Abstract
There is an established relationship between sexual victimization (SV) and disordered eating (DE); however, few theory-based studies exist that adequately examine this relationship. Previous research on objectification theory has insufficiently tested the theory’s application to extreme objectifying experiences such as SV. To address this gap in the literature, we evaluated objectification theory as a means for clarifying the relationship between SV and DE. Specifically, we examined how self-surveillance, a proxy for self-objectification, and two of its psychological correlates, body shame and interoceptive deficits, may clarify the association between SV and DE. Participants were 389 undergraduate women. Path analysis of the model largely supports the application of objectification theory to this relationship but also suggests potential modifications to the theory. Further, the extent to which SV may be uniquely related to DE, above and beyond everyday objectification (e.g., leering, unwanted sexual comments), was explored. Implications are discussed.
Based on DSM-5 criteria, 13.1% of women may develop an eating disorder by the age of 20 (Stice, Marti, & Rohde, 2013). The prevalence is even greater in certain populations and when subthreshold symptoms are considered. For example, in one sample of undergraduate women, 72.7% endorsed some level of symptoms (Tripp & Petrie, 2001). In another sample of undergraduate women without eating disorders, 39.5% of women who were either underweight or in a healthy weight range were actively trying to lose weight, 44% of the sample reported overeating, and participants overall tended to overestimate their BMI (Fayet, Petocz, & Sanman, 2012). It should be noted that men can develop disordered eating (DE) as well, but they do so at a significantly lower rate than girls and women (Fredrickson & Roberts, 1997). Consequently, understanding the etiology of DE in women is essential because it can inform both prevention and treatment. Research examining factors that lead to DE development is diverse, with explanations including biological, developmental, and sociocultural factors (Schmidt, 2003). Particular attention has been paid to models that address cultural attitudes about food and weight, media images of ideal thinness, and body dissatisfaction (e.g., van den Berg, Thompson, Obremski-Brandon, & Coovert, 2002). Although there has been considerable support for these factors (e.g., Preti, Incani, Camboni, Petretto, & Masala, 2006), they do not fully explain DE, and thus it is vital to explore other mechanisms through which DE develops.
Relationship Between Sexual Victimization and Disordered Eating
One risk factor for DE with considerable empirical support is sexual victimization (SV; Fischer, Stojek, & Hartzell, 2010; Thompson & Wonderlich, 2004). Historically, the vast majority of research has exclusively examined childhood SV (Brewerton, 2007); however, a number of more recent studies have found SV that occurs in adulthood to be a risk factor for DE as well (e.g., Dubosc et al., 2012; Messman-Moore & Garrigus, 2007). The inclusion of adult SV increases the number of women known to be at risk for the development of DE, thus creating a heightened need for an understanding of the mechanisms that facilitate the relationship of SV to DE.
Objectification Theory, Sexual Victimization, and Disordered Eating
There are a number of models that are empirically supported as explanations of DE development, yet many do not address SV as a component (e.g., van den Berg et al., 2002). One feminist theory that does account for SV, has been used historically to better understand the development of DE, and is frequently implemented in counseling psychology literature (e.g., Augustus-Horvath & Tylka, 2009; Carr & Szymanski, 2011), is objectification theory (Fredrickson & Roberts, 1997). According to objectification theory, girls and women are often treated in a sexually objectifying manner such that they are treated as a body, rather than a whole person (Fredrickson & Roberts, 1997). Objectifying experiences can present in the form of cultural objectification, which includes but is not limited to, leering and visual media portrayals of sexualized images of the female body, as well as through more direct and extreme routes, such as SV (Fredrickson & Roberts, 1997). Fredrickson and Roberts (1997) posit that women internalize these sexually objectifying experiences, which may result in monitoring, or self-surveillance, which is often used as a proxy to measure self-objectification (e.g., Carr & Szymanski, 2011; Kozee & Tylka, 2006; Tylka & Hill, 2004). Self-objectification is theorized to result in a set of maladaptive consequences (i.e., body shame, appearance anxiety, safety anxiety, interoceptive deficits, and reduced flow, a psychological phenomenon by which one becomes entirely absorbed in a challenging activity that is experienced as rewarding) which, in turn, contribute to negative mental health outcomes (i.e., depressive symptoms, sexual dysfunction, and DE; Fredrickson & Roberts, 1997; Moradi & Huang, 2008).
Body Shame
Body shame and interoceptive deficits are likely most salient constructs in the relationship between SV and DE. In fact, relative to the other theorized psychological consequences, body shame has garnered the most consistent empirical support in the application of objectification theory to DE. Specifically, a number of studies (e.g., Kozee & Tylka, 2006; Noll & Fredrickson, 1998; Tiggemann & Slater, 2001; Tiggemann & Williams, 2012; Tylka & Hill, 2004) have established an indirect relationship between objectification and DE through self-objectification, or self-surveillance, and body shame (see Figure 1, Indirect Path A). According to objectification theory, a culture of sexual objectification creates a climate in which women are more likely to be hyperaware of their bodies and to compare them to the largely unattainable thin ideal, which is likely to result in body shame (Fredrickson & Roberts, 1997). In an effort to alleviate that shame, women may engage in extreme dieting and other weight loss efforts (e.g., compulsive exercise, self-induced vomiting, laxative use).
Hypothesized indirect paths between sexual victimization and disordered eating.
Despite body shame’s success in predicting DE symptoms that are associated with everyday sexual objectification (e.g., leering, unwanted sexual comments), there is a dearth in the literature that specifically addresses more extreme forms of objectification, such as SV. Fredrickson and Roberts (1997) theorized that SV functions in a similar fashion as everyday objectification. That is to say that when women’s bodies are treated, quite literally, as a sexual object to satisfy their perpetrator’s desires, they may internalize their own sexual objectification, which will likely be associated with body shame and DE. Others have posited that, in their attempts to make sense of their own victimization, survivors of SV may become ashamed of their bodies, incorrectly placing blame on their bodies for their attacks (Kearney-Cooke & Striegel-Moore, 1994). This body shame may then result in DE, which may function as punishment or as a means of diminishing perceived attractiveness (i.e., gaining weight via binge eating) or secondary sex characteristics (i.e., losing considerable weight) in hopes of protecting oneself against a future attack (Kearney-Cooke & Striegel-Moore, 1994; Root, 1991).
Despite receiving attention in the conceptual literature dating back several decades, there have been very few empirical studies that test the role of body shame in the relationship between SV and DE. The little research that does exist measures constructs similar to body shame (i.e., body image distortion, negative feelings about one’s body; Byram, Wagner, & Waller, 1995; Tripp & Petrie, 2001) rather than specifically assessing body shame, as theorized by objectification theory. Consequently, they undoubtedly measure a different latent trait and, further, fail to address the role of body shame in the full context of objectification theory. Given both the empirical and theoretical support for the role of body shame in the relationship between SV and DE, it should be explicitly tested within the broader objectification theory framework.
Interoceptive Deficits
Interoceptive deficits refers to a lack of awareness of internal bodily states, both physical and emotional, and is the other psychological correlate of self-objectification that is likely to be relevant in the relationship between SV and DE. In their foundational paper, in which they proposed objectification theory, Fredrickson and Roberts (1997) suggested that when women are hypervigilant with regards to their outer appearance, they might have less cognitive resources with which to attend to their internal bodily states. Thus, they posit that there is an indirect link between sexual objectification and DE through self-objectification, or self-surveillance, and interoceptive deficits (see Figure 1, Indirect Path B).
In the extant literature that assesses interoceptive deficits’ role within objectification theory, interoceptive deficits have been defined and measured in several different ways including decreased awareness of (a) emotions (e.g., Muehlenkamp & Saris-Baglama, 2002); (b) physical internal bodily states broadly defined (i.e., dry throat, heart beating; e.g., Tigemmann & Slater, 2001); and (c) emotions and physical internal body sensations specific to eating (i.e., hunger or satiety; e.g., Tylka & Hill, 2004). Given that women with eating disorders have greater difficulty identifying both emotions and hunger and satiety cues, it may be important for measures of interoceptive deficits to incorporate both facets (Myers & Crowther, 2009). More specifically, with regards to hunger and satiety, women who cannot accurately assess these sensations may end up eating significantly less or more than their bodies need. The same cannot be said of other physical internal bodily states (i.e., dry throat, heart beating), and, as such, these may be less relevant to DE. To this point, studies that have only assessed awareness of emotions (e.g., Muehlenkamp & Saris-Baglama, 2002) and those that only measure physical internal bodily states, broadly defined, have typically not supported the role of interoceptive deficits in objectification theory’s application to DE. However, when decreased awareness in both emotions and hunger or satiety cues is assessed, the role of interoceptive deficits is most often supported (e.g., Myers & Crowther, 2009; Tylka & Hill, 2004) although not always (Tiggemann & Williams, 2012).
Further, there is discrepancy with regards to how interoceptive deficits function as a mediator. Fredrickson and Roberts (1997) posited that it would mediate the relationship between self-surveillance and DE (see Figure 1, Indirect Path B). This has been demonstrated in at least one empirical study (i.e., Myers & Crowther, 2008); however, Tylka and Hill’s (2004) results suggest the role of interoceptive deficits may be more complex. Specifically, although they initially found that interoceptive deficits were positively associated with self-surveillance (i.e., via zero-order correlation), when they tested the full model of objectification theory, they found that the association between self-surveillance and interoceptive deficits was not direct, but rather mediated by body shame (see Figure 1, Indirect Path C). This indirect relationship has been replicated in other studies (e.g., Kozee & Tylka, 2006). If interoceptive deficits’ role in objectification theory really was due to women’s lessened cognitive resources, as was initially theorized (Fredrickson & Roberts, 1997), it might be expected that there would remain a direct association between self-surveillance and interoceptive deficits. However, the fact that the relationship was fully mediated by body shame (Tylka & Hill, 2004) suggests that women may dissociate from their own emotional, hunger, and satiety cues only to the extent to which they experience body shame, a somewhat different process than what was originally posited by Fredrickson and Roberts (1997).
As is the case with body shame, the extant literature has not sufficiently examined the potential clarifying role of interoceptive deficits in the association between SV and DE. This is somewhat surprising given how conceptually relevant interoceptive deficits are in the SV to DE relationship. Specifically, dissociation (i.e., disruptions in awareness, which includes bodily perceptions) is a frequent outcome of SV (Briere & Elliott, 1994). It is no surprise then that interoceptive deficits, which can be conceptualized as a subset of dissociative symptoms, are demonstrated to be higher in survivors of SV than their non-victimized counterparts (Zlotnick et al., 1996). Further, DE symptoms (i.e., bingeing, purging, restricting) can serve as a form of dissociative coping, allowing survivors of SV to distract from painful memories and emotions associated with their victimization (Burns, Fischer, Jackson, & Harding, 2012; Root, 1991). Consequently, when applying objectification theory to the SV to DE relationship, it is reasonable to expect that in addition to the indirect path through self-surveillance and interoceptive deficits (see Figure 1, Indirect Path B) theorized by Fredrickson and Roberts (1997), there may also be another indirect link through interoceptive deficits alone (see Figure 1, Indirect Path D). This path would not be predicated on self-surveillance but rather would represent the dissociative coping that may take place in response to the overwhelming distress associated with SV.
Current Study
To our knowledge, this is the first study to specifically apply objectification theory to the relationship between SV and DE. Of note, several studies evaluating objectification theory (e.g., Augustus-Horvath & Tylka, 2009; Kozee & Tylka, 2006) have utilized the Interpersonal Sexual Objectification Scale (ISOS; Kozee, Tylka, Augustus-Horvath, & Denchik, 2007), which contains several items that assess extreme objectification. However, participants’ responses on these items are averaged with their responses to the remaining majority of items that assess everyday objectification, thus precluding authors from determining the unique predictive ability of SV on self-objectification, the various theorized psychological consequences, and DE. Further, only two studies that we found (i.e., Carr & Szymanski, 2011; Davidson & Gervais, 2015) have ever applied objectification theory to SV, specifically. Although these studies established the relationship between SV, self-surveillance, and body shame, neither incorporated interoceptive deficits or assessed DE as an outcome (Carr & Szymanski, 2011; Davidson & Gervais, 2015).
Given this gap in the literature, the current study was designed to evaluate objectification theory as a means for clarifying the relationship between SV and DE in a sample of undergraduate women by focusing on the two psychological effects of self-objectification with the most theoretical (i.e., interoceptive deficits, body shame) and empirical (i.e., body shame) support. The decision to exclusively sample women is informed by the disproportionately high rates at which women experience SV and report DE symptoms as well as the theoretical context that objectification theory was created in an effort to explain why women experience certain negative mental health outcomes, such as DE, at higher rates than men (Breiding, 2015; Fredrickson & Roberts, 1997). Specifically, we hypothesized that:
Consistent with previous literature (e.g., Fischer et al., 2010; Thompson & Wonderlich, 2004), individuals with a lifetime history of SV will have higher levels of DE than their nonvictimized counterparts.
Self-surveillance and body shame will serially mediate the relationship between SV and DE (i.e., Figure 1, Indirect Path A). This is consistent with Fredrickson and Roberts’s (1997) theory that sexual objectification is associated with DE because women become hyperaware of their bodies’ failure to meet the thin-ideal and that they engage in DE behaviors in an attempt to lose weight.
The relationship between SV and DE will be serially mediated by self-surveillance and interoceptive deficits (i.e., Figure 1, Indirect Path B). This hypothesized path is intended to test Fredrickson and Roberts’s (1997) assertion that high levels of self-surveillance may exhaust resources that would otherwise be allocated to awareness of internal states.
Self-surveillance, body shame, and interoceptive deficits will serially mediate the relationship between SV and DE (i.e., Figure 1, Indirect Path C). This hypothesized indirect path is consistent with positing that interoceptive deficits are the byproduct of distancing from one’s body in reaction to the experience of body shame. It has also been empirically supported in a study that examined everyday objectification (Tylka & Hill, 2004).
Interoceptive deficits will mediate the relationship between SV and DE (i.e., Figure 1, Indirect Path D). Although this indirect path has not received attention in most objectification theory literature, it is consistent with trauma theories that conceptualize dissociation and, by extension, interoceptive deficits, as an experience shared by SV survivors and individuals with DE (Burns et al., 2012; Root, 1991).
Given that the application of objectification theory to the relationship between SV and DE is novel, the current study also sought to explore other potential indirect effects. Specifically, we examined two indirect effects in which SV is directly associated with body shame above and beyond self-surveillance (i.e., relationship between SV and DE mediated by body shame, the relationship between SV and DE serially mediated by body shame and interoceptive deficits), as experiences of SV may cause women to place blame on their own bodies for the attack (Kearney-Cooke & Striegel-Moore, 1994; Root, 1991) without necessarily seeing their own body as an object. Further, a direct effect between SV and body shame, above and beyond self-surveillance, was found in one of the only other studies that explicitly examined extreme objectification (Davidson & Gervais, 2015). The final exploratory indirect effect we tested was the association of SV and DE through self-surveillance alone, a path that has been supported in a number of previous studies that have primarily assessed everyday objectification (e.g., Kozee & Tylka, 2006; Tiggemann & Slater, 2001).
Additionally, we examined these same seven indirect effects when accounting for everyday objectification, to assess the unique contribution of extreme objectification above and beyond that of everyday objectification. There is evidence to suggest that SV predicts unique variance in self-surveillance above and beyond body evaluation (i.e., everyday objectification; Kozee et al., 2007); however, little else is known about their relative role in objectification theory.
Methods
Participants
The participants consisted of 389 women who identified as non-Latina White (74.8%), Black or African American (11.3%), Asian or Pacific Islander (3.9%), Latina or Hispanic (2.6%), Arab (1.8%), Indigenous or Aboriginal (0.3%), Multiracial (3.6%), a race other than what was provided as an option in the survey (1.3%), and 0.5% did not report their race or ethnicity. Participants ranged in age from 18 to 60 years (M = 21.04, SD = 4.79). About a third (34.2%) of participants were first year students, 27.2% were sophomores, 15.4% were juniors, 12.3% were seniors, and 10.8% were in their fifth year or beyond. Of the participants, 15.7% reported a household income of less than $10,000, 18.7% reported between $10,000 and $29,000, 12.0% between $30,000 and $49,999, 10.3% between $50,000 and $74,999, 19.0% between $75,000 and $150,000, 5.1% of greater than $150,000, and 19.0% did not report their household income. With regards to sexual orientation, the majority of the sample (86.1%) identified as exclusively heterosexual, and the remaining participants (13.9%) reported varying levels of same-sex attraction. In the current sample, 58.6% of participants reported experiencing SV, defined as any form of unwanted sexual contact ranging from fondling to attempted or completed rape, at some point in their lifetime.
Procedures
Participants were recruited from a large public university in Ohio through the university’s website, such that students enrolled in psychology courses receive extra credit for taking surveys. Participants were provided a link that directed them to an online software program, Qualtrics. Institutional review board approval was obtained for this study.
Of the initial 474 participants, two were omitted from the analyses because they were 17 years old or younger, 10 because they were men, and two because they indicated their gender as “other.” Three additional participants were removed because they were determined to be multivariate outliers (Tabachnik & Fidell, 2012; see Results section).
Of the remaining 457 participants, 14.9% had either stopped taking the survey partway through or did not complete at least 80% of each primary measure, and thus were omitted from analyses. The decision to use this particular threshold was based on Downey and King’s (1998) finding that when at least 80% of the scale was completed, it did not bias scale reliability. Available item analysis was used for the 389 remaining participants. That is to say that instead of calculating sums for measures, the mean was calculated for each measure providing that at least 80% of the items of that measure were completed. This method of handling missing data has demonstrated approximately equivalent performance to multiple imputations and is particularly suitable for research conducted by counseling psychologists given that it typically involves scale or subscale-level analysis (Parent, 2013).
Measures
Sexual victimization
The Sexual Experiences Survey: Short Form Victimization (SES-SFV; Koss et al., 2007) is a self-report measure that assesses victimization of a wide variety of unwanted sexual experiences. It is extensively used and considered to be the premier measure of sexual aggression (Carr & Szymanski, 2011). Each of the seven primary items describes a scenario and then asks participants to indicate how many times that scenario has occurred (0, 1, 2, 3+) as well as the means by which the perpetrator forced them (Koss et al., 2007). Sample items include “Someone had oral sex with me or made me have oral sex with them without my consent,” and “A man put his penis into my butt, or someone inserted fingers or objects without my consent.” Sample choices for means of force include “by taking advantage of me while I was too drunk or out of it to stop what was happening” and “threatening to physically harm me or someone close to me.” Higher scores indicate that the participant has experienced more SV. The original form of the SES was created by Koss and Oros (1982) and was demonstrated to have adequate psychometric properties. Its internal consistency reliability was adequate (α = .74), and its 1-week test-retest reliability was high (r = .93; Koss & Gidycz, 1985). The SES-SFV has demonstrated convergent validity with the sexual coercion subscale of the Conflict Tactics Scale-2 (Anderson, Cahill, & Delahanty, 2016) and preliminary discriminant validity with various demographic variables (i.e., age, living arrangement; Cecil & Matson, 2006). In the current study, the SES-SFV demonstrated excellent internal consistency reliability (α = .96).
The SES-SFV has been used to assess SV before and after the age of 14 (Ullman, Filipas, Townsend, & Starzynski, 2007). In the current study, we assessed both time frames by having participants respond to all items twice, once pertaining to when they were younger than 14 and the other for when they were 14 and older. Initially, the primary analyses were run separately for SV before 14 years and SV at 14 years and older; however, there were no significant differences in the results and, consequently, the two categories were collapsed to assess experiences of SV across the lifetime.
Although the SES-SFV is most frequently scored using a severity-ranking scheme, there are a number of scoring methods demonstrated to be viable (Davis et al., 2014). One such manner is to score it dichotomously, such that an affirmative response to any of the seven primary items results in the participant being categorized as having a SV history (Anderson et al., 2016). In the current study, we implemented dichotomous scoring (1 = no SV history, 2 = SV history) because it allowed us to retain a greater proportion of our sample size than we would have if we had utilized the severity-ranking scoring. Further, when we ran our primary analyses using the more traditional severity-ranking scheme, results did not significantly differ from those reported here with dichotomous scoring.
Everyday objectification
The Body Evaluation subscale of the Interpersonal Sexual Objectification Scale (ISOS-BE; Kozee et al., 2007) measures the extent to which an individual’s body is sexualized and commented on. It is comprised of 11 items to which participants indicate the frequency of various sexually objectifying experiences on a 5-point Likert-type scale ranging from 1 (never) to 5 (almost always). Sample items include “How often have you been whistled at while walking down the street?” and “How often have you heard a rude, sexual remark made about your body?” Higher scores represent higher levels of body evaluation. The ISOS has strong psychometric properties, with Kozee et al. (2007) reporting an internal consistency reliability ranging from .92 to .94, and a 3-week period test-retest reliability of .89. Further, its convergent validity has been documented with measures of self-objectification, body shame, and sexist degradation; and evidence of discriminant validity was obtained with a measure of unfair sexist events at work/school as well as social desirability (Kozee et al., 2007). The internal consistency reliability of the ISOS-BE for the current study was .94.
Self-surveillance
The Body Surveillance subscale of the Objectified Body Consciousness Scale (OBCS; McKinley & Hyde, 1996) assesses the degree to which women view themselves as an outside observer of their own bodies and has often been used as a means of measuring self-objectification (Calogero, 2011; Carr & Szymanski, 2011; Kozee & Tylka, 2006; Tylka & Hill, 2004). It consists of eight items that participants answer on a 7-point Likert-type scale from 1 (strongly disagree) to 7 (strongly agree; McKinley & Hyde, 1996). Higher scores indicate greater self-surveillance. Sample items include “During the day, I think about how I look many times,” and “I often worry about whether the clothes I am wearing make me look good.” It has been demonstrated to have adequate psychometric properties, with internal consistency reliability ranging from .79 to .89 in samples of undergraduate women, as well as adequate test-retest reliability over a 2-week period (r = .79; McKinley & Hyde, 1996). The Body Surveillance subscale’s convergent validity has been documented with a measure assessing public self-consciousness (McKinley & Hyde, 1996), and its discriminant validity has been established with a measure that assesses enjoyment of sexualization (Smolak, Murnen, & Myers, 2014). In the current study, the internal consistency reliability was .79.
Body shame
The Body Shame subscale of the OBCS (McKinley & Hyde, 1996) assesses the degree to which a woman feels like she is a bad person when she thinks she has not fulfilled the cultural expectations regarding appearance standards, particularly for thinness. It consists of eight items that participants answer on a 7-point Likert-type scale from 1 (strongly disagree) to 7 (strongly agree). Higher scores indicate greater body shame. Sample items include “When I can’t control my weight, I feel like something must be wrong with me” and “I would be ashamed for people to know what I really weigh.” The Body Shame subscale of the OBCS is psychometrically adequate, with internal consistency reliability ranging from .75 to .84, in samples of undergraduate women, and 2-week period test-retest reliability of .79 (McKinley & Hyde, 1996). Its convergent validity has been established with measures that assess internalization of cultural body standards (McKinley & Hyde, 1996) and adherence to the feminine norm of thinness (Parent & Moradi, 2011). The Body Shame subscale has documented discriminant validity with another subscale of the OBCS, Control Beliefs, which measures the extent to which women believe that if they try enough they can attain cultural beauty standards (McKinley & Hyde, 1996). The internal consistency reliability was .85 in the current sample.
Interoceptive deficits
Interoceptive deficits is one of the subscales of the Eating Disorder Inventory-3 (EDI-3; Garner, 2004) that assesses more general psychological symptoms that are relevant to the development and maintenance of disordered eating. More specifically, this subscale measures confusion related to accurately recognizing and responding to emotional states, including a cluster of items that indicate fear of affect and another cluster that indicates affective confusion. It consists of nine forced choice items that ask participants to indicate whether each item applies always, usually, often, sometimes, rarely, or never. Higher scores indicate greater levels of interoceptive deficits. This version is psychometrically strong, with internal consistency reliability ranging from .90 to .97 and test-retest reliability ranging from .93 to .98 (Garner, 2004). The interoceptive deficits subscale has demonstrated convergent validity with emotion dysregulation and personal alienation, a construct that measures poor self-understanding (Garner, 2004). Its discriminant validity has been established with measures of various psychological symptoms other than disordered eating (e.g., anxiety, agoraphobia; Garner, 2004). In the current sample, the internal consistency reliability was .86.
Disordered eating
The Eating Attitudes Test-26 (EAT-26; Garner, Olmsted, Bohr & Garfinkel, 1982) assesses levels of DE symptomology and has demonstrated its validity as a continuous measure of DE in both clinical and nonclinical samples of women (Tylka & Hill, 2004). It consists of 26 items that require participants to indicate how often they experience the described statements. Sample items include “[I] cut my food into small pieces” and “[I] think about burning calories when I exercise.” There are six response options that range from (always) to (never). Higher scores are associated with lower frequency of DE symptoms and, consequently, all items were reversely coded so that results would be more intuitive. The EAT-26 has demonstrated good psychometric properties with high internal consistency reliability (α = .91) and test-retest reliability over a 3-week period (r = .86; Mazzeo, 1999). Additionally, its convergent validity has been documented with the Eating Disorder Risk Composite of the EDI-3 (Garner, 2004), and its discriminant validity has been established with measures that assess symptoms of other psychological disorders (i.e., depression, anxiety; Garner et al., 1982). The current study demonstrated an internal consistency reliability of .87.
Results
Preliminary Analyses
The data were examined to determine whether they met the necessary assumptions for statistical analysis. Histograms were used to assess normality based on Tabachnick and Fidell’s (2012) suggestion to examine the shape of the distribution, rather than relying on formal inference tests, when the sample is relatively large. The data appeared normal for each of the variables. Examination of Mahalanobis distance revealed three multivariate outliers that were removed from the dataset. Zero-order correlations were run, and because there were no correlations that exceeded .70, we concluded that bivariate multicollinearity did not present a statistical problem in the current sample (Tabachnick & Fidell, 2012). To assess multivariate multicollinearity and singularity, collinearity diagnostics were run. The condition index was below .30 for each dimension and thus in the acceptable range (Tabachnick & Fidell, 2012).
Correlations and t-tests were run between demographic variables and scores from the EAT-26. Due to small samples of some racial ethnic groups as well as some points along the Kinsey scale used to assess sexual orientation, both race (1 = White, 2 = Women of Color) and sexual orientation (1 = heterosexual, 2 = sexual minority women) were dichotomized. Only sexual orientation was found to be significantly related to DE, with sexual minority women demonstrating more DE than heterosexual women. Consequently, sexual orientation was controlled for in subsequent analyses. Means, standard deviations, and zero-order correlations are summarized in Table 1.
Correlations, Means, and Standard Deviations Among Measured Variables
Sexual victimization was scored dichotomously (1 = no SV history; 2 = SV history). bSexual orientation was scored dichotomously (1 = exclusively heterosexual; 2 = sexual minority).
p < .05. **p < .01. ***p < .001.
Primary Analyses
To address the first hypothesis, that women with a SV history would have higher levels of DE than their nonvictimized counterparts, an analysis of covariance controlling for sexual orientation was conducted and yielded significant findings F(1, 386) = 13.10, p < .01. Specifically, women with a history of SV (M = .68) had significantly higher levels of DE than their nonvictimized counterparts (M = .54). To addresses the remaining hypotheses, assessing each of the indirect paths from SV to DE, path analysis with 10,000 bootstrap samples was conducted using Model 6 of PROCESS for SPSS (Hayes, 2013). The bias-corrected bootstrap 95% confidence interval for the indirect effect of SV on DE, through self-surveillance and body shame serially, was interpreted to be significant, thus providing support for Hypothesis 2 (i.e., Table 2, Indirect Path A). Results did not support Hypothesis 3, that the relationship between SV and DE would be mediated by self-surveillance and interoceptive deficits, serially (i.e., Table 2, Indirect Path B). However, the other indirect effects that included interoceptive deficits were supported. Specifically, the indirect effect of SV on DE through self-surveillance, body shame, and interoceptive deficits serially (i.e., Table 2, Indirect Path C), as well as the indirect effect of SV on DE through interoceptive deficits (i.e., Table 2, Indirect Path D), can be interpreted as statistically significant. Only one of the exploratory indirect effects was significant. The relationship between SV and DE that was mediated by self-surveillance was significant (i.e., Table 2, Indirect Path E), whereas the paths that included a direct association between SV and body shame were not (i.e., Table 2, Indirect Paths F & G). There was no direct effect of SV on DE above and beyond the aforementioned indirect effects (see Figure 2).
Indirect Effects of Sexual Victimization on Disordered Eating
Note. CI = confidence interval; SV = sexual victimization; SS = self-surveillance, BS = body shame; ID = interoceptive deficits; DE = disordered eating. Indirect effects in bold were hypothesized and the remainder were exploratory.
p < .05.
Relationships among variables in the proposed model of how objectification theory may clarify the relationship between sexual victimization and disordered eating. Numbers in parenthesis represent coefficients when everyday objectification is statistically controlled for.
In an effort to ascertain the ways in which everyday objectification and SV are similar and distinct in their relation to DE, Model 6 was rerun controlling for everyday objectification (i.e., ISOS-BE). The direct association between SV and self-surveillance was no longer significant (see Figure 2) and thus each of the indirect paths that included the SV to self-surveillance relationship (i.e., Indirect Paths A, B, C, E; see Table 2) were no longer significant. Only the indirect effect through interoceptive deficits remained significant (see Table 2, Indirect Path D). Additionally, the indirect effects in which there is a direct association between SV and body shame become significant when everyday objectification is controlled for, although they were not significant in the initial analyses (i.e., Table 2, Indirect Paths F & G).
Discussion
To our knowledge, this is the first study to apply objectification theory framework to the relationship between SV and DE. The results largely support the application; however, they also provide initial evidence for potential additional links above and beyond what Frederickson and Roberts (1997) initially theorized. Further, the current findings help to determine ways in which SV is uniquely related to DE.
Findings
In the current sample, 58.6% of participants reported a history of SV. Initially, this number may appear high; however, it encompasses all unwanted sexual contact and, although disturbing, it does not deviate considerably from the findings of other studies (e.g., Breiding, 2015). For example, results of the National Intimate Partner and Sexual Violence Survey demonstrated that 19.3% of women had been raped at some point in their lifetime and that 43.9% had been victim of some other form of SV (Breiding, 2015). Further, consistent with previous findings (e.g., Fischer et al., 2010; Thompson & Wonderlich, 2004), the current study found that female SV survivors demonstrated greater levels of DE than their nonvictimized counterparts.
The results of the current study largely support the application of objectification theory to the relationship between SV and DE. Fredrickson and Roberts (1997) postulate that women internalize their lived experiences of sexual objectification and that this internalization is associated with psychological correlates which, in turn, are associated with DE. The current results support this indirect relationship with regard to body shame (see Table 2, Indirect Path A). Although the cross-sectional nature of the current study precludes firm conclusions, women who are sexually victimized may begin to see themselves as sexual objects to be compared to the culturally prescribed thin-ideal. Women may then experience shame directed at their own bodies, as they acknowledge the discrepancy between their own body and their desired physique. Consequently, they may engage in DE in an effort to lose weight and more closely approximate the “thin-ideal” (Fredrickson & Roberts, 1997). It is also possible that body shame is associated with DE due to an attempt to decrease attractiveness by gaining weight or to decrease the visibility of secondary sex characteristics by losing weight, in an attempt to prevent revictimization (Kearney-Cooke & Striegel-Moore, 1994; Root, 1991). Notably, the supported indirect relationship between SV and DE, through self-surveillance and body shame, is also consistent with the results of many studies that have tested objectification theory’s application to DE, but did not assess the role of SV, specifically (e.g., Tiggemann & Williams, 2012; Tylka & Hill, 2004).
The current findings deviate from objectification theory, however, with regard to the role of interoceptive deficits. More specifically, there is no support for the hypothesis that interoceptive deficits are the product of diminished cognitive resources associated with self-surveillance, as was originally theorized (see Table 2, Indirect Path B; Fredrickson & Roberts, 1997). That is not to say, however, that interoceptive deficits are not relevant to the relationship between SV and DE. To the contrary, the presence of two significant indirect paths that include interoceptive deficits suggests that this construct is crucial in the link between SV and DE.
Again, the cross-sectional nature of the data prevents determination of causality, but support for the indirect path through self-surveillance, body shame, and interoceptive deficits (see Table 2, Indirect Path C) provides preliminary evidence that women’s shame directed toward their own bodies may be associated with attempts to distance themselves from their bodies. This distancing, which may manifest as interoceptive deficits is, in turn, associated with DE. Although this indirect path has found support in other studies that have not specifically examined SV (e.g., Kozee & Tylka, 2006; Tylka & Hill, 2004), it may be particularly relevant to extreme objectification because of the potential for women to experience shame associated with mistakenly placing the blame for their victimization on themselves and their bodies (Kearney-Cooke & Striegel-Moore, 1994; Root, 1991).
Although not explicitly theorized by Fredrickson and Roberts, this indirect path through self-surveillance, body shame, and interoceptive deficits (see Table 2, Indirect Path C) can be viewed as consistent with objectification theory as it still supports the notion that internalization of objectification (i.e., self-surveillance) is a vital mechanism linking SV to the psychological correlates (i.e., body shame, interoceptive deficits). By comparison, the indirect path between SV and DE through interoceptive deficits alone (see Table 2, Indirect Path D) is a significant deviation from the theory in that interoceptive deficits are not predicated on internalization of extreme objectification. Rather, this path is consistent with trauma literature (i.e., Briere & Elliott, 1994; Root, 1991), which posits that SV survivors engage in various forms of dissociation as a means of coping with the severe distress associated with their victimization. Interoceptive deficits, or lack of internal awareness, can thus be interpreted as evidence of women’s dissociative coping. The fact that the indirect path through interoceptive deficits remained significant above and beyond everyday objectification demonstrates that it is perhaps most salient to extreme forms of sexual objectification.
Several other deviations from objectification theory emerged through the exploratory analyses conducted. First, the indirect path of SV on DE through self-surveillance (see Table 2, Indirect Path E) suggests that the internalization of objectification is associated with DE regardless of the extent to which women experience body shame and interoceptive deficits. However, when everyday objectification is controlled for, the direct association between SV and self-surveillance is no longer significant and, instead, the direct association between SV and body shame is significant when it was not before (see Figure 2). This suggests that the unique portion of the relationship between SV and DE that is not shared with everyday objectification may not be predicated on the extent to which women internalize messages of objectification. Rather, body shame may be a more proximal correlate of SV related to self-blame women may experience as part of their trauma response (Kearney-Cooke & Striegel-Moore, 1994; Root, 1991). Further research is needed to better understand these relationships. Broadly speaking, however, the results of this study provide preliminary evidence that objectification theory may warrant modification in order to adequately address extreme objectification.
Implications
Findings from the current study have important clinical implications for counseling psychologists. Specifically, it can inform prevention efforts as well as therapeutic interventions. College campuses and rape crisis centers could create outreach programming to provide psychoeducation to female survivors of SV on the relationship between SV and DE as well as more proximal outcomes (i.e., body shame, interoceptive deficits) that may arise. Further, clinicians who work with female clients who have been sexually victimized should assess for these proximal outcomes, which would likely result in earlier detection of DE and, consequently, a more timely intervention. Similarly, psychologists working with female DE clients might be attuned to traumatic sexual experiences that may have contributed to their symptomology, which would be important for their conceptualization and work with that client.
More broadly, the application of the social and feminist theoretical framework in this context suggests that treatment modalities that address such dynamics might be especially beneficial for clients with these presenting concerns. For example, empowerment feminist therapy’s (Worell & Remer, 2003) focus on techniques such as consciousness-raising and “separating the external from the internal” may help female clients to name everyday and extreme objectification in their own lives, identify its deleterious impact, and externalize objectification which may protect against psychological correlates of self-objectification (i.e., body shame, interoceptive deficits; p. 68). Further research should test specific treatment outcomes for empowerment feminist therapy and other sociocultural and feminist-based interventions.
Strengths, Limitations, and Future Directions
There are a number of strengths to the current study, including its sound theoretical framework, large sample size, and inclusion of SV that occurs throughout the lifetime, which increases its generalizability relative to the numerous studies that have exclusively examined childhood SV as a risk factor for DE (Brewerton, 2007). It is important, however, to acknowledge the methodological limitations as well. Although the time frame assessed for both extreme and everyday objectification predates the remaining variables in the model, and the order of the variables in the indirect paths evaluated is theoretically based, a shortcoming of the current study is that that data is cross-sectional. Additionally self-report measures were exclusively used. Thus, future research should attempt to replicate the results of the current study through the use of a longitudinal design and utilizing alternative methods of gathering data (i.e., structured interviews). Additionally, although there were significant indirect links illuminating the relationship between SV and DE, it is important to note that those effects are small. As was stated previously, there are numerous factors that contribute to the development of DE and the findings of this study represent only one facet. Finally, the use of an undergraduate sample was relevant to the study given the high rates of both SV and DE in college women (Tripp & Petrie, 2001), yet it resulted in a homogenous sample. Lack of diversity with regard to race and age, in particular, limits the generalizability of the results as well. The current study should be replicated in more diverse samples, particularly with sexual minority women, given that this demographic variable was significantly related to DE.
Conclusion
In conclusion, the results of this study largely support the application of objectification theory to the relationship between SV and DE. However, findings also suggest that the role of interoceptive deficits in the relationship between SV and DE may be more complicated than originally theorized by Fredrickson and Roberts (1997) and may warrant modification to the theory. In particular, it appears that there is a link from SV to DE through interoceptive deficits that is particularly salient when extreme victimization (i.e., SV) is assessed. This link should receive further attention in future research that seeks to examine SV through an objectification theory framework.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.