Let the facts speak, not the pictures: an experimental survey on rape narratives

Abstract

Purpose

According to statistics by the US National Sexual Violence Resource Center, more than 50% of women claim they have been raped by an intimate partner or an acquaintance. In this experiment, the authors test whether exposing individuals to different types of images portraying the victim and their perpetrator influences individuals' perception of (1) the perpetrator's deserved punishment and (2) the role played by the victim in her rape.

Design/methodology/approach

In an experimental survey, the authors randomly treat groups of individuals with manipulated facial and physical expressions of the same photographs.

Findings

The authors find that news about rape are more or less effective in uncovering the seriousness of the crime independent of the pictures accompanying them, suggesting media should avoid their misusage.

Originality/value

While the role of the narrative of rape has been extensively discussed in the literature from a theoretical perspective, this research provides original evidence based on actual behavioural response.

Citation

Focacci, C.N. and Lam, P.H. (2023), "Let the facts speak, not the pictures: an experimental survey on rape narratives", International Journal of Sociology and Social Policy, Vol. 43 No. 13/14, pp. 295-306. https://doi.org/10.1108/IJSSP-08-2023-0182

1. Introduction

The recent news of Gabby Petito's disappearance reached millions everywhere in the world and in a matter of seconds. The fact that her remains were found following travels in the US with her partner Brian Laundrie did not shock readers. In the meantime, newspapers were filled with “happy” pictures portraying the couple. More recently, the body of solider Vanessa Guillén was found dismembered in Texas, after she shared with her family that she was being sexually harassed by an unnamed sergeant. The National Sexual Violence Resource Center reports that 20% of women in the US have experienced rape during their existence. Most of the complete or attempted rape cases were attributed to intimate partners or acquaintances. For more than one in three women, the latter are also responsible of murder (Baker, 2013). Accounting for psychological and physical health costs, the Centers for Disease Control and Prevention in the US claim that, on average, rape corresponds to costs for $122,461 per victim. Victims and families of the victims would probably argue it is more. Indeed, as shown by Perilloux et al. (2012), rape negatively affects survivors' self-esteem, value as a romantic partner, sexual reputation, long-term relationships, social reputation, health, family relationships, as well as social life and work life more in general. While the Me-Too movement has somewhat changed how the media portrays rape, results are only slightly significant. The analysis by journalist Thea Storøy Elnan looked at 400 voldtekt (“rape”) articles published in Norway between October 2017 and March 2018 and found the decline in rape myths was only “slight”, from 44% to 39% [1].

In South Africa, a country not at war, Moffett (2006) argues rape “has become a socially endorsed punitive project for maintaining patriarchal order”. While this might not be the case in the US and the developed rest of the world, scholars agree that the narrative of rape matters for how rape itself is perceived by the more extended society. In an assessment of narratives used by sex offenders in the US, Lea and Auburn (2001) found that there was “no clear distinction between rape and sex”. This connects to the recent experimental findings by Romero-Sánchez et al. (2017) that even exposure to sexist jokes can significantly affect rape proclivity. In parallel, studies shed light on the diverse nature of rape narratives in the news, ranging from describing the crime as “the suburb rape” or “the celebrity rape” to monstering “the lonely pervert” rapist, hiding the true patriarchal problem behind sexual assault (Nilsson, 2019). This does not plainly depart from popular romance, where rape is sometimes romanticised “as a parodic parallel to the violence of falling in love” (Toscano, 2012). Regarding the victims' viewpoints, Abbey et al. (2005) showed how the way in which questions about sexual assault victimisation and perpetration are framed affects reporting of sexual assault incidents in the US, in line with rape mythologising (Brooker and Butler, 2021). Sexual scripting theory also asserts interpersonal concerns play a role for (not) reporting assaults (Masters et al., 2006).

From an experimental point of view, a field experiment by Malamuth and Check (1981) showed how exposing male subjects to violent sexuality increased their acceptance of interpersonal violence against women. Dill et al. (2008) also used experimental methods and found that exposure to sex-stereotyped video game characters increased tolerance of real-life sexual harassment in men. An experiment more focused on media usage suggests exposure to sexually violent media positively affect violence against women (Malamuth and Briere, 1986). Another experiment by Franiuk et al. (2008b) also used experimental methods to test the effect of rape myths in print media. They found that subjects exposed to the myth-endorsing article (“she's lying”) were more likely to believe that Kobe Bryant was not guilty and the alleged victim was lying. Their second study confirmed that those exposed to myth-endorsing headlines were more likely to hold rape-supportive attitudes (Franiuk et al., 2008a). While the recent literature lacks in experimental studies on this subject, a survey by Maes et al. (2019) on Finnish adolescents observed that exposing them to sexually explicit Internet material, without receiving negative appearance feedback on social media, was related to more resistance towards the Me-Too movement and the acceptance of rape myths through notions of women as sex objects.

“Media perpetuates institutional myths, myths drive policy, and policy leads to media reporting” (Zatkin et al., 2021). With the aim of correcting rape culture, it seems relevant to understand what affects people when reading about rape and how this changes perceptions of rape. The lack of recent research output using experimental methods to answer these questions encouraged us to carry out our analysis, with particular reference to the effects of print media and, particularly, the use of photographs in addition to text. Our study departs from Strange et al. (2011), according to which people's memories are more strongly affected by imagery than narrative per se and is justified by the increasingly growing (mis) usage of pictures when reporting news about rape, which distract readers from the objectivity and severity of rape. on In particular, it builds on Dodge (2015)'s argument that the digitalisation of rape culture—which is normalised in the Western culture—can have clear damaging effects for both the victim and the perpetrator. This is also in line with the findings by Ali et al. (2020) that news channels often follow unethical guidelines to report cases of sexual assaults.

In this article, we investigate the reactive power that pictures may have on individuals' opinion on (1) the type of punishment deserved by the perpetrator and (2) the role of the victim in her rape. We do so by artificially manipulating the same image of the victim and her perpetrator, varying the intensity of physical contact and facial expressions. We provide experimental evidence that the nature of pictures shown to people when reading a piece of news on a rape case does not influence people's perceptions of rape and that, therefore, their usage should not be exploited by the media at the cost of the subjects involved.

Section 2 describes the data and the experimental design used in our analysis, Sections 3 and 4 present and discuss the main findings.

2. Data and methods

2.1 Sample

In an attempt to investigate the implications of visually communicating information about sexual violence, we randomly assign participants to a specific treatment. The experiment was conducted online on Qualtrics and n = 744 participants were hired through the platform of Prolific, where they were paid 5.02 pounds per hour. Data were collected on April 6th, 2022. The experiment obtained ethical approval by the Ethics Committee at Erasmus School of Law. Participants were informed they would be exposed to sensitive images and had to give consent to take part in the experiment.

Regarding our sample, participants are all residing in the US, are on average 41 years old, and mostly female (62%). 46.8 and 29.1% hold, respectively, less than a high-school degree and a professional degree (JD, MD) (Table 1). Participants declare a household income level of $10,000–$19,999 most of the time (14.8%), followed by more than $150,000 (10.9%) and $40,000–$49,999 (10.4%). With respect to political preferences, 48% of subjects in our sample are Democrats, while 26.2% are Republicans. Political view is reported following the recent finding by Ortiz and Smith (2022) that US partisans engage in greater victim blaming and sexual assault myth acceptance to defend their political identities in the Me-Too era. Because we are also interested in family characteristics, we collect data on the number of members of the participants' family: on average, participants indicate a total of 2.5 members, excluding themselves; if they have a brother or more than one brother: 37.3% of them do; if they have a sister or more than one sister: 40.4% say yes (Table 2).

We observe that the groups are well-balanced across treatments with respect to key demographics. As regards age, the average age in each treatment is, respectively, 41.9 (Treatment 1), 41.1 (Treatment 2), 40.5 (Treatment 3) and 40.1 (Treatment 4), with a standard deviation that ranges between 14.4 and 13.5. The sample is also balanced with respect to gender, where females represent the majority in all treatments. In Treatment 1, 59.5% are women; in Treatment 2, 62.6% are women; in Treatment 3, 55.4% are women; in Treatment 4, 54.4% are women. Finally, as regards education, we find equivalent proportions of low and high levels of education across all groups. Doctorate holders remain the minority in all treatments, while most hold a Bachelor degree: 44.3%, 44.9%, 50.5%, 47.8% in, respectively, Treatment 1, 2, 3, 4.

2.2 Experimental design

Through the experimental platform, the treatments were automatically randomly assigned to participants to have a balanced sample. The treatments designed for this analysis differ by the type of picture shown to participants when they read the news. The fictitious news illustrates a rape case experienced by a woman on behalf of her ex partner. This is the text that all participants are carefully asked to read.

“Following numerous episodes of psychological abuse, Sarah left her fiancé Adrien in March 2019 and moved to […], after which they lost contact. Yesterday, police officially reported Ms Sarah was raped in her apartment by Mr Adrien. The ex partner, which Sarah coincidentally met at a concert at a local bar a couple of nights before the distressing event, had apparently approached Sarah during the concert and learned about her address. After misleading her to believe someone had a delivery for her, he allegedly threatened Sarah with a kitchen knife, plugged her mouth, hit her and raped her multiple times. […] Mr Adrien is now in custody, while Ms Sarah is receiving medical treatment and social support.”

Based on the many examples of rape cases reported by the media using different types of images of the victim and the perpetrator, we artificially replicate four versions of the same image. The fourth case can be considered our control group as we purposefully avoid actors having expressions or identifiable physical gestures. For this purpose, a professional photographer and two actors are hired to make sure each picture only differs by the one specific variable we manipulate.

In the first picture, the victim and the perpetrator are photographed hugging and smiling. This type of picture was used multiple times by newspapers in recent times, such as for the case of Gabby Petito [2]. This picture is assigned to Group 1. We then proceed to vary the intensity of physical contact between victim and perpetrator, as well as their facial expressions. The piece of news reported to Group 2 is accompanied by a picture with the same two individuals, hugging but not smiling. The third picture is characterised by the victim and the perpetrator neither smiling nor hugging. Group 4 is also provided with a picture of the victim and the perpetrator not smiling and not hugging. However, in this case, the victim also presents clear physical marks of violence on her body. This type of image is more frequent for cases of abuse, followed by murder [3].

After seeing the picture assigned to their group of treatment, individuals are asked two questions. The validity of these questions is justified by both theoretical and experimental evidence in their use. First, “how many years of prison do you think the perpetrator deserves?”, with the possible answers being: None, less than 1, between 1–5, 5–10, 10–20, 20–30, more than 40, life in prison. This is in line with the argument that individuals are sensitive to actions committed by sexual predators and may have different preferences regarding the punishment to be attributed to them depending on the gravity of their crime (Carlsmith et al., 2007), also based on whether they are low or high in rape myth acceptance (Bohner et al., 2002). It also connects to the work of Carlsmith et al. (2007) on the civil commitment expected by sexual predators following punitive or lenient sentences. Similar experimental questions were used by Davies (1997), Bilz (2016) and Schwarz et al. (2020).

Second, they are asked about the role played by the victim in her rape. Namely, whether and to what extent they believe she was also accountable for the event (“Sarah should also be considered accountable for the event”), as well as if she could have avoided the event had she been more responsible and careful (“Sarah could have avoided the event if she had been more responsible and careful”). Similar questions can be found in the works of McCaul et al. (1990), Koppelaar et al. (1997) and Grubb and Harrower (2008). Answers are to provided according to a five-point Likert scale that goes from “Strongly disagree” to “Strongly agree”. This is in congruence with past research by Smith and Frieze (2003), which showed gender differences for rape victim empathy, lower for men and perceived victim responsibility, higher for men. It also connects to one of the first theories advancing the implications of victim blaming by Winkel et al. (1991) and of wrong conceptualisations of rape itself (Maung, 2021).

Participants were then asked to fill out a questionnaire with respect to their gender identity, educational background, political preferences and income. They were also asked whether they had a sister, or more than one sister. This conforms to earlier research by Bunting and Reeves (1983), who identified a positive effect of the sister variable on the strength of the general rape belief relationship.

3. Findings

After having collected our data, we investigate whether treating individuals with different images of the victim of the rape case and their perpetrator influences individuals' perception on the perpetrator's deserved punishment (Table 3) and and the role played by the victim in her rape (Tables 4 and 5). Below, we show mean results from our experiment by treatment, where Group 1 includes people exposed to a picture with positive facial and physical expressions (Hug, smile), with Groups 2 and 3 as variations of the same expressions (Hug, no smile; No hug, no smile). In Group 4 subjects were exposed to the less cheerful version of the same image of the victim and her perpetrator; including clear signs of physical violence on the victim.

As regards the perpetrator's deserved punishment (Table 3, Figure 1), we observe that most participants—an average of 28.38% per treatment—opt for a punishment of 10–20 years in prison, followed by 5–10 years. A very small minority—less than 0.54% per treatment—of the participants opts for either no punishment or a punishment of less than 1 year. On the other hand, a small yet relevant proportion thinks the perpetrator deserves to spend the rest of his life in prison—about 13.66%. This is in line with the findings by Sommer et al. (2016), who found increased victim blame in re-victimisation conditions. A previous study suggested that, in rape cases, defendants were more likely to receive greater punishments especially if from low socioeconomic status (Mazzella and Feingold, 1994). Overall, we observe that patterns remain constant across treatments. This suggests that using a specific picture over another does not affect people's perception on the punishment deserved by the perpetrator.

When we investigate patterns related to the victim's accountability (Table 4, Figure 2) and lack of cautiousness (Table 5, Figure 3), we find similar results. Most of the participants, independent of treatment, disagree with a statement affirming that the victim should also be held accountable for her rape—about 94.75% per treatment. While the proportion of individuals agreeing that the victim is to be held accountable is low, the proportion is the lowest for individuals in Group 4, exposed to a picture characterised by physical distance, lack of positive facial expressions, and clear violent marks on the victim (1.63%). Our findings do not present significant evidence of victim blaming. This could be explained by the fact that no expressions were used in the text to suggest the victim was deserving of her misfortune. Similarly, higher levels of blame are usually found for women who deviate from traditional gender roles (Masser et al., 2010).

Patterns do not differ across treatments as regards a further statement that the victim should have been more careful and responsible, as this could have prevented the event from happening. However, the proportion of participants who agree with this statement is higher compared to the proportion of participants who agree on the victim's accountability. While the majority of the participants does not agree with the statement that the victim should have been more cautious (81.10%, on average), 10.86% does. As regards the latter, the proportion is higher for people in Group 1, treated with a picture where there is no physical distance between the perpetrator and the victim and facial expressions are positive (12.43%). Results do not significantly vary based on political preferences, income level, gender identity or presence of female or male siblings in the family. On the one hand, society has become more aware of the existence and the negative implications of emotional abuse on women, especially during the pandemic (Sediri et al., 2020). On the other hand, the Me-Too movement has created a more empathetic environment where women are heard and believed (Alaggia and Wang, 2020). Our findings in favour of the victim may reflect these sociological changes.

4. Discussion

Media are filled with news on rape cases. In most cases, the rapist is identified by the victim as an intimate partner, a previous intimate partner or an acquaintance. The role of the narrative of rape has been extensively discussed in the literature. This has occurred due to the often non-significant impact of practical instruments, including mandatory distribution of sexual assault kits (Mourtgos et al., 2021) or programmes in favour of readiness to help, usually more effective for women (Hoxmeier et al., 2018). Based on our findings, it is likely that media does not influence the way people thing about the victim, the rapist and rape itself. This supports the argument that media are not to exploit pictures portraying the victim unless it is necessary to do so, as this does not affect support for (1) a more severe punishment of the rapist or the (2) nullification of victim co-responsibility in her rape. Our study is limited in that it is exploratory in nature, is limited to US participants, and exclusively looks at print media and photographs. Further studies could produce inferential analyses; explore results for countries other than the US, especially those were gender differences prevail in terms of culture; and investigate different and more interactive types of media, such as videos.

While media have full freedom of choice in the images they select when reporting cases of sexual assault, we appeal to them not to fall into the millions-of-clicks trap for the sake of publicity. Our findings suggest more attention should be dedicated to community-based interventions, who often and effectively recur to images to spread awareness against the culture of violence (Baker, 2013). With this experiment, we show that news covering cases of rape are not more effective in highlighting the gravity of rape when accompanied by images that portray either joyful facial and physical expressions of the victim with her rapist (e.g. the victim and her rapist hugging and smiling) or sensitive content such as remarks on the victim's body. Where necessary, policy makers should step in to make sure that the culture of rape is not encouraged but, instead, repressed. This can be achieved by encouraging sex education in schools, especially as regards consent; promoting laws that clarify the consequences of sexual assault to potential future perpetrators; creating community-based initiatives to spread awareness; supporting victims through legal, psychological, medical support; and training authorities to handle sexual assaults sensitively. As regards the media, policy makers should work closely with media organisations to ensure portrayals of rape cases do not reflect or encourage rape culture, but rather provide an objective and respectful representation of sexual assaults and violence. As regards national and regional broadcast programmes, for instance, ethical guidelines could be developed for the media with respect to sexual assault; programmes could be reviewed with experts and regulators before being broadcast to avoid victim blaming or re-victimisation; media professionals should be educated on such topics, including the effect of broadcasting of rape cases on victims and relatives. International standards should also be designed so as to create a safe space for victims everywhere in the world. This also means accounting for the misusage of pictures in the media, which often contribute to creating narratives that downplay rape, deeply affecting their readers and most importantly the victims of sexual violence.

Figures

Figure 1

Perpetrator's deserved punishment (years in prison), by treatment

Figure 2

Victim's accountability, by treatment

Figure 3

Victim's lack of cautiousness, by treatment

Table 1

Descriptive statistics

	Mean	St. dev.
Gender
Male	0.41	0.49
Female	0.61	0.59
Age	40.9	14.0
Education
Less than high school degree	0.47	0.50
High school degree (or) equivalent including GED	0.03	0.18
Professional degree (JD, MD)	0.29	0.45
Bachelor's degree	0.01	0.08
Master's degree	0.17	0.38
Doctorate	0.03	0.16
Political preference
Democrat	0.48	0.50
Republican	0.26	0.44
Independent	0.04	0.19
Other	0.02	0.14
No preference	0.20	0.40

Source(s): Author's own work

Table 2

Descriptive statistics

	Mean	St. dev.
Income
Less than $10,000	0.06	0.24
$10,000 – $19,999	0.15	0.36
$20,000 – $29,999	0.15	0.36
$30,000 – $39,99	0.07	0.26
$40,000 – $49,999	0.08	0.28
$50,000 – $59,999	0.10	0.30
$60,000 – $69,999	0.09	0.28
$70,000 – $79,999	0.09	0.29
$80,000 – $89,999	0.05	0.21
$90,000 – $99,999	0.09	0.29
$100,000 – $149,999	0.02	0.16
More than $150,000	0.11	0.31
Family
Number of family members	2.5	1.5
One or more than one brother	0.37	0.48
No brother	0.63	0.48
One or more than one sister	0.40	0.49
No sister	0.59	0.49

Source(s): Author's own work

Table 3

Deserved punishment by treatment: years in prison

Treatment	None	<1	1–5	5–10	10–20	20–30	>40	For life
Hug, smile	0.54	0.00	5.41	23.78	25.41	21.62	9.19	14.05
Hug, no smile	0.00	0.00	6.95	21.39	31.02	21.39	6.95	11.23
No hug, no smile	0.54	1.07	2.72	26.09	27.72	18.48	8.15	15.22
No hug, no smile, violent marks	1.09	0.00	5.98	17.93	29.35	23.37	8.15	14.13

Source(s): Author's own work

Table 4

Role of victim in rape by treatment: accountable

Treatment	Disagree	Neutral	Agree
Hug, smile	95.14	2.70	2.16
Hug, no smile	93.05	4.81	2.14
No hug, no smile	95.65	1.63	2.72
No hug, no smile, violent marks	95.11	3.26	1.63

Source(s): Author's own work

Table 5

Role of victim in rape by treatment: not careful enough

Treatment	Disagree	Neutral	Agree
Hug, smile	82.16	5.41	12.43
Hug, no smile	77.01	9.09	13.90
No hug, no smile	84.24	8.15	7.61
No hug, no smile, violent marks	80.98	9.78	9.24

Source(s): Author's own work

Notes

1. Please see: https://kjonnsforskning.no/en/2023/03/metoo-has-changed-how-media-portrays-rape

2. Examples of newspapers that used this type of image include the BBC (https://www.bbc.com/news/world-us-canada-58629192) and The Independent (https://www.independent.co.uk/news/world/americas/crime/gabby-petito-brian-laundrie-death-b1943203.html).

3. An example of media reporting a case of abuse, and murder, using separate pictures of the victim and the perpetrator can be found in The Daily Mail (https://www.dailymail.co.uk/news/article-9601077/Missing-Arkansas-woman-dead-wrapped-inside-sheet-abusive-exs-bed.html).