Abstract
National and international research has repeatedly identified the specific gendered nature and context of women’s victimisation of violence, whereby women are disproportionately victims of sexual and partner violence and overwhelmingly at the hands of known male perpetrators. As such, violence against women warrants a targeted and substantial focus, within overall violence reduction and prevention efforts. In the Australian policy context, there is an emerging and influential focus on attitudes towards violence against women as key targets for primary prevention and as foci for monitoring progress in reducing this violence. The Australian National Community Attitudes Towards Violence Against Women Survey was established to estimate community-level understanding of and attitudes towards violence against women. It has arguably evolved into an important instrument both for monitoring shifts in Australians’ knowledge and attitudes, as well as for directing primary prevention efforts. The purpose of this article is to provide an integrative review in relation to one of the key dimensions of the national community attitudes survey: violence supportive attitudes (see Webster et al., 2014). Here, we seek to identify patterns in defining and measuring attitudes that support violence against women, as well as advance the field by offering recommendations for progressing the measurement of violence supportive attitudes in Australia in the future.
Introduction
Violence against women is a significant and pervasive global health and human rights problem (World Health Organization, 2005). In Australia, figures from the Australian Bureau of Statistics (ABS) report that one in five women have experienced sexual violence since the age of 15, while one in six report having experienced physical or sexual violence from an intimate partner (ABS, 2014). Overall both nationally and internationally, research evidence regarding the prevalence, under-reporting, impacts and costs of violence against women all make evident that its prevention is crucial to improving health, wellbeing and quality of life for women and families. The World Health Organization (WHO, 2005, 2009) has highlighted the importance of primary prevention in particular and asserts that this area has received too little attention in public health and health promotion. To this end, there have been significant advances in policy frameworks, strategic prevention plans, and funded projects focused on primary prevention over the last 10 years in particular and most notably in the United States (CDC, 2013; DeGue et al., 2014), the European Union (Council of Europe, 2014), and Australia (Commonwealth of Australia, 2011; OurWatch, 2015; State Government of Victoria, 2010; VicHealth, 2007). Indeed, the Victorian Health Promotion Foundation (hereafter, VicHealth) developed one of the first comprehensive frameworks identifying the opportunities for primary prevention of violence against women (VicHealth, 2007). Drawing on a public health approach, the VicHealth framework, Preventing Violence Before it Occurs, has proven to be influential in subsequent government policy and prevention programme development and funding in Australia. Most recently the framework has substantially informed a subsequent national plan for the primary prevention of violence against women and their children launched in November 2015 (OurWatch, 2015). The rationale behind these frameworks and indeed much Australian primary prevention policy work, as elsewhere, is that violence against women is reduced in communities where knowledge of violence against is accurate, where there is low attitudinal and normative support for violence, and where there is a high level of normative, structural and practical support for gender equality in public and personal life (e.g. Department of Premier and Cabinet, 2013; OurWatch, 2015; State Government of Victoria, 2010; VicHealth, 2007).
As part of an ongoing policy commitment to reducing violence against women, successive Australian federal governments have funded three national surveys of community attitudes since 1995, with a further survey scheduled in 2016–2017. The most recent national community attitudes towards violence against women survey (hereafter, NCAS) includes questions that seek to measure (a) understandings or knowledge of violence against women, (b) attitudes towards violence against women (including attitudes that blame victims and/or excuse perpetrators), and (c) attitudes towards gender equality (VicHealth, 2014; Webster et al., 2014). The NCAS arguably acts as an important benchmarking tool which, at a broad societal level, provides an indication of change over time in each of the three key dimensions: knowledge of violence against women, attitudinal support for violence against women, and endorsement of gender equality. While the NCAS does not serve as an evaluation of the range of primary prevention strategies and programmes operating across Australia, the data are indicative of which aspects of knowledge and attitudes towards violence against women may benefit most from continued prevention efforts. It can also help to identify groups within the population that ought to be prioritized for prevention efforts. In these ways it also serves as a monitoring tool on the basis of which strategic decisions about investment in primary prevention may be made. For these reasons, as both a national benchmarking and monitoring tool, it is crucial that the NCAS questions and constructs hold up and reflect the substantial development of empirical work in recent years. Yet the NCAS itself was first developed in 1994 and, though there have been some changes to questions in subsequent surveys (in 2006, 2009, 2012), many of the original questions were developed at a time when the scholarly literature concerning community attitudes towards violence against women was comparatively scant. It is timely then, to reconsider the NCAS in light of developments made in the field particularly in the last decade.
The purpose of this article is to provide an integrative review in relation to one of the key dimensions of the NCAS: violence supportive attitudes (see Webster et al., 2014). Here, we seek to identify patterns in defining and measuring attitudes that support violence against women, as well as advance the field by offering recommendations for progressing the measurement of violence supportive attitudes in Australia in the future. To meet this aim, this review encompasses (1) empirical quantitative studies with adult community and university student samples, (2) studies over the 10-year period of 2006 to 2016 (original scales used in these studies are also included and some of these may be recorded in articles predating this period), (3) studies that sought to measure ‘attitudes’, ‘myths’, ‘perceptions’ or ‘beliefs’, and (4) which were concerned with four forms of interpersonal violence against women, namely rape, sexual harassment, domestic violence and stalking.
In the following section, we first provide some further background to the concept of violence supportive attitudes, and provide an overview of the empirical evidence regarding the relationship between such attitudes and victimization the prevalence of violence itself. In the subsequent sections, we review the empirical literature concerning definition and measurement of community attitudes and perceptions in relation to each of: rape or sexual assault, sexual harassment, domestic or intimate partner violence, and stalking. Finally, we discuss common underlying dimensions of these constructs as well as suggested ways forward in measuring community attitudes towards violence against women in Australia.
Violence supportive attitudes: Background to the concept
Many factors contribute to violence against women and these interact with and reinforce one another (OurWatch 2015; VicHealth 2007; WHO, 2010); among this range of factors are attitudes supportive of violence against women. The NCAS defines violence supportive attitudes as those which: Justify, excuse, minimise or trivialise physical, sexual and other forms of violence against women, or blame or hold women at least partly responsible for violence perpetrated against them …. such attitudes expressed by influential individuals or held by a substantial number of people can create a culture in which violence is at best not clearly condemned and at worst condoned or encouraged. (VicHealth, 2014, p. 8)
Further, collectively held attitudes influence the development of social norms and institutional and organizational structures and practices, which in turn are associated with violence against women (OurWatch, 2015). At the same time attitudes are themselves socially determined, being shaped by practices, structures and norms in family, community, organizational and broader institutional environments. This evidence suggests that attitudes, while not the only factor, play an influential part in the aetiology of violence against women and hence are among the targets of interventions to prevent the problem. Both in their own right and because they reflect broader social norms, structures and practices implicated in violence against women, they are also arguably an important barometer of our progress in addressing the problem. Yet, most large-scale general community surveys are focused on establishing prevalence of violence against women, rather than measuring psychometric dimensions such as beliefs, attitudes, perceptions, tolerance, and cultural support as correlates of violence (see Basile Hertz, & Back, 2007; Schwartz et al., 2012).
Attitudinal support for sexual violence
Since the 1970s, sociologists and feminists have employed the concept of ‘rape myths’ to refer to a number of widely held cultural beliefs that misrepresent rape, rape victims, and rapists. Numerous measures have been developed to assess: rape myth acceptance (Burt, 1980; Gerger, Kley, Bohner, & Siebler 2007; McMahon & Farmer, 2011; Payne, Lonsway & Fitzgerald, 1994), rape supportive attitudes and beliefs (Barnett & Feild, 1977; Burgess, 2007) as well as perceptions of rape (Cowan & Quinton, 1997), and rape victims (Ward, 1988). Despite some measures emerging as dominant (discussed below), inconsistencies in the measurement of attitudinal support for sexual violence preclude analyses that can compare various populations or changes over time (see Edwards, Turchik, Dardis, Reynolds, & Gidycz, 2011; Suarez & Gadalla, 2010). Many existing scales measure concepts which, while related to each other, may also be conceptually distinct (Suarez & Gadalla, 2010); ‘rape myths’ may not be equivalent to ‘attitudes towards rape victims’ for instance. Moreover, as discussed by Edwards et al. (2011), there is a lack of consistency in language and definitions used in many rape myths and attitudes measures as compared with prevalence surveys of sexual violence victimisation. For example, many measures specifically use the word ‘rape’ (e.g. ‘If a woman doesn’t physically fight back, you can’t really say that it was rape’, Lonsway & Fitzgerald, 1994), while victimisation scales more commonly refer to descriptive definitions (e.g. ‘Has any man/woman ever forced you, or tried to force you, into sexual activity against your will?’, Australian Bureau of Statistics, 2014). Additionally, the vast majority of survey research into rape myths has been conducted in the United States, and often with college ‘convenience’ samples (such as first year undergraduate students), meaning that few studies have validated measures across countries and with general populations. These considerations are particularly important when reviewing measures that might be suited to adoption in the Australian context (see Xenos & Smith, 2001).
Internationally, much research surveying attitudinal support for sexual violence employs one of three measures: The Rape Myth Acceptance Scale (RMAS; Burt, 1980), the Sexual Beliefs Scale (SBS, Muehlenhard & Felts, 1998), and the Illinois Rape Myth Acceptance Scale (IRMAS; Payne et al., 1999). 1 The RMAS comprises 11 statements (examples include ‘Any healthy woman can successfully resist a rapist if she really wants to’ and ‘In the majority of rapes, the victim is promiscuous or has a bad reputation’), as well as three open-ended questions (Burt, 1980). Responses are recorded on a 7-point Likert scale ranging from strongly agree to strongly disagree, with an originally reported Cronbach alpha of .88 indicating good internal consistency. Meanwhile, the SBS comprises 40 items (long form), and a 20-item short-form version, with statements such as ‘When girls say no, they often mean yes’, and ‘Girls say No so that guys don’t lose respect for them’ (Muehlenhard & Felts, 1998). The instrument also specifically seeks to measure five sub-domains of individuals’ beliefs about rape, these are: Token Refusal, Leading on Justifies Force, Women Like Force, Men Should Dominate, and a positive sub-domain that No Means Stop. Each sub-domain includes eight items in the long form and four items in the short form. Examples of items forming the latter positively framed domain of No Means Stop include: ‘When girls say No, guys should stop’, and ‘If a girl doesn’t want sex, the guy has no right to do it’. Each item is rated on a 4-point likert scale from 0 (disagree strongly) to 3 (agree strongly). Internal reliability (Cronbach’s alpha) is reported by the authors as slightly higher for the long-form version of the instrument, across each of the five sub-domains (e.g. for Token Refusal, .84 and .71 respectively, and for No Means Stop .96 and .94, respectively), though the authors’ note that participants report the long form as repetitive and that correlations between the two scales were high. They therefore recommend using the short-form version (see Muehlenhard & Felts, 1998). Finally, the original IRMAS comprises 45 items with a 20-item short-form version (IRMA-SF) using a 7-point scale ranging from 1 (not at all agree) to 7 (very much agree), thus higher scores indicating a greater acceptance of rape myths. The IRMAS includes statements such as ‘Rape happens when a man’s sex drive gets out of control’, ‘A lot of women lead a man on and then they cry rape’, and ‘If a woman is raped while she is drunk, she is at least somewhat responsible for letting things get out of control’, with a previously reported Cronbach alpha of .93 indicating high internal consistency (Payne et al., 1999).
Despite being widely used, the RMAS, SBS, and IRMAS arguably have a number of potential limitations for the Australian context. For instance, some of the language in the RMAS is outdated or rarely used in contemporary Australia (such as ‘first date’, ‘promiscuous’, ‘necking’, and ‘petting’). Both the RMAS and the IRMAS also use the word ‘rape’ as opposed to ‘sexual violence’, ‘sexual assault’, or ‘forced sex’, and with no studies that have validated either measure with an Australian general population, the relevance of such language in the Australian context is unclear. The SBS meanwhile does use language such as ‘forced sex’, which might have the advantage of being capturing attitudes which might not explicitly support ‘rape’ per se, but in fact support aggressive or non-consensual sex regardless of whether the individual respondent would define this as ‘rape’. This is arguably an important distinction to make when measuring community attitudes towards sexual violence, since there is much variation in understandings of what constitutes ‘rape’ (Kahn, Jackson, Kully, Badger, & Halvorsen, 2003; Larcombe, Fileborn, Powell, Henry, & Hanley, 2015; Peterson & Muehlenhard, 2004).
In recent years, a number of scholars have argued that there is a need to update measures of rape myth acceptance, suggesting (a) that respondents may be more aware of the ‘politically correct’ answers in response to overtly hostile and victim blaming attitudes, (b) that the content of rape myths may have changed as societies have become more gender aware (see Gerger et al., 2007; McMahon & Farmer, 2011), and (c) that there is conceptual overlap between current RMASs and a more generalised hostility or aggression towards women (Grubb & Turner 2012; Hockett, Smith, Klausing, & Saucier, 2016; Lonsway & Fitzgerald 1994, 1995). In response to such observations, Gerger et al. developed the acceptance of modern myths about sexual aggression scale (AMMAS), which is available in English, German (Gerger et al., 2007), and Spanish (Megías, Romero-Sánchez, Durán, Moya, & Bohner 2011). Examples from the 30-item scale, which they developed to incorporate less overt rape myths, include: ‘When it comes to sexual contacts, women expect men to take the lead’, ‘Alcohol is often the culprit when a man rapes a woman’, and ‘It is a biological necessity for men to release sexual pressure from time to time’ (Gerger et al., 2007).
Some studies have further found support for a multidimensional factor structure of rape myths. For example, Briere et al. (1985) report four factors: disbelief of rape claims; victim responsible for rape; rape reports as manipulation; and rape only happens to certain kinds of women (based on analyses of Burt’s (1980) RMAS). Similar themes emerged in Payne et al.’s (1999) analysis of the IRMAS, which identified seven conceptually distinct dimensions of rape mythology: she asked for it; it wasn’t really rape; he didn’t mean to; she really wanted it; she lied; rape is a trivial event; and rape is a deviant event. As discussed above, Muehlenhard and Felts (1998) report five sub-domains of attitudes towards rape: Token Refusal, Leading on Justifies Force, Women Like Force, Men Should Dominate, and No Means Stop.
However, there is not widespread agreement in the literature regarding the extent to which rape supportive attitudes might represent a multidimensional as opposed to a single coherent construct (see e.g. Gerger et al., 2007). Indeed overall, attitudes that minimise, deny or justify sexual violence, blame rape victims and/or excuse perpetrators are also strongly associated with: hostile attitudes towards women generally; sexism and heterosexism; acceptance of interpersonal violence (AIV); and adversarial sexual beliefs (Grubb & Turner, 2012; Hockett et al., 2016; Suarez & Gadalla, 2010). This suggests that even the subdomains or multi-dimensions of rape supportive attitudes are each tapping into a common underlying construct. For example, the findings of several reviews suggest a potential overlap in the concepts of rape myths with attitudes that are more generally hostile or aggressive towards women (see Grubb & Turner, 2012; Hockett et al., 2016; Lonsway & Fitzgerald 1994, 1995; Suarez & Gadalla, 2010). In other words, individuals who hold rape supportive attitudes are more likely to also endorse other attitudes supportive of violence and/or aggression towards women. This suggests the need to consider whether it is necessary or valuable for an attitudinal survey to ask separate and comparable questions regarding distinct forms of violence against women (such as those reviewed here), or whether a consolidated violence supportive attitudes construct is as predictive of, for example, rape supportive attitudes as it is attitudes that support sexual harassment, domestic violence, or stalking.
Attitudinal support for sexual harassment
Measures regarding attitudinal support for sexual harassment are somewhat less developed than for rape mythology. Despite a number of scales having been developed (see e.g. Cowan, 2000; Cowan & Ullman, 2006; Dekker & Barling, 1998; Lott, Reilly, & Howard, 1982; Reilly Lott, & Gallogly, 1986), two measures emerge as dominant in the international literature: the Sexual Harassment Attitude Scale (SHAS, Mazer & Percival, 1989), and the Illinois Sexual Harassment Myth Acceptance Scale (ISHMA, Lonsway et al., 2008). Mazer and Percival’s (1989) SHAS has been used extensively in previous research (see e.g. Kennedy & Gorzalka, 2002; Russell & Trigg, 2004), and is an extension of the earlier Tolerance for Sexual Harassment Inventory (TSHI, Lott et al, 1982; Reilly et al 1986). The 19-item SHAS assesses beliefs and tolerance of sexual harassment in academia, using a 5-point Likert response scale ranging from 1 (strongly agree) to 5 (strongly disagree). Higher scores indicate greater tolerance of sexual harassment and internal consistency is good with a Cronbach’s alpha of .84 (Mazer & Percival, 1989). Example items include ‘A lot of what people call sexual harassment is just normal flirtation between men and women’ and ‘It is only natural for a man to make sexual advances to a woman he finds attractive’.
A likewise widely used, yet more contemporary measure, is the 20-item ISHMA Scale (Lonsway et al., 2008; see also Diehl et al., 2014; Page et al., 2015). Example items include: ‘If a woman is sexually harassed, she must have done something to invite it’ and ‘Women can usually stop unwanted sexual attention by simply telling the man that his behavior is not appreciated’. Responses are rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) with a higher score indicating a greater endorsement of sexual harassment myths. Lonsway et al. (2008) report very good internal consistency (Cronbach’s alpha, .91) for the 20-item scale, with subscales representing four factors: fabrication/exaggeration, ulterior motives, natural heterosexuality, and woman’s responsibility. The scale has also been adapted and validated in both Spanish (Expósito et al., 2014) and German (Diehl et al., 2014) with likewise high internal consistency with general community populations in these sites.
As is the case for sexual violence, attitudinal support for sexual harassment has also been correlated with sexist ideology, hostility towards women, AIV, adversarial sexual beliefs, and rape myth acceptance (e.g. Cowan & Ullman, 2006; De Judicibus & McCabe, 2001; Mazer & Percival, 1989; Russell & Trigg, 2004). While endorsement of sexual harassment myths is strongly associated with gender, such that men typically report higher endorsement than women (see Lonsway et al., 2008), some research suggests that hostility towards women is a much greater predictor of tolerance of sexual harassment than gender (Russell & Trigg, 2004). Lonsway et al. (2008) further suggest that the core functions of sexual harassment mythology are justification and denial; themes which they note are shared with rape myths and thus illustrative of a common and underlying sub-domain across different forms of violence against women.
Attitudinal support for domestic and intimate partner violence
One of the earliest scales seeking to measure attitudes supportive of violence against women was Burt’s (1980) AIV Scale. The 6-item scale includes one generic item regarding acceptance of violence: ‘People today should not use “an eye for an eye and a tooth for a tooth” as a rule for living’, while the remaining items focus more specifically on sexual and partner violence. Other example items include: ‘A wife should move out of the house if her husband hits her’, and ‘Sometimes the only way a man can get a cold woman turned on is to use force’. While the AIV scale has been widely used in subsequent research (e.g. Malamuth & Check, 1981; Malamuth, Hald, & Koss, 2012), it was originally reported by Burt (1980) to have poor internal consistency (Cronbach’s Alpha, .586), suggesting that the items may not each measure the same construct that they purport to measure. Additionally, some studies have reported slightly higher, moderate international consistency such as with college student samples (e.g. Lonsway & Fitzgerald, 1995). However, similar to the critiques levelled at Burt’s (1980) RMAS, some studies have suggested that the constructs of attitudes supporting or endorsing particular ‘myths’ in relation to violence against women may be better explained by a common underlying construct of ‘hostility towards women’.
Perhaps surprisingly, given the attention to domestic violence research in recent years and scale development regarding prevalence and nature of partner abuse in particular (such as the Conflict Tactics Scale developed by Straus, 1979), there appears to be less agreement or consistent adoption of a set of measures for domestic violence ‘myths’, ‘attitudes’, or ‘beliefs’ (as compared with the widely adopted ‘rape myths’ scales for example). Rather, a wide variety of researcher-designed instruments have been developed in separate studies (see e.g. Briere, 1987; Carlson & Worden, 2005; Nayak, Byrne, Martin, & Abraham, 2003; Schwartz, Kelley, & Kohli, 2012; Worden and Carlson, 2005), often with little available information on the psychometric properties of the measure proffered. Another common approach is to ask a small number of questions regarding how serious and/or acceptable partner violence is (see e.g. Gracia & Tomas, 2014). This may reflect, at least in part, the focus of many contemporary studies on individual-level explanatory variables, such that attitudes or beliefs are measured with specific respect to violence perpetration, while research into community-level attitudes, and associated scales for their measurement, is comparatively lacking (see Scott & Straus, 2007).
Notable exceptions to this state of the research exist foremost in the justifications of wife beating items that have been developed in the Demographic Health Survey (DHS) and World Health Organisation (WHO) multi-country studies, resulting in their widespread adoption particularly in developing countries. For example, in a systematic review, Waltermaurer (2012) sought to identify measures of general community attitudes that justify partner violence. She identified both the DHS and WHO items as widely adopted in published research (see e.g. Hayati et al., 2011; Pierotti, 2013; Waltermaurer, 2012; Yount et al., 2011). Both measures adopt a similar set of items, and ask respondents to indicate their agreement or disagreement. Examples from the DHS include: ‘A husband is justified in beating his wife if … she goes out without telling him’, ‘if she neglects the children’, ‘if she refuses to have sex with him’ and ‘if she burns the food’. Meanwhile, example WHO items include: ‘A husband is justified in beating his wife if … she does not do the household chores well’, ‘she disobeys him’, and ‘she is unfaithful’. However, neither of these measures have been adopted in Western countries. Items designed to gauge support for justifying violence in the NCAS have shown that in contrast to some countries participating in the WHO and DHS studies, only a very small proportion of Australians (less than 4%) is inclined to believe that violence against women can be justified in any circumstance (Webster et al., 2014). Arguably, the measure is not appropriately sensitive for the purposes of an Australian general community attitudes survey on violence against women, firstly because it focuses only on wife abuse (itself a very narrow definition of violence against women), and secondly because it asks respondents only about severe and physical forms of abuse as indicated by the language of ‘beating’.
A further exception is the Domestic Violence Myth Acceptance Scale (DVMAS) developed by Peters (2003, 2008) with a sample of 290 university students, faculty and staff. The 18-item DVMAS seeks to measure the extent to which individuals accept certain myths about domestic violence on 6-point scale (1 = strongly disagree, 6 = strongly agree). Example items include ‘Domestic violence does not affect many people’ and ‘Women instigate most family violence’. The scale consists of four subscales: two subscales that ‘hold the victim responsible for the abuse’ (p. 3) by ascribing blame either to the woman’s character or her behavior, one subscale that ‘exonerates the perpetrator’ (p. 3), and one subscale that ‘minimizes the seriousness and scope of the problem’ (p. 3). The DVMAS has demonstrated good internal consistency with a reported Cronbach’s alpha of .88 (Peters, 2008). Subsequent studies have likewise reported good internal consistency, including Jankowsky and colleagues (2011) (Cronbach’s alpha of .88), with a college student sample in the United States; and Ferrer-Perez and Bosch-Fiol (2014) with a Spanish general community sample.
Finally, Yamawaki, Ochoa-Shipp, Pulsipher, Harlos, and Swindler (2012) developed a Domestic Violence Myth Scale. The measure comprises five items: ‘Domestic violence is easily resolved when the victim leaves the situation’, ‘Victims of domestic violence can leave the situation whenever they want to’, ‘If a domestic violence victim has financial resources, she can leave’, ‘If a woman doesn’t like it, she can leave’, and ‘Any healthy woman can successfully leave her abuser if she really wants to’. Items are rated by participants on a 7-point likert scale (1 = strongly agree, 7 = strongly disagree), and the authors report good internal consistency (Cronbach’s alpha, .84). Yamawaki et al. (2012) and Yamawaki, Ostenson and Brown (2009) also have developed a set of related scales comprising a perceived seriousness of domestic violence measure (five items), victim-blame attribution measure (five items), and an excuse-perpetrator measure (four items). Each scale is reported on a 7-point Likert ranging from 1 (strongly agree) to 7 (strongly disagree) and Yamawaki et al. (2009) report moderate to good internal consistency with US and Japanese university student samples (with Cronbach’s alphas ranging from .58 to .84). Example items, which follow the presentation of a scenarios about ‘Steve’ and ‘Marci’, include: ‘This incident should be considered domestic violence’, ‘Marci has some responsibility for creating this situation’, and ‘Hitting his wife is wrong no matter what’ (Yamawaki et al., 2009). The measures thus specifically sought to explore three related dimensions of domestic violence attitudes, namely, minimising the violence, blaming the victim, and excusing the perpetrator (discussed further below).
Attitudinal support for stalking
Empirical studies regarding attitudinal support for stalking are perhaps less developed than for either sexual or partner violence more broadly. Rather than scale development, community surveys more commonly employ short vignettes and seek to establish participants perceptions of stalking with respect to key areas of interest (seriousness and gender-based differences for example). Within this well-established sub-field, several studies emerge as highly influential, such as by Scott and colleagues (Scott & Sheridan, 2011; Scott, et al., 2010, 2014), Sheridan (Sheridan et al., 2001, 2002), Dennison (Dennison, 2007; Dennison & Thomson, 2000, 2002), and Dunlap and colleagues (Dunlap et al., 2012).
With respect to attitudinal scale development however, two such measures have been established and validated in the international literature: the Stalking Myths Scale (SMS, Sinclair, 2006, 2012), and the Stalking-Related Attitudes Questionnaire (SRAQ, McKeon, 2010; McKeon et al., 2015). Sinclair’s (2006) SMS comprises 21 items and is intended as a gender neutral measure of attitudes towards the impacts of stalking as well as the characteristics of ‘typical’ victims and perpetrators. Example items include: ‘Many instances of stalking by would-be-lovers could be avoided if the alleged victim would have just told his/her stalker clearly that s/he was definitely not interested in a romantic relationship’, and ‘Being in love in not justification for stalking someone (reverse-scored)’. Responses are rated on a 6-point Likert scale ranging from 1 (strongly disagree) to 6 (strongly agree) with higher scores indicating a greater endorsement of stalking myths. Sinclair (2012) reports moderate to good internal consistency with a Cronbach’s alpha of .76 (.80–.82 in a previous study, Sinclair, 2006).
Meanwhile McKeon’s SRAQ (McKeon, 2010; McKeon et al., 2015) encompasses 34 items, of which 19 are written to be gender neutral, with 15 referring to a male perpetrator and female victim. Responses are rated on a 7-point Likert scale from 1 (absolutely untrue) to 7 (absolutely true), with higher scores indicating greater agreement with stalking myths (four items are reverse scored). The SRAQ demonstrates high internal consistency (Cronbach’s alpha = .91). McKeon et al. (2015) surveyed a general community and a police sample, reporting that police generally endorsed stalking as ‘more serious’ than general community members and that overall men endorsed stalking myths to a greater extent than women. The study also found that those participants who obtained higher stalking myth endorsement scores were significantly more likely to make ‘not guilty’ judgements in response to a fictional stalking vignette.
Studies employing these instruments have also variously examined whether stalking myths are best understood as a unidimensional or multidimensional construct. For example, McKeon et al. (2015) report a three-factor structure: ‘stalking isn’t serious’, ‘stalking is romantic’, and ‘victims are to blame’. Dunlap et al. (2015) report a similar three-factor structure, emerging from their sample of US university students, which they describe as ‘minimising’, ‘flattery’, and ‘victim blaming’. Meanwhile, in their Italian university student sample, De Fazio and colleagues (2015) report two-factors emerging; one that taps into the notion that stalking is a nuisance or even flattering behaviour rather than a criminal act. The second factor they report as encapsulating attitudes that normalise stalking behaviours as normal aspects of courtship pursuit (De Fazio et al., 2015). When comparing the descriptions and item-wording of these factors with the previously described studies examining rape and sexual harassment myths, it is arguable that they each share commonalities of minimising violence against women, blaming victims, and excusing or justifying the actions of perpetrators.
Dimensions of violence supportive attitudes: Minimise, blame and excuse
Though there is not universal agreement, the research instruments reviewed here lend support to the multidimensional nature of a violence supportive attitudes construct. For example, several of the measures reviewed here have been subject to factor and/or principal components analyses revealing sub-domains including: denying or minimising the seriousness of violence against women, blaming the victims, excusing or otherwise justifying or minimising the actions of perpetrators, as well as sub-domains related to gender or gender-roles (e.g. Bohner et al., 1998; Briere, Malamuth, & Check, 1985; Burgess, 2007; Burt, 1980; Chapleau et al., 2008; Gerger et al., 2007; Lonsway & Fitzgerald, 1994, 1995). Furthermore, in their review, Grubb and Turner (2012) suggest that while rape myths vary among societies, they tend to follow consistent themes ‘whereby, they blame the victim for their rape, express a disbelief in claims of rape, exonerate the perpetrator and allude that only certain types of women are raped’ (Grubb & Turner, 2012, p. 445 emphasis in original).
The importance and relevance of understanding whether violence supportive attitudes are a single or multidimensional construct is not however merely an academic issue or question of statistical relevance. As others have suggested, there may be important implications of these various sub-domains as correlates both of violence proclivity itself, as well as the responses of others in the community towards both victims and perpetrators. For example, Scott and Straus (2007) identify three sub-domains of ‘minimise’, ‘blame’, and ‘deny’, which they further describe as underlying cultural norms that support men’s use of violence against women. Scott and Strauss (2007) likewise highlight that high levels of victim-blaming attitudes are associated with an individual’s increased risk of partner violence perpetration (see also Dutton & Starzomski, 1997), while minimization of the extent and impacts of abusive behaviours is more common among those with prior perpetration histories (see also Ehrensaft & Vivian, 1999). Victim-blaming and minimisation attitudes are also frequently reported as an important aspect of professional responses to victims and perpetrators disclosing violence including among legal, medical, and health professionals (see e.g. Jackson, Witte, & Petretic-Jackson, 2001), making the understanding of these dimensions particularly important for directing policy and programme work.
Conclusion and future research directions
Attitudes supportive of violence against women are a key correlate of violence perpetration and victimisation itself, as well as responses towards victims and perpetrators who disclose violence. Collectively, held attitudes reflect and reinforce social norms, structures and practices implicated in violence against women. Such attitudes thus represent an important focus for prevention education and for monitoring cultural change in Australian attitudes towards violence over time. Compared with the United States and Europe, Australian research into community attitudes towards violence against women is lacking, both with regard to empirical studies and scale development. Recognising that the NCAS has been developed in the absence of existing Australian validated measures, it is timely to review the instruments’ coverage and construct development, as well as the extent to which the NCAS represents consistency and/or divergence with validated measures emerging in the international research.
What emerges from this integrative review of measures of attitudinal support for violence against women, is that measures of ‘myths’, ‘attitudes’, and ‘beliefs’ regarding some forms of violence are far more developed than for others. In particular, there is an extensive literature regarding the development and validation of ‘rape myths’ and associated measures, as well as updated instruments that seek to modify existing scales for use in contemporary western societies. By comparison, there are fewer scales addressing sexual harassment and stalking myths, although the available measures draw heavily from the robust literature on attitudes towards rape. ‘Domestic violence myths’ meanwhile are an emerging area for scale development, and just two measures reviewed appear relevant to adaptation to the Australian context. To date, none of these measures have been tested and validated with an Australian general community sample, representing an enormous gap in existing research. This further raises the need to consider whether the language, terms, and constructs used in these measures are readily transferable to the Australian general community. Though some evidence for the applicability of these various measures can perhaps be gleaned from their relative consistency with items included in the NCAS in its various versions (e.g. in 2006, 2009, 2012).
Overall, further research within the Australian community may be required to test and refine a suitable composite measure of violence supportive attitudes that: addresses each of sexual violence, sexual harassment, stalking and domestic violence; considers adaptation from existing items and scales wherever possible, rather than generating new items in isolation, to enable cross-sample comparability; undertakes further statistical exploration of the unidimensional or multidimensional nature of a violence supportive attitudes construct; and undertakes further and more extensive analyses of the underlying and related constructs of endorsement of gender inequality, rigid or stereotypical gender roles, as well as hostility and sexism towards women. What the research reviewed here does ultimately indicate is that to understand and measure community attitudes supportive of violence against women (or ‘violence supportive attitudes’), it is also necessary to understand and measure community attitudes regarding hostility towards women, traditional gender roles, and/or endorsement of gender equality more broadly.
Footnotes
Declaration of conflicting interests
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: The authors are currently working with Australia’s National Research Office for Women’s Safety (ANROWS) as part of a team of researchers and project staff delivering the 2017 National Community Attitudes Survey, which is due to report in 2018. The views expressed here are those of the authors and do not represent the views of either ANROWS or the Australian Government which funds the NCAS.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The background research to this publication was undertaken with funding and project support from the Victorian Health Promotion Foundation (VicHealth), Melbourne. The views expressed here are those of the authors and do not represent the views of VicHealth.
