It is true rape is a most detestable crime, and therefore ought severely and impartially to be punished with death; but it must be remembered, that it is an accusation easily to be made and hard to be proved, and harder to be defended by the party accused, tho never so innocent. (Hale 1847, p. 634)

Sexual assault affects one in five U.S. women during their lifetime (Black et al. 2011). Among crimes, sexual assault is unique in that the accuser in a rape case is examined for fault in the situation as much as, and sometimes even more than, the accused. Although some anti-feminist critics assert that women who are victims of sexual violence enjoy a privileged position in society (e.g., Will 2014), there are several examples to counter this claim. For example, though 60 women have officially come forward to accuse comedian Bill Cosby of sexually assaulting them, Cosby continues to have a contingent of supporters behind him whereas the women are accused of lying and attacking him for fame or money (Mallenbaum et al. 2018; Malone 2015).

One explanation for such effects is that U.S. culture is rife with myths about rape, rapists, and rape victims. These myths generally shift blame from perpetrators to victims (Lonsway and Fitzgerald 1994), and they minimize or trivialize sexual aggression (Edwards et al. 2011). Furthermore, these rape myths are linked to other attitudes. For example, Burt (1980) documented a link between rape myths and other attitudes about sexuality and violence. Acceptance of these myths was predicted by endorsement of: (a) traditional gender stereotypes, (b) adversarial sexual beliefs (i.e., beliefs that each party in a romantic relationship tries to manipulate the other), (c) sexually conservative beliefs, and (d) the acceptance of interpersonal violence.

Such attitudes can influence behaviors in ways that cause problems for rape victims. Research indicates that individuals in positions of power (e.g., law enforcement officers, medical staff, clergy) have reported negative attitudes toward rape victims, which can color the support and care they provide to victims. For example, Sheldon and Parent (2002) assessed attitudes toward rape victims and attributions of blame in a sample of interdenominational clergy members. Participants who scored higher on measures of sexism and fundamentalism were more likely to blame the female victim strongly and to cite common rape myths as the reasons for their judgments (e.g., the woman was provocative, the woman did not resist). Sleath and Bull (2012) explored the links in police officers among rape myth endorsement with gender role knowledge, victim blame, and perpetrator blame. The extent to which the officers endorsed rape myths predicted their level of victim blame, particularly myths reflecting ideas such as “She wanted it” and “He didn’t mean to” (Sleath and Bull 2012). Moreover, when police officers believe rape myths, such as the myth that women often lie about being raped, they may not be motivated to investigate a charge of rape (Temkin and Krahe 2008).

Even when rape cases come to trial, rape myths can continue to cause problems for rape victims. For example, jurors who serve in rape trials may be influenced by rape myths in their assignments of blame. Research suggests that mock jurors who report higher rape myth acceptance are especially likely to blame the accuser and especially unlikely to blame the accused perpetrator (Eyssel and Bohner 2011; Paul et al. 2014; Sussenbach et al. 2013). Such effects occur even when the facts of the case do not conform to the traditional rape script, which includes elements such as a physically brutal attack, a stranger rapist, and a virginal victim who cooperates with the police (Frese et al. 2004; Krahe et al. 2008; McKimmie et al. 2014; Sussenbach et al. 2013). When these elements are not present, individuals may be especially unlikely to perceive what happened as a “real rape” (Temkin and Krahe 2008).

The studies we conducted attempt to add to this existing body of work in several ways. First, in the context of mock jury studies, they will measure mock juror participants’ levels of rape myth acceptance (RMA) and will examine the extent to which these levels predict various trial-related judgments. Prior research results lead to the expectation that when responding to possible rape case scenarios, higher mock juror RMA will be linked to higher levels of accuser blaming and higher leniency for the accused (Eyssel and Bohner 2011; Frese et al. 2004; Sleath and Bull 2012; Sussenbach et al. 2013; van der Bruggen and Grubb 2014).

However, our studies will also explore the extent to which these predictive effects might be moderated by the nature of some of the information presented in the case scenarios. In pursuit of this goal, in our studies we manipulated information that was presented to mock jurors. In contrast to prior studies that manipulated details of the rape scenario, our studies varied what was said in the trial about the rape scenario. This idea is of importance because it is believed that juror decisions are influenced not only by the actual details of a case, but also by case-related information that is offered during trial.

One of these kinds of information comes from so-called “expert” testimony that is often used by lawyers to provide perspective on a case. For example, one can imagine that defense attorneys might call an expert witness whose purpose is to influence jurors by confirming rape myths (e.g., most rape accusers are lying). Alternatively, one can imagine that prosecuting attorneys might call an expert witness whose purpose is to influence jurors by debunking rape myths (i.e., rape accusers rarely lie). Accordingly, two studies that we present explore the potential impact on mock juror participants’ judgments of an expert witness who either confirms rape myths (e.g., accusers often lie about being raped) or contradicts rape myths (e.g., accusers lie about being raped only rarely). We explore not only whether this information influences juror judgments, but more importantly, whether such influence depends on a mock juror’s level of RMA. One possibility is that mock juror participants who are higher than others in RMA might assign higher levels of blame to rape victims, even when the expert contradicts the rape myth. This is supported anecdotally, as when in his reporting of a rape trial in Missoula, Montana, Krakauer (2015) found that jurors were inclined to dismiss expert testimony on the impact of rape myths because the testimony did not conform with their personal beliefs.

Two of our studies address the impact of in-trial information on juror judgments in a different manner: via the presentations of the lawyers. One can imagine that prosecution lawyers can try to bolster their case by saying that “there is nothing about the victim’s behaviors that induced or led on the rapist,” explicitly framing the scenario as inconsistent with rape myths. On the other hand, one can imagine that defense lawyers might try to bolster their case by saying that “of course the accuser wanted it, and looking at their behavior, the accuser did all they could to get it,” explicitly framing the scenario as consistent with rape myths. Not only do we explore whether this manipulation influences jurors’ judgments, but more importantly, we explore whether such influence depends on a mock juror’s level of RMA. As reflected in the Krakauer (2015) example, one possibility is that mock juror participants who are higher than others in RMA might assign higher levels of blame to rape victims, even when a prosecution team explicitly uses framing that contradicts the rape myth.

Research results show that when individuals evince high levels of RMA they are especially likely to blame rape victims and especially unlikely to blame rapists (Paul et al. 2014; Sussenbach et al. 2013; Whatley 2005). Extrapolating from such results, we expected that in all of our studies, participants’ strength of, and endorsement of, rape myths should predict participants’ responses to a rape event. In terms of the specific measures used in our studies, we hypothesized that rape myth acceptance (RMA) would positively correlate with judgments of: (a) accuser culpability (Hypothesis 1a) and (b) accuser pleasure (Hypothesis 1b) as well as negatively correlate with judgments of (c) accuser credibility (Hypothesis 2a), (d) accuser trauma (Hypothesis 2b), (e) accused perpetrator culpability (Hypothesis 2c), (f) accused perpetrator guilt (Hypothesis 2d), and the severity of sentence recommendations given to the accused (Hypothesis 2e).

The studies that we report also included some conditions in which rape myths were supported, and other conditions in which rape myths were debunked. Because people tend to preferentially search for, and are especially prone to accept, information that fits their existing beliefs and biases (e.g. they exhibit a confirmation bias, Nickerson 1998), we expected that the judgments of those higher in RMA would be especially affected by rape myth-supportive presentations and the judgments of those lower in RMA would be especially affected by rape myth debunking information. Thus, for higher RMA participants, rape myth confirming information should increase judgements of: (a) accuser culpability (Hypothesis 3a) and (b) accuser pleasure (Hypothesis 3b), and decrease judgments of (c) accuser credibility (Hypothesis 4a), (d) accuser trauma (Hypothesis 4b), (e) accused perpetrator culpability (Hypothesis 4c), (f) accused perpetrator guilt (Hypothesis 4d), and the severity of sentence recommendations given to the accused (Hypothesis 4e). In comparison, for lower RMA participants, rape myth confirming information should decrease judgements of: (a) accuser culpability (Hypothesis 5a) and (b) accuser pleasure (Hypothesis 5b), and decrease judgments of (c) accuser credibility (Hypothesis 6a), (d) accuser trauma (Hypothesis 6b), (e) accused perpetrator culpability (Hypothesis 6c), (f) accused perpetrator guilt (Hypothesis 6d), and the severity of sentence recommendations given to the accused (Hypothesis 6e).

Study 1

As we noted, when individuals display higher levels of RMA, they are especially likely to blame rape victims and especially unlikely to blame rapists. Our first study was designed to reproduce this finding and, in the context of a mock jury study, to examine whether this relationship could be moderated by the presentation of expert witness testimony designed to either confirm rape myths (e.g., women often lie about being raped) or to debunk rape myths (women rarely lie about being raped). Participants’ judgments of responsibility made about both the victim and the perpetrator were examined (a) for evidence of elasticity to this informational manipulation and (b) whether the degree of elasticity observed was dependent on participants’ levels of RMA.

Method

Participants

Study 1 sampled undergraduate students from the introduction to psychology subject pool at a major U.S. university. The initial sample size was 102, but five participants were excluded because they failed to complete more than one measure. Thus, the final sample size was 97, and it included 63 women (65%) and 31 men (32%), with three not reporting gender. Most participants were straight (n = 88, 91%), with two identifying as gay (2%), three identifying as bisexual (3%), three identifying as pansexual (3%), and one not reporting sexual orientation. Almost two-thirds of participants were White (n = 59, 61%), with 15 identifying as Latinx (16%), 11 identifying as African American (11%), 8 identifying as Asian American (8%), 3 identifying as other (3%), and one not reporting race. The mean participant age was 22.43 years (SD = 3.95, range = 18–45).

Procedure and Materials

Psychology undergraduate students were recruited through the university’s SONA recruitment system. At least 2 weeks prior to completing the rape scenario protocol, potential participants first completed a rape myth acceptance scale. After at least a 2-week delay, participants then came to the lab and completed the study online. When they entered the lab, participants were greeted by either a male or a female research assistant, who explained the informed consent form. After signing the informed consent form, participants read a scenario in which a college-aged man named Robert was accused of raping a college-aged woman named Erika. The scenario included a description of the situation from both Erika’s and Robert’s perspectives, and it showed the facts as perceived by each actor.

The scenario indicated that Erika and Robert met when Robert transferred to Erika’s college. Robert was sexually inexperienced and learned that Erika had a reputation for sleeping around. Erika asked Robert out to dinner as a thank you for helping with a project, and throughout the evening, Robert convinced himself that she had been flirting with him. When she invited him to her house after the dinner, he believed that she expected to have sex with him. Robert attempted to kiss her, and she responded by laughing uncomfortably and pushing him away. At this point, Robert believed that Erika was playing hard to get, so he pushed her into the couch and raped her; Erika froze, not saying or doing anything until her friends returned home and she reported the rape to the police. At the end of the narrative of events, participants read that Robert was arrested and that the case went to trial.

Next, participants were randomly assigned to one of three conditions. In the control condition, participants simply read that Robert and Erika’s case went to trial (no other information). In the rape myth confirmation condition, participants read that the defense asserted that women often fabricate rape claims, offering expert witness testimony that 50–90% of rape allegations are false, while the prosecution disputed that testimony. In the rape myth debunking condition, participants read that the prosecutor asserted that women do not often lie about rape, offering expert witness testimony about the actual rate of false rape allegations: 2–10% (Lisak et al. 2010), whereas the defense disputed that testimony. Following the scenario and experimental manipulation of presentation of rape myth information, participants reported how much blame they assigned to both Robert and Erika, as well as their perceptions of other victim- and perpetrator-related attributes. Participants also responded to demographic items assessing gender, age, sexual orientation, relationship status, and relationship length.

When finished, participants were debriefed carefully. The research assistant explained to all groups that the rate of false allegations is low and that women do not often lie about being raped. The debriefing form included extensive sources for research on false rape allegations. Participants were then thanked and released.

Illinois Rape Myth Acceptance Scale

Participants reported their RMA using the Updated IRMAS (McMahon and Farmer 2011). This scale purports to measure individuals’ endorsement of common rape myths. Participants responded to items on a scale of 1 (strongly disagree) to 5 (strongly agree) wherein higher scores indicate greater endorsement of rape myths. The scale comprises 22 items, which were averaged into a composite RMA score. Sample items include: “If a girl goes to a room alone with a guy at a party, it is her own fault if she is raped” and “Rape happens when a guy’s sex drive goes out of control.” Although the RMA contains several subscales, the current study only used the overall scale mean (α = .86).

Scenarios

The scenarios depicted the story of Erika and Robert, two college students who meet, like each other, and agree to go on a date. In the Study 1 scenario, Erika asks Robert out to dinner, then back to her place afterward. A sexual encounter occurs, and Erika subsequently accuses Robert of rape (for all scenarios used in all studies, see the online supplement). To simulate a real-life situation, participants read the scenario from both Erika’s and Robert’s perspectives (the order was counterbalanced to equalize any order effects). Depending on the condition, participants also received prosecution information or defense information regarding the frequency of false rape allegations. In the rape myth confirmation condition, participants read that women frequently lie about rape, whereas in the rape myth debunking condition, participants read that the rate of false rape allegations is 2–10%. The scenarios were pilot-tested to ensure that pertinent details were noticed and remembered.

Judgments about the Accuser and the Accused

Participants responded to questions adapted from Angelone et al. (2015). With the exception of the sentencing recommendation item, all items incorporated a response scale of 1 (not at all) to 10 (very much). For Erika, participants responded to items on subscales probing four judgment content areas: (a) accuser culpability, comprising seven items (e.g., “How much was it Erika’s fault that Robert engaged in sexual activity with her?”; α = .84); (b) accuser credibility, comprising seven items, three of which were reverse coded (e.g., “How likely is it that Erika only called the police so that Robert would not think she was too ‘loose’ or ‘easy’?”; α = .76); (c) accuser pleasure, which was only one item (i.e., “How much pleasure would you estimate Erika experienced during the incident?”); and (d) accuser trauma, also only one item (i.e., “How much trauma would you estimate Erika experienced because of the incident?”). Questions also probed three judgment content areas for Robert, with subscales assessing: (a) accused perpetrator culpability, comprising seven items (e.g., “How much choice did Robert have about what happened in the scenario you read?”; α = .82); (b) accused perpetrator guilt, comprising three items (e.g., “How guilty do you think Robert is of committing rape?”; α = .88); and (c) a sentencing recommendation made about the accused, which was one item (i.e., “If Robert were convicted of committing rape, how much time do you believe he should spend in prison?”) and for which participants responded to on a response scale of 1 (No time at all) to 10 (More than 40 years). (The complete list of items is available in the online supplement.)

Results

Analysis Strategy

Bivariate correlations assessed the relationship between each dependent variable and each participant’s rape myth acceptance score. One-way analyses of variance (ANOVAs) examined the effect of the rape myth confirmation/debunking information on each of the dependent measures. Moderation analyses were conducted using the SPSS PROCESS macro (Hayes 2013). For each moderation analysis, the experimental conditions were dummy coded using indicator codes, where the control condition was coded as 0, the rape myth confirmation condition was dummy code 1, and the rape myth debunking condition was dummy code 2. The full model for each PROCESS analysis included as the independent variables the rape myth acceptance variable, the dummy code variables for the rape myth condition, and the accurate condition (both contrasted against the control condition), and two interaction terms, one reflecting RMA x Rape Myth Confirmation condition and the other reflecting RMA x Rape Myth Debunking condition. Independent analyses were conducted for each of the dependent variables.

Analyses

The means and standard deviations for the dependent variables appear in Table 1a. Table 2 presents correlations between RMA and responses to the various judgment scales, as well as inter-correlations among the scales. The data from Table 2 clearly show that RMA was a significant predictor of most of the judgments that were assessed, supporting many hypotheses. Individuals who were higher in RMA reported significantly: (a) greater accuser culpability (Hypothesis 1a), (b) lower accuser credibility (Hypothesis 2a), (c) lower accuser trauma (Hypothesis 2b), (d) lower accused perpetrator culpability (Hypothesis 2c), and (e) lower accused perpetrator guilt (Hypothesis 2d). However, Hypotheses 1b and 2e were not supported: There was not a significant relationship between RMA and either accuser pleasure or recommended prison sentence for the accused perpetrator. Examination of both the ANOVA results (Fs < 2.23, ps > .113, ηp2s < .05) and the PROCESS results indicated that the judgments were not significantly affected by the differing rape myth-relevant information that was included in the scenarios, indicating a lack of support for all parts of Hypotheses 3–6.

Table 1 Descriptive statistics for study variables, Studies 1–4
Full size table
Table 2 Correlations for dependent variables, Studies 1 and 2
Full size table

Discussion

In summary, then, the results of Study 1 yielded strong and reliable individual differences in judgments. After reading a rape scenario, those who were higher in RMA reported greater accuser culpability as well as lower accuser credibility, accuser trauma, accused perpetrator culpability, and accused perpetrator guilt. However, there was no significant relationship between RMA and accuser pleasure or between RMA and the recommended prison sentence for the accused perpetrator. These results are consistent with previous research examining the relationship between rape myth acceptance and victim blame (see Suarez and Gadalla 2010, for a meta-analysis).

Participants’ judgments were not moderated by the information presented during the trial. This lack of moderation might be viewed by some with apprehension, especially given that the rape myth debunking condition presented the truth (i.e., that women lie about being raped with relative rarity). Pre-testing results suggested that participants were aware of the confirming/debunking scenario information. Hence, it seems more plausible that participants’ judgments were simply not influenced by the information presented.

One potential reason for this possible lack of influence is that participants’ beliefs about rape may be so strong that they are difficult to influence. This possibility is daunting, suggesting that it will be difficult to alter people’s existing beliefs about rape (e.g., in a single anti-rape presentation). An alternative possibility is that our manipulation was not powerful enough. Such impotence may be the consequence of the use of abstract statistical information, which is often perceived by message recipients as not very informative (Borgida and Nisbett 1977; Gemberling and Cramer 2014; Kovera et al. 1997; Parrott et al. 2015).

Study 2

We conducted a second experiment to further explore participants’ responses to a possible rape scenario and how those might be altered by information presented at trial. This second experiment mostly replicated the methods and design used in Study 1. However, in Study 2 we used a different rape myth information manipulation. In one version (rape myth confirmation condition), the defense attorneys explicitly argued from rape stereotypes (e.g., as described in Edwards et al. 2011) that certain facts of the case invalidated the accuser’s rape claim. In another version (rape myth debunking condition), the prosecutor explicitly debunked the stereotypic ideas about rape by pointing out (correctly) that rape can occur in many ways, with many different kinds of people, and in many different situations. Hence, the prosecutor asserted that despite the fact that the encounter did not match the stereotypic scenario of a rape, the encounter was still a rape. The study again examined whether a mock juror participant’s RMA score predicted judgments made about the accuser and the accused perpetrator as well as whether these judgments were moderated by the rape myth-relevant information presented in the scenarios.

Method

Procedure and Materials

Except for the use of different experimental manipulations, the procedure and measures used in Study 2 duplicated those used in Study 1. Reliability for the dependent variables was good: (a) rape myth acceptance (α = .90), (b) accuser culpability (α = .89), (c) accuser credibility (α = .77), (d) accused perpetrator culpability (α = .73), and (e) accused perpetrator guilt (α = .94). As in Study 1, in the control condition participants simply read that Robert and Erika’s case went to trial. In the rape myth confirmation condition, the defense lawyers argued that certain facts of the case (e.g., Robert and Erika knew each other, Erika had a history of promiscuity) did not fit the typical idea of rape, and thus Robert did not rape Erika; the prosecution disputed this argument. In the rape myth debunking condition, the prosecution lawyers argued that victims often know their assailants and that there is no typical idea of what constitutes rape; the defense disputed this argument.

Participants

The initial sample size was 90 U.S. undergraduates, but six people were excluded because they failed to provide responses for more than one dependent variable. Thus, the final sample size was 84: 61 women (73%) and 19 men (23%), with 3 not reporting gender. Most participants were straight (n = 78, 93%), with two identifying as gay (2%), three identifying as bisexual (3%), and one identifying as pansexual (1%). Almost two-thirds of participants were White (n = 54, 64%), with 12 identifying as African American (14%), 11 identifying as Latinx (13%), two identifying as Asian American (2%), and five identifying as other (6%). The mean participant age was 22.43 years (SD = 3.14, range = 19–40).

Results

The means and standard deviations for the dependent variables appear in Table 1b. Table 2 presents correlations between RMA and responses to the various judgment scales as well as inter-correlations among the scales. As in Study 1, participants’ RMA levels predicted many participant judgments, supporting multiple hypotheses. Participants who were higher in RMA judged Erika, the accuser: to be more culpable (Hypothesis 1a), to experience greater pleasure (Hypothesis 1b), to be lower in credibility (Hypothesis 2a), and to experience lesser trauma (Hypothesis 2b). Individuals who reported higher RMA also: (a) reported lower accused perpetrator (Robert) culpability (Hypothesis 2c), (b) reported lower accused perpetrator guilt (Hypothesis 2d), and (c) recommended shorter sentences for the accused perpetrator (Hypothesis 2e).

In contrast, examination of both the ANOVA results (Fs < 1.79, ps > .173, ηp2s < .04) and the PROCESS results (see analysis strategy in Study 1) indicated that the judgments were not significantly affected by the differing rape myth information that was included in the scenarios. These results were contrary to all sub-hypotheses we proposed in Hypotheses 3–6.

Discussion

In summary, the results of Study 2 replicated the results obtained in Study 1. The data from Study 2 yielded strong and reliable individual differences in judgments. After reading a rape scenario, those who were higher in RMA reported greater accuser culpability, lower accuser credibility, higher accuser pleasure, lower accuser trauma, lower culpability of the accused, lower guilt of the accused, and recommended a lower prison sentence for the accused. There are two differences between the results of Study 2 and those of Study 1: in Study 1, there was no relationship between RMA and (a) accuser pleasure judgments or (b) recommended sentence length. This could reflect a difference between studies in the effect of the rape myth information presented. This may suggest that participants may be more familiar with the real rape stereotype (used in Study 2) than they are with statistics regarding false rape reports (used in Study 1). Additionally, these judgment tendencies were not significantly moderated by the rape myth-relevant information presented during the trial, suggesting that the experimental manipulation had no effect.

Study 3

Study 3 attempted to replicate and extend the finding of Study 1 showing that judgments made about rape accusers and accused rapists were relatively unaffected by the expert testimony that was presented during a rape trial. It did so by altering the rape scenario so that it was a man who was raped by a woman. This is an interesting extension of Study 1 because the myths about man-raping-woman encounters may differ from those in woman-raping-man encounters. However, as with the rape of women, most rape-of-man-by-woman myths center around delegitimizing the phenomenon (Turchik and Edwards 2012). For example, the myth that men are always ready for sex discounts the idea that men can be victims.

Importantly, there are individual differences in the extent to which people accept these beliefs about women raping men (Melanson 1999). As in Studies 1 and 2, we expected that in a mock jury study these individual differences would predict the judgments rendered about accusers and about the accused. As in Study 1, of interest was whether such judgments are moderated by the rape myth-relevant information (different than that provided in Study 1) provided by an expert witness.

Method

Participants

As in the previous studies, Study 3 used a sample of students recruited from the undergraduate psychology subject pool at a major U.S. university. The initial sample size was 102, but four participants were excluded for failing to complete more than one measure. Thus, the final sample size was 98, and included 57 women (58%) and 41 men (42%). Almost all participants were straight (n = 96, 98%), with one participant identifying as gay (1%) and one identifying as bisexual (1%). Just under half the participants were White (n = 47, 48%), with 22 identifying as African American (22%), 18 identifying as Latinx (18%), 8 identifying as Asian American (8%), 2 identifying as other (2%), and one not reporting race (1%). The mean participant age was 20.25 years (SD = 2.19, range = 18–28).

Procedure and Materials

With three exceptions, the procedure and measures used in Study 3 duplicated those used in Study 1. One exception was that in the rape scenario the genders of the accuser and accused perpetrator were reversed. The second exception was that Study 3 used an RMA scale that differed from the one used in Study 1. This new scale was needed to capture people’s views about a man being raped by a woman. The third exception concerned the timing of completion of the rape myth acceptance scale. All participants completed the scale during the experimental session; this change was due to external time constraints on data collection. However, the presentation of the rape myth scale and the experimental manipulation was counter-balanced to equalize order effects. Preliminary analyses suggested that the ordering manipulation did not affect responses to either the rape myth acceptance scale or the judgment measures.

Male Rape Myth Scale

Participants revealed their degree of male rape myth acceptance on Melanson’s (1999) Male Rape Myth Scale (MRMS). Similar to the IRMAS used in Studies 1 and 2, this measure comprises 22 items assessing participants’ endorsement of male rape myths on a scale of 1 (strongly disagree) to 6 (strongly agree), with higher scores indicating more acceptance. As with the IRMAS, a composite MRMA score was created by averaging the scale items. Sample items include: “If a man obtained an erection while being raped it probably means that he started to enjoy it” and “I would have a hard time believing a man who told me that he was raped by a woman.” Melanson found good internal consistency for the MRMS (α = .90) and test-retest reliability (after 4 weeks) was also high (r = .89, p < .001). Reliability was good in the current study (α = .84).

Scenarios

The scenarios for Study 3 were the same as the ones used for Study 1, with the exception that Erika’s and Robert’s actions were reversed. Recall that in the Study 1 scenario, Erika asked Robert out to dinner, then back to her place afterward, where a sexual interaction that might be a rape of Erika by Robert occurred. In the Study 3 scenario, Robert asked Erika out, then invited her back to his place, where a sexual encounter occurred that might be a possible rape of Robert by Erika.

In one condition (control), participants received no additional information in the scenario that they received. In the rape myth confirmation condition, in their scenario participants read a defense statement disbelieving Robert’s accusation of rape because Robert’s erection signaled his consent for the sexual activity, indicating that men cannot be raped by women, while the prosecution disputed that claim. In the rape myth debunking condition, participants read a prosecution statement about how erections reflect an automatic physiological response. Thus, an erection does not imply consent, a state of affairs suggesting that women can indeed rape men. The defense disputed that claim.

Dependent Variables

As we mentioned, the judgment measures for accuser and accused perpetrator were the same as in Studies 1 and 2. Reliability for these variables was good: (a) accuser culpability (α = .84), (b) accuser credibility (α = .73), and (c) accused perpetrator guilt (α = .91). Due to unacceptably low reliability (α = .58), the measure of accused perpetrator culpability was excluded from Study 3.

Results

The means and standard deviations for the dependent variables can be found in Table 1c. The correlations among the dependent measures, as well as between the MRMA measure and the dependent measures, appear in Table 3. The results showed that responses to the MRMA scale predicted judgments made after reading the rape scenario, supporting several hypotheses. Individuals who reported higher MRMA also reported: (a) greater accuser culpability (Hypothesis 1a), (b) greater accuser pleasure (Hypothesis 1b), (c) lower accuser credibility (Hypothesis 2a), (d) lower accuser trauma (Hypothesis 2b), and (e) lowered accused perpetrator guilt (Hypothesis 2d). There was no significant relationship between MRMA and the sentence recommendation for the accused perpetrator, thus Hypothesis 2e was not supported.

Table 3 Correlations for dependent variables, Studies 3 and 4
Full size table

In contrast to results from Studies 1 and 2, the ANOVAS conducted on the various judgment indices yielded one main effect for the rape myth information manipulation. This effect showed that judgments of accuser pleasure varied across conditions, F(2, 95) = 3.82, p = .025, ηp2 = .074, and were lower in the rape myth debunking condition (M = 2.87, SD = 2.54) than in the control condition (M = 4.46, SD = 2.99, p = .007, d = .57). This effect fits with expectations about how judgments of accuser pleasure ought to be affected by rape myth debunking information.

However, more crucial to our article is whether the rape myth information presented to participants moderated the relation between RMA and the various measures assessing perceptions of accused rapists and their accusers. The moderation analyses conducted using PROCESS yielded two effects (see analysis strategy in Study 1).

One was a significant interaction on the accuser pleasure variable between MRMA and the rape myth debunking condition (vs. control), b = −1.88, t(89) = −2.11, p = .038, 95% CI [−3.66, −.11], d = .45 (see Fig. 1). A simple slopes analysis conducted within condition indicated that MRMA predicted accuser pleasure more strongly in the control condition (b = 3.30, t(89) = 4.99, p < .001, 95% CI [1.99, 4.62], d = 1.06), than in the rape myth debunking condition (b = 1.42, t(89) = 2.37, p = .02, 95% CI [.23, 2.62], d = .50). The simple slope in the rape myth confirmation condition was also significant, b = 2.68, t(89) = 2.93, p = .004, 95% CI [.86, 4.49], d = .62), and resembled the effect in the control condition. These results provide back-door support for Hypothesis 5b, which suggested that the judgments of those lower in MRMA should be especially affected by the debunking information. Instead, the data suggest that those higher in MRMA were especially unaffected by the rape myth debunking information, which can be viewed as one way that a confirmatory bias tendency might affect judgments.

Fig. 1
figure 1

Simple slopes for male rape myth acceptance within condition on accuser pleasure in Study 3. All three slopes are significant

Full size image

The PROCESS analyses also yielded a significant interaction between MRMA and the rape myth confirmation condition on the accuser trauma measure, b = 2.52, t(89) = 2.13, p = .036, 95% CI [.17, 4.86], d = .45 (see Fig. 2). A simple slopes analysis conducted within each condition indicated that in the control condition MRMA inversely predicted victim trauma, b = −2.66, t(89) = −3.84, p < .001, 95% CI [−4.03, −1.28], d = .84), and this inverse predictive relation held in the rape myth debunking condition, b = −6.45, t(89) = −2.80, p = .02, 95% CI [−12.01, −.90], d = .59). However, this relation was not significant in the rape myth confirmation condition, b = −2.62, t(89) = −1.18, p = .24. An examination of the data suggests that it was the judgments of those who were lower in MRMA that were especially affected by the rape myth confirming information. Thus, these data do not fit the confirmatory bias idea that drove the formulation of Hypotheses 4b and 6b.

Fig. 2
figure 2

Simple slopes for male rape myth acceptance within condition on accuser trauma in Study 3. Only the control condition’s and rape myth debunking condition’s slopes are significant

Full size image

Discussion

In summary, then, the results of Study 3 generally replicated the results from Study 1, but extended those results to a male rape accuser and a female accused rapist. As in Study 1, Study 3 yielded strong and reliable individual differences in judgments. After reading a rape scenario, those who were higher in MRMA reported greater accuser culpability, lower accuser credibility, higher accuser pleasure, lower accuser trauma, and lower accused perpetrator guilt. However, there was no significant relationship between MRMA and the recommended prison sentence length.

As in Study 1, the sentence participants recommended for the accused perpetrator was unrelated to their reported level of MRMA. However, while Study 1 participants did not demonstrate a relationship between accuser pleasure and RMA, in Study 3, participants reporting higher MRMA also reported higher accuser pleasure. Recall that the variable of accused perpetrator culpability was excluded from analysis due to the measure’s low reliability. The fact that both Study 1 and Study 3 used novel rape myth information (i.e., women often falsely report rape and men’s erections are equivalent to sexual consent), could account for the lack of significant relationships for RMA/MRMA and all DVs. This receives some support from the results of Study 2, where all DVs were significantly correlated in expected directions with RMA.

A difference from Study 1’s results was that in Study 3, for two judgments (accuser pleasure and accuser trauma), these individual differences were moderated by the kind of rape myth information (rape myth confirming, rape myth debunking, no information) presented during the scenario. Information indicating that men’s erections are not under conscious control and do not indicate consent were especially ineffective in altering the accuser pleasure judgments of those who were higher in MRMA. Judgments of victim traumatization were also moderated by MRMA in Study 3: Somewhat puzzlingly, lower MRMA participants in the rape myth debunking condition reported lower victim trauma.

As with Study 1, the absence of moderation effects on the other judgment variables might be viewed by some with trepidation, especially given that the rape myth debunking condition presented the truth about how an erection does not imply consent. As in Study 1, pre-testing results suggested that participants were aware of the confirming/debunking scenario information. Hence, it seems more plausible that participants’ judgments were simply not influenced by the information presented.

Study 4

In the same manner that Study 2 sought to replicate and extend the results of Study 1, Study 4 tried to replicate and extend the results of Study 3 by altering the nature and source of the rape myth-relevant information presented in the trial transcript. As in Study 2, in Study 4 the rape myth-relevant information presented to participants was offered by the trial lawyers and either overtly endorsed male rape myths (rape myth confirming condition) or overtly debunked male rape myths (rape myth debunking condition).

Method

Procedure and Materials

Except for the experimental manipulations, all methods and measures used in Study 4 duplicated those used in Study 3. In the Study 4’s control condition, as in the Study 3’s control condition, participants merely read that Robert and Erika’s case went to trial. In the Study 4’s rape myth confirmation condition, the defense argued that the facts of the case (e.g., Robert and Erika knew each other, they’re both straight) did not fit the typical idea of rape (e.g., men are only raped in prison by other men), and thus Erika could not have raped Robert; the prosecution disputed this argument. In the rape myth debunking condition, the prosecution argued that there is no typical manner in which male rape occurs and that male rape victims often know their assailants; the defense disputed this argument. As in our prior studies, reliabilities for the dependent variables were good: (a) rape myth acceptance (α = .83), (b) accuser culpability (α = .77), (c) accuser credibility (α = .71), (d) accused perpetrator culpability (α = .68), and (e) accused perpetrator guilt (α = .95).

Participants

The initial sample contained 118 people, but two participants were excluded because they provided incomplete data. Thus, the final sample size was 116 undergraduates at a major U.S. university, and included 77 women (66%), 38 men (33%), and one participant identifying as other (1%). As in the other studies, most participants were straight (n = 109, 94%), with five identifying as bisexual (4%), and two identifying as pansexual (2%). Forty-eight participants were White (42%), with 34 identifying as African American (29%), 22 identifying as Latinx (19%), 6 identifying as Asian American (5%), and 6 identifying as other (5%). The mean participant age was 19.91 years (SD = 1.73, range = 18–26).

Results

The means and standard deviations for the dependent variables can be found in Table 1d. The correlations among the dependent measures, as well as between the MRMA measure and the dependent measures, appear in Table 3. As in Studies 1 through 3, participant rape myth acceptance predicted judgments made of the male accuser and the female accused rapist. Participants displaying higher MRMA also reported: (a) greater accuser culpability (Hypothesis 1), (b) greater accuser pleasure (Hypothesis 1b), (c) lower accuser credibility (Hypothesis 3), (d) lower accuser trauma (Hypothesis 2b), (e) lower accused rapist culpability (Hypothesis 2c), (f) lower accused rapist guilt (Hypothesis 2d), and a shorter accused prison sentence for the accused rapist (Hypothesis 2e).

The ANOVA results revealed two main effects for the rape information presentations. First, for the significant condition main effect of accuser credibility, F(2, 120) = 3.64, p = .029, ηp2 = .061, the mean for the rape myth confirming condition (M = 6.82, SD = 1.59) was significantly higher than the mean for the rape myth debunking condition (M = 6.09, SD = 1.45, p = .035, d = .48) and for the control condition (M = 5.93, SD = 1.43, p = .013, d = .59). Second, for the significant condition main effect of accused perpetrator guilt, F(2, 112) = 4.04, p = .020, ηp2 = .067, the mean for the rape myth confirming condition (M = 7.43, SD = 2.74) was significantly higher than the mean in the rape myth debunking condition (M = 5.79, SD = 2.87, p = .015, d = .58) and the mean in the control condition (M = 5.71, SD = 3.18, p = .014, d = .58). Both findings partially fit expectations about how the information presentations ought to affect judgments.

However, more important to our article was whether there was evidence that the relations between MRMA and the dependent measures were moderated by the rape information presentations. The moderation analyses conducted using the PROCESS tool yielded only one instance in which the correlation between the MRMA and a judgment measure was moderated by the manipulation of the rape myth information (see analysis strategy in Study 1). This occurred for the sentence recommendation measure, and it involved the interaction between MRMA and the rape myth debunking (vs. control) condition, b = 2.12, t(106) = 2.18, p = .032, 95% CI [.19, 4.05], d = .42 (see Fig. 3). Simple slopes analyses separately examining within each condition the relation between MRMA and sentence recommendations yielded a significant slope for the control condition, b = −2.68, t(106) = −3.53, p < .001, 95% CI [−4.19, −1.18], d = .69, but not for the rape myth confirmation condition or the rape myth debunking condition. The data patterns suggest that, in comparison to the judgments rendered in the control condition, judgments of higher MRMA participants were more punitive in response to both the rape myth confirmatory information and the rape myth debunking information. Similarly, the data patterns suggest that in comparison to the judgments rendered in the control condition, judgments of lower MRMA participants were more lenient in response to both the rape myth confirmatory information and the rape myth debunking information. Hence, the confirmatory bias idea that drove the formation of Hypotheses 4e and 6e did not receive much support from these data.

Fig. 3
figure 3

Simple slopes for male rape myth acceptance within condition on sentence recommendation in Study 4. Only the control condition’s slope is significant

Full size image

Exploratory Gender Analyses

Due to the sample sizes for each study, we were not able to include participants’ gender as a variable in our analyses. However, we did conduct post hoc t-tests examining gender differences on our dependent variables for all four studies. Each study, with the exception of Study 2, yielded significant differences by gender for at least some of the dependent variables.

In general, men displayed more negative evaluations of accusers and more positive evaluations of accused perpetrators than did women. Men also reported significantly higher rape myth acceptance than did women (Studies 1, 3, and 4, ps = .001–.003, ds = .62–.64). Compared to women, men reported greater accuser culpability (Studies 1, 3, and 4, ps = .002–.045, ds = .41–.63) and greater accuser pleasure (Study 3, p < .001, d = .73). Compared to men, women reported greater accuser credibility (Studies 1 and 4, ps = .007–.013, ds = .55–.58), accuser trauma (Studies 1, 3, and 4, ps = .017–.043, ds = .43–.49), accused perpetrator guilt (Studies 3 and 4, ps = .003–.013, ds = .50–.63), and accused perpetrator culpability (Study 4, p < .001, d = .82). Although we cannot draw conclusions from these post-hoc analyses, they do indicate that inclusion of participants’ gender as a study variable would be important in future work.

Discussion

In summary, then, the results of Study 4 replicated and extended the results that emerged from Studies 1 through 3. The data from Study 4 yielded strong and reliable individual differences in judgments. After reading a rape scenario, those who were higher in MRMA reported greater male accuser culpability, lower male accuser credibility, higher male accuser pleasure, lower male accuser trauma, lower female accused culpability, lower female accused guilt, and a lower prison sentence for the accused woman. Thus, as in Study 2, participants’ endorsement of gender-appropriate rape myths was significantly correlated with their judgments of accusers and accused perpetrators. The difference in results between Studies 1 and 3 and Studies 2 and 4 supplies further support for the idea that these differences are due to the rape myth information presented in each study.

In only one case (sentence recommendations) were these judgment tendencies moderated by the rape myth-relevant information presented during the trial. The pattern of data evinced for this moderation was not consistent with the idea that people will be especially influenced by presented information that fit their preconceptions. Moreover, given the absence of similar interactions on the other judgment variables, we recommend that the interaction be viewed with caution until it is replicated in future research.

General Discussion

In four studies, U.S. college students played the role of mock jurors in a case of possible rape. The data from all four studies yielded strong and reliable individual difference effects in judgments of the accused rapist and the accuser. Those participants who were higher in rape myth acceptance generally reported greater accuser culpability, lower accuser credibility, higher accuser pleasure, lower accuser trauma, lower accused perpetrator culpability, and lower accused perpetrator guilt, and they offered a more lenient prison sentence for the accused perpetrator than those who were lower in rape myth acceptance. A relatively novel contribution of our studies is that these effects emerged regardless of whether judgments were being made of male accused rapists and female victims/accusers, or of female accused rapists and male victims/accusers. Collectively, these individual-difference results converge with long history of prior work examining the relationship between rape myth acceptance and judgments about alleged victims and alleged perpetrators. Whether participants were clergy (Sheldon and Parent 2002), lawyers (Krahe et al. 2008), law enforcement officers (Sleath and Bull 2012, 2015), students (Angelone et al. 2015), or online adults (Sussenbach et al. 2013), prior studies reported consistently positive relationships between high endorsement of rape-supportive beliefs and negative judgments of sexual assault victims.

One additional purpose of our research was to attempt to alter these judgments via the presentation of information relevant to rape myths. In two of the studies, the scenarios read by participants presented information from an “expert witness” who either confirmed or debunked rape myths. In the other two studies, the lawyers involved in the case either explicitly used rape myths (this scenario did not fit the rape stereotype) in an attempt to exonerate the accused rapist or debunked the rape stereotype (there are lots of ways in which rape occurs) in an attempt to establish the guilt of the accused rapist.

These attempts were generally ineffective, although they seemed to be a bit more successful in the female-raping male scenarios (Studies 3 and 4) than in the male-raping female scenarios (Studies 1 and 2). Specifically, we found only three main effects of the experimental manipulations on the judgments of accusers and accused perpetrator. In Study 3, participants in the rape myth debunking condition reported significantly lower evaluations of accuser pleasure than did the participants in the control condition. In Study 4, manipulated information differences emerged for the measures of accuser credibility and accused perpetrator guilt. For accuser credibility, participants in the rape myth confirming condition reported higher credibility than did participants in both the control condition and the rape myth debunking condition.

However, for our purposes, more important than the emergence of main effects for the information presentation manipulations was whether these manipulations moderated the relations between individual differences in rape myth acceptance and the dependent measures. Only three such interactions emerged. In Study 3, for two judgments (accuser pleasure and accuser trauma), individual differences in MRMA were moderated by the kind of rape myth information (rape myth confirming, rape myth debunking, no information) presented during the scenario. In the first of these interactions, information indicating that erections are not under conscious control and do not indicate consent was especially ineffective in altering the accuser pleasure judgments of those who were higher in MRMA. In the second of these interactions, lower MRMA participants in the rape myth debunking condition reported especially low victim trauma. The third significant interaction came from Study 4. In this study, in comparison to the judgments rendered in the control condition, judgments of the higher MRMA participants were especially severe in response to both the rape myth confirmatory information and the rape myth debunking information. However, the data patterns suggest that in comparison to the judgments rendered in the control condition, judgments of the lower MRMA participants were especially lenient in response to both the rape myth confirmatory information and the rape myth debunking information.

Our inclination is to view all significant effects of the information presentation manipulation with extreme caution. It seems to us most likely that the manipulations were ineffectual and that the main effects and interactions that involve the information presentation manipulations reflect Type I error. There are two reasons underlying this judgment. The first is that each of the effects described emerged in only one of the studies. Moreover, effective manipulations should have affected all of the measures in a given study in conceptually similar ways (because the dependent measures were strongly correlated). Hence, we are skeptical of these effects because we see both lack of replication across studies and lack of convergence within studies. The second reason underlying our caution is that many of the information presentation effects that did emerge were not consistent with the confirmatory bias idea that drove many of the hypotheses. Although lack of consistency with a theoretical idea is not by itself a reason to discount a significant effect, the fact that the significant effects that did emerge did not seem to coalesce around any theoretical idea suggests that these significant effects are not to be trusted.

However, we recognize that our assessment is not definitive. We know this because we conducted an additional series of equivalence testing (i.e., TOST; Lakens 2017) analyses. The technique of equivalence testing tests the null hypothesis that groups are not equivalent; rejection of the null suggests that groups are equivalent (i.e., implying that our manipulations were ineffectual). We examined some of the critical comparisons from all four of our studies using the equivalence testing technique. The results of all tests suggested that we are unable to confidently conclude that our groups are equivalent (i.e., we could not reject the null hypothesis that the groups that we statistically compared were different).

Given these considerations, we are comfortable in using our data to offer two conclusions. First, our individual difference results clearly reinforce and extend prior work examining the relationship between endorsement of rape myths for men and women and the judgments individuals make in sexual assault scenarios: Those who strongly believe rape myths are especially likely both to judge accusers harshly and to be lenient in judgments of the accused. Our second conclusion is that our studies did not provide persuasive evidence that presentation of rape myth confirming or rape myth debunking information can influence these judgments. However, the results from the equivalency analyses hint at the possibility that such effects might be demonstrable in future research.

Limitations and our Manipulations

The fact that the equivalency analyses provide a glimmer of hope for producing information presentation manipulations that alter judgments made from a rape scenario prompts speculation for why the effects of the manipulations that we used were not detectable using the usual null hypothesis testing strategy (in which the null hypothesis is that there is no difference between groups or among groups). We generated five explanations (this list is not intended to be inclusive; there may be more): (a) individuals’ rape myth acceptance is too deeply entrenched in a participant’s value system and cannot be overcome by experimental manipulations; (b) participants may have already made their judgments of guilt prior to encountering the manipulations; (c) when reading the information in the manipulations, participants may have experienced psychological reactance; (d) the strength and/or style of the manipulations may not have been powerful enough to influence the participants’ judgments; and (e) participant scenario processing tendencies, such as low attention or low reading comprehension, may have reduced the impact of the manipulations. The sections that follow briefly discuss each of these ideas and some empirical implications of each.

Entrenched Value Systems

One reason the experimental manipulations may have failed is due to the ingrained nature of rape myths in individuals’ perceptions of sexual violence. Rape myths appear in virtually all types of media and discourse, from movies and television to news and advertisements to conversations with friends and family. Early and consistent exposure to rape myths may lead to an internalization of rape myth-supportive beliefs, an internalization that is difficult to alter. Indeed, several theories of attitudes and attitude change discuss how strong attitudes are especially resistant to change (Bohner and Dickel 2011). Strong attitudes and beliefs are easily accessible, so they are especially likely to influence attitude-relevant judgments.

This position was supported by results of an exploratory study that asked participants to explain their reasoning for allocating blame for Erika and Robert (Klement 2018). No participant mentioned anything about the defense’s presentation or prosecution’s presentation. Several participants indicated that they relied on their own values or morals to make the decisions (e.g., “Base[d] on the moral outcome of them”), or that they relied on the facts of the case (e.g., “The fact that Erika kept quiet during the rape was a large factor”). Many participants indicated that the blame should be shared equally (e.g., “I think Erika was mostly guilty for the crime. However, I believe Robert could have pushed her off or stopped her from being raped.” and “I’ve heard about these kind of situations very often. Both parties are equally guilty. They are just asking for trouble and to be hurt.”). Thus, it appears that the participants used their own perceptions of Erika and Robert, the situation, and their prior experiences with sexual violence (whether as victims or not) to make decisions.

Therefore, the manipulations used in our studies may have failed because it is difficult for such manipulations to overcome these kinds of strongly entrenched rape-relevant beliefs. This position implies that lawyers may be wasting their time by trying to frame or bolster their cases by presenting expert testimony or by using or debunking rape myths in trials. This position also implies that even strong laboratory manipulations that debunk stereotypic rape beliefs may be relatively impotent in altering those beliefs. Finally, this would further suggest that because rape myth relevant beliefs are so strongly entrenched, it would be relatively useless to embark on public service persuasion campaigns designed to debunk rape myths. The difficulty that climate scientists have in trying to persuade many in the public about the reality of global warming may serve as prototype example of this kind of impotence (see Graham et al. 2009).

Predetermined Judgments

A second reason for the relative impotence of the rape myth information manipulations used in our studies relates to the fact that our experimental manipulations were encountered after the participants read the scenario. It is thus possible that participants’ judgments about the sexual encounter were already made from the “facts” of the sexual encounter before they encountered the rape myth-relevant information contained in some of the scenarios. This idea is supported by a large body of research suggesting that people will often make on-line or spontaneous judgments about others. When such processing is used, information encountered first has an especially large influence on participant judgments (see Hogarth and Einhorn 1992).

However, results from an exploratory study that replicated Study 1 and Study 3 (e.g., data collected online, one study had a female victim, one had a male victim), but that presented the manipulations prior to the scenario, did not support this idea (Klement 2017). Instead, results were similar to those obtained in Study 1 and Study 3. Hence, in our view, the manipulation timing explanation does not hold much promise as an explanation for the relative impotence of the rape myth information manipulations used in Study 1 through Study 4.

Psychological Reactance

Another reason the manipulations may have been relatively impotent might stem from psychological reactance against both the rape myth confirming information and the rape myth debunking information. Reactance may happen when individuals feel that their freedoms are being threatened (Brehm 1981). In our studies, participants may have felt pressured to accept the rape myth-relevant information presented in the scenarios, so they may have rejected the manipulation information and responded in a manipulation-neutral manner, or even in a manipulation-contrary manner. This possibility is further supported by the idea that overt or memorable attempts to influence participants often produces contrast effects in social judgments (see Bless and Schwarz 2010).

This speculation suggests that it might be possible to induce people to use rape myth-relevant information if it presented in a somewhat “gentle” manner that is not an obvious part of an attempt to influence. If subsequent research shows this to be the case, it would represent a practical problem for legal teams. Much of the work of such teams might generally be perceived by jurors as part of attempts to influence. Thus, although it may be theoretically possible to induce perceivers to respond to messages that include rape myth-relevant information, doing so may be difficult when jurors may see almost any presentation made by lawyers or witnesses as a part of an attempt at persuasion.

Strength and Style of Manipulation

A fourth explanation for the results of our studies could lie in the nature of the manipulations. Other studies that have investigated differences in victim blaming and perpetrator blaming have used details of the actions that appeared in the manipulation. For example, scenarios may differ in whether the assailant uses force or not (Krahe et al. 2008), whether or not alcohol is present (Eyssel and Bohner 2011), or whether the victim has a prior relationship with the assailant (McKimmie et al. 2014). These kinds of manipulations do influence judgments. In contrast, our studies tried to manipulate judgments via presentations from expert witnesses or from lawyers. It may be that such presentations might need to be especially strong or extreme to impact judgments.

It also may be the case that perceivers might have a general tendency to discount information coming from lawyers or lawyer-produced “expert witnesses.” That is, because such sources might be perceived to have vested interests in the outcome of the case, the information provided by such sources might be discounted or totally ignored. This dismissal may occur because participants may use stereotypes about lawyers and expert witnesses, stereotypes that may contain the belief that lawyers and expert witnesses will say whatever they need to say to win the case. This view suggests that the kinds of information that was provided in Studies 1 through 4 might indeed affect participants’ judgments if the information was perceived as coming from an unbiased source. For example, a future study might instead have the rape-relevant information come from the judge, whom participants might perceive as an impartial party, thereby removing perceptions of bias from the information presentation.

Using a trial context may also have muted the effects of the information presentation. In the scenarios that we used, the case had already gone to trial. Participants may have inferred a certain level of guilt for the accused rapist from the fact that sexual predators do not often appear in the justice system (Rape Abuse and Incest National Network 2016), and so one who does must be at least somewhat guilty. Thus, a potential future study could remove the at-trial context and present the information as emerging pre-trial. Alternatively, the legal context could be removed altogether by using an informal disclosure context, such a friend-to-friend interaction.

Participant Processing Tendencies

The last of our explanations for the lack of significant effects from the information presentation manipulations used in all studies is that the results may have been influenced by participant processing tendencies. Our studies did not include attentional checks or manipulation checks so we cannot say with certainty that participants paid attention to the manipulations. Reading comprehension may also have played a role in the relative impotence of the manipulations. If participants were not able to understand the content of the scenarios and the manipulations, those manipulations would not be able to influence their responses. Clearly, future replications of our studies ought to include items assessing attention to, and understanding of, the scenarios (especially the content of the manipulations).

A similar case can be made for the variable of social desirability. Presentation of the rape myth confirming information was especially impotent in our studies, and this may have been caused by participants’ reluctance to increase the harshness of their judgments of the accuser because of the belief that it would make them look bad. Future studies might account for this possibility by including a measure of social desirability (such as the Balanced Inventory of Desirable Responding, Li and Bagger 2007), and by using this measure to account for social desirability biases in participants’ responses.

Other Future Research Directions

Our discussion of the possible reasons for the weakness of the manipulations that we used in Studies 1–4 incorporated numerous possible directions for future research. A few other potentially fruitful research directions also seem reasonable. One is to collect data that directly examines participants’ evaluations of the expert witness testimony. Previous research and anecdotal accounts indicate that jurors can evaluate information from such testimony differently depending on their existing beliefs, motivation, and ability (Gemberling and Cramer 2014; Krakauer 2015). It is possible that an examination of such reactions might guide the construction of more effective information presentations. Indeed, such research may suggest that, to be maximally effective, presentations with different characteristics may need to be developed for different kinds of presentation recipients.

One might also include in research an examination of how participants’ gender is related to judgments derived from the rape scenario. This is desirable because men both tend to commit more acts of sexual aggression than women (Black et al. 2011; Breiding et al. 2014) and they also are more likely to believe rape myths (Suarez and Gadalla 2010). Hence, men might be an especially important target for presentations that effectively debunk rape myths (e.g., Taschler and West 2017). Because all of our samples contained skewed gender ratios, we were unable to validly explore our findings by gender. However, in exploratory analyses, we did find significant differences between men and women on many of the dependent variables, with men generally reported greater rape-supportive beliefs and judgments. These preliminary findings should encourage future studies that systematically incorporate gender into research designs.

Practice Implications

Our results have potential implications in two settings: sexual assault prevention programs and the courtroom. Given continuing concern about rape-relevant issues, such as mitigating sexual assault on college campuses, from a practice perspective, our inability in the current studies to reduce victim blaming is frustrating. However, other studies have demonstrated reductions in rape-supportive beliefs and attitudes and behavioral intentions. Several literature reviews focused on secondary and higher education have summarized the results of sexual violence prevention and risk reduction programming (Lonsway et al. 2009; Vladutiu et al. 2011). These findings have supported the idea that interventions can work to reduce rape-supportive beliefs, but long-term attitude change is more likely when the programming is spread over several sessions, uses a variety of administrations (e.g., videos, skits, worksheets), and is targeted at specific audiences (e.g., fraternities, athletic teams). In this domain, Vladutiu et al. (2011) suggested that peers are better at prompting changes in rape myth-related beliefs than are professional speakers.

A path to effective interventions can also be derived from research conducted by Holz et al. (2018). They found that men who scored higher on all four our factors of Bolton and DiLalla’s (2007) Fear of Unintentional Rape Inventory—(a) personal legal concerns about rape accusations, (b) concerns about sexual communication, (c) concerns that rape laws unfairly target men, and (d) personal concerns that alcohol use leads to unintentional rape—reported the highest amount of RMA before and after the prevention programming, compared to men who only reported a subset of the factors. Holz et al. concluded that sexual violence interventions may be especially effective when targeting men’s specific concerns regarding rape.

In the courtroom, the path to effectively mitigating the problematic rape-related beliefs that jurors bring to deliberation is difficult to see. Individuals who serve on juries can come from very different backgrounds; can vary widely in age, racial identity, gender identity, and class status; and can bring with them a variety of different rape myth-related beliefs. Thus, jurors may vary widely in their tendency to use expert testimony from witnesses attempting to debunk common rape myths. However, there are a few reasons for optimism. Using the theory that intergroup contact can reduce prejudice, Taschler and West (2017) found that both men’s and women’ rape-related beliefs were especially low when they reported frequent and high-quality contact with counter-stereotypical women. These women included friends, coworkers, and acquaintances who were in senior positions of authority. This correlational research indicates that having relationships with socially powerful women can have a positive impact on individuals’ perceptions of women and sexual violence. This linkage suggests that lawyers might be able to debunk juror rape myths by exposing jurors to powerful women during courtroom presentations.

Conclusions

Rape accusers often are disbelieved or are perceived to be responsible for being raped. Our data suggest that these tendencies are especially powerful in those who strongly endorse stereotypical beliefs about rape, regardless of whether the scenario involves male accused perpetrator/female accuser or involves female accused perpetrator/male accuser. It would clearly be desirable to be able to dispel tendencies toward accuser blaming or disbelief by the presentation of information that debunks rape myths. Our data suggest that this may not be easily done, especially in the context of a legal proceeding. However, as we note in our discussion of our research, there remain reasons for hope. It is incumbent on future researchers to discover those conditions that would effectively reduce perceiver endorsement of rape myths and a reduction in the tendency to disbelieve rape accusers or to blame them for being raped.