most rapes are committed
spontaneously in a deserted area by a stranger. The fact is that many rapes take place in the victim's home and most are committed by a relative or acquaintance (Parrot, 1985). Most rapists appear to be average American men and, while many are married and young, rapists can be of any age, race, or class. Rapists, thus, are generally not deranged, sexstarved persons. Rather, they are more accurately characterized by an acceptance of misogynous attitudes, rape-supportive beliefs, and the use of aggression in a sexual context. In addition, they are likely to demonstrate low levels of responsibility and social conscience (Rapaport
In addition to the above misconceptions about rape, it is also untrue that rape occurs only to certain types of people. The fact is that rape victims include individuals of any age, race, class, religion, occupation, education, or physical characteristic (Grossman & Sutherland, 1983). Rape victims are usually female, but can also be male (Burgess &
Holmstrum, 1974). The majority of rapes involves persons of the same ethnic background (Davis, 1981; Friedman, 1979; Grossman & Sutherland, 1983; Martin et al., 1984). The belief that most rapes are committed by black men against white women is untrue and is thought to be associated with the maintainance of white male property rights and the control of people of color (Friedman, 1979).
Finally, it is a myth that it is not really possible to rape a nonconsenting adult. It is false that any healthy, fully functioning woman can resist a rapist if she really wants to. Fear of death or physical injury often causes victims to say and do things against their will. Often victims will exhibit little resistance just in order to survive (Grossman & Sutherland, 1983). Because rape myths have been internalized by men (Beneke, 1982), women (Katz, 1984), laypeople and professionals (Burt, 1980), it is imperative that educational efforts aimed at increasing awareness continue to be made in order to avoid further perpetuation of this crime.
In studies investigating sex differences in beliefs
about rape, males are consistently more likely than females to accept rape myths (Ashton, 1982; Barnett & Feild, 1977; Feild, 1978; Sandberg et al., 1987), less likely to have knowledge about rape trauma, less likely to perceive the rape experience as aversive for the victim (Hamilton & Yee, 1990), and more likely to believe that rape will not occur
if the victim fights back (Krulewitz, 1981). Men, compared to women, have also been found to hold less favorable attitudes toward women (Spence, Helmreich, & Stapp, 1973).
The degree of rape myth acceptance has also been found to vary among rape crisis counselors, citizens, police officers, and convicted rapists (Feild, 1978). Results of a survey of 1,448 subjects revealed that (1) counselors endorsed the fewest number of rape-supportive beliefs and
(2) citizens and police officers were more similar to rapists than to counselors on attitudes toward rape. Results of this study suggest that educational efforts may be necessary in order to increase awareness among persons who may have contact with rape victims, but who may not have accurate information for dealing effectively with these individuals.
Fischer (1986) investigated the predictive value of
five variables (attitudes towards women, sexual knowledge, sexual experience, tolerance of socially unapproved sexual behavior, and religiosity) on acceptance of forcible date rape. Results suggested that higher acceptance of forcible date rape is related to relatively more traditional attitudes about women and greater sexual selfpermissiveness. Also, persons who were more accepting of forcible date rape were less sure that the interaction really was rape, had slightly less accurate knowledge about sexuality, and, although most did indeed blame the male
rapist, were more likely to blame society or the situation. overall sexual experience (i.e., number of partners, frequency of masturbation, positions attempted, oral sex, etc.) and religiosity (defined in terms of a biblical basis for morality and frequency of church attendance) were not found to be significant predictor variables.
In a study investigating predictors of rape myth
acceptance, Burt (1980) found that the higher one's sex role stereotyping, adversarial sexual beliefs, and acceptance of interpersonal violence, the higher was one's acceptance of rape myths. Variables that were found to be significant predictors of rape myth acceptance included age and education, with younger and better educated persons revealing less stereotypic, adversarial, and proviolence attitudes. Variables that were not found to significantly predict rape myth acceptance included self-esteem, personal experience of attempted or actual rape, knowledge of a rape victim, and conservative sexual attitudes.
Burt's (1980) findings support the notion that rape myth acceptance is widespread and strongly connected to a host of other beliefs about men, women, and relationships. As mentioned previously, attitudes about rape have also been associated with self-reported likelihood of raping (Briere & Malamuth, 1983). Expressed likelihood of raping, in turn, has been related to aggression towards women in a laboratory setting (Malamuth, 1981).
Burt (1980) suggests that rape prevention efforts should include education about sex-role stereotypes and promote the idea that sex is a mutual engagement that is freely chosen by its participants. Additionally, she recommends that rape prevention efforts challenge the societal values which tolerate and reinforce violence in our culture. These urgings are based on the finding that acceptance of interpersonal violence was the strongest predictor of rape myth acceptance among all the variables tested (explaining 27.9% of the variance in rape myth acceptance).
Sex-Role socialization and Cultural Norms Regarding violence
Feminist theory views rape as a logical extension of a patriarchal, competitive, sex-role stereotyped society (Burt, 1980). According to feminist theory, it is the culture in which we live that plays the primary role in establishing and maintaining rape by reinforcing pro-rape attitudes and behaviors (Brownmiller, 1975; Groth, 1979; Herman, 1984). Prevention efforts developed from this perspective focus on raising consciousness about prescribed gender roles that dictate male dominance over women and define the male identity in terms of the traditional "macho" ego.
In our society, male and female development generally takes place along fairly rigidly defined, sex-stereotypic
lines. Male development emphasizes the processes of separation and individuation, whereas female development emphasizes the processes of attachment and connectedness. Male identity is formed through roles, position, and individual achievement, whereas female identity is gained through the development of relationships and cooperative achievements. The cognitive styles of males and females generally differ, with males encouraged to develop basically rational cognitive styles and females encouraged to develop basically intuitive cognitive styles (Gilligan, 1982; Lerner, 1988).
The traditional male role, thus, is characterized by reason and intellect, an instrumental or task orientation, success gained through individual achievement and dominance over others, and demonstrations of strength and invulnerability. The traditional female role, in contrast, is characterized by emotional expression, a peopleorientation, recognition gained through association with others, relationships based on nurturance, helpfulness, and agreeableness, and demonstrations of deference and selfeffacement (Bloom, Coburn, & Pearlman, 1975).
Regarding sexual attitudes and behavior, men and women are socialized along very different lines. Men are taught to view sex as a performance in which an erection, intercourse, and orgasm are essential for success. Men are socialized to assume responsibility for initiating sexual
contact and to believe that a man should always be ready for and desiring of sex. Other lessons on male sexuality include the myths that all physical contact must lead to sex, that men shouldn't express certain feelings, that sex equals intercourse (Zilbergeld, 1978), and that a dating man can expect a sexual return on his financial and social investment (Sandberg et al., 1987).
Women, on the other hand, are taught that it is not feminine to show an interest in or initiate sex (Heiman & LoPiccolo, 1988), that it is best to defer to men in relationships, that it is not proper to hurt another's feelings, and that men will protect them from harm (Griffin, 1979; Lerner, 1988). These roles and attitudes, while complementary, support a male-dominated societal structure and create a game-like atmosphere in which sexual coercion, and possibly rape, are likely to occur.
Whereas much has been written about the limits placed on the traditional female role, it is only relatively recently that the restrictions placed on the traditional male role have been considered (O'Neil, 1981; Shiffman, 1987). The traditional male role is characterized by restricted emotional expression (allowing for expressions of anger and aggression, but not vulnerability), an emphasis on control, power, and competition (resulting in a need to dominate and succeed over others), homophobic attitudes (fearing intimacy with other men), restricted sexual and
affectionate behavior (behavior characterized by performance expectations and dominance), an obsession with achievement and success, and health care problems (resulting from a failure to attend to physical and emotional signs of distress) (O'Neil, 1981). Gender roles that limit men from developing themselves to their full human capacity result in attempts at proving masculinity through competition, sexual conquest, and rejection of men who display nontraditional male behavior. Role stress may also be manifested by alcohol and drug abuse, avoidance of intimacy, and an overinvolvement in work or studies.
Congruent with the notions of competition, hierarchical relationships, and an appreciation for dominance is a cultural acceptance of violence and aggression. Aggression is regarded as a satisfactory method for obtaining goals in our society, a message made clear to us through our history, our media, our advertising (Sheffield, 1984), and occasionally, our laws (Roark, 1987). We learn to make cowboys, cops, and military men our heros; we learn to glorify aggression and war (Kokopeli & Lakey, 1983). Violence is often tolerated, especially against groups who are unlike those in power. It is often tolerated within relationships. Traditionally, the closer the relationship, the more acceptable the violence (Roark, 1987).
In national surveys on the use of violence as a socialcontrol measure (e.g., spankings, capital punishment, and
military action), males were found to be more supportive of using force or violence as a way to achieve compliance than were females (Smith, 1984). Other research shows that males also admit to more hostility and aggression, particularly if the aggressive behavior is physical (e.g., electric shocks) as opposed to psychological (e.g., insults) and when the behavior causes the other person to feel anxious, guilty, or unsafe (Eagly & Steffen, 1986). In real-world situations, men are arrested for violent crimes eight times more often than women (FBI, 1985). In college dating situations, men, compared to women, are twice as likely to engage in severe, expressive violence and are four and a half times more likely to assault their partners with lethal weapons (Makepeace, 1983).
Rape Prevention Research and Practice
Approaches to the prevention of rape take a variety of forms. Some are remedial; some are proactive. Some focus on the potential victim; some target potential offenders. Some are directed towards administrative and program changes. All strategies are considered to be important in a multifaceted approach to preventing sexual assault. In this section, an overview of the current theory and practice regarding sexual assault prevention will be provided, focusing on outreach efforts aimed at decreasing date and
acquaintance rape. Research on the effects of rape prevention strategies will also be presented.
Prevention of campus violence, including rape, can occur on three levels: (1) tertiary prevention (direct services to victims in the aftermath of violence), (2) secondary prevention (policy and procedural development and local research on the nature and extent of the problem), and
(3) primary prevention (actions aimed at preventing further instances of violence from happening by addressing causes and by changing attitudes and behaviors that support violence) (Roark, 1987).
Primary prevention efforts include changes made to the physical environment (e.g., increased lighting, provision of nighttime escorts, accessible phones, and trimming of shrubbery near buildings), skill-building workshops (on assertiveness, self-defense, alcohol awareness, communication skills, conflict-resolution, and sexual decision-making), and programs that address topics such as sex-role socialization, sexuality and violence, and personal power. Because the vast majority of rapes occur among acquaintances or on dates,* efforts aimed at altering the physical environment may achieve only limited success in combatting sexual assault. It is argued that educational efforts and programs designed to increase awareness regarding sex-role socialization and sexual assault may be most effective in preventing date and acquaintance rape.
Many traditional rape prevention strategies focus solely on the potential victim and what women can do to avoid rape. Such strategies include assertiveness and selfdefense training (Kidder, Boell, & Moyer, 1983; Parrot, 1985; Sandberg et al., 1987), and efforts to increase a woman's sense of empowerment, knowledge of personal rights, and comfortability in discussing sexual topics (Parrot, 1985).
There has been a great deal of research investigating rape victim response, assault outcome, and causal attributions about rape. These studies generally reflect beliefs that women can prevent rape by altering their behavior. In one study, subjects assigned greater responsibility to the victim and less responsibility to the rapist for completed rape than attempted rape. Interestingly, men attributed less fault and more intelligence, while women attributed more fault and less intelligence to rape victims who resisted rape more forcefully. Subjects also were more sure that a rape had occurred when the victim showed more resistance (Krulewitz Nash, 1979).
In a later study, Krulewitz and Kahn (1983) found that, in general, subjects rated strategies which placed the locus of responsibility for change on women as more effective than strategies which placed the responsibility on men and society. Approaches that conformed to sex-role stereotypes
(women avoiding rape with passive behavior and men and society attempting to stop rape by behaving aggressively towards the rapist) were perceived as more effective than strategies that transcended sex-role stereotypes. Feminists, however, differed from nonfeminists in that they viewed nonstereotypic strategies and strategies that placed the locus of responsibility on men and society as more desirable. one acknowledged limitation of these studies is that no empirical evidence of the actual effectiveness of rape reduction strategies was provided.
In a study that compared the response strategies of
rape victims and rape avoiders to determine whether or not there were any differences in emotional, cognitive, or behavioral responses between groups (Levine-MacCombie & Koss, 1986), results revealed that, compared to rape victims, avoiders were more likely to have run away and screamed, less likely to have quarrelled with the assailant, and more likely to have viewed the assault as less violent. No differences were found in the use of physical resistance among groups to avoid rape. one limitation of this study is that there is no way of knowing whether or not the completed rapes were more serious or threatening than the avoided rapes. Avoiders' perceptions of the rape attempts as being less violent may be accurate reflections of reality.
Only one study was found which explored the effects of an acquaintance rape prevention program designed for women
(Gray et al., 1990). This study investigated the effects of a personalized rape prevention program on women's selfreported likelihood of high-risk dating behavior and perceptions of vulnerability to acquaintance rape. Results showed that personalized programs, achieved through the use of local data and examples, resulted in perceptions of increased vulnerability and intentions to reduce risk-takinq behavior. The authors suggest that prevention programs targeted towards women use local data and examples in addition to providing generic information about rape and effective avoidance strategies.
The above research focuses on "prevention" strategies aimed at women. Although the objective of these efforts is admirable, i.e., the reduction of the likelihood of rape, they are, unfortunately, limited in that they do not address the central problem of rape. "Prevention" strategies that focus on changing women's behavior do not prevent rape. At best, they may help to avoid it, or reduce its likelihood of occurring. Furthermore, such efforts say nothing of the male offender's responsibility for rape; they do nothing to challenge the notion that "boys will be boys."
Recognizing that rape is a societal problem rather than a concern of women only, several programs have been designed to include both male and female participants. Objectives of these programs include increasing male and female students'
awareness of each other's feelings and perceptions and encouraging more open communication between the sexes.
The typical date rape awareness workshop involves a presentation of information including legal terms, definitions, incidence rates, common myths, rape prevention strategies, and available counseling resources (Borden et al., 1988; Buhrke & Lustgraaf, 1988; Pace & Zaugra, 1988). Workshops may also include information on the effects of rape, videotaped portrayals of typical date rape scenarios or interviews with rape survivors, small group discussions, and anonymous surveys of participant knowledge and experience. unique strategies such as "fishbowl exercises" are also occasionally used (Pace & Zaugra, 1988). Many of these programs are developed with the idea that peer educators can be trained to facilitate the workshops (Buhrke & Lustgraaf, 1988; Pace & Zaugra, 1988; C.P. Walsh, personal communication, January 28, 1991).
There is a great deal of discussion and investigation of the relative importance of various program topics and relative effectiveness of program formats. Some researchers recommend that educational efforts include information regarding rape myths, incidence rates, and the effect of sex-role socialization on dating practices (Sandberg et al., 1987). Others support education regarding sexual functioning and the effects of alcohol and drug use on sexual interactions (Miller & Marshall, 1987). Many support
communication skills training for males and females (especially around sexuality) (including assertiveness and accepting no for an answer) (Miller & Marshall, 1987; Sandberg et al., 1987).
Hamilton and Yee (1990) investigated the relationship between knowledge about the after-effects of sexual assault, beliefs about the aversiveness of rape, attitudes toward rape, and self-reported likelihood of raping among a sample of 276 undergraduate students. Results indicated that, for both males and females, greater knowledge about the social, psychological, and behavioral effects of rape was significantly correlated with perceptions of rape as more aversive, fewer rape-supportive attitudes, and, for men, less self-reported likelihood of committing rape. Their findings suggest that educational intervention programs that inform participants of the negative consequences of rape may help to reduce the incidence of sexual assault, particularly those assaults that are motivated by the goal of attaining sexual access (instrumental aggression), as opposed to those that are motivated by anger or the desire to harm the victim (hostile aggression) (Hamilton & Yee, 1990).
one study was found that was designed to compare the effects of two different strategies designed to increase male and female college students' knowledge about date and acquaintance rape (Nelson and Torgler, 1990). This experiment compared a 30-minute videotape entitled, "Someone
You Know: Acquaintance Rape," and a brochure on the same topic. No other prevention strategy was included in either intervention group, for example, a workshop presentation or discussion. A control group was included which received a brochure on career planning.
Results of the experiment revealed that attitudes
towards date rape changed significantly for all three groups from pretest to posttest and that posttest attitude scores did not differ between groups. Possible explanations for these results include (1) that the pretest measure itself sensitized participants to the topic of acquaintance rape, and (2) that subjects may have chosen socially desirable responses after having correctly guessed the nature of the study. It is suggested that further research assess the effects of combined prevention strategies (Nelson & Torgler, 1990). It also seems sensible to measure the degree of social desirability responding in order to test the veracity of the second possible explanation of the results.
An investigation of the effects of a didactic
presentation on empathy and attitudes toward rape also failed to produce significant results (Borden et al., 1988). The authors suggested that future research assess the impact of more dynamic, interactive formats which may include roleplays, videotapes, or live actors. They also recommended the comparison of different workshop formats to
assist in the development of successful rape prevention programming.
Although many rape education/prevention efforts are
targeted towards women and mixed audiences of men and women, few published reports were found that described programs aimed explicitly towards men. Some have argued a need for this type of intervention because (1) males are, in almost all cases, the ones who rape, and (2) men may be more willing to discuss certain issues in an all-male group (Lee, 1987).
Three reports were found which described rape education interventions targeted specifically towards males (Johnson, 1978/1979; Lee, 1987; Gilbert et al., 1991). The purposes of these educational efforts were to increase men's awareness about rape, change rape-supportive attitudes, and/or increase men's empathic understanding of rape victims.
The first of these interventions was a relatively early study investigating the effects of four rape education videotapes on fraternity members' attitudes toward rape and women (Johnson, 1978/1979). Johnson presented four successive films on rape education or drug education, using a pretest, posttest, and delayed posttest design. Johnson was interested in assessing the effects of the videotapes, sex-role identity, and Machiavellianism on male attitudes. Results of the experiment indicated that men who viewed the
rape videotapes became more liberal (profeminist, egalitarian) in attitudes toward rape, but not towards women. Neither sex-role identity nor Machiavellianism was found to have a significant main effect on change scores for attitudes toward women and rape.
Johnson discusses her results in terms of the
inconsistency-based motivational approach to cognitive dissonance theory, an approach that recently has been challenged on conceptual grounds (Scher & Cooper, 1989). The study is limited in that three of the four rape videotapes that were shown focused only on what women can do to avoid rape; the fourth was designed to demonstrate an effective police interview with a rape victim. None of the videotapes addressed the effect of sex-role socialization on males and females. The intervention, thus, may have been inadvertently reinforcing some of the rape-supportive beliefs that it was designed to change.
A more recent account of a rape awareness program
targeting males was described by Lee (1987). Lee's 2-hour experiential program involved a 20-minute didactic presentation on rape myths and facts, a male's reading of a detailed account of being raped, a guided imagery exercise of a date rape situation, and discussion following each of these three parts. Lee included ideas about what men can do to help prevent rape, including discouraging friends from telling jokes about violence toward women, volunteering for
a night escort service, crossing the street when passing a woman walking alone at night, and getting involved in community education regarding rape prevention. Results from very preliminary research suggested that the workshop was effective in changing subjects' attitudes toward rape in the desired direction. These results are limited, however, in that they were based on a sample of only 24 students and the study did not include a control group.
Furthering the research on rape prevention programming for men, Gilbert et al. (1991) designed a study to assess the effectiveness of a psychoeducational intervention in changing the rape-supportive attitudes of male college students. In building upon the research initiated by Lee (1987), Gilbert and colleagues used a slightly larger sample of male undergraduates (n = 61), included a control group, and provided a theoretical framework for their research (Petty and Cacioppols (1986) elaboration likelihood model (ELM) of attitude change). Results showed significantly tore attitude change among males who received the intervention than among males who did not receive any intervention at all. The authors conclude that the elaboration likelihood model of attitude change is a useful framework for developing rape education interventions and understanding attitudinal change processes regarding rape.
While this research is recognized as a significant
first step in applying social psychological theory to rape
prevention research, the study is limited in that it did not include a treatment control group, i.e., a group that received a psychoeducational intervention that was not based on the principles of ELM. The inclusion of such a group would have allowed the researchers to assess whether a presentation based on the principles of the ELM was significantly more effective in changing attitudes than a presentation not based on the ELM.
one other limitation of this study is the sole focus on attitude change as opposed to behavior change. No data were reported on the effects of the intervention on behavior or behavioral intent. The researchers apparently collected data on behavioral intent (self-reported likelihood of raping and using force in a sexual interaction), but chose not to discuss the data as they did not relate to the stated hypotheses. What is needed is a theory that encompasses behavioral change to guide future research on the effects of date rape education programs.
No research was found which compared the effectiveness of male and female presenters of rape education workshops. Lee (1987) used only male presenters with the assumption that participants would be more open to discussion. However, no data were provided to support this assumption. A male and a female team was used by Gilbert et al. (1991) to present their sexual assault educational workshop. No rationale or data were supplied to support this approach.
Research is currently being conducted, however, to compare the effectiveness of male-male and male-female teams in programs targeted solely to males (Lee, 1987).
There seems to be some general consensus that increased attention needs to be placed on targeting males for rape education and prevention programming. In a research study investigating male sex-role orientation, beliefs about rape, and self-reported likelihood of acquaintance and stranger raping, Quackenbush (1989) found that (1) males, in general, viewed stranger rape as more deleterious than acquaintance rape and (2) males who more closely adhered to traditional masculine roles tend to hold more rape-supportive beliefs than do androgynous males. Implications of these findings for rape prevention include (1) increasing males, awareness of the negative consequences of date/acquaintance rape and
(2) increasing male's awareness of and access to feelings, desires, and needs that traditionally have been viewed as appropriate only for females.
In their needs assessment study of services of rape victims in Florida, Martin and colleagues (1984) call for prevention efforts targeted towards potential rape offenders. Such efforts should include adult males or respected male authority figures to reinforce anti-rape messages (especially if the main presenter is female) and education about the consequences of rape. Other recommended activities include panel discussions, public service
announcements, and mass public forums. The use of mass media to inform the public about the facts of rape has received support elsewhere (Youn, 1987).
Among the ten major findings of the needs assessment
study was the fact that little attention was placed on rape prevention in Florida. It was noted that much of the current focus in existing prevention efforts was on rape avoidance, rather than on the conditions which foster rape. Efforts that focus solely on potential victims were criticized on the grounds that they reinforce the myth that victims are responsible for the crime committed against them (Martin et al., 1984).
The researchers noted that rape victim-blaming is
pervasive in Florida, not only among laypeople but among many service providers as well. Their recommendations include not only increased education/prevention efforts, but further research on rape prevention, with particular focus on assessing the differential success of strategies and materials aimed at specific target groups, including potential offenders as well as victims (Martin et al., 1984).
The systematic evaluation of rape education workshops is lacking in the literature. Some authors offer a description of their program as a model for others to emulate (Buhrke & Lustgraaf, 1988) while others may also provide a summary of feedback obtained informally from
workshop participants (Pace & Zaugra, 1988). Few studies have attempted to assess empirically the effectiveness of workshops in changing rape-supportive attitudes, increasing empathy towards rape victims, and altering interpersonal behavior.
Social Learnina Theory and Human Aggression
Although the majority of rape research has been guided by feminist philosophy, few studies on rape have employed social psychological theory to provide direction for the research. There is a great need for a social psychological model of sexual aggression (Malamuth, 1988). Such a model should encompass the general component of aggression, factors specific to violence towards women, and factors that promote aggression towards those regarded as "inferior" or "weak" (Malamuth, 1988).
A review of the literature reveals that the social
psychological theory of human aggression that has received the most abundant and most consistent empirical support to date is social learning theory as proposed by Albert Bandura (Myers, 1987). Briefly, social learning theory postulates that aggression is primarily learned by observing others behaving aggressively and witnessing the reinforcement of such behavior. Aggressive behavior is also shaped and maintained by positive reinforcement from external and internal sources (Bandura, 1973, 1977). A more detailed
account of the principles of social learning theory is outlined below.
Principles of Social Learning Theory
Human aggression is a complex phenomenon originating from a multitude of sources and serving a number of purposes. It can be personally initiated or collectively sanctioned (Bandura, 1979). It is usually defined as "behavior that results in personal injury and physical destruction" (Bandura, 1979, p. 198), although not all acts that result in injury or damage are perceived as aggressive.
To label an act as aggressive depends on perceptions of injurious intent and attributions of responsibility (Bandura, 1973, 1979). The greater the attribution of personal responsibility and perception of injurious intent, the more likely an act will be seen as aggressive (Bandura, 1973). Causal attribution and judgments about harmful intent vary depending on the sex, race, age, attractiveness, and status of the harmdoer. In general, the harmful behavior of highly regarded persons is seen as less intentional and prompted by external circumstances, whereas the harmful acts of disfavored persons are often seen as more intentional and internally motivated (Bandura, 1979). origiins of aggression
Social learning theory emphasizes two mechanisms for acquiring aggressive behavior: observational learning and
reinforced performance (Bandura, 1973, 1979). The principle mechanism of learning aggression is through observation, i.e., watching others model aggressive behavior and then witnessing the consequences of their acts (Bandura, 1973, 1979). observational learning occurs much faster than reinforced performance alone. This process of learning is particularly important in the acquisition of aggressive behavior because a large amount of aggressive behavior requires intricate and complex skills (e.g., sparring with opponents and military combat) and mistakes can be fatal.
Bandura (1973, 1979) outlines four interrelated
subprocesses that comprise observational learning. The first subprocess is attention and involves sensing, perceiving, and exploring the environment. The second subprocess serves a memory function and involves the formation of symbolic representations of sensed experiences. It is the process of coding and storing information. The third subprocess is the motor reproduction of the stored memory, or the integration of the observed acts with actual response repertoires. The fourth and final subprocess of observational learning serves a regulatory function. Essentially, this is the process of deciding whether or not the learned behavior is performed by analyzing the costs and benefits of the enactment. Each of these subprocesses operate in the learning of specific and general behavioral skills (Bandura, 1979).
Bandura (1973, 1979) contends that aggression is
learned by observing models that exist in everyday society and outlines three sources of these models: (1) one's family, (2) one's subculture, and (3) the mass media. Parents are often potent models of behavior for children. Children whose parents use physical aggression as a means of gaining compliance tend to use similar strategies in their interactions with others (Bandura & Walters, 1959). Parents who abuse their children are often survivors of child abuse themselves (Silver, Dublin, & Lourie, 1969).
The subculture in which one lives or with which one has repeated contact can also be a source of aggressive models. Violent gangs and certain all-male enclaves are examples of such subgroups. Subcultures that value traditional male roles and the "macho" image teach younger generations that aggressive behavior is acceptable, functional, and rewarding.
A third source of aggressive models is the mass media. Television, for example, has been shown to be an excellent vehicle for the symbolic modeling of aggression. Research has shown that children and teens who watch televised violence are more interpersonally aggressive in everyday life (Bandura, 1973). Television is also thought to be a vehicle for teaching criminals new skills, a way for them to perfect their crimes (Aronson, 1988; Bandura, 1979). Research on the effects of television reveals that
television can also influence the amount of positive social behavior exhibited. Prosocial programming has been shown to increase cooperation and sharing and to decrease aggression among children (Leiffer, Gordon, & Graves, 1974).
In addition to teaching aggressive styles of behavior, television can also disinhibit persons from behaving aggressively. This occurs when television watchers witness few negative consequences for modeled aggression. For example, when heros are portrayed as winning a great deal and losing little by killing, robbing, and taking revenge, observers learn that violence is not only acceptable, but preferred as a way of achieving goals and solving conflict (Bandura, 1973, 1979).
Television influences aggressive behavior by
desensitizing and habituating persons to violence. Research shows that heavy viewers of television respond with less emotion to violence (Bandura, 1979) and decreased emotional responsiveness can occur after watching just one violent program (Thomas, Horton, Lippincott, & Drabman, 1977). In addition to decreased emotionality, those who watch television a great deal are less likely to intervene when observing human conflict and aggression (Bandura, 1979).
Television also affects our sense of reality. Heavy watchers, compared to light viewers, are less trusting of others and perceive a greater likelihood of personal victimization. Heavy viewers are more apt think that
societal violence is more common than not. Apparently, this difference is not affected by sex, age, educational level, or amount of newspaper reading (Bandura, 1979).
Exposure to modeled aggression influences attitudes and values as well as behavior. It is suggested that modeled aggression influences attitudes towards groups that differ in occupation, race, and other demographic variables. Attitude formation is thought to occur through modeled associations, i.e., by making evaluations of others based on observations of their behavior (Bandura, 1973).
Although modeling can account for the majority of learning of complex human behavior including aggression, behavior is also learned, or shaped, by direct reinforcement (Bandura, 1973, 1977, 1979). In humans, the process of reinforcement is considered primarily to be an informative and motivational operation, rather than a mechanistic one. It is basically a cognitive process of learning response consequences; it allows one to make judgments about the future likelihood of behavioral reinforcement. Bandura (1973, 1977, 1979) suggests that reinforcement and modeling operate together in daily life. Behavior may be learned first through observation and then later refined through reinforcement.
Observational learning and reinforced performance
account for most aggressive behavior in humans. The amount and type of aggression exhibited, however, is somewhat
limited by biological and structural factors (e.g., physical strength and genetic factors that influence speed of learning) (Bandura, 1973). The effects of biology, however, are relatively less among humans than among other animals primarily because of our cognitive capacities. For example, stimulation of the hypothalamus (a part of the brain that helps mediate aggression) is controlled by the central processing of environmental stimuli. How we perceive an event determines whether or not activity in the hypothalamus occurs. Humans also have the capacity to design and build weapons of aggression which further decreases our dependence on biological structure (Bandura, 1979). Instigators of aggression
According to Bandura, there are two basic types of instigators of aggressive behavior: those that are biologically based and those that are cognitively based. Biologically based motivators include tissue defects and pain felt as a result of aversive experience. Cognitively based motivators include mental representations of future consequences and self-generated inducements affected by personal goal-setting and individual performance standards (Bandura, 1979). Cognitively based motivators are considered to account for more aggression than biologically based motivators.
One type of biologically based instigator of aggression is aversive experience. This includes pain, deprivation,
delay or reduction of rewards, personal insults, failures, and obstructions. Aversive experiences can lead to emotional arousal that, together with the pull of anticipated consequences, can facilitate any number of responses including aggression, dependency, achievement, withdrawal, problem-solving, and drug use. The type of behavioral response chosen depends on how the source of arousal is evaluated, what responses one has learned, and the effectiveness of the response. Aggression is more likely if the instigating event is regarded as intentionally injurious and if the individual believes that he or she can do the behavior and that the behavior will lead to the desired consequence (self-efficacy) (Bandura, 1979).
The most common source of cognitively based instigators of aggression are the actions of others. Modeled behavior instigates aggression (1) by informing observers about what behavior leads to desired consequences (directing), (2) by showing observers what behaviors result in a lack of social censorship (disinhibiting), (3) by generating emotional arousal, and (4) by directing observer's attention to particular instruments used in aggressive acts (stimulus enhancing) (Bandura, 1973, 1979). The effect of modeling influences is greater when observers are angered and when the modeled aggression is socially justified and rewarded (Bandura, 1979).
In addition to modeling cues and aversive experience,
aggression can be influenced by instruction. Because people learn to obey orders (through rewarding compliance and punishing disobedience) aggression on command is possible. In fact, some of the most horrific acts in human history occurred due to obedience to authority. Obedient aggression declines when the harmful consequences of aggression become more salient and personalized (Bandura, 1979).
A less common antecedent cue for aggression is delusional thought. Examples of delusional types of instigators include divine inner voices, paranoid beliefs, and grandiose convictions about one's heroic duty to eliminate evil (Bandura, 1973, 1979). These types of instigators occur much less frequently than modeled cues and are basically limited to persons with psychotic thinking.
Finally, aggressive behavior can be motivated by selfgenerated inducements influenced by personal goals and performance standards. The anticipation of self-rewards based on successful goal attainment can serve as a pull for aggression (if aggressive behavior is considered an acceptable means of accomplishing one's goals). Likewise, self-praise may be an effective inducement if aggressive behavior is a valued part of one's role or identity. Regulators of aggression
Bandura identifies three types of behavioral
consequences that serve to regulate behavior: external
consequences, observed consequences, and self-produced consequences (Bandura, 1973, 1977, 1979). One can, thus, experience, witness, and create consequences. Behavioral consequences may be positive or punishing. They may also be relatively concrete or abstract. Following is a discussion of the various maintaining mechanisms, or regulators, of aggression.
External consequences of aggression include positive reinforcement and punishment. Positive reinforcement may include tangible rewards (money, food, and drink), social and status rewards (approval, promotions, awards), reductions of aversive treatment (relief from misery), and expressions of pain among victims. Punishment can originate from one's social environment or from within (i.e., selfcondemnation). Both positive reinforcement and punishment influence aggression by creating expectations of similar outcomes on future occasions. Aggression is more likely when the expected benefits are great and the anticipated punishments are few (Bandura, 1973, 1979).
Like external consequences, observed consequences of
aggression also may be positive or negative. The difference here, of course, is that learning occurs vicariously rather than through direct experience. As with external regulators of aggression, observed reinforcement operates primarily through its informative function. The likelihood of aggression is increased when one witnesses the reinforcement
of aggression and is decreased when one witnesses the punishment of aggression (as long as the punishment occurs infrequently). The frequent use of punishment may inadvertently promote aggressive behavior by modeling punitive methods of control (Bandura, 1973, 1979).
The third type of regulator of aggressive behavior is self-produced reinforcement. Self-generated reinforcement results from a cognitive process of judging one's own behavior against referential standards. The first stage of this process is self-observation. Particular aspects of one's own behavior that may be considered are quality, quantity, rate, originality, deviancy, and ethicalness. The second stage involves making a judgment about one's own behavior. This occurs by comparing one's own behavior to personal standards, the standards of others, or the behavior of others; making a value judgment of the behavior (i.e., positive or negative); and making an attribution of the performance (i.e., internal or external locus of control). The final stage is making a self-response. This may include positive or negative self-evaluations, rewarding or punishing tangible consequences, or choosing no response at all (Bandura, 1979).
It is theorized that aggressive acts may occur through a process of disengagement of internal control (Bandura, 1973, 1979). There are several ways in which this may happen. The aggressive behavior may be morally justified,
contrasted with more horrible acts, or disguised by the use of euphemistic language. Responsibility may be displaced onto others or diffused among a crowd. The impact of the aggressive act may be minimized, ignored, or distorted. And, finally, internal control may be disengaged by dehumanizing or blaming the victim. It is suggested that humanizing or personalizing victims may be an effective strategy in counteracting aggression (Bandura, 1979).
Psychological functioning is characterized by a continuous reciprocal interaction between cognitive, behavioral, and environmental determinants. All three aspects are considered interdependent and mutually influential. People, thus, have the capacity to play active roles in affecting their environment, their own behavior, and the behavior of others. Social learning theory considers behavior change to be possible, and compared to more traditional theories of aggression, social learning theory is relatively optimistic about the chance of decreasing aggression in our society.
Social Learning Theory and Aggression Research
There is substantial empirical evidence to support a social learning model of human aggression. The earliest experiments were conducted by Bandura in a series of "Bobo" doll studies that were designed to investigate the impact of modeled aggression on children's behavior. Bandura and
colleagues (Bandura, Ross, & Ross, 1961) noted that children who observed an adult model aggressive behavior towards an Bobo doll and who were then intentionally frustrated by an experimenter, were more likely to behave aggressively than were similarly frustrated children who did not observe an adult behave aggressively. moreover, those children who watched an aggressive adult often imitated the very same actions and words used by the adult. Bandura's experiments support the hypothesis that the observation of aggressive behavior not only lowers inhibitions, but also teaches specific strategies for behaving aggressively (Bandura et al., 1961).
Interestingly, male subjects who were exposed to a nonaggressive male model in this experiment exhibited significantly less physical and verbal aggression than did males who were not exposed to any model (control group males). Furthermore, subjects who watched a nonaggressive model engaged in significantly more nonaggressive play with dolls than did subjects who either observed an aggressive model or did not observe any model at all (Bandura et al., 1961). One clear implication of these results for interventions aimed at reducing male violence and aggression is to provide nonaggressive or nonviolent male models for potentially aggressive males.
Bandura et al. (1961) also noted sex differences in the learning of aggressive behavior. They found that an
aggressive male model was more effective in influencing male and female subjects than an aggressive female model. They suggested that this was due to the fact that aggressive behavior is more "masculine-typed." They suggested further that, for behavior that is less clearly sex-linked, the greatest amount of imitation occurs by observing a same-sex model. Following this line of reasoning, one might suppose that for behavior that is more traditionally "femininetyped" (e.g., relationship-building skills) the greatest amount of learning may occur by observing a female model. Such a hypothesis for male learners of relationship skills, however, contradicts the notion that learning is greatest when observers perceive themselves to be similar to the model.
The effects of watching violence are not limited to aggressive behavior exhibited towards "Bobo" dolls. They extend to interpersonal aggression as well. In one study, children were shown a violent episode of "The Untouchables" (a "cops-and-robbers" show), and then were allowed to play with a group of other children (Liebert & Baron, 1972). Results showed that compared to a control group who watched an action-oriented sporting event, those children who watched the violent television program displayed far more aggression towards the other children.
In another study, juvenile delinquent males in
detention centers were shown either violent or nonviolent
movies. Those boys who watched the violent films displayed significantly more physical and verbal aggression towards others than did the boys who watched the nonviolent films (Parke, Berkowitz, Leyens, West, & Sebastian, 1977). These results support the conclusion that viewing filmed or televised violence can lead to increased displays of aggression.
Research shows that observation of aggressive models
not only increases aggression, but it also provides cues as to the characteristics of potential victims of future violence. In an investigation of the effects of highly publicized prizefights, Phillips (1983) found that in the days following the prizefights, there was a significant increase in the number of homicides. The race of the losers of these prizefights was related to the race of the victims of subsequent murders. When white boxers lost the matches, there was an increase in the number of white men murdered, but not in the number of black men killed. When black boxers lost the fights, there was an increase in the number of black men murdered, but not in the number of white men slain.
Social Learning Theory and Ra2e
There is a growing body of research to support a social learning model of sexual assault. Much of this support comes from investigations of the modeling effects of violent
pornography. Recent studies of the reinforcing properties of all-male groups on male aggression against women lend further support to a social learning analysis of sexual violence. Below is a discussion of this research. Effects of Modeling
There is increasing evidence to suggest that sexual violence against women is symbolically modeled through violent pornography. Malamuth, Donnerstein, and colleagues have conducted a series of studies which, taken together, support the conclusion that exposure to violent pornography promotes greater acceptance of sexual violence toward women and is associated with increased aggression toward women in natural and laboratory settings.
In a study comparing the effects of exposure to
aggressive-erotic, erotic, and neutral films on aggressive behavior in a laboratory setting, Donnerstein (1980) found that the men who watched an aggressive-erotic film (rape film) later displayed the most intense aggression, but only towards female confederates. Apparently, the aggressiveerotic film not only modeled aggressive behavior, but provided information about the gender of "appropriate" victims as well. No differences in aggression levels were obtained between nonangered subjects who watched the erotic and neutral films. These findings are consistent with those of the President's Commission on Obscenity and Pornography
(1970) which concluded that exposure to sexually explicit material (in and of itself) does not contribute to antisocial behavior.
In a field experiment, the effects of exposure to violent-sexual films that portray sexual aggression as having positive consequences were investigated (Malamuth & Check, 1981). Male and female college students were randomly assigned to watch either violent-sexual or control films shown as part of the regular campus film program. one week later, subjects were surveyed regarding their attitudes about rape, violence, and sexual relations. Findings showed that males who watched the sexually violent films were more accepting of interpersonal violence towards women and more accepting of certain rape myths. Women had tendencies in the opposite direction. These results provide some evidence to support the contention that the observation of modeled aggression can influence attitudes in a real-world setting.
Further support for the influence of symbolic modeling on attitudes was provided by the results of an investigation of the effects of aggressive pornography on men's beliefs about rape (Malamuth & Check, 1985). During this experiment, subjects first listened to one of eight audiotapes passages and then listened to another passage depicting either rape or mutually consenting sex. Afterwards, subjects' perceptions of the second passage and their attitudes about rape were assessed. Results provided
support for the hypothesis that media depictions can influence men's beliefs in rape myths. Men with relatively higher inclinations to aggress against women were particularly likely to be affected by media depictions of rape myths.
Demare and Briere (1988) examined the relationship between pornography use, attitudes, and self-reported likelihood of raping and using sexual force among male undergraduate students. Results showed that use of sexually violent pornography (as measured by self-report) and acceptance of interpersonal violence against women were both associated with self-reported likelihood of engaging in sexual aggression against women. Although the correlational nature of the study precludes any conclusions regarding causal antecedents of sexual aggression, the findings do point out a relationship between observed sexual violence and inclinations toward future sexual aggression.
In addition to the media as a source for the modeling
of sexual violence, peers can also serve as models of sexual aggression. It is suggested that inclusive male groups such as fraternities may provide a context for modeling and reinforcing rape-supportive behavior (Boeringer, Shehan, & Akers, 1991). Because fraternity members, compared to nonmembers, are more likely to have friends who engage in verbal coercion and who use drugs or alcohol to gain sexual access, it is likely that fraternity members have more
opportunities to observe such behaviors modeled and reinforced.
One's parents or caretakers can also be a source of modeled sexual aggression. In a study investigating the relationship between child sexual abuse and sexual coercion experienced as a young adult, researchers found a significant relationship between histories of child sexual abuse and reported acts of sexual coercion in college. No relationship was found between histories of child sexual abuse and reported sexual victimization at college (Miller & Marshall, 1987).
Effects of Reinforcement
One potential source of reinforcement for sexually
aggressive behavior is one's peer group. Such reinforcement can take a variety of forms. Positive consequences for sexual aggression may include social approval, affirmation of one's masculinity and manhood, and increased social status. Negative consequences for refusing to engage in sexual aggression may include ostracization, rejection, and ridicule.
Evidence to suggest that the social learning principles of differential association and differential reinforcement do indeed play significant roles in the development and maintenance of sexually aggressive behavior among college males was provided by Boeringer and colleagues in a study
conducted at the University of Florida (Boeringer, Shehan, & Akers, 1991). These researchers designed a study to assess the impact of fraternity membership in the adoption of sexually aggressive behavior. They compared fraternity members' and nonmembers' perceptions of their friends use of alcohol to gain sexual access, anticipated social approval for sexually aggressive behavior, self-reported pornography use, and acceptance of rape myths. Results showed that fraternity members, compared to nonmembers, anticipated a greater likelihood of positive reinforcement for sexual promiscuity and sexual aggression and less likelihood of disapproval for the use of drugs or alcohol to obtain sex.
Boeringer and researchers (1991) demonstrated that when the social learning variables (i.e., differential association, differential reinforcement, modeling, and beliefs about sexual aggression) were statistically controlled, the effect of fraternity membership on selfreported likelihood of using force or committing rape was nonsignificant. Similarly, when controlling for the social learning variables, the effect of fraternity membership on the use of alcohol and drugs, verbal coercion, and physical force in gaining sexual access was nonsignificant. The researchers conclude that social learning is the process through which fraternity members come to engage in sexual coercion and aggression.
Implications for Date Rave Prevention Programming
Results from research supporting the application of social learning theory to the understanding of sexual coercion and aggression in males has several implications for date rape prevention programming. In this discussion the focus will be on date rape interventions for male college students.
one important implication of social learning theory for rape prevention efforts is the provision of appropriate models for target audiences. Because observational learning is important in the learning of complex behavior patterns, and because responsible, mature, and appropriate interpersonal behavior requires complex behavioral skills, target audiences must have the opportunity to observe models demonstrate these behaviors. It may also be useful to have models demonstrate alternative coping skills and strategies for dealing with anger. Theoretically, the most successful models will be peers. Regardless of who the model is, it is imperative that the observed behavior be positively reinforced.
Following the demonstration of the new behaviors,
observers (males) should have the opportunity to practice the newly learned skills. Such practice efforts need to be appropriately reinforced so that the learned behavior can be perfected. As part of this learning process, males must be educated regarding what constitutes sexual aggression
(including verbal coercion, the use of alcohol as a "weapon," and the use of physical force). In this way, they will learn to discriminate better between behaviors that are acceptable and unacceptable. Boeringer et al. (1991) suggest that the differential reinforcement of anti-rape attitudes and behaviors may be effective in reducing rapesupportive behavior.
Negative consequences for rape and other forms of sexual aggression need to be made more salient. It is important that persons observe that rape will not be tolerated. The consistent and firm use of punishment for sexual aggression may help change referential standards held by potential assailants, and perhaps eventually, alter internal self-standards of behavior.
Attention must be paid to the existing social
environment in which students reside. Faculty advisors and leaders of fraternities and other all-male groups are strongly encouraged to become aware of and modify any existing aspects of the social environment that may positively reinforce sexual aggression (Boeringer et al., 1991). To encourage change in the social environment, it is recommended that sanctions be levied against organizations (including fraternities) that provide social support for individual members who engage in sexually aggressive behavior (Boeringer et al., 1991). Punishment for violencesupporting organizations may help shape the social context
in which individual members' behavior is learned and maintained. Positive reinforcement for organizations that have clear and effective anti-rape policies may also be effective.
Strategies aimed at blocking the process of
disengagement from internal control may also be useful tactics in preventing rape. Such methods include personalizing and humanizing the rape victim (perhaps by educating males on rape trauma syndrome), increasing awareness regarding the effects of alcohol on dating behavior (to reduce the likelihood of diffusing responsibility), increasing awareness regarding rape language (to decrease the use of euphemisms that conceal the reality of rape), and increasing awareness about rape (to decrease victim-blaming).
In summary, social learning theory provides a
reasonable model for understanding the development and prevention of sexually aggressive behavior. Bandurals model for understanding the origins, instigating factors, and regulatory mechanisms of human aggression give us a framework for understanding how sexual aggression is developed and maintained. It also provides ideas as to how such behavior may be prevented. It is now up to research to test empirically the adequacy of social learning theory as a framework for understanding rape and rape prevention approaches.
Subjects were 189 male undergraduate students enrolled in introductory psychology classes at the University of Florida. The sample consisted of Caucasians (n = 148, 79.6%), Hispanic Americans (n = 15, 8.1%), African Americans (D = 7, 3.8%), International students (n = 5, 2.7%), and Native Americans (n = 3, 1.6%). A small number (n = 8,
4.3%) did not report their ethnicity. Almost half of the sample were freshmen (D = 88, 47.6t), while smaller numbers were sophomores (D = 50, 27.0t), juniors (n = 28, 15.1%), and seniors (11 = 19, 10.3%). The average age for subjects was 19.2 years (S.D. = 1.67); ages ranged from 17 to 27.
Fraternity membership was reported by 40 (21.5%) subjects and previous attendance at a workshop on date/acquaintance rape was reported by 34 (18.3%) subjects. A substantial number (n = 72, 38.9%) indicated that they knew a rape survivor. For those that knew a survivor, the survivor was said to be an acquaintance (n = 45, 58.4%), a classmate (n = 25, 32.9%), a date 24, 31.6%), or a
family member (n = 6, 7.8%).
The vast majority of subjects reported no previous
sexual activity (including fondling, kissing, petting, and intercourse) that was performed or attempted without a woman's consent (n = 143, 77.7%). A full 22.3% (D = 41), however, reported having engaged in or attempted some sexual activity against a woman's will. More specifically, of these males, 27 (14.5%) indicated that they had engaged in sex play (fondling, kissing, and/or petting) with a woman by coercion or threat of physical force; 21 (11.3%) reported having engaged in sexual intercourse with a woman by verbal pressure or position of authority; 9 (4.8%) stated that they had engaged in activities that met the definition of attempted rape; and 7 (3.8%) acknowledged having engaged in behaviors that met the criterion for rape. Because some subjects had engaged in more than one type of sexually coercive behavior, the total percent above exceeds 100%.
A Demographic Information Form (Appendix A) was used to collect demographic data on the research subjects. This form consisted of items regarding subject age, race, academic class level, fraternity membership, personal knowledge of rape survivors, and previous attendance at a date/acquaintance rape workshop.
The Sexual Experiences Survey (SES) (Koss & Gidycz, 1985; Koss & Oros, 1982) was used in the sample selection
procedure to ensure that intervention groups were matched on previous experience with sexually coercive behavior (Appendix B). The Sexual Experiences Survey is a 10-item instrument designed to assess varying degrees of sexual aggression and victimization. It was developed to aid in the identification of "hidden" rape victims and undetected offenders and to support a dimensional, as opposed to a typological, view of sexual aggression (Koss & Oros, 1982).
Each of the ten items on the SES are to be answered either "yes" or "no." Respondents are then classified according to the most severe type of sexual aggression reported. "Sexual contact" is the label given to those whose response is "yes" to items 1, 2, or 3, but not to any higher numbered items. "Sexual coercion" is the category for those who answer "yes" to items 6 or 7, but not to any higher numbered items. "Attempted rape" is the group label for those giving "yes" responses to items 4 or 5, but to no higher item. And finally, "rape" is the classification for those saying "yes" to items 8, 9, or 10.
Internal consistency for the SES was found to be .89 for males and test-retest reliability with a one-week interval was found to be .93 (Koss & Gidycz, 1985). Initial investigation of the validity of the SES revealed a Pearson correlation between a male's self-report on the SES and responses given in the presence of a male psychologist/ interviewer of .61 (Koss & Gidycz, 1985). More recently, a
93% agreement was found between male subjects' responses on the SES completed in private and responses given in the presence of a male interviewer. These subjects rated their honesty at 95% and explained that the reason for a lack of full honesty was due to time pressure to complete the questionnaire (Koss, Gidycz, & Wisniewski, 1987).
The Rape Myth Acceptance Scale (RMAS) (Burt, 1980) is a 19-item measure designed to assess adherence to rapesupportive beliefs (i.e., false, prejudicial, and stereotyped beliefs about rape, rape victims, and rapists) (Appendix C). Items on the RMAS are to be answered on 5- or 7-point Likert scales ranging from "strongly agree" to "strongly disagree," "almost all" to "almost none," and "always" to "never." Responses are converted (to correct for item direction) and summed to create an index from 19 (low rape myth acceptance) to 117 (high rape myth acceptance). Mean scores for a sample of 598 adults, aged 18 and over, was 49.4 with a standard deviation of 11.9. Cronbach's alpha with this sample was calculated to be .875 (Burt, 1980).
In a previous study, validity testing of the RMAS
yielded predicted results. That is, scores on the Rape Myth Acceptance Scale were found to be significantly and positively correlated with scores on a dogmatism scale (r.51, p < .05), and significantly and negatively correlated with scores on a scale measuring trustworthiness (r =-.46,
R < .05) (Ashton, 1982). To some extent, acceptance of rape myths is associated with a closed-minded inability to evaluate information critically and objectively, and a tendency to view others with suspicion and, perhaps, as dishonest (as many rape victims are perceived).
The Acceptance of Interpersonal Violence Scale (AIV) (Burt, 1980) is a 6-item measure that was developed to assess one's belief that force and coercion are legitimate ways to gain compliance, particularly in intimate, sexual relationships (Appendix D). Responses to this scale are made on a 7-point Likert scale ranging from "strongly agree" to "strongly disagree". Responses are converted and summed to create an index from 6 (low acceptance of interpersonal violence) to 42 (high acceptance of interpersonal violence). Mean scores for a sample of 598 adults was 18.2 with a standard deviation of 5.9. With this sample, Cronbach's alpha was found to be .586 (Burt, 1980).
Two items were administered to assess subjects' selfreported likelihood of using force in a sexual interaction. These items were adaptations from items used in previous research to investigate self-reported likelihood of using force (LF) (Briere & Malamuth, 1983). The items were prefaced with the following question: "If you could be assured that no one would know and that you could in no way be punished for engaging in the following behaviors, how likely is it that you would do them?" The items of interest
were: (1) "pressuring a woman with continual verbal arguments to obtain sexual intercourse," and (2) "using some degree of physical force (i.e., twisting a woman's arm, holding her down, etc.) to obtain sexual intercourse." These items were chosen to distinguish between instances of sexual assault achieved by way of verbal coercion and instances of sexual assault gained by way of physical force. Each of the two items was to be rated on a 5-point Likert scale ranging from 1 = "not at all likely" and 5 = "very likely."
Test-retest reliability for the likelihood of force
item (LF) developed by Malamuth was found to be .74. Tests of construct validity yielded significant relationships between the LF item and behavioral measures of previous sexual aggression (.34) and future intent (.60) (Malamuth, 1989).
The Marlowe-Crowne Social Desirability Scale, shortform (M-C (20]) (Strahan & Gerbasi, 1972), is a 20-item measure designed to assess the need for positive selfpresentation via culturally acceptable and approved behaviors that are unlikely to occur (Appendix E). Based on the original 33-item instrument (Crowne & Marlowe, 1960), the 20-item scale was developed in order to find a measure that was more practical for research, yet still valid and reliable. Kuder-Richardson formula 20 (K-R 20) reliability coefficients for the short version closely approximated
those for the original 33-item inventory, and were .78 and .83 respectively for university males. Correlations between the 20-item scale and the original scale were all in the .90's (Fraboni & Cooper, 1989; Strahan & Gerbasi, 1972) suggesting adequate construct validity for the shorter scale.
A 10-item information test (Appendix F) was used after each presentation intervention to ensure that subjects attended to and understood the material that was presented. The test was an assessment of subjects' knowledge regarding date/acquaintance rape. All questions were designed to be answerable upon attentive listening to the presentation. Each correctly answered item was given a score of I'1", thus yielding a possible score range from 0 (no correct answers) to 10 (all correct answers).
Following each presentation intervention, a 1-item
question was also administered to assess subjects' perceived similarity to the presenter. Responses to the item, "How similar do you perceive yourself to be to the main presenter of this workshop?" were made on a 5-point Likert scale ranging from 1 = "not at all similar" to 5 = "very similar." This item served as a validity check for the male (more similar presenter) vs. female (less similar presenter) manipulation.
A preliminary screening measure that included
demographic items and a measure of previous experience with sexual coercion (The sexual Experiences Survey (SES), Koss & Oros, 1982; Koss & Gidycz, 1985) was administered to 453 male college students enrolled in introductory psychology classes. Test administration was conducted by the experimenter at the beginning of two different semesters during which these classes were offered and was conducted during the class hour designated as pretesting time for psychology experiments using this pool of potential subjects.
In order to protect students' anonymity, all response forms were coded with subject-generated 6-digit code numbers. These code numbers were used in subsequent phases of the study for those students who opted to participate further in the study. These code numbers were the only means by which subject data could be tracked across data collection points.
Using a block randomization sampling procedure, all 453 male college students who completed the preliminary screening measure were assigned to one of six groups such that groups were matched on race and previous use of sexual coercion. Scores on the preliminary screening measure were used to match the groups.
of the six different groups involved in the study, four were intervention groups and two were control groups. The four intervention groups involved only two different types of interventions, but each type was implemented separately by a male and a female, thus resulting in four groups. Each of the two intervention control groups was also run separately by a male and a female. Because of logistical considerations (e.g., having a workshop size that allowed interaction between and reinforcement of participants), each of the six groups was subdivided into four subgroups with an average size of about eight subjects. The four subgroups within a group received the same intervention.
The resulting twenty-four subgroups (four subgroups for each of the six larger groups) were randomly distributed across time in such a manner as to ensure that (1) an equal number of subgroups representing each larger group occurred during any one semester of data collection and that (2) when more than six subgroups occurred in a semester (i.e., when 18 subgroups occurred during the fall semester), then subgroups representing each of the larger groups occurred with equal frequency at the beginning, middle, and end of that semester. Different numbers of subgroups (i.e., 6 vs. 18) occurred across semesters due to subject availability.
Of the 453 students who completed the preliminary screening measure, those who were interested in participating further were asked to "sign up" using a
department-assigned 4-digit code number that allowed them to receive course credit for their participation. Everyone who completed the preliminary screening measure was eligible to continue in the study.
Subjects who indicated an interest in participating
further in the study were asked to meet (by subgroup) in a conference room located in the Psychology Department. After a brief overview of the study given by the experimenter, all subjects who agreed to participate further in the study were asked to sign an Informed Consent Form (Appendix G). This form and all other assessment instruments used in the study were administered by the experimenter in the same conference room for every group involved in the experiment.
Subjects were then given the following assessment battery: the Rape Myth Acceptance Scale (RMAS), the Acceptance of Interpersonal Violence Scale (AIV), and the Likelihood of Force items (LF). This assessment battery constituted the baseline dependent measures. At the same time, a 20-item short version of the Marlowe-Crowne Social Desirability Scale was administered to measure the extent to which subjects were responding in a socially favorable manner.
In order to control for a possible order effect to the dependent measures, the RMAS, AIV scale, and the LF items were administered in one of three different sequences to each subject. The three different sequences were: (1) RMAS,
AIV scale, and LF items, (2) AIV scale, LF items, and RMAS, and (3) LF items, RMAS and AIV scale. This procedure was performed at each of the three different testing intervals (described below).
The independent variables involved in the study were
(1) workshop/presentation type and (2) gender of presenter. (There were three workshop/presentation types; each was implemented separately by a male and a female presenter, thus resulting in six groups/conditions.) The three levels of the first independent variable (workshop/presentation type) were the following:
(1) Information only. This intervention was a 1 hour and
15 minute presentation of information regarding date rape myths, rape trauma syndrome, and strategies for
reducing the likelihood of date rape and aggressive
interpersonal behavior (Intervention I). The
information was presented in verbal and written form.
(2) Information Plus modeling and reinforcement of rape
Preventive behaviors and attitudes. This
intervention involved Intervention I plus (a) the
videotaped modeling of assertive nonaggressivee)
interpersonal dating behaviors by a male and female, (b) the presenter's positive verbal reinforcement of nonaggressive, anti-rape behaviors (including verbal statements) modelled by the videotaped persons, and (c) the presenter's positive verbal reinforcement of
subjects' comments conveying anti-rape behaviors and attitudes (Intervention II). The subjects comprised
the audience. This intervention was also about 1
hour and 15 minutes long.
(3) Videotapes unrelated to rape or interpersonal
violence (Intervention Control). This group did not
receive a workshop/presentation on date rape during
the data collection period. Instead, subjects in
this group watched two half-hour videotaped
presentations on topics unrelated to date rape and
interpersonal violence. Between 5 and 10 minutes
were reserved for comments and discussion about the
The information presented during Interventions I and II included rape myths and facts, the effects of rape on victims, and rape prevention strategies for men and women. Two examples of the rape myths discussed were, "Rape happens only rarely" and "Rapists are motivated by a need for sexual release." Rape trauma syndrome was presented as a stage model that outlined the emotional, cognitive, and behavioral impact of rape on victims. The effects of rape on family, friends, and partners of victims was also acknowledged. Strategies for reducing the likelihood of sexual aggression included: "Know your sexual desires and limits" (men and women), "Assume that 'no' means 'no,' not maybe" (men), "When in doubt, ask" (men), "Say "no, if you mean 'no"
(women) and "Know your rights" (women). The information in Interventions I and II was presented in a manner that allowed subjects the opportunity to make comments or ask questions throughout the presentation.
The videotape shown during Intervention II included a male and a female modeling assertive interpersonal behavior in a dating situation. The modelled behavior included an honest expression of feelings regarding a potential sexual interaction with the date, assertive limit-setting, and a show of respect for each other's feelings and desires. Examples of the presenter's positive verbal reinforcement for this behavior were statements such as, "In this video, the male seemed very relaxed, confident, honest, and respecting" and "You know, what I really like about this way of doing it is that it relieves the male of the responsibility of guessing what his date really wants."
The Intervention Control group with the male presenter/ facilitator was shown two half-hour videotaped presentations by a male on topics unrelated to date rape and interpersonal violence. Similarly, the Intervention Control group with the female presenter/facilitator was shown two half-hour videotaped presentations by a female also on topics unrelated to the dependent measures.
Control group subjects were invited to attend a date rape workshop/presentation conducted by the principal investigator after the study was completed. Participation
in this "extra" workshop was voluntary. No data were to be collected during this presentation. Only one subject (1.6%) from the control groups expressed interest in attending such a workshop and was met with individually to share information presented in Intervention I and to address specific questions and concerns raised.
The second independent variable in this study was
gender of presenter. As stated earlier, each of the three workshop/presentation types was conducted separately with a male presenter and a female presenter; therefore, the total number of groups was six (four intervention groups and two intervention control groups). The male and female presenters were matched on age, race, and level of expertise (e.g., similar educational backgrounds and experience).
Immediately following the interventions (excluding the volunteer workshop for Intervention Control subjects at the end of the study), subjects were asked to complete the same assessment battery administered at baseline (the RMAS, AIV scale, and LF items), thus providing immediate posttest data. In addition, subjects were asked to report their perceived similarity to the presenter on a one-item Likert type question. Subjects in Interventions I and II were tested on their acquisition of workshop information to ensure that the intervention was effectively implemented (i.e., that the subjects attended and learned most of the information that was common to Interventions I and II).
Subjects in the Intervention Control group were not required to complete the information test. The duration of each intervention and control group experimental session plus baseline and immediate posttest assessments was approximately I hour and 45 minutes.
one week after immediate posttest data collection, all subjects were asked to attend a follow-up meeting (1) to complete anonymously the same assessment battery administered at baseline and at immediate posttesting (RMAS, AIV scale, and LF items) and (2) to be debriefed. This delayed posttest (i.e., the third administration of the assessment battery) was designed to assess any change maintained over time and to control for the demand effect of immediate evaluation of the workshop. Subjects who may have had personal concerns as a result of participating in the study were invited to discuss these concerns with the principal investigator individually. Subjects would have been referred to the Counseling Center if the principal investigator deemed this to be necessary; no subjects were viewed as requiring this referral. All subjects received course credit for their participation after completion of the study.
The design for the experiment was a 2 X 3 repeated measures factorial design. The independent variables
involved in the study were (1) type of presentation (information only vs. information plus modeling and positive verbal reinforcement vs. two videotapes unrelated to rape and interpersonal violence) and (2) gender of presenter. The dependent variables were (1) acceptance of rape myths (as measured by the RMAS), (2) acceptance of interpersonal violence (as measured by the AIV scale), and (3) selfreported likelihood of using force in a sexual interaction (as measured by the LF items).
A model of the experimental design including the distribution of subjects across groups is as follows: Intervention I
Condition A (male presenter) Group 1 (n = 35) Condition B (female presenter) Group 2 (n = 31)
condition (male presenter) Group 3 (n = 29)
condition B (female presenter) Group 4 (n = 33)
condition A (male presenter) Group 5 (n = 30) Condition B (female presenter) Group 6 (n = 31)
Each of the six groups was comprised of 4 subgroups ranging in size from 3 to 14 subjects.
A summary of the assessment battery administration schedule is as follows:
Prescreening Demographic Form SES
Pretest RMAS AIV LF M-C (20)
Immediate Posttest RMAS AIV LF IT PSP
Delayed Posttest RMAS AIV LF
In the above summary, IT = Information Test and PSP Perceived Similarity to Presenter item. The Information Test was not given to the Intervention Control groups.
Specific hypotheses tested were:
(1) Subjects who are exposed to Intervention II will have
lower scores on all dependent measures (i.e., accept fewer rape myths, be less accepting of interpersonal
violence, and report less likelihood of using force in a
sexual interaction) at immediate and delayed posttest
assessments than subjects exposed to either Intervention
I or to the Control Intervention only.
(2) subjects who are exposed to Intervention I will have
lower scores on all dependent measures at immediate and
delayed posttest assessments than subjects exposed to
the Control Intervention only.
(3) Subjects receiving Interventions I or II from a male
presenter will have lower scores on all dependent
measures at immediate and delayed posttest assessments
than subjects receiving the same intervention from a
(4) Baseline measures of self-reported likelihood of using
force, acceptance of rape myths, and acceptance of
interpersonal violence will all be significantly
of the 453 students who completed the Demographic
Information Form and the Sexual Experiences Survey as part of the sample selection procedure, 189 (41.7%) chose to participate in the experiment. No significant differences were found between those who completed the experiment and those who did not on any of the following preliminary test measures: age, race, academic class level, fraternity membership, personal knowledge of a rape survivor, previous attendance at a date/acquaintance rape workshop, and previous experience with sexual coercion. Among students who knew a rape survivor, those who participated in the experiment were less likely to describe the survivor as an acquaintance (58.4%) than were persons who did not participate in the experiment (74.0%), X2 (1, N = 173)
4.66, R < .03.
Of the 189 subjects who participated in the experiment, only 15 failed to return for the follow-up (1-week delayed posttest assessment), yielding a completion rate of 92.1%. One additional subject did not finish the entire RMAS (at
baseline) and thus, was omitted from analyses involving this variable. The overall usable data ratio, therefore, was .915.
Subjects were fairly evenly distributed across
intervention groups with 66 (34.9%) receiving Intervention 1, 62 (32.8%) receiving Intervention II, and 61 (32.3%) receiving the Intervention Control condition. Approximately half of the sample was exposed to a male presenter (n 94, 49.7%) and half was exposed to a female presenter (n 95, 50.3%). (Cell sizes for each of the six groups are listed in the Research Design section of Chapter 3.) Fifty-two (27.5%) subjects participated during the summer semester and 137 (72.5%) subjects participated in the fall.
The three different test sequences for the RMAS, AIV scale, and LF items were evenly distributed across the six intervention groups at pretest, posttest, and delayed posttest assessments. A manova used to test the hypothesis of no overall order effect revealed no significant differences between groups, indicating that test sequence did not account for a significant amount of variance in scores on the dependent measures.
Tests of the Hypotheses
Four repeated measures ancovas were computed in order to test hypotheses 1, 2, and 3. As stated previously, hypothesis 1 was that subjects receiving Intervention II
would have lower scores on the dependent measures at both posttest assessments than would subjects exposed to Intervention I or to the Control Condition; hypothesis 2 was that subjects receiving Intervention I would have lower scores on the dependent measures at both posttest assessments than would subjects in the Control Conditions; and hypothesis 3 was that subjects in the male presenter condition would have lower scores on the dependent measures at both posttest assessments than would subjects in the female presenter condition. The covariates for these analyses were baseline scores on the RMAS, AIV scale, and LF items. The independent variables were intervention type, gender of presenter, and semester. The semester variable was included to assess potential score differences between students enrolled in psychology classes during different semesters. The semester variable was to be discarded if preliminary analyses indicated no overall effect for semester and no interaction between semester and any other independent variable.
Results from the repeated measures ancova with rape myth acceptance (RMA) as the dependent variable showed a significant presenter by semester interaction effect, E(l, 161) = 4.04, R < .05, with less disparity in RMAS scores between the two presenter conditions during the fall semester than during the summer term. There was no overall effect for intervention type or any other 2- or 3-way
interaction on RHAS scores. The interaction between presenter and semester suggests that the data were affected by the semester during which the data were collected, thus semester remained in the model for further data analysis.
Univariate tests for within subjects effects showed a significant overall time effect for RMAS scores, F(2, 322) 20.71, R < .0001, but no significant differences in scores over time due to presenter by semester interaction. Thus, changes in RMAS scores over time did not affect the nature of the presenter by semester interaction. These results suggest that changes in RMAS responses across time were due to factors not tested in the present model.
Results from the repeated measures ancova with
acceptance of interpersonal violence (AIV) as the dependent variable revealed no significant differences between groups. There was no overall effect for presenter, intervention type, semester, or any 2- or 3-way interaction on the AIV scores.
Similarly, results from the repeated measures ancova with likelihood of using verbal coercion (LFl) as the dependent variable also revealed no significant differences between groups. There was no overall effect for presenter, intervention type, semester, or any 2- or 3-way interaction on this LF item.
In addition, results from the repeated measures ancova with likelihood of using physical force (LF2) as the