Dangerous Safe Havens: Institutional Betrayal Exacerbates Sexual Trauma


Correspondence concerning this article should be addressed to Carly Parnitzke Smith, Department of Psychology, 1227 University of Oregon, Eugene, OR 97403-1227. E-mail: carlys@uoregon.edu


Research has documented the profound negative impact of betrayal within experiences of interpersonal trauma such as sexual assault (Freyd, 1994, 1996; Freyd, DePrince, & Gleaves, 2007). In the current study of college women (N = 345, 79% Caucasian; mean age = 19.69 years, SD = 2.55), we examined whether institutional failure to prevent sexual assault or respond supportively when it occurs may similarly exacerbate posttraumatic symptomatology—what we call “institutional betrayal.” Almost half (47%) of the women reported at least one coercive sexual experience and another 21% reported no coercion, but at least one unwanted sexual experience (total reporting unwanted sexual experiences, N = 233). Institutional betrayal (e.g., creating an environment where these experiences seemed more likely, making it difficult to report these experiences) was reported across different unwanted sexual experiences (47% and 45% of women reporting coercion and no coercion, respectively). Those women who reported institutional betrayal surrounding their unwanted sexual experience reported increased levels of anxiety (R2 = .10), trauma-specific sexual symptoms (R2 = .17), dissociation (R2 = .11), and problematic sexual functioning (R2 = .12). These results suggest that institutions have the power to cause additional harm to assault survivors.

Traditional and Simplified Chinese Abstracts by AsianSTSS


撮要:以往研究紀錄了諸如性侵犯之類人際創傷的經歷中背叛的震撼負面影響(Freyd,1994;1996; 2009, Freyd, DePrince & Gleaves, 2007)。這個專上學院婦女(N = 345, 79%白種人,平均年齡19.69歲, SD = 2.55)的研究樣本中,「體制背叛」即體制未能防止性侵犯,或提供其後的支援和回應,可加劇創傷後症狀。約半數(47%)婦女報告曾有最少一次強制性經驗,而有21%則有最少一次非自願性經驗但沒有任何強制性經驗(總報告非自願性經驗,N = 233)。在不同非自願性經驗中(47%婦女報告被強制而45%則非強制)都有「體制背叛」(如環境令此類經驗似較易發生,或令人難以舉報此類經驗)。對報告遭遇非自願性經驗及有「體制背叛」的女士而言,她們有較多焦慮(R2 = .10),創傷專属性方面症狀(R2 = .17),解離(R2 = .11)和性功能障礙(R2 = .12)。結果展示體制對創傷生還者能帶來額外的傷害。


撮要:以往研究纪录了诸如性侵犯之类人际创伤的经历中背叛的震撼负面影响(Freyd,1994;1996; 2009, Freyd, DePrince & Gleaves, 2007)。这个专上学院妇女(N = 345, 79%白种人,平均年龄19.69岁, SD = 2.55)的研究样本中,「体制背叛」即体制未能防止性侵犯,或提供后续的支持和回应,可加剧创伤后症状。约半数(47%)妇女报告曾有最少一次强制性经验,而有21%则有最少一次非自愿性经验但没有任何强制性经验(总报告非自愿性经验,N = 233)。在不同非自愿性经验中(47%妇女报告被强迫而45%则非强迫)都有「体制背叛」(如创造一个令此类经验似较容易发生的环境,或令人难以举报此类经验)。对报告遭遇非自愿性经验及有「体制背叛」的女士而言,她们有较多焦虑(R2 = .10),创伤专属性方面症状(R2 = .17),解离(R2 = .11)和性功能障碍(R2 = .12)。结果展示体制对创伤生还者能带来额外的伤害。

Not all traumatic experiences are equal. Although all traumatic events have the potential to leave lasting scars, traumatic events that are interpersonal in nature tend to be most damaging (Briere & Spinazzola, 2005; Cloitre et al., 2009). Interpersonal traumas that involve betrayal of a trusted or depended upon relationship have been repeatedly shown to be uniquely harmful (Goldsmith, Freyd, & DePrince, 2012).

Betrayal trauma theory (Freyd, 1994, 1996; Freyd et al., 2007) explains the unique posttraumatic sequelae of traumatic experiences that involve betrayal as stemming from the maintenance of attachment relationships necessary for survival. The clearest case of this is betrayal blindness—the state of being consciously unaware of interpersonal abuse committed by a trusted or depended upon other (Freyd, 1994, 1997). This blindness may be as extreme as a complete lack of memory of episodes of abuse or more minor such as the tendency to overlook indications of infidelity (Gobin & Freyd, 2009). Although it is typically maladaptive to remain unaware of these transgressions as it increases the likelihood they will reoccur, the need to maintain a necessary relationship takes precedence (Gobin & Freyd, 2009; Goldsmith, Barlow, & Freyd, 2004).

A betrayed individual may maintain little or no conscious awareness of abuse, but the traumatic nature of these events cannot often be erased entirely. The effects of the abuse are instead made known by physical and psychological symptoms that are often perplexing and upsetting. Higher incidences of physical complaints such as irritable bowel syndrome, chronic pelvic pain, and musculoskeletal pain have been documented in individuals with a history of interpersonal abuse (Beck, Elzevier, Pelger, Putter, & Voorham-van der Zalm, 2009; Ross, 2005), including those maintaining little or no memory of documented abuse (Courtois, 1997). The effects of this abuse are seen psychologically as dissociation, anxiety, depression (Briere & Spinazzola, 2005), and interpersonal difficulties that are sometimes extreme enough to fit a diagnosis of borderline personality disorder (Kaehler & Freyd, 2009; Trippany, Helm, & Simpson, 2006). Even if memory of abuse is largely intact, these memories are often of a disjointed and “unshareable” nature, which leads to difficulty seeking satisfying psychological help (van der Kolk & Fisler, 1995).

The research in this area has largely focused on emotional, physical, and sexual abuse committed by one individual against another individual (e.g., incest within a parent–child relationship, domestic violence between romantic partners, sexual harassment or assault between an authority figure and subordinate). Institutional involvement in this type of abuse is often indirect and occurs around individually perpetrated violence (e.g., the recent allegations of sexual abuse occurring at Penn State University, military sexual assault, or clergy sexual abuse). Larger institutions, however, often elicit similar trust and dependency from their members as is found in interpersonal relationships (Baker, McNeil, & Siryk, 1985; Cardador, Dane, & Pratt, 2011; Somers, 2010; Tremblay, 2010). Often, as with trusted interpersonal relationships, these institutional environments are expected to be safe (Platt, Barton, & Freyd, 2009; Tremblay, 2010) and indeed, may be quite literally depended upon for survival as in the military (Suris, Lind, Kashner, & Borman, 2007).

At the very least, abuse experienced within institutional environments seems to carry the same ill effects as interpersonal abuse. For example, physical difficulties following military sexual assault include pelvic pain, gastrointestinal distress, chronic fatigue, and headaches (Suris & Lind, 2008). Sexual and physical abuse experienced while in institutionalized child care is predictive of adult anxiety, depression, borderline personality disorder, and substance abuse (Carr et al., 2010). Yet research suggests that interpersonal abuse experienced in these settings may be more harmful than can be explained by the traumatic events themselves. Even when previous childhood and adult sexual abuse are controlled for, military sexual trauma explains unique variance in physical and psychological difficulties (Kimerling et al., 2010; Luterek, Bittinger, & Simpson, 2011; Suris & Lind, 2008; Suris et al., 2007). This is true even when sexual assault experienced while in the military is not described as the subjectively worst lifetime trauma experienced (Luterek et al., 2011). A similar pattern is seen in the outcomes of abuse experienced in institutionalized childcare: Even after controlling for other childhood sexual and physical abuse, abuse experienced in an institutional setting was predictive of adult psychological distress (Carr et al., 2010).

We posit that these more severe outcomes associated with trauma experienced in institutional settings can be explained by betrayal trauma theory. For example, betrayal trauma theory would predict that sexual assault occurring in a context where one's safety is dependent upon an institution (e.g., the military) would be associated with more difficulties as one continues to try and function in that environment (e.g., continuing to serve in the military). In this way, we predict that sexual assault occurring in a context where an important institution acts in a way that betrays its member's trust will be especially damaging—what we call “institutional betrayal.”

The current study had two main objectives. First, a measure of institutional betrayal was introduced and used to assess the occurrence of violations of members’ trust surrounding incidents of sexual assault. Second, we tested the hypothesis that institutional betrayal interacts with experiences of sexual assault, leading to increased posttraumatic symptomatology as compared to experiencing sexual assault without institutional betrayal. This is the proposed mechanism by which institutional betrayal would exacerbate experiences of sexual trauma.


Participants and Procedure

A sample of undergraduate students at a large, public northwestern university were recruited via the human subject research pool to complete an online measure consisting of several self-report surveys commonly known as the “general survey” for research participation credit. Participants did not self-select into this study based on knowledge of the content, rather the sign-up was based on schedule availability. The university's Office of Research Compliance approved the study and participants provided informed consent electronically by agreeing to participate. The sample completing the measures associated with the current study consisted of 514 students, mostly female (68%), Caucasian (80%), and college-aged (M = 19.82, SD = 2.92), which reflects the demographics of the human subject pool at this university. The focus of the current study was on the experiences of female participants (N = 345), so only data from females were included in analyses. The female sample was similar to the complete sample demographically (79% Caucasian; mean age = 19.69 years, SD = 2.55).


Sexual experiences

The Sexual Experiences Scale (SES; Koss & Oros, 1982) is a 12-item scale that assesses the occurrence (yes/no responses) of a range of unwanted sexual experiences such as, “Have you ever been in a situation where someone used some degree of physical force (twisting your arm, holding you down, etc.) to try to make you engage in kissing or petting when you didn't want to?” Scores on the SES can be used as a measure of total unwanted sexual experiences (scores can range from 0 to 12, with higher scores representing more unwanted sexual experiences) or broken down into three types of coercive experiences: verbally coercive (scores ranging from 0 to 3), physically coercive (scores ranging from 0 to 6), or instances where alcohol or drugs were used to facilitate unwanted sexual contact (scores can be zero or one). Women who had only experienced noncoercive sexual assault (e.g., unwanted sexual experiences where no resistance was reported as a sexual partner becoming so aroused that it seemed useless to stop them even if intercourse was not desired by the woman) are represented in the noncoercive group (scores ranging from 0 to 2). In the current study, the SES proved a reliable measure of total unwanted sexual experiences (α = .78), verbal coercion (α = .66), and physical coercion (α = .79). The SES has been demonstrated to be a valid measure of unwanted sexual experiences in a sample of college women (Franklin, 2010).

Institutional betrayal

The Institutional Betrayal Questionnaire (IBQ), created by Smith and Freyd for this study, is a 10-item questionnaire that assesses institutional betrayal and involvement (Smith & Freyd, 2011). The questionnaire introduces the concept of institutional involvement in unwanted sexual experiences in the following way: “This section will ask you to think about larger institutions to which you belong or have belonged, which may or may not call to mind specific individuals. This may include large systems such as a university, the military, the Greek System (i.e., the Fraternity/Sorority System as a whole), or organized religion. Additionally, this can refer to parts of these systems such as a campus dormitory, a military unit, a specific fraternity or sorority, or a particular church.” Participants are then provided a list of seven potential experiences (e.g., creating an environment in which sexual assaults seem like no big deal, covering up experiences; the full measure is available by request from authors) and provided the following instructions: “In thinking about the events described in the previous section, did an institution play a role by….” In the current study, responses options were a dichotomous Yes or No, which were used to compute a summed institutional betrayal score that ranged from 0 to 7. Women in the current study endorsed all forms of institutional betrayal and a principle components factor analysis did not indicate multidimensionality (eigenvalue for one component = 1.96, with 28.03% of variance explained). This questionnaire was designed to quantify the number of ways in which an individual might experience institutional betrayal and these different experiences, while perhaps related, are not necessarily expected to typically occur in concert. The IBQ is a tool for gathering a participant's experience rather than a scale meant to measure an underlying trait or state. Therefore, this questionnaire is more like a census or checklist. This type of scoring is consistent with other checklists of traumatic life experiences such as the Life Events Checklist (Gray, Litz, Hsu, & Lombardo, 2004). Additionally, the questionnaire includes asks about current involvement with the institution (scored on a Yes or No scale of current involvement) and the type of institution involved.

Trauma symptoms

The Trauma Symptom Checklist (TSC) is a 40-item scale that assesses the presence of six categories of trauma-related difficulties: anxiety, depression, dissociation, sexual abuse trauma index, sexual problems, and sleep disturbance (Elliot & Briere, 1992). Reliability of subscales typically ranges from α = .62 to .77 (Elliott & Briere, 1992); in the current study they ranged from α = .72 to .80. In keeping with typical scoring, each category was measured with several items that were averaged together to create a scale score. Higher scores for each category indicated more severe problems in that area.

Data Analysis

The association between sexual assault, institutional betrayal, and posttraumatic symptoms was examined using multiple linear regression. No participants had missing data for these variables. Prior to completing these analyses, the data were examined for normality to determine whether multiple regression would be appropriate. Scores on the IBQ were positively skewed (Skew = 2.37, SE = 0.16), indicating that lower levels of betrayal were reported most frequently with few women reporting higher levels of betrayal. To address this skewness, IBQ scores were transformed using a square root transformation that made this distribution more normal (Skew = 0.96, SE = 0.16). This transformation did not result in different outcomes in any of the multiple regression analyses when compared to those using the untransformed variable (i.e., the same relationships and interactions were significant). Therefore, we elected to report the untransformed results as these units are more readily interpretable (i.e., for every one point above the mean a woman is on a given measure, we would expect her posttraumatic outcome to increase by the unstandardized beta coefficient).

Unwanted sexual experiences as measured by the SES and posttraumatic symptoms were all found to be relatively normally distributed (Skew = 0.66 to 1.61, SE = 0.16), which is within the range acceptable skewness (i.e., < 2.00), particularly for a variable that is essentially a count of experiences which tend to be somewhat positively skewed (Tabachnick & Fidell, 2007).

In examining six separate multiple regression analyses (i.e., one for each of the six subscales on the Trauma Symptom Checklist), concerns about power and potential Type II errors (i.e., failing to detect significant associations among variables where they do exist) arose. Post hoc tests of power of five of the models with three independent variables, a sample size of 233, α = .05, and R2 ranging from .08 to .17 led to power estimates of .97 to .99, which are adequately powered to detect effects. The only model that appeared underpowered was the one predicting sleep problems, with an R2 of .02; it had a power estimate of .43, which is lower than recommended (Cohen, 1992).

The current study aims to understand the range of experiences of women reporting any unwanted sexual experiences, which includes 68% of the sample (N = 233). The associations between the six subscales on the Trauma Symptom Checklist (TSC) and unwanted sexual experiences and institutional betrayal were examined simultaneously using multiple linear regression. First, SES and IBQ scores were entered as individual variables (both centered around their means to increase interpretability; Tabachnick & Fidell, 2007). These two variables were found to be only slightly correlated with each other, r(231) = .13, p = .017, indicating that multicollinearity between independent variables would not be a problem. Second, the interaction between these centered variables (SES × IBQ) was included to examine the extent to which the effect of SES on each of the TSC symptoms depended upon institutional betrayal as measured by the IBQ.


The majority of women (68%, N = 233) reported at least one unwanted sexual experience and many reported several such experiences (M = 3.06, SD = 2.17). This rate is similar to that found in other samples of college-aged women reporting unwanted sexual experiences using the SES (Franklin, 2010). Responses on the SES can be categorized as coercive or noncoercive experiences (Koss & Oros, 1982). In keeping with the original scoring of the SES, women who reported any coercive sexual assault were categorized into the coercive group, even if they had also experienced noncoercive assault. Across the sample, the most frequently reported types of unwanted sexual experience were those that involved verbal coercion (35% of women), such as threatening to end the relationship if the woman did not have sexual intercourse, followed by alcohol facilitated coercion (19% of women) and physical coercion (18% of women). As many women in the sample reported multiple types of coercion, these figures do not sum to 68% (the total percentage of women experiencing any unwanted sexual experience). Among women reporting unwanted sexual experiences, 70% experienced some form of coercion and over half (52%) reported verbally coercive experiences.

Of the women who reported any unwanted sexual experiences, many (46%) also reported experiencing at least one form of institutional betrayal (see Table 1). Experiences of institutional betrayal did not significantly differ across types of unwanted sexual experience (e.g., noncoercive, verbally, physically, and alcohol/drug coerced), χ2(3, N = 233 = 0.83, ns). This offers support of institutional betrayal as an independent construct rather than an indicator of the subjective severity of a traumatic sexual experience. Institutional betrayal was positively correlated with institutional closeness, r(231) = .18, p = .023, indicating that women did report identifying with the institutions prior to experiencing an unwanted sexual experience. The most frequently reported forms of institutional betrayal involved institutions that had created environments where unwanted sexual experiences seemed common (21%) and where unwanted sexual experiences seemed more likely to occur (17%). Of the institutions identified as betrayers, the most common was a university or related institution such as a residence hall (56%) followed by a fraternity or sorority (19%). Many women reported still being associated with the institution they described (46%), which is reasonable given that this was a university sample and most participants were naming a university as the source of institutional betrayal. Interestingly, women who reported still being associated with the institution did not differ in the amount of betrayal they reported (M = 0.98, SD = 1.25) from women who reported no longer being involved with the institution (M = 1.08, SD = 0.89), t(201) = 0.55, ns.

Table 1. Percentage Endorsing Each Item of the Institutional Betrayal Questionnaire
  1. Note. N = 233. Items have been abbreviated.

Experience/s seemed like no big deal?21.550
Experience seemed more likely16.739
Not taking proactive steps10.725
Making it difficult to report the experience/s?9.021
Covering up the experience/s?8.620
Responding inadequately3.99
Punishing you in some way2.15

In these analyses, participants’ total SES and IBQ scores were used as continuous measure of unwanted sexual experiences and institutional betrayal, respectively. Perhaps unsurprisingly, unwanted sexual experiences accounted for a large amount of variance in each of the posttraumatic symptoms given the strong zero-order correlations between these two measures (see Table 2).

Table 2. Correlations of Trauma Symptom Checklist Scales With Institutional Betrayal and Sexual Trauma
  1. Note. N = 233. SATI = sexual assault trauma index; IBQ = institutional betrayal questionnaire; SES = sexual experiences scale.

  2. *p > .05. ***p > .001.

1. Dissociation0.670.52      
2. Anxiety0.570.40.70***     
3. Depression0.710.42.63***.70***    
4. SATI0.530.48.85***.71***.67***   
5. Sleep problems1.210.58.50***.51***.68***.55***  
6. Sexual problems0.500.45.56***.61***.68***.74***.38*** 
7. IBQ0.721.05.14*.13*.03.16*.05.12
8. SES3.062.17.32***.28***.29***.38***.15*.33***.17***

Although institutional betrayal was not a unique predictor in any of the models, significant interactions between sexual assault and institutional betrayal were observed for four TSC subscales: Sexual Abuse Trauma Index (SATI), Anxiety, Sexual Dysfunction, and Dissociation (see Table 3). Women who experienced institutional betrayal had more severe posttraumatic symptoms in these four areas following a sexual assault.

Table 3. Sexual Assault and Institutional Betrayal Predicting Posttraumatic Symptoms
 SATIAnxietySexual dysfunctionDissociationDepressionSleep Problems
  1. Note. N = 233. Each regression model is described by total variance in posttraumatic symptom explained (R2). SATI = sexual assault trauma index; SES = Sexual Experiences Scale; IBQ = Institutional Betrayal Questionnaire.

  2. *p < .05.

SES × IBQ0.17*0.07.340.16*
R2 Model0.17  0.10  0.12  0.11  0.08  0.02  


Consistent with betrayal trauma theory, sexually assaulted women who also experienced institutional betrayal experienced higher levels of several posttraumatic symptoms. This pattern of results may offer an explanation for the increased difficulties observed following abuse experienced in institutional settings such as the military (Luterek et al., 2011; Suris & Lind, 2008; Suris et al., 2007), institutionalized childcare (Carr et al., 2010), and cases of domestic violence involving failed attempts to seek help from the justice system (Platt, Barton, & Freyd, 2009). It appears that the added betrayal surrounding sexual assault exacerbates what is already a traumatic experience for most women.

This exacerbative effect of institutional betrayal is particularly interesting as the betrayal necessarily occurred apart from the sexual assault itself—either in events leading up to the sexual assault or in those following it. Institutions such as universities expend much public effort to respond to sexual assault and support survivors (e.g., U.S. Department of Defense, 2006). These results suggest that it is necessary to widen the scope of institutional action to prevent and respond to sexual assault. Many current programs and research focus on responses to reported assaults (Campbell, 2008; Campbell & Raja, 2005; Martin & Powell, 1994), but our results suggest that attention should also be focused on the events leading up to sexual assault. In fact, betrayals occurring in events leading up to sexual assault such as creating an environment that is conducive to sexual assault were more commonly reported than insufficient responses to assault in this sample. This perception may be more damaging to members because it creates a sense that the institution could have done something to prevent the experience from occurring. Given the sample of university women reporting their own experiences on a campus, the results of this study provide direction for reform around campus sexual assault policies.

This study introduced a novel measure to capture betrayal by institutions, the IBQ. Future research should focus on developing this questionnaire and applying it in new settings. Potential developments could include additional items to better capture the relationship between the individual and the institution and more nuanced information about reporting and its outcomes. Comparatively few participants reported institutions outside of university life and this suggests our understanding of institutional betrayal would benefit from studying more diverse populations. For example, the military is arguably one of the more recognizable institutions in the United States yet it was not named by any participant as a source of institutional betrayal. This is of particular note as women in the current study tended to report relatively low overall levels of identification with the institutions they described. Institutional betrayal may act as more than an exacerbative force in populations such as the military where identification with the institution is much deeper. Research currently under way is examining many of these same questions in a sample with a military background with the goal of understanding institutional betrayal surrounding military sexual trauma.

There were at least two important limitations to the current study: a narrow focus on sexual assault as the sole traumatic experience measured and including only women in the sample. These two limitations are intertwined in their implications and potential solutions. Although sexual assault is a common and harmful traumatic experience, focusing on it exclusively may limit our potential understanding of institutional betrayal. Although sexual assault does not solely affect women, it does affect them at a disproportionate rate (Herman, 1992). By casting a wider net for measuring traumatic events, both men and women could be more readily studied. Sexual harassment would be a logical extension and research suggests that gender differences could be examined (Shipherd, Pineles, Gradus, & Resick, 2009). Another direction that lends itself readily to the application of institutional betrayal is understanding the role of schools in events leading up to and following bullying. Experiencing bullying is associated with posttraumatic difficulties at rates similar to interpersonal violence and is common across genders (Carney, 2008). Both of these areas have the added benefit of more readily incorporating men into the ideal sample as well as understanding institutional betrayal across more diverse settings.