Rape Myth Scale: Factor structure and relationship with gender egalitarianism among Japanese professionals

Authors


Masayo Uji, PhD, Department of Clinical Behavioral Sciences, Kumamoto University Graduate School of Medical Sciences (Psychological Medicine), 1-1-1, Honjo, Kumamoto 860-8556, Japan. Email: ujimasayo@excite.co.jp

Abstract

Abstract  Services provided to rape victims by human service professionals are usually helpful but are occasionally very harmful in that the victim ends up feeling ‘re-victimized’. This may be caused by the attitudes of the professionals towards the victims based on beliefs regarding rape and gender roles. Japanese human service professionals were solicited for responses to the Rape Myth Scale (RMS) and the short form of the Scale of Egalitarian Sex Role Attitudes (SESRA-S). One interpretable factor was extracted according to an exploratory factor analysis. Impact of the participants' age, sex, residential area, and type of profession on rape myth acceptance were examined by four-way layout anova. Nurses had significantly higher rape myth acceptance than any other professional group. Furthermore, a structural equation model showing the contribution of sex role egalitarian attitude to rape myth acceptance was established.

INTRODUCTION

Attitudes toward sexual offences differ from one society to another and are dependent on culture. In many societies the issue of sexuality is taboo, and the prevalence of sexual offences has been underestimated. In 1979 Finkelhor pointed out similarities between rape and child sexual abuse, stating that ‘society has in the past treated both offences similarly, in effect, denying that they were important and blaming the victim for their occurrence’.1 In 1999 Myers et al. examined how rape and sexual abuse were treated by professionals before 1975 and concluded that rape and sexual abuse were underrated, allegations of sexual offences were sometimes judged to be fabricated in the courts, and sexual offences were thought to be harmless.2 Moreover, there was an idea that victims' seduction provoked their own victimization. Infants who were born as a result of rape used to be killed in some societies.3

Sexual offences did not become an important issue until 1970 when, reflecting the women's movement, many researchers began to discuss the fact that the prejudicial, stereotyped beliefs about rape, rape victims, and rapists played a part in creating a climate hostile to rape victims.4–7 Burt coined the term ‘rape myths’ in 1980 to describe false beliefs or social perceptions about rape that deny or reduce perceived injury or blame victims for their own victimization.4 Recent studies concerning sexual victimization have often focused on the meaning of the experience in the context of the victim's culture. As in any other situation, the victim's individual illness is thus constructed through a blend of sociocultural meaning and individual experience or context.

The well-being of rape victims is influenced by the societal response, including the response of professionals who provide post-assault care. Services provided by community agencies such as police, hospitals, mental health clinics, and rape crisis centers are helpful. However, these services are occasionally distressing because of professionals' insensitivetreatment, including victim-blaming, victim-doubting, and victimization-neglecting attitudes. Several terms have been coined to describe such experiences: the second rape,8 the second assault,9 and the secondary victimization.10 One factor associated with the second rape is the rape myth.11 Even professionals often feel disgust regarding such unusual sexual intercourse. They are reluctant to face, intervene, and deal with the sexual abuse or victimization. The higher the rape myth attitudes of community service agency professionals, the more frequently the second rape occurs.

The first purpose of the present study was to assess professionals' attitudes toward rape and rape victims. However, in Japan we do not have an inventory to assess those attitudes.

Burt developed the Rape Myth Scale (RMS) to assess societal victim-blaming and victim-doubting attitudes.4 The RMS consists of only 14 items. This measure has been used in many studies11–26 but its factorial structure has not been widely investigated.

An important research and clinical topic related to the rape myth is its determinants. Burt reported that the rape myth was predicted by sex stereotyping, sexual conservatism, and adversarial sexual beliefs.4 Several other researchers have focused on the relationship between sex role stereotyping and acceptance of rape myth. Marciniak also pointed out the contribution of traditional gender role stereotyping to the rape myth acceptance level.20 Several researchers have focused on the relationship between sex role stereotyping and the acceptance of rape myth. Thus Jimenez and Abreu compared Latino and Caucasian undergraduates in terms of rape myth acceptance and empathy towards victims using a written acquaintance rape vignette.27 They showed that Caucasian women had more positive attitudes towards the victims and less rape myth acceptance level compared with the Latino counterpart. They speculated that this was because of the higher traditional set of sex-role values in Latinos. They thought that these traditional sex-role values privilege the men at the expense of women. Check and Malamuth examined how respondents evaluated vignettes of three different rape situations (stranger rape condition, acquaintance rape condition, and mutually consenting condition) and presented that high sex-role-stereotyping individuals were more likely to perceive the rape victim reacting more favorably to the assault than were low sex-role-stereotyping ones, particularly in the case of acquaintance rape.28 Abrams et al. examined how the victim-blaming attitude and proclivity to commit rape are related to sexism by discerning the benevolent (e.g. women should be cherished and protected by men) and hostile sexism.29 Higher benevolent sexists blamed the victims more. They noted that the benevolent sexists perceived the victim as having behaved in a manner that is inappropriate for a woman and were more likely to view a woman who invited a relationship with a man as being responsible for anything unfortunate that may happen to the victim. White and Kurpius reported that the victim-blaming attitude was higher among men than in women and that victim-blaming attitude was accounted for by the gender role attitudes.30 Simonson and Subich examined the perception of four types of rape: stranger, acquaintance, date, and marital rape, using scenarios.31 They showed that respondents holding less traditional gender-role stereotypes perceived rape scenarios as more serious and were less likely to blame the victim. Additionally, they presented the result that marital rape was less perceived as rape. Sex role stereotyping is concerned with people's minimizing attitudes, particularly in the case of rape within intimate relationships, including the marital relationship.

These results suggest that there are fixed standards among people with high stereotyped conservative gender roles in that women should be coy, submissive and stay at home, and if women violate these expectations then they are supposed to take responsibility for negative experiences including rape. Particularly, even in modern Japanese society, a husband introduces his wife as my kanai, which means ‘person in the house’, and people call a woman who is married as okusan, which means ‘behind’. A wife introduces her husband as shujin, which means ‘master’. Therefore, women are supposed not to work actively but stay at home and take responsibility for the household and support her husband from behind. The women's attitudes that are described in the RMS may be opposite to the social expected attitudes towards women. People's image of sex role may precede the rape myth acceptance. We hypothesized that the higher the sex role egalitarian attitude, the lower the rape myth acceptance. Therefore, sex role egalitarianism may explain rape myth acceptance.

In Japan, attitudes toward rape and rape victims have rarely been the subject of empiric studies. This is coupled with a lack of reliable and valid measures of such attitudes. We translated the RMS and used it among different types of Japanese populations. We wanted to assess the contribution of sex role egalitarianism to the rape myth acceptance level in Japan. The purpose was to go beyond calculating correlations between the two concepts and determine whether sex role stereotyping attitudes constitute a part of the RMS using structural equation analysis. This will support the construct validity of the RMS.

The purposes of the present study were (i) to eliminate from the RMS items that are inadequate to assess the rape myth acceptance level in Japanese culture; (ii) to examine the influence of gender, age, type of profession, and residential area in the score of rape myth acceptance; and (iii) to assess the contribution of attitude of egalitarianism to rape myth acceptance using a structural equation model.

METHODS

Participants

Human service professionals (nurses, health visitors, social workers, psychiatrists, clinical psychologists, schoolteachers, pediatricians, gynecologists, pharmacists, nutritionists, telephone counsellors, speech therapists, and occupational therapists) were solicited to participate in a questionnaire survey. Participation was requested at one general social service counselling center, four health centres, six psychiatric hospitals, three psychiatric clinics, one junior high school, one technical college, two psychiatric departments at general hospitals, and one infants' and neglected children's home. We asked the head of each institution to distribute the questionnaire to employees and confirm an autonomous decision as to participation. The questionnaire was answered anonymously and returned to the researchers. Of a total number of 774 distributed questionnaires, 662 (85%) were returned and 578 were usable for further analyses (Table 1). As expected, there were a greater number of female nurses and health visitors (χ2 = 55.65, P < 0.001, d.f. = 6). More health visitors, psychiatrists, and psychologists than other professionals lived in urban areas (χ2 = 80.15, P < 0.001, d.f. = 6). There was a significant difference in the mean age across different professions (F = 3.25, P < 0.01).

Table 1.  Participant details according to profession
 nFemale/male***Urban/rural area***Mean age ± SD**
  • ** 

    P < 0.01;

  • *** 

    P < 0.001.

Nurses354268/86 92/26241.49 ± 10.69
Health visitors4343/033/1039.53 ± 11.71
Social workers31 15/1616/1534.55 ± 10.90
Psychiatrists27 8/1918/8 40.33 ± 13.68
Psychologists13 8/57/634.62 ± 9.89
School teachers13 9/44/937.15 ± 9.16
Others97 21/7658/3938.29 ± 12.68
Total578 426/152228/34740.10 ± 11.39

Measurement

Rape Myth Scale

The RMS (Burt, 1980) consists of 14 items: 11 items measured on a 7-point scale and three items measured on a 5-point scale.4 Among these items, two are reverse questions. A higher score indicates higher rape myth acceptance. With the original author's permission, the RMS was translated into Japanese by the senior author. To verify the translation, the Japanese was back-translated into English by a person who was unaware of the original wordings. Three items were eliminated from the Japanese version of the RMS. The first item, ‘A woman comes to you and claims she was raped. How likely are you to believe her statement if she were: Your best friend? An Indian woman? A neighborhood woman? A black woman? A white woman?’, was eliminated because it did not fit Japanese culture. We did notice a substantial cultural difference regarding the remaining items. Two questions, ‘What percentage of women who report a rape would you say are lying because they are angry and want to get back at man they accuse?’ and ‘What percentage of reported rapes would you guess were merely invented by women who discovered they were pregnant and wanted to protect their own reputations?’, were eliminated because they differed from the other 11 items for the following reasons. First, these two items were measured on a Likert-type 5-point scale, whereas the other items were measured on a 7-point scale. Furthermore, in these two items, the participants were asked to designate a level of estimation on a percentage scale as follows: almost 100%, approximately 75%, approximately 50%, approximately 25%, almost 0%. In contrast, the other items asked the participants to choose one of seven degrees ranging from strongly agree to strongly disagree. This might produce an artificial factor that is not interpretable. To obtain the soundness of the inventory, all items should be measured on the same Likert type.

Short version of the Scale of Egalitarian Sex Role Attitudes

The Scale of Egalitarian Sex Role Attitudes (SESRA) was developed by Suzuki to assess the degree ofegalitarianism in sex role attitudes.32 An English version of this scale was introduced in the United States to examine cross-cultural differences; the results of such studies have been previously reported.33,34 The SESRA consists of 40 items measured on a Likert-type 5-point scale from strongly disagree (1) to strongly agree (5). Among the items in the inventory, 15 are reverse questions such as ‘Bringing up children is the most important job for a woman’. Twenty-five non-reverse items are included in the inventory, including ‘For a woman, the roles of wife and mother are important, but working outside is equally important.’ Higher scores are indicative of more egalitarian attitudes. The short version of the Scale of Egalitarian Sex Role Attitudes (SESRA-S) was invented by Suzuki for convenience.35 Fifteen items that met both of the following requirements were selected from the original SESRA: (i) carrying >0.5 factor loading on the first factor according to the exploratory factor analyses (varimax rotation); and (ii) lack of significant factor loading on any other factors. The reliability and construct validity of the SESRA-S were shown using a Japanese population. The test–retest reliability of the SESRA-S was 0.91 and the correlation between the total scores of the SESRA and SESRA-S was 0.94.35 Uji et al. conducted exploratory factor analysis of the SESRA-S and found two factors of the inventory.36 The description of the items belonging to the first factor generally addresses the idea that women are supposed to take traditional women's role. They thought that the opposite meaning of these items addresses the release of women from being a home-maker and named this ‘Women’s Right'. As to the items belonging to the second factor, they named this ‘Women’s Independence' because those items address women's independence and equality to men.

Demographic variables

Demographic data included age, gender, type of profession, residential area, and institute of employment.

Statistical analysis

We performed an exploratory factorial analysis on the items of the RMS. The number of factors was determined by the scree test.37 The maximum-likelihood method (ML) estimation was adopted because of its robustness as an estimation method. PROMAX rotation, an oblique rotation, was used because we presumed that the extracted factors would be related.

The influence of demographic variables such as age, sex, residential area, and type of profession on the score of the short version of the RMS (RMS-S) was examined by four-way layout anova.

Relationship between the SESRA-S and RMS-S

We constructed a model based on the presumption that the RMS-S could be explained by the subscales of the SESRA-S. The validity of this model was evaluated by structural equation modeling. The goodness of fit of the model to the data was expressed by the goodness of fit index (GFI), adjusted goodness of fit index (AGFI), and root mean square error of approximation (RMSEA).

RESULTS

Characteristics of the utilized variables

The mean ± SD of the score for each item of the RMS is shown in Table 2.

Table 2.  Factor loading on each item of the RMS
 Mean ± SD2-factor solution
Factor 1
Single factor Factor 2Item to total solutioncorrelation
  1. RMS, Rape Myth Scale.

RMS1 A woman who goes to the home or apartment of a man on their first date implies that she is willing to have sex.3.90 ± 1.820.400.690.480.59
RMS2 Any woman can get raped.3.69 ± 2.04−0.230.20Not analyzedNot calculated
RMS3 One reason that women falsely report a rape is that they frequently have a need to call attention to themselves.3.62 ± 1.730.270.470.330.45
RMS4 Any healthy woman can successfully resist a rapist if she really wants to.3.17 ± 2.160.500.290.490.59
RMS5 When women go around braless or wearing short skirts and tight tops, they are just asking for trouble.3.19 ± 1.760.660.590.690.73
RMS6 In the majority of rapes, the victim is promiscuous or has a bad reputation.2.82 ± 1.820.610.290.580.62
RMS7 If a girl engages in necking or petting and she lets things get out of hand, it is her own fault if her partner forces sex on her.4.59 ± 1.770.530.640.590.65
RMS8 Women who get raped while hitchhiking get what they deserve.3.11 ± 1.940.680.490.690.71
RMS9 A woman who is stuck-up and thinks she is too good to talk to guys on the street deserves to be taught a lesson.2.75 ± 1.670.790.350.720.72
RMS10 Many women have an unconscious wish to be raped, and may then unconsciously set up a situation in which they are likely to be attacked.1.72 ± 1.300.550.250.520.55
RMS11 If a woman gets drunk at a party and has intercourse with a man she's just met there, she should be considered ‘fair game’ to other men at the party who want to have sex with her too, whether she wants to or not.3.73 ± 1.930.630.530.660.69
Percentage of variance explained 36.912.540.6 

Factor structure of the RMS

Exploratory factor analysis was carried out on the 11 items of the RMS. Two factors were extracted using the scree test (Table 2). Even if we adopt factors for which eigenvalues are more than unity, two factors were extracted again. Together, these two factors explained 49.3% of the variance of the RMS. Only item 2 had a small factor loading on both factors (<0.3), indicating that this item was not appropriate for inclusion in the RMS. All the other items had positive large factor loading (>0.4) on either or both of two factors. The two extracted factors had a relatively strong correlation (r = 0.61). Furthermore, the differences of these factors were not interpretable. Hence, we thought that a single factor solution (excluding item 2) would be more appropriate, and determined the number of factors as 1.

We conducted another exploratory factor analysis on these items fixing the number of factors at 1 (Table 2). Kaiser–Meyer–Olkin analysis was 0.879. Minimum factor loading acceptable for main factor was 0.3. Thus, all items except item 2 were adopted in the RMS-S.

Determinants of the RMS-S

Older age correlated with higher rape myth acceptance (r = 0.182, P < 0.01). The total score of the RMS-S did not differ significantly between women and men (Table 3). The total score was significantly higher in people living in rural areas than those living in urban areas (Table 3). There were significant differences among people in different professions. Scheffe's post-hoc comparison showed that the RMS-S total score was significantly higher among nurses as compared to health visitors (P < 0.01), social workers (P < 0.01), and psychologists (P < 0.01).

Table 3.  RMS-S score by gender, residential area, and profession
 RMS-S score ± SD
  • *** 

    P < 0.001.

  • RMS, Rape Myth Scale.

Sex
 Women 33.6 ± 11.3
 Men 33.1 ± 10.8
 t-testt = −0.367
Residential area
 Urban 30.7 ± 10.7
 Rural 35.0 ± 11.2
 t-testt = −4.030***
Profession
 Nurses 36.0 ± 10.8
 Health visitors 27.6 ± 10.3
 Social workers26.3 ± 8.7
 Psychiatrists28.2 ± 6.5
 Psychologists22.5 ± 9.1
 School teachers23.3 ± 9.5
 anovaF (5, 474) = 16.71***

Because the four variables (age; gender; residential area; and profession) were interrelated, the effects of the predictor variables on the RMS-S score were examined by four-way layout anova: the total RMS-S score was regressed on the participants' age (<21, 0; 21–30, 1; 31–40, 2; 41–50, 3; 51–60, 4; 61–70, 5; 71–80, 6; >81, 7), gender (male, 0; female, 1), residential area (urban, 0; rural, 1), and profession (nurse, 0; health visitor, 1; social worker, 2; psychiatrist, 3; psychologist, 4; schoolteacher, 5). Only profession (F = 5.47, P < 0.001) had a significant effect on the RMS-S score. Scheffe's post-hoc comparison showed that nurses had significantly greater rape myth acceptance attitudes than any other group of professionals. No interaction was observed between any two, three, or four variables.

Based on the hypothesis that, to some extent, egalitarian sex role attitudes contribute to rape myth acceptance, a structural equation model was built (Fig. 1). There were three latent variables: Rape Myth Acceptance, a factor that was abstracted from the RMS-S; and two subscales of the SESRA-S: Women's Rights and Women's Independence. Furthermore, we presumed a correlation between Women's Rights and Women's Independence because an exploratory factor analysis produced a correlation between these two factors of 0.312.36 The fitness of this model was as follows: GFI (0.850), AGFI (0.821) and RMSEA (0.074). Women's Rights contributed to the Rape Myth Acceptance with fairly negative value (causal coefficient: −0.55, P < 0.00) and Women's Responsibilities contributed to the Rape Myth Acceptance with fairly positive value (causal coefficient: 0.12, P < 0.05).

Figure 1.

Relationship between the two factors extracted from the short version of the SESRA (SESRA-S) and the factor extracted from the short version of the RMS (RMS-S). RMS, Rape Myth Scale; SESRA, Scale of Egalitarian Sex Role Attitudes.

DISCUSSION

This study has demonstrated that, excluding item 2, the items of the RMS constitute a single factor structure. Item 2 asks if the respondent agrees that ‘any women can get raped’. This is a reverse item: if the respondent answers ‘no’, it is interpreted that the subject believes that not all women are likely to be raped, but rather that there is a specific group of women who are likely to be raped. In the Japanese linguistic text, this item may be interpreted as either ‘any women can get raped’ or ‘no women can get raped’. Although the translation of this item was precise as demonstrated by back-translation, the wording failed to convey the implicit meaning of the original question.

The two factors initially extracted from the RMS correlated relatively strongly with each other and we failed to identify differences in the meanings of the two factors. Thus, we believed that the RMS represented a single factor structure. Because the internal structure of the RMS has not been well studied, we believe that our results support the scale's utility as a dimensional measure of the rape myth.

Burt has reported that age and education level influence the level of rape myth acceptance and stereotyped sex role attitudes: younger and more well-educated persons have lower levels of rape myth acceptance and lower stereotyped sex role attitudes.4 Because participation in the present study was limited to professionals, educational levels did not appear to diverge widely. Hence, participants were not asked about their education level.

Varelas and Foley reported that women responded more negatively to rapists and more positively to victims than men.26 Contrary to these results, the present results did not demonstrate gender differences in rape myth attitudes. This may be because Japanese women are not as strongly influenced by the women's movement as women in Western societies and have rape myth attitudes that are as strong as those held by men. Alternatively, the men in the present study may have less stigmatizing attitudes towards rape victims because of their professions. In regard to the type of profession, nurses were the most likely to have stigmatizing attitudes towards rape victims. This may be because the nurses in the present study were employed at psychiatric hospitals and had little opportunity to be educated about how to treat and deal with victims of sexual violence. Japanese psychiatric hospitals are relatively isolated from society, which may contribute to the traditionally conservative attitudes of mental hospital nurses. Generally, older age is linked to stronger stigmatizing attitudes towards rape victims. The present results did not support this, probably because greater experience as professionals offsets the effects of age.

Contrary to our expectation that the two subscales of the SESRA-S, Women's Rights and Women's Independence, would be linked to lesser stigmatizing attitudes towards rape victims, the present study has shown that rape myth acceptance was greater among those low in Women's Rights and those high in Women's Independence. The Women's Rights subscale includes items such as ‘Women at home and men at work’, ‘Important issues should be decided by husbands’, and ‘Bringing up children is the most important job for women’ (all reverse items). Attitudes against these statements imply the release of women from the traditional conservative social value system. These attitudes are easily linked to antistigmatizing attitudes towards rape victims. In contrast, the Women's Independence subscale includes items such as ‘Working outside is equally important for women’ and ‘Women should work even after having a child’ (all non-reverse items). These items regard women as responsible individuals. Such egalitarian attitudes towards women's independence may have created a new viewpoint that regards rape victims as those who have failed to take on their own responsibilities.

Several limitations of the present study should be discussed. First, our participants were not from the general population. Moreover, the distribution of the participants' sex and type of profession was skewed. It was not possible to solicit questionnaires from police officers, who are first responders and have many opportunities to meet rape victims, but they should be included in a future study. We excluded some items from the study, including ‘A woman comes to you and claims she was raped. How likely are you to believe her statement if she person were: Your best friend? An Indian woman? A neighborhood woman? A black woman? A white woman?’ Because Japan is an ethnically less diversified country, this question was not applicable. In contrast, in Japan, people's norm violation is easy to be identified, and women who do not meet the social expectation are likely to be evaluated as ‘inappropriate women’. Thus, instead of this item, in future studies we should discuss victim's attributes that may elicit Japanese people's negative attitudes about being raped, and develop RMS that reflects Japanese culture. However, in the present study we did not develop the Japanese rape myth. The reason why we only translated the items of RMS and used them in the present study was that we wanted to compare the results across countries and cultures. According to Burt, the respondents' gender and the age affect the level of rape myth acceptance. In the present study this was not the case. If we had developed a scale that was unique to Japanese culture we would not have been able to determine whether the difference between Burt's and our results was because of the differences between the scales or other factors such as cultural difference or sample difference.

We used egalitarian sex role attitudes as a construct of rape myth acceptance. There may be many other concepts that can be used as constructs. These include past personal experience as professionals or individuals related to victims. Zurbriggen and Yost reported that greater acceptance of the rape myth was associated with fantasies of dominance among men, whereas it was associated with emotional and romantic fantasy themes among women.38 Thus, different constructs may exist between men and women.

It is a very important research topic to examine the extent to which rape myth acceptance attitudes, as measured by the RMA scale, reflect the actual behaviors of mental health professionals who deal with rape victims. The attitudes of these professionals may result in divergent treatments of victims of sexual offences. It is not as important to forcefully control and change individual beliefs about rape, but to notice that we are all influenced by surrounding multidimensional beliefs and we are likely to unconsciously introject such beliefs. We must always examine our own beliefs and be aware that our counter-transference toward clients originates from these beliefs.

ACKNOWLEDGMENT

The authors would like to thank all the participants in the present study.

Ancillary