The Effects of Portrayals of IPV in Popular Written Fiction

(1)

THE EFFECTS OF PORTRAYALS OF INTIMATE PARTNER

VIOLENCE (IPV) IN POPULAR WRITTEN FICTION ON

PERCEPTIONS OF IPV WARNING SIGNS AND BELIEFS

SURROUNDING SEXUAL CONSENT

Aleksandra Szczerba

11844590

Master’s Thesis submitted to the University of Amsterdam in partial fulfilment of the requirements for the completion of the Master’s Programme in Communication Science:

Entertainment Communication Supervised by Susanne Baumgartner

June 2018 Word count: 7867

(2)

Popular erotic romance literature, such as “Fifty Shades of Grey” has been heavily criticized over the years for its unhealthy portrayals of romantic relationships, including the possible normalisation of intimate partner violence (IPV). However effects of these texts on attitudes regarding gender, violence, and relationships have rarely been studied. This study aimed to investigate the effects of content from “Fifty Shades of Grey” on people’s perceptions of IPV and their gender-stereotypic sexual beliefs, as well as the effects of fictionality of the text and the role of transportation in eliciting attitude change. A total of 168 participants read a short excerpt from “Fifty Shades of Grey” labelled as fiction or non-fiction, with some reading an unrelated control text. The results showed that, contrary to past findings, exposure to negative media content did not have a negative effect on people’s attitudes. In addition, the perceived fictionality of the text did not affect the outcomes; it did however have an effect on the experience of transportation, with the non-fiction text being the most transporting. Although the effects of text condition were non-significant, the mediation analyses did show that transportation was predictive of enjoyment of the text, and that in some cases transportation and enjoyment were predictive of the measured outcomes.

(4)

2

Introduction

“Fifty Shades of Grey” is one of the best-selling book series of the past few years, spawning a trend in written fiction in the form of other BDSM-themed romantic novels, as well as its own movie trilogy. ‘BDSM’ refers to a set of practices including bondage, discipline, dominance and submission, and sadomasochism, which people engage in within sexual contexts. The “Fifty Shades” franchise has been targeted mainly at women and has enjoyed success across all age groups, nonetheless men may constitute as much as 20-30 percent of the readership (Bowker, 2012). Despite its popularity, the series has been heavily criticized for its portrayal of the central relationship between the protagonist Anastasia Steele and her love interest Christian Grey, which is often described as not only an inaccurate portrayal of BDSM, but also simply as an abusive relationship, with criticism coming not only from readers and activists (e.g. Roper, 2017), but also from academics (Bonomi, Altenburger, & Walton, 2013), as many believe that the books may contribute to a normalisation of abusive behaviours and unsafe sexual practices among its readers.

The ongoing criticism of the “Fifty Shades” franchise coincides with an increase in public discourse surrounding sexual abuse and assault. Recent months have seen high engagement in movements such as “#MeToo”, which has allowed women to shed light on their experiences with harassment and assault, while also bringing attention to issues surrounding sexual consent, such as coercion and intoxication. When we consider these recurring Hollywood abuse scandals, debates regarding consent, and researchers expressing concern over the overwhelming rates at which women suffer from sexual and intimate partner violence (e.g. according to a 2011 US-based survey, over 40 percent of women have been victims of sexual violence in their lifetimes; Breiding, 2014), it comes as no surprise that media representations of unhealthy or abusive romantic relationships are being scrutinised by audiences and academics alike.

(5)

3 We learn a lot about the world around us from the media we consume, although there is some dispute over how influential media are when it comes to our attitudes and behaviour, and what the exact process is (for a comprehensive review of media effects theories, see Valkenburg, Peter, & Walther, 2016). Pop culture is often accused of normalising violence, in particular violence against women – but these studies are largely limited to audiovisual media formats such as video games (Beck, Boys, Rose, & Beck, 2012), television (Kahlor & Eastin, 2011) or music videos (Rhodes, Potocki, & Masterson, 2018). Research on literature is primarily limited to critical textual analyses (Bonomi et al., 2013; Borgia, 2014) and some correlational studies (Altenburger, Corotta, Bonomi, & Snyder, 2017). There is relatively little research on the effects of written fiction, and there is no known research specifically on the causal effects of literary portrayals of abusive romantic relationships.

As “Fifty Shades of Grey” is a hugely popular series, the central romantic

relationship of which is characterised by IPV (Bonomi et al., 2013), it warrants academic interest. Considering the lack of research on the effects of portrayals of abuse in written fiction, as well as the aforementioned increased public debate about sexual consent and assault, this study will not only begin to fill a gap in the literature on media effects, but is also relevant to society today. The first aim of this study is therefore to examine the effects of exposure to excerpts from “Fifty Shades of Grey” on people’s attitudes towards IPV, as well as their beliefs surrounding sexual consent. In addition, it has been suggested in literature that fiction narratives may be more powerful in eliciting attitude change than non-fiction (e.g. Fong, Mullin, & Mar, 2015), due to the way we become transported into the story-based world we construct (Busselle & Bilandzic, 2008). Thus the second aim of this study is to examine if exposure to a fiction text produces different outcomes than exposure to a non-fiction text.

(6)

4

Theoretical Background

A number of academic articles focusing on “Fifty Shades of Grey” have been published in recent years. Bonomi et al. (2013) carried out a content analysis of the first “Fifty Shades” book, identifying various themes of abuse in the texts, including sexual violence (coercion, intimidation) and emotional abuse (stalking, isolation, intimidation). In a different study, women who had seen the first “Fifty Shades of Grey” movie took part in focus groups in which they expressed finding the central relationship romantic and exciting, but also concerning and unhealthy (Bonomi, Nichols, Carotta, Kiuchi, & Perry, 2016). A correlational survey study which investigated the possible negative health outcomes (e.g. binge drinking, diet pill use) related to “Fifty Shades” readership, identified significant relationships between readership and history of having experienced IPV, as well as unhealthy eating practices (Bonomi et al., 2014). Another survey-based study associated women’s readership of the books with higher levels of reported sexism, with the effects being even stronger when the women found the books romantic (Altenburger et al., 2017). Even though these studies point to the books being potentially problematic, there are no studies from which we can learn of any casual relationships between readership and outcomes.

There is academic consensus that fiction narratives can influence people’s real life attitudes and beliefs (Braddock & Dillard, 2016; Green & Brock, 2000; Hoeken & Hustinx, 2007). Unfortunately, there generally is limited literature on the effects of written romance fiction, particularly fiction containing sexual content. This may be due to the fact that more attention is paid to the effects of audiovisual media as opposed to plain text, since imitation of presented behaviours has often been at the core of media effects studies, for example based on Bandura’s social cognitive learning theory (1994). Romance novels are also rarely

criticised for problematic portrayals of relationships or violence compared to other media – in fact, according to one study romance fiction has largely remained consistent in its portrayals

(7)

5 of ‘western sexual scripts’, which do not include violent or deviant sexual scenarios (Ménard & Cabrera, 2011). It is really in this decade that academics have raised concerns about potential abusive themes within literature, for example the “Twilight” saga (Borgia, 2014; Collins & Carmody. 2011), and more recently “Fifty Shades of Grey” (Bonomi et al., 2013).

This relative lack of interest in written media is surprising – we spend longer engaged with a book than with a music video or a movie, and reading requires active construction of a story-based world in one’s mind based on the narrative (Busselle & Bilandzic, 2008). These features of reading suggest that it may indeed be influential, and there is meta-analytic

evidence to support this (Braddock & Dillard, 2016). In terms of long-term fiction readership, both positive and negative outcomes have been identified. Fiction readership has been related to reduced gender stereotyping and increased gender egalitarianism, regardless of genre (Fong et al., 2015). On the other hand, negative attitudes towards condom use have been associated with romance novel readership (Diekman, McDonald, & Gardner, 2000). However this evidence is correlational, and even though genres were considered – neither study

considered sexually explicit texts. Instead, sexually explicit materials have been incorporated into experimental studies. A study by Van Oosten, Peter, and Boot (2015) showed that women’s gender orientation was predictive of their responses to sexually explicit texts, with hyperfemininity being predictive of less critical responses. Another study found that reading an erotic story featuring a dominant man and a submissive woman increased women’s benevolent sexism and men’s rape myth acceptance (Harris, Thai, & Barlow, 2017).

Nevertheless, there is no notable research on the effects of sexually explicit texts on readers’ attitudes beyond this, and more importantly there is no research on the effects of texts containing descriptions of IPV.

Influences on people’s perceptions of sex, consent, etc. have been studied extensively based on other media formats. Pornography has been a focal point, due to its explicit

(8)

6 representation of sex and due to adolescents’ tendency to seek it out (e.g. Peter &

Valkenburg, 2016). Some academics have expressed concern over pornography becoming more violent towards women, and thus this violence becoming normalised (Bonomi et al., 2014). Much research has also paid attention to mainstream media, including music videos (Rhodes et al., 2018; Van Oosten, Peter, & Valkenburg, 2015), television (Kahlor & Eastin, 2011; Lee, Hust, Zhang, & Zhang, 2010), video games (Beck et al., 2012), women’s

magazines (Kim & Ward, 2004), and men’s magazines (Romero-Sanches, Toro-Garcia, Horvath, & Megias, 2017; Ward, Vandenbosch, & Eggermont, 2015). Although studies have focused on somewhat different outcomes, findings have consistently pointed to media, in particular sexualised and objectifying media content, reinforcing gender-based stereotypical beliefs about sexuality, and increasing rape myth acceptance, sexually coercive behaviour, and minimization of perceptions of IPV (for a review, see Ward, 2016).

As the “Fifty Shades of Grey” series depicts a relationship characterised by IPV (Bonomi et al., 2013), the increased minimization of perceptions of IPV (as previously studied in the context of music videos by Rhodes et al., 2018) as a result of exposure to content from the book is of interest in this study. In addition, the books contain descriptions of sexual situations where consent is ambiguous or coercion is implied; this may affect attitudes towards what is known as female token refusal (sometimes referred to as token resistance), i.e. the stereotype that women say no when they mean yes, which have also been found to be enforced by exposure to music videos (Van Oosten, Peter, & Valkenburg, 2015). Finally, the relationship in “Fifty Shades” is based on gender-based sexual norms, wherein the man is dominant and the woman is submissive, which is further emphasised by the BDSM aspects of the book. Based on these aspects of the narrative, the leading research question was as follows:

(9)

7

RQ1: What effect does exposure to content from the book “Fifty Shades of Grey” have on people’s perceptions of real life IPV, their acceptance of female token refusal, and support for gender-based sexual stereotypes?

There is no academic consensus on the exact mechanisms of media effects. Theories have evolved over time, as models have incorporated more elements. One of the main current models was developed in 2013 by Valkenburg and Peter: the Differential Susceptibility to Media Effects Model (DSMM). The model proposes a sequence of influences starting at dispositional moderating personal factors, through media selection/use, a variety of response states, and finally resulting media effects. Compared to previous models, the DSMM captures a higher number of factors and processes, thus capturing the dynamic, complex nature of media influence. In a number of studies indirect relationships reminiscent of those proposed in the DSMM have been identified. In the Rhodes et al. (2018) study, the authors concluded that exposure to violent music videos resulted in minimization of perceptions of IPV, which occurred via enjoyment: violent videos were rated as more enjoyable (a mediating affective response as per the DSMM), and this predicted IPV minimization. Similarly, another study which utilised music videos (Van Oosten, Peter, & Valkenburg, 2015), found that exposure to sexual music videos influenced acceptance of female token resistance via affective

enjoyment. These findings suggest that enjoyment of media, including written media which is the focus of the current study, should increase their influence on attitudes and beliefs.

The DSMM is not the only model that can be applied to media effects. Another theory to consider is Green and Brock’s (2002) Transportation Narrative Theory, based on which as one feels “transported” into a narrative, they become fully focused on its events and lose contact with the real world around them. This “turns off” one’s critical thinking, and as they become unable to construct counterarguments, belief change can occur. This theory has been applied to studies of audiovisual media, such as music videos, where transportation was

(10)

8 identified to be predictive of both enjoyment of the content and of its effects on minimization of IPV warning signs (Rhodes et al., 2018). More importantly, transportation is often

identified as vital part of the persuasive effects of written narratives (e.g. Bal & Veltkamp, 2013), and it is usually linked to enjoyment (Green, 2004; Nabi & Krcmar, 2004), making it even more effective. In the current study, participants were exposed to excerpts from “Fifty Shades of Grey” labelled as fiction or non-fiction. As fiction is thought to be more

transporting, and transportation is associated with enjoyment, the first two hypotheses were as follows:

H1: Participants exposed to content labelled as "fiction" will experience greater transportation into the narrative than participants exposed to content labelled as “non-fiction”.

H2: Participants exposed to content labelled as “fiction” will experience greater enjoyment of the text than participants exposed to content labelled as “non-fiction”.

There is neuroscientific evidence that fiction and non-fiction texts are processed differently as we read them (Altmann, Bohrn, Lubrich, Menninghaus, & Jacobs, 2012). Previously it has been largely believed that non-fiction narratives are superior to fiction in terms of persuasiveness, however this has changed over time (Green, Garst, & Brock, 2004). When reading fiction, we come to accept the narratives within the story-based world we construct (Busselle & Bilandzic, 2008). Fiction is associated with higher transportation and less critical processing of the information presented (Green et al., 2004), which increases the effectiveness of fiction narratives. Long-term readership of fiction has been associated with positive outcomes whereas non-fiction has not (Fong et al., 2015), and fiction has been associated with increases in empathy (Bal & Weltkamp, 2013). Beyond this, meta-analytic evidence for the moderating role of fictionality on the effects of narratives on beliefs and

(11)

9 attitudes is mixed (Braddock & Dillard, 2016), however considering the increased

transportation and decreased critical processing associated with fiction, a narrative labelled as fiction should be more persuasive than a non-fiction one.

Further to this, studies on print magazines have shown that perceived source plays a role in how content is processed as well. In a number of studies, quotations about women from British “lads’ magazines” were placed alongside quotes from convicted rapists.

Horvath, Hegarty, Tyler, and Mansfield (2012) found that people could not tell the difference between magazine quotes and rapists’ quotes; when quote labels were manipulated and men were asked how much they identified with the quotes, they identified more with quotes labelled as magazine quotes, regardless of content. Hegarty, Stewart, Blockmans, and

Horvath (2018) similarly found that jokes which were presented in the context of a “lad mag” were rated as less hostile. These studies suggest that when a message is presented in the context of an entertaining media product, the message may be seen as more acceptable and harmless, which could lead to higher acceptance of the message. Taking this into

consideration, as well as the suggested differences between fiction and non-fiction effects, a research question and hypotheses regarding fictionality were proposed:

RQ2: Do the effects of exposure to ‘Fifty Shades of Grey’ on measured attitudes differ when the content is perceived as either fictional or non-fictional?

H3a: Participants exposed to content from “Fifty Shades” labelled as fiction will exhibit higher minimization of IPV as well as higher acceptance of female token refusal and gender-based sexual beliefs, as compared to the non-fiction and control conditions

H3b: Participants exposed to content from “Fifty Shades” labelled as non-fiction will exhibit higher minimization of IPV as well as higher acceptance of female token

(12)

10

refusal and gender-based sexual beliefs, as compared to the control condition, but lower in comparison to the fiction condition.

In addition, based on the previously identified mediating roles of transportation and enjoyment in attitude change (e.g. Rhodes et al., 2018), fourth and fifth hypotheses were proposed:

H4: Experienced transportation will be predictive of enjoyment of the text.

H5: The effect of the fictional text on the measured outcomes will be mediated by both transportation and enjoyment of the texts.

Research into media effects has also identified gender as an important factor. In romantic and sexual media contexts, the genders are often assigned stereotyped roles of the dominant man and the submissive woman (e.g. Farvid & Braun, 2014), which is also the case with “Fifty Shades”. Some media affect males and females differently – for example in the 2012 study by Burgess and Burpo on the effects of sexualised music videos on perceptions of rape, sexualized videos had a stronger effect on men in lowering empathy for the victim and on strengthening perceptions of perpetrator’s guilt. In a recent study in which participants read short erotica (Harris et al., 2017), reading a story with a dominant man and submissive woman increased men’s rape myth acceptance, but increased women’s benevolent sexism. In addition, after reading a story in which the woman was the dominant party, women showed a similar preference for partner dominance to men – in other conditions women favoured more dominant partners than men did (Harris et al., 2017). Further to this, although “Fifty Shades” is a franchise aimed at women, male readership has been estimated at around 20 percent (Bowker, 2012) – but literature tells us little about how women and men interact with media targeted at the opposite gender, including experienced transportation or their enjoyment of the media. Women have been found to be more critical of sexually explicit texts aimed at

(13)

11 men (Van Oosten, Peter, & Boot, 2015), and there is evidence to suggest both women and men prefer texts with male protagonists (Bortolussi, Dixon, & Sopčák, 2010), but we know little beyond this. Considering the lack of conclusive findings regarding gender, the following research question was also proposed:

RQ3: Does gender have a moderating effect on how exposure to content form "Fifty Shades" affects participants' attitudes, as well as their transportation into the narrative and enjoyment of the text?

Method Participants

Recruitment of participants took place directly via social media networks Facebook, Twitter, Instagram, and Reddit and via snowball sampling through personal networks. There was no remuneration for participation. Young adults aged 18-35 were recruited. A total of 172 individuals initially participated. However, one person did not disclose their gender and three identified as non-binary, and were excluded from the sample as the study sought to

investigate binary gender differences. The resulting sample was 168. The age range was 18-34 (MAge = 24.31, SDAge = 3.56). One hundred and twelve of the participants were female, 56 male. The majority of participants were heterosexual (n = 121), 16 were gay or lesbian, and 22 were bisexual. Three participants did not disclose their sexual orientation, and six described themselves as “other”. Participants were asked to describe their ethnicity via text entry. Four participants did not to disclose their ethnicity. The majority of the participants in the sample described themselves as some variant of “white”, “Caucasian” or “European” – 128 in total. Seventeen participants described themselves as “Latin” or “Hispanic”, eight described their ethnicity as “mixed”, and 11 were of other non-white ethnicities. The majority of the sample was therefore made up of participants with a western or European background, which suggested similar access to the type of media being investigated in this study.

(14)

12 When asked about past experience with intimate partner violence (i.e. if they or

anyone close to them had ever been a victim), participants were given the option to answer “yes” or “no” or to leave the question blank if they did not feel comfortable disclosing this information. Eight participants left the question blank, 67 declared they had no personal experience with IPV, whereas 93 disclosed they or someone close to them had experienced IPV. Of those who had direct experience with IPV, 62 were female and 31 were male. In terms of familiarity with the “Fifty Shades” franchise, 48 participants had read at least one book, and 73 had seen at least one film. Of those who had read at least one book, only 22 participants recognised that the stimulus text was from the book – 12 participants in the fiction condition, and 16 in the non-fiction condition.

Procedure

The study had a 2 (participant gender: female, male) x 3 (text condition: fiction, non-fiction, control) between-subjects experimental design. The study was set up using Qualtrics.

Participants completed the survey using desktop or mobile devices. Qualtrics randomly assigned participants to one of the three conditions.

Participants were presented with an information sheet which included a vague

description of the study aim; they were informed that the study investigated how young adults respond to literary texts; disclosing where the source of the texts would have primed the participants. They were informed that the texts may contain sexual content and descriptions of abuse and that they could quit the study at any point if they wished. Once they confirmed their consent, participants were shown one of the three texts (Appendix A). Each stimulus page included a short description explaining what the text was about and what literary genre it was from. The fictionality of the stimuli was manipulated. The story content of the fiction and non-fiction texts was the same: the fiction texts were taken directly from “Fifty Shades of Grey”, and then edited for the non-fiction condition to sound like a personal memoir. The

(15)

13 characters’ names were changed to avoid bias. Two excerpts from “Fifty Shades” were

chosen to show a range of IPV behaviours: a longer text including an argument and the build up to a sexual encounter, and a shorter text comprising a post-coital exchange in which the intercourse is discussed. The passages were chosen based on the content analysis carried out by Bonomi et al. (2013), in which both were identified as examples of abusive behaviours.

After reading the stimulus, participants were presented with an online survey. At the end participants completed demographics as well as questions regarding their familiarity with the books and movies in the “Fifty Shades” franchise. A final question asking about

participants’ familiarity with IPV in their lives was included: a yes or no question asking if they, or anyone close to them, had been a victim. Participants could leave this blank if they felt uncomfortable answering. Following this participants were shown a short debrief page explaining the aim of the study and the experimental manipulation. Completion of the study took on average 10 minutes.

Measures

Enjoyment. The participants expressed their enjoyment of the text by rating their

agreement on a scale of 1 (“strongly disagree”) to 7 (“strongly agree”) with three statements: “I enjoyed reading the text.”, “I found the text interesting.” and “I would want to keep on reading the text.” The scale was reliable (α = .90; M = 4.29, SD = 1.60); the responses were averaged to provide a mean score, with a higher value representing higher enjoyment.

Transportation. The transportation measure consisted of nine items (Appendix B).

The measure was adapted from Green and Brock (2000); out of the original 11 items, four were removed as they were more fitting for a longer text. Two new items were added to the scale, asking specifically about the participant’s empathy and sympathy for the protagonist of the text. These items were added as the stimulus texts were first-person narratives with a lot

(16)

14 of dialogue; it was therefore important that the transportation items did not just ask about engagement with the events but also with the protagonist. Participants rated their agreement with each item on a scale of 1 to 7, with 7 indicating higher agreement; two items were reverse coded. An exploratory factor analysis was carried out on the items, which loaded onto two factors with eigenvalues greater than one, explaining 55.6% of variance; the measure was kept as 11 items, as the measure has consistently been used as such across research, and the newly added items loaded highly onto the main first factor. The scale was reliable (α = .76; M = 4.42, SD = 0.95), and the items were averaged to produce a mean transportation score.

Intimate Partner Violence (IPV) Minimization. The measure of IPV minimization

was taken from the study by Rhodes et al. (2018). Items from the original measure were edited to be uniform in terms of syntax and language. As the original list of items did not represent a wide enough range of IPV warning signs, five extra items were added to the original eight, creating a 13 item scale (Appendix C). The added items were based on the abuse warning signs described on the website of the Center for Relational Abuse Awareness (http://stoprelationshipabuse.org/educated/warning-signs-of-abuse/) and the website for Break The Cycle, an organisation which promotes healthy relationships

(https://www.breakthecycle.org/warning-signs). An exploratory factor analysis was carried out; two factors of an eigenvalue above one were extracted (8.18 and 1.32 respectively), explaining 73.0% of variance. However the second factor accounted for only 10.14% of variance, and all items loaded onto the first factor with loadings above 0.7, hence the measure was kept as a complete 13 item measure. It was highly reliable (α = .95; M = 5.85, SD = 1.12). Participants rated their assessment of each of the listed IPV warning sign behaviours on a scale of 1 (“not at all abusive”) to 7 (“extremely abusive”), and the ratings were averaged to create a mean score of IPV minimization. Upon initial analysis, it became clear the measure was non-normally distributed across genders and conditions, being extremely

(17)

15 leptokurtic and negatively skewed. The values were reflected and log transformed to

normalise the distributions. As the scores were reflected, higher scores indicated greater minimization of IPV warning signs (M = 0.29, SD = 0.18).

Sexual Beliefs Scale. To assess sexual beliefs including consent, an edited version the

Sexual Beliefs Scale (Muehlenhard & Felts, 1998) was employed. The original full-length scale consists of 40 items and five subscales; for this study, this was shortened to 15 items due to length concerns (Appendix D). The majority of the items used are part of the short form SBS; a number of other items were switched out to match the aims of the study. The items represented four subscales of the SBS: “Token Refusal” (five items; α = .80, M = 1.83, SD = 0.77), “Women Like Force” (four items; α = .88; M = 2.70, SD = 0.90), “Men Should Dominate” (three items; α = .83; M = 1.77, SD = 0.81), “No Means Stop” (three items; α = .74; M = 4.91, SD = 0.36). No items from the fifth “Leading on Justifies Force” subscale were kept, as they were repetitive and their content was more aggressive and forceful (e.g. “A man is justified in forcing a woman to have sex if she leads him on”), when the aim was for this measure to capture more subtle and implicit attitudes surrounding consent.. The chosen items were edited to ensure wording and sentence structure were constant across the scale, e.g. referring to ‘men’ and ‘women’ throughout, instead of ‘guys’, ‘girls’, and ‘boys’.

Upon initial data analysis, it was decided to exclude the “No Means Stop” subscale due to the pattern of responses to the subscale being highly inconsistent with other items, potentially due to the reversed phrasing of the three items causing confusion. An exploratory factor analysis with direct oblimin rotation was carried out on the remaining 12 items. Three factors with an eigenvalue greater than one were identified; items loaded highly onto their respective subscales. The three components explained 68.2% of the variance. Participants were asked to mark their agreement with the items on a 1-7 scale (“strongly disagree to “strongly agree”), and a mean response for each subscale was calculated.

(18)

16

Results

The descriptive statistics for female and male participants across the three text conditions are presented in Table 1. A number of variables were non-normally distributed, with skeweness and kurtosis z-scores exceeding ±1.96. Upon inspection, no suspicious responses could be identified. Polarised outcomes were to be expected due to the content of the measures, i.e. gender dynamics and IPV, due to some persons’ more stereotypic attitudes, or differing definitions of what constitutes abuse. As the ANOVA is a robust test even when assumptions of normality are violated (Blanca, Alarcón, Arnau, Bono, & Bendayan, 2017) the planned analyses went ahead.

Hypotheses 1 and 2: Transportation and Enjoyment

H1 stated that participants exposed to the text labelled as fiction would experience greater transportation than those who read the text labelled as non-fiction, whereas H2 stated that participants in the fiction text condition would experience greater enjoyment than those in the non-fiction condition. RQ3 – the effects of participants gender - was also investigated. Two individual two-way ANOVAs with transportation and enjoyment as the outcome variables, and the text condition (fiction versus non-fiction) and gender as the independent variables were carried out. The assumption of homogeneity of variances was met for all conditions analysed.

For transportation, there was no significant main effect of text condition, F(1, 107) = 3.71, p = .057, ηp2 = .03. However, it should be noted the p-value was nearing significance level, with the non-fiction text (M = 4.75, SD = 0.92) being more transporting than the fiction text (M = 4.37, SD = 0.83). There was no significant main effect of gender, F(1, 107) = 1.66,

p = 0.201, ηp2 = .02, or a significant interaction effect, F(1, 107) = 0.06, p = .804, ηp2 = .00, hence gender of participant did not affect transportation experienced when reading the text.

(19)

17 H1 was not supported, as the results indicated a different direction of results, with the text labelled as non-fiction resulting in higher transportation, although not significantly so.

For enjoyment, a main effect of text condition was found, F(1, 107) = 5.07, p = .026, ηp2 = .05, with the fiction text (M = 3.84, SD = 1.71) being significantly less enjoyable than the non-fiction text (M = 4.65, SD = 1.35). The was no effect of gender of participant, F(1, 107) = 1.27, p = .263, ηp2 = .01, nor an interaction effect, F(1, 107) = 1.09, p = .299, ηp2 = .01. H2 was therefore not supported, with the non-fiction text being significantly more enjoyable than the fiction text, the reverse of the direction predicted.

Hypothesis 3: The effects of perceived fictionality of the text

Based on H3a, it was predicted that participants exposed to the text labelled as fiction would exhibit the greatest IPV minimization and score higher on subscales of the SBS than participants in the non-fiction and control conditions; according to H3b, participants exposed to the non-fiction text were predicted to score higher on the outcome measures in comparison to those in the control condition. In addition, based on RQ3, the aim was to also investigate if participant gender would moderate the effects of the fictionality of the text. Two-way

ANOVAs were run for all four outcome variables to investigate the main and interaction effects of gender and experimental condition on the participants’ responses. The assumption of homogeneity of variances was met on the measures analysed.

There was no main effect of text condition on the measure of IPV minimization, F(2, 162) = 0.16, p = .856, ηp2 = .00, and there was no interaction effect between condition and gender F(2, 162) = 0.30, p = .740, ηp2 = .00. There was however a main effect of gender, F(1, 162) = 3.90, p = .050, ηp2 = .02, with male participants (M = 0.33, SD = 0.16) exhibiting greater minimization of IPV warning signs than female participants (M = 0.27, SD = 0.18). For the ‘Token Refusal’ subscale of the SBS, there was no main effect of text condition, F(2, 162) = 0.03, p = .974, ηp2 = .00, or an interaction effect, F(2, 162) = 1.36, p = .260, ηp2 = .02,

(20)

18 but there was a significant large effect of gender, F(1, 162) = 15.38, p < .001, ηp2 = .09, with males (M = 2.15, SD = 0.84) expressing higher agreement with the subscale items than females (M = 1.67, SD = 0.68). The same pattern of effects was found for the ‘Men Should Dominate’ subscale: no significant effect of text condition, F(2, 162) = 2.59, p = .078, ηp2 = .03, and no interaction effect, F(2, 162) = 0.80, p = .449, ηp2 = .01, but a significant main effect of gender, F(1, 162) = 12.32, p = .001, ηp2 = .07, with males (M = 2.06, SD = 0.84) again expressing higher agreement than females (M = 1.62, SD = 0.76). Finally, for the ‘Women Like Force’ subscale no significant effects were found: effects of text condition,

F(2, 162) = 0.70, p = .500, ηp2 = .01, participant gender, F(1, 162) = 12.32, p = .501, ηp2 = .00, and the interaction, F(2, 162) = 0.80, p = .859, ηp2 = .00, were all non-significant.

Neither H3a nor H3b were therefore confirmed, as the fictionality of the text did not produce any significant effects. Gender did not moderate the effects of experimental

condition, however for three of the four outcomes male participants’ responses were significantly higher than female participants’.

Hypotheses 4 and 5: Transportation and enjoyment as mediators

H4 stated that transportation would be predictive of enjoyment of the text, whereas H5 stated that the predicted effects of exposure to the texts on IPV minimization and SBS items would be mediated by experienced transportation and enjoyment. The text condition variable was recoded into a dichotomous fiction versus non-fiction variable for the purposes of testing these hypotheses (n = 111). Although previous analysis results did not indicate a significant effect of perceived text fictionality on the outcomes, serial mediation analyses were carried out using Model 6 in the PROCESS macro for SPSS (Hayes, 2013) with 1,000 bootstrap samples to estimate the bias-corrected bootstrap confidence intervals for each of the outcome variables. Four mediation models were constructed.

(21)

19 For H4, the effect of transportation on enjoyment was evaluated. In each mediation model, text condition was a significant predictor of experienced transportation (b = .38,

t(109) = 2.29, p = .024, [.05, .71]), with the non-fiction condition being more transporting.

Transportation was a significant predictor of enjoyment (b = .86, t(108) = 5.80, p < .001, [.56, 1.15]). Higher transportation predicted higher enjoyment, therefore H4 was supported.

For H5, the four serial mediation models were evaluated one by one. First the direct and indirect effects of text condition on IPV minimization were investigated; these are pictured in Figure 1. The total model was not significant (b = .01, t(109) = .36, p = .723, [-.06, .08]) . Text condition was not a significant direct predictor of IPV minimization (b = .01,

t(107) = .29, p = .770, [-.06, .08]), and the indirect effect of text condition was also

non-significant (b = .00, [-.02, .03]). Therefore even though text condition non-significantly predicted transportation, and transportation predicted text enjoyment, transportation and enjoyment did not mediate the relationship between text condition and IPV minimization.

The mediation model for Token Refusal is shown in Figure 2. The total model again was not significant (b = .08, t(109) = .58, p = .565, [-.20, 36]). Text condition was not a significant direct predictor of Token Refusal (b = .06, t(107) = .41, p = .684, [-.22, .33]), and thus the indirect effect of text condition was also non-significant (b = .02, [-.10, .15]). Interestingly, both transportation (b = -.26, t(107) = -2.95, p = .004, [-.44, -.09]) and enjoyment (b = .15, t(107) = 3.04, p = .003, [.05, .25]) were individually significant predictors of Token Refusal, even though the complete model was not significant.

Figure 3 shows the mediation model for the Men Should Dominate (MSD) subscale. Transportation was not a significant direct predictor of MSD (b = -.17, t(107) = -1.74, p = .085, [-.36, .02]), but the effect of enjoyment on MSD was significant (b = .14, t(107 = 2.55,

(22)

20 1.94, p = .055, [-.01, .59]) and the direct effect was non-significant also (b = .24, t(107) = 1.58, p = .116, [-.06, 0.55]).

Finally, the mediation model for the Women Like Force (WLF) items is shown in Figure 4. Like in the TR and MSD models, enjoyment was a significant predictor of WLF (b = .20, t(107) = 3.23, p = .002, [.08, .22]), however both the total (b = .21, t(109) = 1.22, p = .227, [-.13, .54]), and direct (b = .09, t(107) = .52, p = .606, [-.25, .43]), effects of text

condition on the WLF outcome were non-significant. H5 was therefore not supported for any of the four outcome variables.

Discussion

The current study investigated the effects of perceived fictionality of excerpts from “Fifty Shades of Grey” containing descriptions of intimate partner violence on people’s perceptions of real life IPV and their sexual beliefs, including sexual consent, as well as the mediating roles of enjoyment and transportation. This is the first study to examine the effects of “Fifty Shades” in an experimental setting, and the first to test the effects of fictionality of a narrative on outcomes regarding IPV. Contrary to the hypotheses, the perceived fictionality of the text had no significant effects on the outcomes; in fact, the attitudes of participants exposed to the content from the book did not differ significantly from those exposed to a control text. However, the results did show that male participants exhibited significantly greater minimization of IPV behaviours and exhibited significantly greater agreement with beliefs on the Token Refusal and Men Should Dominate subscales of the SBS. The results also supported previous findings that transportation is predictive of enjoyment of a media product, whilst also contradicting research which has suggested that fiction texts result in higher transportation than non-fiction texts.

(23)

21 The main prediction of this study was that exposure to content containing depictions of IPV would negatively affect people’s attitudes regarding IPV and their sexual beliefs, with content labelled as fiction eliciting stronger effects than the non-fiction fiction. No evidence to support these hypotheses was found, as the texts did not have any significant effects on the outcomes. This contradicts past findings where exposure to media content featuring IPV did affect perceptions of real-life IPV (Rhodes et al., 2018). The lack of effects may be the result of the use of two short book excerpts as stimuli. The passages were chosen in order to reflect both abusive and romantic aspects of the main relationship in “Fifty Shades of Grey”, and participants were given some context, but the passages still lacked the context of the entire narrative from the book. In comparison to book passages, a short story has a clearer start-to-finish narrative, which could explain why short stories have been effectively used in research (e.g. Harris et al., 2017). Therefore testing the effects of exposure to content taken from books in an experimental setting may not advisable. Reading an entire book requires engagement over a longer period of time, during which messages from the text can elicit attitude changes. There are numerous studies which show that reading fiction does affect beliefs and attitudes (Braddock & Dillard, 2016; Fong et al., 2015), so clearly the period of engagement needs to be considered. There is also the possibility that attitude change based on written narratives does not occur instantly, but instead requires a longer time; there is indeed evidence that time strengthens the effects of narratives on belief change, referred to as “sleeper effect” (Appel & Richter, 2007).

Contrary to our hypotheses, it was also found that the non-fiction condition resulted in greater transportation than the fiction text, although this effect wasn’t significant. Even though fiction narratives are often associated with higher transportation, the fact that the fiction text was not more transporting is partially supported by some studies which have not found differences in how transporting fiction and non-fiction texts are (e.g. Green & Brock,

(24)

22 2000). It is possible that participants reading a text described as real events were more

invested in the events and people described. The texts were also not just labelled as fiction or non-fiction, but the non-fiction text was rewritten to appear to be taken from a different literary genre; perhaps the writing style caused the higher transportation as opposed to the label. In addition, the description of the fiction text specified the genre, potentially priming participants who did not like romance novels to disengage from the text, resulting in not only lower transportation, but also preventing the predicted exposure effects from occurring. It is possible that labelling the materials in a different way may help elicit effects. Further, the enjoyment of the text differed between the two conditions, with the non-fiction text being significantly more enjoyable than the fiction text. This effect is surprising, as one would think fiction would be more enjoyable, and the excerpts were taken directly from “Fifty Shades of Grey” which is a popular book. However, the result can be explained by the non-fiction text being more transporting, as transportation is associated with enjoyment (Green, 2004).

Based on past research, the mediating role of transportation and enjoyment on the outcomes was tested; however, as no significant effects of text fictionality were found, no significant mediation relationships were identified, besides the relationship between transportation and enjoyment of the text, wherein higher transportation predicted higher enjoyment, supporting evidence for this relationship found in other studies (Rhodes et al., 2018). The results of the mediation analyses are contrary to findings of the study by Rhodes et al. (2018) who found that transportation and enjoyment of music video narratives predicted an increase in IPV minimization. There was some indication from the mediation models that enjoyment and transportation may be predictive of the sexual beliefs outcomes, however the interpretation of these is limited by the lack of effects of text fictionality. Since affective response is an important part of the DSMM (Valkenburg & Peter, 2013), and enjoyment is related to transportation which in of itself is thought to allow belief change to occur (Green &

(25)

23 Brock, 2002), it is possible that enjoyment is necessary for a media product to have an effect – and even though differences in enjoyment across conditions were identified in this study, the enjoyment experienced by participants was overall low.

The significant effects of gender on the outcomes are interesting. It is possible that men generally hold more gender stereotypic beliefs than women, although there is no

academic evidence to support this specifically. Nonetheless, women across cultures are more likely to reject hostile sexism, for example (Glick et al., 2000). Also, although transportation into the narrative, including two items regarding empathy for the protagonist, was measured and did not differ between the genders, male participants may have identified with the protagonist less, as identification and transportation are separate constructs. Identification, had it been measured, may have shown different results. Based on past research both females and males have shown a preference for male protagonists over female protagonists

(Bortolussi et al., 2010), so perhaps in general the female protagonist was not liked by participants. The texts used in the study only presented a female victim and an abusive male, which could affect identification further – 31 male participants had direct experience with IPV in their lives, and twice as many female participants. Based on other studies we also know that manipulating the gender of the characters in a narrative can alter the effects it has (Harris et al., 2017). Based on all this, similar research where the gender of the protagonist who experiences IPV is manipulated would be an interesting avenue for further study.

Two main weaknesses of the present study can be identified. One is the sample, which was not only comprised of a larger proportion of female participants, but of these female participants many were recruited from feminist networks on social media. These women were therefore likely to exhibit more egalitarian views regarding gender, as well as more negative attitudes regarding IPV, and their attitudes were very likely to be resistant to change. It is possible that replicating this study with a less political sample may produce different

(26)

24 outcomes. Similarly, the majority of the sample was western or European and exhibited low agreement with the outcome measures; taking into consideration that cultural differences in sexism and gender inequality exist (Glick et al., 2000), perhaps comparative media effects research of this sort should be encouraged. Secondly, the study suffered from the lack of a manipulation check. It is unclear how many participants truly paid attention to the description of the text they were given and knew whether they were reading fiction or non-fiction. This could have influenced all of the measures, including transportation. It is also possible some participants did not engage with the text actively, quickly skimming it instead. It may have been useful to add a minimum time setting to the stimulus page, so that participants would have to spend a set minimum time reading.

This study does suggest some paths for future research. Specifically studies into long-term exposure are recommended, preferably going beyond the correlational studies already in existence. Perhaps long-term experimental studies where participants get new chapters or stories sent every day, or every few days, would be effective, allowing researchers to monitor attitude change at the beginning and end of exposure. Similar experimental formats have already been employed successfully within research examining long-term neurological effects of reading (Berrns, Blaine, Prietula, Pye, 2013). There is also literature to suggest that repeated exposure increases transportation (e.g. having read a book before seeing a movie; Green et al., 2008). In the present study we did not delve into the effects of prior exposure as not many participants had read the “Fifty Shades of Grey” books.. Perhaps looking at the effects of repeat exposure, both in terms of “Fifty Shades”, and other media, is something to be considered. Finally, although the perceived fictionality of the text did not produce

significant effects in this study, this too remains an avenue for further research. In particular since fictionality can be manipulated not just by changing the paratextual information in the form of a “fiction” or “non-fiction” label, but also by manipulating writing style, the

(27)

25 described source of the material, or the genre. Notably genre (Fong et al., 2015) and source of material (Hegarty et al., 2018) do influence how participants respond to the content they are presented with, and should be considered in the future.

Conclusion

The present study failed to identify negative effects of exposure to content from “Fifty Shades of Grey”, and it did not find the perceived fictionality of the content to have an effect on attitude change. Analyses of transportation and enjoyment showed that the text labelled as non-fiction was significantly more enjoyable, and more transporting, although not

significantly so. Also, transportation was predictive of enjoyment of the text. The lack of effects of text manipulation prevented any conclusions about the mediating roles of

transportation and enjoyment from being drawn. The main reason for the lack of effects of exposure may stem from the fact that content from the book was presented to participants without much context, in a short-term experimental setting, which is not representative of how books are read and are likely to shape attitudes and beliefs. As other media effects studies have often identified violent and sexual media to produce negative outcomes, and there is correlational support for various effects of long-term engagement with fiction

literature, this study should not be considered to be proof that written media have no negative effect, but as a valuable contribution to the growing field of media effects research.

(28)

26

References

Altenburger, L. E., Carotta, C. L., Bonomi, A. E., & Snyder, A. (2017). Sexist attitudes among emerging adult women readers of fifty shades fiction. Archives of Sexual

Behavior, 46(2), 455-464.

Altmann, U., Bohrn, I. C., Lubrich, O., Menninghaus, W., & Jacobs, A. M. (2012). Fact vs fiction—how paratextual information shapes our reading processes. Social

Cognitive and Affective Neuroscience, 9(1), 22-29.

Appel, M., & Richter, T. (2007). Persuasive effects of fictional narratives increase over time. Media Psychology, 10(1), 113-134.

Bal, P. M., & Veltkamp, M. (2013). How does fiction reading influence empathy? An experimental investigation on the role of emotional transportation. PLOS One,

8(1), e55341.

Bandura, A. (1994). Social cognitive theory of mass communication. In. J. Bryant & D. Zillmann (eds.), Media effects: Advances in theory and research (pp. 61–90). Beck, V. S., Boys, S., Rose, C., & Beck, E. (2012). Violence against women in video

games: A prequel or sequel to rape myth acceptance? Journal of Interpersonal

Violence, 27(15), 3016-3031.

Berns, G. S., Blaine, K., Prietula, M. J., & Pye, B. E. (2013). Short-and long-term effects of a novel on connectivity in the brain. Brain Connectivity, 3(6), 590-600.

Blanca, M. J., Alarcón, R., Arnau, J., Bono, R., & Bendayan, R. (2017). Non-normal data: Is ANOVA still a valid option? Psicothema, 29(4), 552-557.

(29)

27 Bonomi, A. E., Altenburger, L. E., & Walton, N. L. (2013). “Double crap!” Abuse and

harmed identity in Fifty Shades of Grey. Journal of Women's Health, 22(9), 733-744.

Bonomi, A. E., Nemeth, J. M., Altenburger, L. E., Anderson, M. L., Snyder, A., & Dotto, I. (2014). Fiction or not? Fifty Shades is associated with health risks in adolescent and young adult females. Journal of Women's Health, 23(9), 720-728.

Bonomi, A. E., Nichols, E. M., Carotta, C. L., Kiuchi, Y., & Perry, S. (2016). Young women's perceptions of the relationship in Fifty Shades of Grey. Journal of

Women's Health, 25(2), 139-148.

Borgia, D. N. (2014). Twilight: The glamorization of abuse, codependency, and white privilege. The Journal of Popular Culture, 47(1), 153-173.

Bortolussi, M., Dixon, P., & Sopčák, P. (2010). Gender and reading. Poetics, 38(3), 299-318.

Bowker. (2012, November 29). Who’s Really Reading 50 Shades? [Press release]. Retrieved from http://www.bowker.com/news/2012/Whos-Really-Reading-50-Shades.html

Braddock, K., & Dillard, J. P. (2016). Meta-analytic evidence for the persuasive effect of narratives on beliefs, attitudes, intentions, and behaviors. Communication

Monographs, 83(4), 446-467.

Breiding, M. J. (2015). Prevalence and characteristics of sexual violence, stalking, and intimate partner violence victimization-national intimate partner and sexual

(30)

28 violence survey, United States, 2011. American Journal of Public

Health, 105(4), E11.

Breiding, M. J., Basile, K. C., Smith, S. G., Black, M. C., & Mahendra, R. R. (2015).

Intimate Partner Violence Surveillance: Uniform Definitions and Recommended Data Elements, Version 2.0. Atlanta (GA): National Center for Injury Prevention

and Control, Centers for Disease Control and Prevention.

Burgess, M., & Burpo, S. (2012). The effect of music videos on college students' perceptions of rape. College Student Journal, 46(4), 748-763.

Busselle, R., & Bilandzic, H. (2008). Fictionality and perceived realism in experiencing stories: A model of narrative comprehension and engagement. Communication

Theory, 18(2), 255-280.

Collins, V. E., & Carmody, D. C. (2011). Deadly love: Images of dating violence in the “Twilight Saga”. Affilia: Journal of Women and Social Work, 26(4), 382-394.

Diekman, A. B., McDonald, M., & Gardner, W. L. (2000). Love means never having to be careful. Psychology of Women Quarterly, 24(2), 179-188.

Farvid, P., & Braun, V. (2014). The “sassy woman” and the “performing man”: Heterosexual casual sex advice and the (re) constitution of gendered subjectivities. Feminist Media Studies, 14(1), 118-134.

Fong, K., Mullin, J. B., & Mar, R. A. (2015). How exposure to literary genres relates to attitudes toward gender roles and sexual behavior. Psychology of Aesthetics,

(31)

29 Glick, P., Fiske, S. T., Mladinic, A., Saiz, J. L., Abrams, D., Masser, B., . . . López, W.

L. (2000). Beyond prejudice as simple antipathy: Hostile and benevolent sexism across cultures. Journal of Personality and Social Psychology, 79(5), 763-775. Green, M. C. (2004). Transportation into narrative worlds: The role of prior knowledge

and perceived realism. Discourse Processes, 38(2), 247-266.

Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79(5), 701. Green, M. C., & Brock, T. C. (2002). In the mind's eye: Transportation-imagery model

of narrative persuasion. In M. C. Green, J. J. Strange, & T. C. Brock (Eds.), Narrative impact: Social and cognitive foundations (pp. 315-341). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers.

Green, M. C., Garst, J., & Brock, T. C. (2004). The power of fiction: Determinants and boundaries. In L. J. Shrum (Ed.), The Psychology of Entertainment Media (pp. 161-176). Mahwah, NJ: Lawrence Erlbaum.

Green, M. C., Kass, S., Carrey, J., Herzig, B., Feeney, R., & Sabini, J. (2008). Transportation across media: Repeated exposure to print and film. Media

Psychology, 11(4), 512-539.

Harris, E. A., Thai, M., & Barlow, F. K. (2017). Fifty shades flipped: Effects of reading erotica depicting a sexually dominant woman compared to a sexually dominant man. The Journal of Sex Research, 54(3), 386-397.

Hayes, A. F. (2013). Introduction to Mediation, Moderation, and Conditional Process

(32)

30 Hegarty, P., Stewart, A. L., Blockmans, I. G., & Horvath, M. A. (2018). The influence

of magazines on men: Normalizing and challenging young men’s prejudice with “lads’ mags”. Psychology of Men & Masculinity, 19(1), 131.

Hoeken, H., & Hustinx, L. (2007). The impact of exemplars on responsibility

stereotypes in fund-raising letters. Communication Research, 34(6), 596-617. Horvath, M. A., Hegarty, P., Tyler, S., & Mansfield, S. (2012). “Lights on at the end of

the party”: Are lads’ mags mainstreaming dangerous sexism?. British Journal of

Psychology, 103(4), 454-471.

Kahlor, L., & Eastin, M., S. (2011). Television's role in the culture of violence toward women: a study of television viewing and the cultivation of rape myth

acceptance in the United States, Journal of Broadcasting & Electronic Media,

55(2), 215-231.

Kim, J. L., & Ward, L. M. (2004). Pleasure reading: Associations between young women's sexual attitudes and their reading of contemporary women's magazines. Psychology of Women Quarterly, 28(1), 48-58.

Lee, M. J., Hust, S., Zhang, L., & Zhang, Y. (2010). Effects of violence against women in popular crime dramas on viewers' attitudes related to sexual violence. Mass

Communication and Society, 14(1), 25-44.

Ménard, A. D., & Cabrera, C. (2011). ‘Whatever the approach, tab B still fits into slot A’: Twenty years of sex scripts in romance novels. Sexuality & Culture, 15(3), 240-255.

(33)

31 Muehlenhard, C. L., & Felts, A. S. (1998). Sexual beliefs scale. In C. M. Davis, W. L.

Yarber, R. Bauserman, G. Schreer, & S. L. Davis (Eds.), Handbook of Sexuality

Related Measures. Thousand Oaks: Sage

Nabi, R. L., & Krcmar, M. (2004). Conceptualizing media enjoyment as attitude: Implications for mass media effects research. Communication Theory, 14, 288– 310.

Peter, J., & Valkenburg, P. M. (2016). Adolescents and Pornography: A Review of 20 Years of Research. The Journal of Sex Research, 53(4-5), 509-53.1

Rhodes, N., Potocki, B., & Masterson, D. S. (2018). Portrayals of intimate partner violence in music videos: effects on perceptions of IPV warning signs. Media

Psychology, 21(1), 137-156.

Romero-Sánchez, M., Toro-García, V., Horvath, M. A., & Megías, J. L. (2017). More than a magazine: exploring the links between lads’ mags, rape myth acceptance, and rape proclivity. Journal of Interpersonal Violence, 32(4), 515-534.

Roper, C. (2017, February 10). Fifty Shades Darker Isn't Empowering, It's Abuse. Retrieved from https://www.huffingtonpost.co.uk/caitlin-roper/fifty-shades-darker-is-fi_b_14674132.html

Valkenburg, P. M., & Peter, J. (2013). The differential susceptibility to media effects model. Journal of Communication, 63(2), 221-243.

Valkenburg, P. M., Peter, J., & Walther, J. B. (2016). Media effects: Theory and research. Annual Review of Psychology, 67, 315-338.

(34)

32 Van Oosten, J. M., Peter, J., & Boot, I. (2015). Women's critical responses to sexually

explicit material: the role of hyperfemininity and processing style. The Journal

of Sex Research, 52(3), 306-316.

Van Oosten, J. M., Peter, J., & Valkenburg, P. M. (2015). The influence of sexual music videos on adolescents’ misogynistic beliefs: The role of video content, gender, and affective engagement. Communication Research, 42(7), 986-1008.

Ward, L. M. (2016). Media and sexualization: State of empirical research, 1995– 2015. The Journal of Sex Research, 53(4-5), 560-577.

Ward, L. M., & Rivadeneyra, R. (1999). Contributions of entertainment television to adolescents’ sexual attitudes and expectations: The role of viewing amount versus viewer involvement. Journal of Sex Research, 36, 237–249

Ward, L. M., Vandenbosch, L., & Eggermont, S. (2015). The impact of men's

magazines on adolescent boys' objectification and courtship beliefs. Journal of

(35)

33 Table 1

Descriptive Statistics for Female and Male Participants across the Three Text Conditions

Females Males Fiction text condition (n = 37) Non-fiction text condition (n = 39) Control text condition (n = 36) Fiction text condition (n = 20) Non-fiction text condition (n = 15) Control text condition (n = 21) M (SD) M (SD) M (SD) M (SD) M (SD) M (SD) Enjoyment 3.59 (1.67) 4.64 (1.41) 4.41 (1.68) 4.28 (1.74) 4.67 (1.20) 4.38 (1.58) Transportation 4.43 (0.94) 4.83 (1.00) 4.31 (0.88) 4.24 (0.59) 4.55 (0.67) 3.91 (1.16) IPV Minimization 5.91 (1.36) 5.80 (1.27) 6.06 (0.89) 5.61 (1.23) 5.79 (0.79) 5.77 (0.76) IPV Minimization (log

transformed values) 0.26 (0.21) 0.30 (0.19) 0.26 (0.16) 0.34 (0.17) 0.32 (0.14) 0.32 (0.16) Token Refusal 1.59 (0.51) 1.83 (0.77) 1.57 (0.71) 2.18 (0.76) 2.01 (0.91) 2.23 (0.90) Women Like Force 2.58 (0.87) 2.85 (0.99) 2.56 (1.05) 2.75 (0.88) 2.83 (0.77) 2.71 (0.59) Men Should Dominate 1.50 (0.76) 1.78 (0.84) 1.57 (0.64) 1.78 (0.77) 2.18 (0.72) 2.24 (0.95)

Note. The log transformed IPV values were reflected before being transformed, thus reversing the direction of results. For the original values,

(36)

34

Figure 1. Serial mediation of the effect of exposure to a fiction versus a non-fiction text

on minimization of IPV behaviours, as mediated through transportation and enjoyment of the text. Total model b = .01 [-.06, .08].

on the Token Refusal subscale of the SBS, as mediated through transportation and enjoyment of the text. Total model b = .08 [-.20, .36].

Transportation

Text condition Minimization of IPV

.38 [.05, .71] Enjoyment .86 [.56, 1.15] .02 [-.01, .04] .01 [-.06, .08] -.03 [-.08, .01] .48 [-.04, 1.01] Transportation

Text condition Token Refusal

.38 [.05, .71] Enjoyment .86 [.56, 1.15] .15 [.05, .25] .06 [-.22, .33] -.26 [-.44, -.09] .48 [-.04, 1.01]

(37)

35

on the Men Should Dominate subscale of the SBS, as mediated through transportation and enjoyment of the text. Total model b = .29 [-.01, .59].

on the Women Like Force subscale of the SBS, as mediated through transportation and enjoyment of the text. Total model b = .21 [-.13, .54].

Transportation

Text condition Men Should Dominate

.38 [.05, .71] Enjoyment .86 [.56, 1.15] .14 [.03, .25] .24 [-.06, .55] -.17 [-.36, .02] .48 [-.04, 1.01] Transportation

Text condition Women Like Force

.38 [.05, .71] Enjoyment .86 [.56, 1.15] .20 [.08, .32] .09 [-.25, .43] -.12 [-.32, .10] .48 [-.04, 1.01]

(38)

36 Appendix A

Stimulus Texts Fiction condition

The following excerpts are taken from a chapter of an erotic romance novel. Charlotte has just graduated from university; Nicholas is a wealthy business man in his late 20s. The characters have been dating for a just a few weeks at the time of the events below.

Excerpt 1

Parked outside is a red hatchback car, a two-door compact Audi.

“It’s for you. Happy graduation,” he murmurs, pulling me into his arms and kissing my hair.

He’s bought me a damned car, brand-new by the looks of it. Jeez … I’ve had enough trouble with the books. I stare at it blankly, trying desperately to determine how I feel about this. I am appalled on one level, grateful on another, shocked that he’s actually done it, but the overriding emotion is anger. Yes, I’m angry, especially after everything I told him about the books … but then he’d already bought this. Taking my hand, he leads me down the path toward this new acquisition.

“Charlotte, that Beetle of yours is old and frankly dangerous. I would never forgive myself if something happened to you when it’s so easy for me to make it right …” His eyes are on me, but at the moment I cannot bring myself to look at him. I stand silently staring at its awesome bright red newness.

“I mentioned it to your stepfather. He was all for it,” he murmurs. Turning, I glare at him, my mouth open in horror.

“You mentioned this to Steve? How could you?” I can barely spit the words out. How dare he? Poor Steve. I feel sick, mortified for my dad.

“It’s a gift, Charlotte. Can’t you just say thank you?” “But you know it’s too much.”

“Not to me it isn’t, not for my peace of mind.”

I frown at him, at a loss what to say. He just doesn’t get it! He’s had money all his life. Okay, not all his life—not as a small child—and my worldview shifts. The thought is very sobering, and I soften toward the car, feeling guilty about my fit of pique. His intentions are good, misguided, but not from a bad place.

“I’m happy for you to loan this to me, like the laptop.”

(39)

37 “No, not indefinitely, but for now. Thank you.”

He frowns. I reach up and kiss him on his cheek.

“Thank you for the car, sir,” I say as sweetly as I can manage.

He grabs me suddenly and yanks me up against him, one hand at my back holding me to him and the other fisting in my hair.

“You are one challenging woman, Charlotte.” He kisses me passionately, forcing my lips apart with his tongue, taking no prisoners.

My blood heats immediately, and I’m returning his kiss with my own passion. I want him badly—in spite of the car, the books, the soft limits … the caning … I want him. “It’s taking all my self-control not to fuck you on the hood of this car right now, just to show you that you are mine, and if I want to buy you a fucking car, I’ll buy you a fucking car,” he growls. “Now let’s get you inside and naked.” He plants a swift rough kiss on me.

Boy, he’s angry. He grabs my hand and leads me back into the apartment and straight into my bedroom … no passing go. He switches on the sidelight and halts, staring at me. “Please don’t be angry with me,” I whisper.

His gaze is impassive; his eyes cold shards of smoky glass.

“I’m sorry about the car and the books …” I trail off. He remains silent and brooding. “You scare me when you’re angry,” I breathe, staring at him.

He closes his eyes and shakes his head. When he opens them, his expression has softened. He takes a deep breath and swallows.

“Turn around,” he whispers. “I want to get you out of that dress.”

Another mercurial mood swing; it’s so hard to keep up. Obediently, I turn, and my heart is thumping, desire instantly replacing unease, coursing through my blood and settling dark and yearning, low, low in my belly. He scoops my hair off my back so it hangs down my right side, curling at my breast. He places his index finger at the nape of my neck and achingly slowly drags it down my spine, his fingernail grazing my skin. “I like this dress,” he murmurs. “I like to see your flawless skin.”

His finger reaches the back of my halter dress midway down my spine, and hooking his finger beneath the top, he pulls me closer so that I step back against him so that he’s flush against my body. Leaning down, he inhales my hair.

“You smell so good, Charlotte. So sweet.” His nose skims past my ear down my neck, and he trails soft, featherlight kisses along my shoulder.