Sexual Violence and Major Depression among Women: Evidence for Reciprocal Relationships

Abstract

Using two waves of the National Comorbidity Survey, this study examines the relationship between sexual violence and major depression among women, focusing on rape. It uncovers evidence for two-way relationships that, among other things, shed light on who is at risk of sexual assault. Although rape increases the risk of developing major depression in a prospective fashion, depressed women are also at a higher risk of being raped. These pathways inflate the unadjusted prospective relationship between rape and major depression, though they do not explain it altogether. Comparisons between disorders indicate that major depression plays as significant a role as alcohol abuse disorder, and that depressed women are not at a higher risk only because of associated alcohol consumption. This study encourages further research not only on the effects of sexual violence on depression but also the role of common psychiatric disorders in the stress generation process, including how perpetrators target victims.

Keywords

sexual violence depression sex and gender

Sociologists have devoted very little attention to sexual violence, and research on the social determinants of mental health is no exception (Armstrong, Gleckman-Krut, and Johnson 2018). Sexual violence is common among women and may contribute to their especially high risk for major depression (Jordan, Campbell, and Follingstad 2010; National Research Council 2014). Other characteristics of sexual violence point to a potentially powerful role. For one, the vast majority of sexual violence is perpetrated against women by men, indicating a highly unequal risk (Mellins et al. 2017). Furthermore, sexual violence is especially common at younger ages, overlapping with the early onset of many cases of major depression (Breiding 2014; Tjaden and Thoennes 1998).

Despite the potentially powerful role of sexual violence, research on its effects on major depression is incomplete. Existing research has suffered from several limitations. To be sure, research has generally revealed that sexual violence increases the risk of major depression and depressive symptoms (Campbell, Dworkin, and Cabral 2009; Herman 1992), as well as other forms of psychopathology (Dworkin et al. 2017). Yet, the relationship between sexual violence and depression in particular is more complex than it might seem. For one, the relationship between sexual violence and major depression is often weaker than for other psychiatric disorders, including posttraumatic stress disorder (PTSD; Briere and Jordan 2004; Campbell et al. 2009; Clum, Calhoun, and Kimerling 2000; Dworkin et al. 2017; Kilpatrick et al. 2003; Kilpatrick et al. 1987). Furthermore, the magnitude of the association between sexual violence and major depression varies considerably from study to study (Campbell et al. 2009; Jordan et al. 2010; Koss and Figueredo 2004). Moreover, research has generally neglected to consider the role psychiatric disorders themselves in increasing the risk of sexual violence. This omission is perhaps surprising given that research has convincingly demonstrated that those who suffer from psychiatric disorders are at a much higher risk of being the victim of a crime. In general, those who suffer from severe mental illness are victims of violent crime at a rate more than 11 times higher than the general population and are victims of rape specifically at a rate more than 22 times higher (Teplin et al. 2005). In addition, alcohol use is implicated in a large amount of sexual violence, especially on college campuses, and alcohol abuse is strongly comorbid with major depression among women (Grant and Harford 1995; Kendler et al. 1993). A separate literature points to the potentially important role of depression itself in stress generation, as well as the role of depression in how perpetrators target victims (see Wolpert 2001 on the stigma of depression).

This study explores the effects of sexual violence on major depression using two waves of nationally representative data, spaced a decade apart. Among the most important features of the dataset is its ability to address timing. It permits an analysis of the relationship between sexual violence and depression allowing for the possibility that sexual violence occurs in part as a consequence of depression. It also allows for the assessment of the short- and long-term impact of sexual violence on major depression.

Background

Given its frequency and significance, sexual violence is a potentially powerful determinant of major depression. Although sexual violence may be decreasing over time, it is not uncommon (Wheeler 2015). Nearly one in five (19.3 percent) women report having been raped during their lifetimes and many more (43.9 percent) report other forms of sexual violence, including unwanted sexual contact and sexual coercion (Breiding 2014; Tjaden and Thoennes 1998). Sexual violence is associated with a wide variety of negative outcomes, including ongoing difficulty in personal relationships and diminished physical health (Cherlin et al. 2004; Golding 1994), though considerable evidence has focused on mental health specifically (see Dworkin et al. 2017 for a recent meta-analysis).

Sexual Violence and Major Depression

Most evidence indicates that sexual violence adversely affects emotional well-being. Although research on sexual violence has explored a variety of psychiatric disorders, major depression remains among the most frequently discussed psychiatric outcomes, in large part because it is common (Logan et al. 2006). Of women who have experienced sexual violence, between 13 and 51 percent meet the diagnostic criteria for major depression (Campbell et al. 2009; Jordan et al. 2010). The symptoms of depression are especially common shortly after the assault and tend to recede with time, though some evidence points to effects that linger for up to two years, a long lag relative to other stressful events (Koss and Figueredo 2004). There are a variety of explanations for this association, but the most prominent pertains to sexual violence as a form of trauma with especially broad consequences. In general, traumatic events increase stress and social isolation, both of which are in turn related to depression (Koss and Harvey 1991). The association between sexual violence and major depression may be even stronger than other traumas, though, in the sense that sexual violence can be isolating. Women who have experienced sexual violence often report declines in the quality of their other social relationships, in part because women who report such experiences are often discredited, shunned, or met with skepticism, experiences that can elevate the harm of sexual violence relative to other traumatic experiences (Campbell et al. 2009).

Despite compelling evidence and theory linking sexual violence to depression, the range of estimates in previous empirical studies is large and, indeed, sexual violence is occasionally not significantly associated with major depression on its own (e.g., Kilpatrick et al. 2003). Part of this inconsistency reflects the wide variety of research designs employed in prior studies, as well as differences in the types of sexual violence studies have focused on, including differences in perpetrator and context (see Dworkin et al. 2017 on methodological heterogeneity). Studies of rape employing nationally representative samples will often estimate the effects of acquaintance rape, whereas other studies are often designed to focus on rape by a family member (Fedina, Holmes, and Backes 2018). Another limitation in prior research is inconsistency in the use of control variables. One recent meta-analysis of the relationship between intimate partner violence and depression noted that few studies control for confounding variables other the basic demographic variables, casting doubt on the magnitude of prior estimates if not their overall significance (Devries et al. 2013). Depression and sexual violence share at least some common risk factors, especially in the context of early-life experiences. For instance, sexual violence is more common among those growing up in a non-traditional household (Freeman and Temple 2010). In addition, parental mental illness has been related both to depression in offspring and to the risk of abuse (Walsh, MacMillan, and Jamieson 2002; Weissman et al. 2006).

Major Depression and the Risk of Sexual Violence

Perhaps an even more significant limitation in research on sexual violence and depression is the failure to account for alternative pathways between the two, something that has been explored with respect to PTSD and trauma exposure but has yet to be applied in the same fashion to major depression and sexual violence specifically (Orcutt, Erickson, and Wolfe 2002). Given its traumatic nature, sexual violence almost certainly plays some role in the development of major depression, leading to its initial onset and perhaps also its recurrence years later. Nonetheless, there is also theory and some evidence to suggest that depression increases the risk of sexual victimization and revictimization. Some of this comes from general evidence that psychiatric disorders increase the risk of criminal victimization, including sexual violence (Khalifeh and Dean 2010; Littleton and Ullman 2013; McCart et al. 2012). Part also, however, comes from research on the role of depression in stress generation. Prior research and theory suggests that depression could increase the risk of sexual violence in two ways, by changing the behavior and cognition of the person suffering from depression and/or by changing how others respond to her.

Regarding the former, stress generation theory posits that depression itself plays a uniquely powerful role in the subsequent generation of stress (Conway, Hammen, and Brennan 2012; Hammen 1991). Depressed people experience more stressful events than those who are not depressed, especially in the interpersonal domain (Hammen 1991). In particular, evidence indicates that depressed women report significantly more conflict in their relationships, including disagreements with partners, family members, and friends. Stress generation theory has focused on two general explanations for understanding this process (Hammen 2006). The first pertains to the personal characteristics of the person with depression, including their cognitions, beliefs, and expectations. The second pertains to the context of their lives as reflected in their decisions and actions, including mate selection. In general, depression acts to burden individuals with environmental contexts and cognitions that are stressful, as when a depressed person enters a relationship with a partner who is abusive. Depressed women, for instance, are more likely to marry men with psychiatric disorders (Hammen 1999). Although much of the evidence regarding stress generation theory has focused on interpersonal disagreements and conflict, some evidence goes further and links symptoms to interpersonal violence. Some evidence, for instance, links emotional numbing to the subsequent risk of re-abuse among women already exposed to interpersonal violence (Krause et al. 2006). By the same token, some evidence links hyperarousal to later sexual revictimization, based on the idea that the symptom prevents the individual from responding to situation-specific danger cues (Risser et al. 2006). Other studies link depression and distress to subsequent sexual violence. A study using two waves of the National Longitudinal Study of Adolescent Health found that, among women, depressive symptoms at Wave 1 increased the odds of subsequent partner violence by 86 percent (Lehrer et al. 2006). The same study found that even low-grade depressive symptoms, short of a depressive disorder, increased the risk. Another study found that women with depression were more likely to become involved in an abusive relationship (Ehrensaft, Moffitt, and Caspi 2006). Psychological distress also provides a link between childhood sexual abuse and adult revictimization (Cuevas et al. 2010; Miron and Orcutt 2014).

The second pathway is also consistent with the general idea that psychological distress generates additional stress but focuses less to how the person suffering from depression evaluates her environment and more to what depression signals to others. Stress generation theory is transactive in nature, though mutual influences, including how those suffering from depression are perceived by others, have generally not been a strong empirical focus (Hammen 2006). Other literatures point to the potential significance of such influences, especially in the context of sexual violence. Feminists have argued that rape is a process of intimidation through which men keep women in a state of fear (Brownmiller 1976). In this way, rape is fundamentally about one person’s domination and power over another (Armstrong et al. 2018). Although the connection is usually not made explicit in the psychiatric literature, depression could play a role in shaping the risk of sexual violence by conveying weakness. The stigma surrounding depression centers around depression as a sign of weakness and vulnerability and depressed people ordinarily experience considerable social isolation because they are perceived as weak (Schnittker 2000).

Despite suggestive evidence and theory of this sort, it is unclear whether depression alone is a critical factor in sexual violence. For one, major depression is often accompanied by alcohol abuse. Growing evidence indicates that the stress generation process is not limited to major depression and therefore that it should be expanded to include disorders with which depression is associated (Connolly et al. 2010). Absent such evidence research risks a misplaced specificity on depression. Among those who experience major depression, about 40 percent also report alcohol abuse or dependence compared with only 16 percent among those who do not experience major depression (Grant and Harford 1995, Table 3). The relationship between alcohol use and sexual violence is well known. Alcohol use by the perpetrator, victim, or both is involved in approximately 50 percent of cases of sexual assault among women (Abbey et al. 2001; Collins and Messerschmidt 1993). Estimates of alcohol use by the victim range from 30 to 79 percent (Abbey, Ross, and McDuffie 1994; Abbey et al. 2001; National Research Council 1996). If alcohol abuse follows depression, it could account for the increased risk of sexual violence.

The analysis that follows explores these issues in detail. It explores reciprocal relationships between major depression and sexual violence using longitudinal data with a wide variety of control variables, including the presence of alcohol abuse disorder. Figure 1 presents the basic structural equation model informing the analysis. Research has focused on Pathway 3 (from sexual violence to major depression), though Pathway 1 (from early onset major depression to subsequent sexual violence) is potentially significant and may inflate the magnitude of Pathway 3 if it is assumed to be zero. The data used in this study allow for the estimation of such a model.

Figure 1.

Path diagram of relationships between sexual violence and major depression.

Data and Method

Data for this study are drawn from two waves of the longitudinal component of the National Comorbidity Survey, a family of studies focused on assessing the prevalence and correlates of psychiatric disorders in the United States (Kessler 2008, 2015). The baseline longitudinal survey, referred to as NCS 1, was conducted between 1990 and 1992 in a sample of 8,098 respondents. The second wave, NCS 2, was a re-interview of 5,001 Wave 1 respondents, conducted approximately a decade later, between 2001 and 2002. NCS 1 employed a stratified multistage area probability sample, targeting the non-institutionalized civilian population age 15 to 54 years living in the continuous United States. The analysis presented here is limited to women. In addition, the analysis is limited to the subset of respondents who completed a risk assessment interview in the first wave, which included many of the key covariates necessary for this study. After eliminating a small amount of missing data on the key covariates (less than 1 percent), the final sample size for NCS 1 is 3,048. The sample that also completed the Wave 2 interview is 2,657. All the descriptive statistics and structural equations are based on weighted data.

The primary purpose of the NCS 1 and 2 was to collect longitudinal data on the prevalence and onset of common psychiatric disorders. NCS 1 evaluates the presence of Diagnostic and Statistical Manual of Mental Disorders (3rd ed., rev.; DSM-III-R; American Psychiatric Association 1987) psychiatric disorders using a modified version of the World Health Organization Composite International Diagnostic Interview (CIDI), developed at the University of Michigan (Kessler and Üstün 2004; Robins et al. 1988). The CIDI is a fully structured lay-administered diagnostic interview, evaluating whether respondents meet the diagnostic criteria for specific DSM disorders. The instrument provides lay diagnoses: respondents meeting the diagnostic criteria for a disorder need not have been diagnosed with the disorder by a professional. Like NCS 1, NCS 2 also assessed the presence of psychiatric disorders using a version of the CIDI, though it drew on diagnostic criteria from Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; American Psychiatric Association 1994) rather than DSM-III-R, as per the publication of a new manual in the intervening years.

The current study focuses on major depression, the most common mood disorder and one of the most common psychiatric disorders in the United States (Kessler, Chiu et al. 2005). When asked about episodes of major depression, respondents were asked when the first episode occurred, as well as information on whether they are currently (or recently) suffering from an episode. The analysis focuses on 12-month major depression, referring to an episode of major depression occurring in the preceding 12 months (often referred to as “current” major depression). Major depression was assessed in both waves, and the structural equation model allows for Wave 1 major depression causing Wave 2 major depression.

The models also include early onset major depression and alcohol abuse as determinants of Wave 1 depression, as well as a determinant of subsequent sexual violence. These variables are defined as the onset of the specific disorder prior to the age of 15 unless the disorder occurred subsequent to sexual violence. This age cut-off was used in light of the average age of first sexual violence, which, in the NCS, was just above 16 years. Alcohol abuse is defined as the presence of the disorder with or without dependence.

Sexual Violence and Other Variables

In a series of questions included in Wave 1, respondents were presented with a list of traumatic life events they may have experienced. If they answered affirmatively to at least one of the events, they were then asked specifically which one(s) from the list, followed by additional questions on those experiences. Included on the list were “you were raped.” Respondents were then asked a series of questions about the circumstances surrounding the experience, including the age at which it first occurred. This information is used to insure proper sequencing in the structural equation model.

The NCS also provides a wide assortment of control variables. The models used here control for basic demographic characteristics, including age (at the time of the first interview), race/ethnicity (non-Hispanic white, black, Hispanic, other), and education (0–11 years, 12 years, 13–15 years, and 16 or more years). The models control for some important current environmental conditions related to depression, including employment status and whether the respondent is married or not. The models also control for several family background factors implicated in both major depression and sexual violence. The models control for whether either the mother or father likely suffered from major depression, an anxiety disorder, or a substance abuse disorder (based on a series of questions regarding the presence of specific symptoms associated with those disorders). The model also include controls for whether the respondent lived with both parents up to the age of 15 and the level of education of the primary breadwinner in the household, an indicator of socioeconomic resources.

The models presented below are based on generalized structural equations, allowing for dichotomous outcomes (Skrondal and Rabe-Hesketh 2004). The models are based on a logit link function and are estimated using maximum likelihood. Structural equation models are especially well-suited to this study because they allow for the simultaneous estimation of a series of equations and allow for the transparent comparison of estimates when one or another pathway is assumed to be zero. In Table 2 a baseline structural equation model is presented wherein there are no Pathways 1 and 2, as depicted in Figure 1. In Table 2, these two pathways are added, first adjusting for early onset major depression and next including early onset alcohol abuse.

Results

The results begin with basic descriptive statistics, presented in Table 1 and Figure 2. Approximately 8 percent of the women report having been raped. The prevalence of major depression in Wave 1 was 13.1 percent and the prevalence in Wave 2 was 10.7 percent, though among those who have been raped, the prevalence is 26.9 and 18.2 percent, respectively. Among respondents who reported major depression in Wave 1, 19.5 percent report major depression in Wave 2. Early onset major depression was apparent in just under 4 percent of the sample.

Table 1.

Summary Statistics, National Comorbidity Survey I and II (N = 3,048 [2,657]).

Variables	Mean
Rape	0.080
Psychiatric disorders
Major depression (Wave 1)	0.131
Major depression (Wave 2)	0.107
Early onset major depression	0.038
Early onset alcohol abuse	0.010
Control variables
Age	33.397
Employed	0.718
Married	0.617
Race/ethnicity (ref. non-Hispanic white)
Black	0.131
Hispanic	0.090
Other	0.034
Education (ref. less than 12)
12 years of schooling	0.360
13–15 years of schooling	0.252
16 years or more of schooling	0.185
Family background characteristics
Lived with both Parents	0.710
Parental years of schooling	11.368
Any parental psychiatric disorder	0.531

Figure 2.

Kernel density function of sexual violence and major depression over age, national comorbidity survey.

In order to illustrate the potential overlap between sexual violence and major depression over the life course, Figure 2 presents information on the age of first occurrence for rape and the age of onset for major depression. The figure presents kernel density functions for both, corresponding to the probability density of both experiences over age. Several things are notable and relevant to the issue of timing. For one, the density function for major depression is more flat, though its peak is approximately the same as the peak for the density function for rape: both peaks in the late teens, pointing to the frequency with which depression precedes rape and, therefore, highlighting the potential threat of reverse causality.

Tables 2 and 3 explore this issue directly. Table 2 presents two structural equation models. The outcomes are modeled simultaneously, so Model 1 refers to the same structural equation model encompassing both major depression in Wave 1 and major depression in Wave 2. The first model presents bivariate relationships between rape and major depression. The second model adds basic controls. In both models, rape is significantly related to major depression in both waves. Rape is associated with the onset of major depression in Wave 2 and the recurrence of major depression in Wave 2, controlling for Wave 1 major depression. The structural equation models presented in Table 2, however, do not include Pathways 1 and 2, as depicted in Figure 1. The models assume there is no relationship from earlier psychiatric disorders to subsequent sexual violence. Although the relationship between rape reported in Wave 1 and major depression reported in Wave 2 would not be sensitive to the inclusion of such pathways given that sexual violence is measured in Wave 1, it is notable that the magnitude of the Wave 2 relationship is somewhat weaker than that of Wave 1, especially in the unadjusted model.

Table 2.

Generalized Structural Equation Models of Major Depression and Sexual Violence, National Comorbidity Survey I and II.

	Major depression Wave 1		Major depression Wave 2
Independent variables	Model 1	Model 2	Model 1	Model 2
Rape	1.009***	0.762***	0.681**	0.559*
	(0.159)	(0.133)	(0.230)	(0.236)
Major depression Wave 1				0.738***
				(0.186)
Employed		−0.421*		0.129
		(0.181)		(0.223)
Married		−0.551**		−0.016
		(0.163)		(0.201)
Age		0.003		−0.0015
		(0.006)		(0.008)
Black		−0.207		−0.128
		(0.216)		(0.223)
Hispanic		0.206		−0.227
		(0.149)		(0.268)
Other		−0.379		−0.306
		(0.328)		(0.451)
12 years of schooling		0.106		−0.342
		(0.213)		(0.235)
13–15 years of schooling		0.008		−0.351
		(0.230)		(0.209)
16 years or more of schooling		−0.478		−0.800**
		(0.284)		(0.291)
Family characteristics
Parental schooling		0.002		−0.006
		(0.023)		(0.028)
Any parental disorder		1.336***
		(0.118)
Constant	−2.008	−2.329	−2.185	−1.430
N	3,048		2,657

Note. 95% confidence intervals in parentheses. Reference category of race/ethnicity is non-Hispanic white and for years of schooling less than 12.

p < .05. **p < .01. ***p < .001.

Table 3.

Generalized Structural Equation Models of Major Depression and Sexual Violence with Pathways from Major Depression to Sexual Violence, National Comorbidity Survey I and II.

	Major depression Wave 1		Major depression Wave 2		Rape
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
Rape	0.675***	0.635***	0.559*	0.559*
	(0.149)	(0.152)	(0.236)	(0.236)
Major depression Wave 1			0.738***	0.738***
			(0.186)	(0.186)
Employed	−0.457*	−0.436*	0.129	0.129
	(0.189)	(0.190)	(0.223)	(0.223)
Married	−0.479**	−0.468**	−0.016	−0.016
	(0.156)	(0.153)	(0.201)	(0.201)
Age	0.011	0.011	−0.015	−0.015	0.000	0.001
	(0.006)	(0.006)	(0.008)	(0.008)	(0.008)	(0.008)
Black	−0.082	−0.056	−0.128	−0.128	0.278	0.313
	(0.203)	(0.206)	(0.223)	(0.223)	(0.239)	(0.237)
Hispanic	0.376*	0.388*	−0.227	−0.227	−0.552	−0.538
	(0.154)	(0.155)	(0.268)	(0.268)	(0.280)	(0.284)
Other	−0.471	−0.444	−0.306	−0.306	−0.715	−0.677
	(0.470)	(0.469)	(0.451)	(0.451)	(0.433)	(0.428)
12 years of schooling	0.113	0.115	−0.342	−0.342
	(0.231)	(0.229)	(0.235)	(0.235)
13–15 years of schooling	0.018	0.032	−0.351	−0.351
	(0.242)	(0.242)	(0.209)	(0.209)
16 years or more of schooling	−0.384	−0.364	−0.800**	−0.800**
	(0.283)	(0.282)	(0.291)	(0.291)
Family characteristics
Both parents at 15					−1.020***	−0.993***
					(0.198)	(0.205)
Parental schooling	−0.009	−0.011	−0.006	−0.006	−0.007	−0.007
	(0.024)	(0.024)	(0.028)	(0.028)	(0.030)	(0.030)
Any parental disorder	1.149***	1.143***			0.874***	0.867***
	(0.127)	(0.125)			(0.169)	(0.166)
Early psychiatric disorders
Major depression	2.652***	2.629***			0.732*	0.674*
	(0.192)	(0.192)			(0.301)	(0.324)
Alcohol abuse		1.069*				1.124*
		(0.487)				(0.493)
Constant	−2.612	−2.628	−1.430	−1.430	−2.316	−2.376
N	3,048		2,657		3,048

Note. 95% confidence intervals in parentheses. Reference category of race/ethnicity is non-Hispanic white and for years of schooling less than 12.

p < .05. **p < .01. ***p < .001.

Table 3 includes the models that estimate a pathway from psychiatric disorders to sexual violence. Once again two structural equation models are estimated, the first allowing for the impact of early onset major depression on rape (as well as a pathway from early onset major depression to Wave 1 major depression, Pathway 2 from Figure 1) and the second allowing for the impact of early onset alcohol abuse as well. Altogether, these two models, in comparison with the models presented in Table 2, allow for the evaluation of how much the relationship between sexual violence and major depression is elevated by assuming there is no impact of earlier psychiatric disorders on the risk of sexual violence (i.e., assuming that Pathway 1 in Figure 1 is zero), as well as what role major depression plays in the risk of sexual violence once its comorbidity with alcohol abuse is considered.

The model indicates that major depression is related to subsequent sexual violence. Major depression increases the risk of rape and increases the risk of major depression in Wave 1. But including both pathways does little to diminish the impact of rape on major depression. The coefficient drops from 0.762 to 0.635, corresponding to a drop in the odds ratio of 2.143 to 1.887. Both early onset major depression and early onset alcohol abuse increase the odds of rape. The inclusion of alcohol abuse reduces the relationship between early onset major depression and subsequent rape, though it does not explain it. Furthermore, major depression is as important as alcohol abuse: a Wald test revealed no evidence for a statistically significant difference between the two coefficients. Of note, the two variables also rival in relevance other family background factors, including growing up with both parents and the presence of a psychiatric disorder in either parent. By assumption, the long-run relationship between rape and major depression remains unchanged, as the impact of early onset disorders on the risk of rape operates only through Wave 1 outcomes. This assumption is warranted: in a structural equation model including a direct pathway from early onset psychiatric disorders to Wave 2 major depression (and also including a pathway from Wave 1 depression to Wave 2), neither coefficient for an early onset psychiatric disorder was significant.

Discussion

The results of this study reveal two things at once. On one hand, the results reveal that the pathway from sexual violence to major depression may be inflated in studies that do not consider the role of depression in increasing the risk of sexual violence. Although major depression has a later age of onset than many other psychiatric disorders and, therefore, is less subjected to a reverse sequence (Kessler, Berglund, et al. 2005), it is not uncommon for depression to emerge before sexual violence, especially rape. Both happen relatively early. In NCS 1, the average age for the onset of a first episode of major depression is just above 23 years, whereas the average age for a first rape is just above 16 years. In models that include pathways from early-onset depression to subsequent sexual violence, the path from sexual violence to major depression is reduced. Nonetheless, the impact of sexual violence remains large, increasing the odds of subsequent major depression by 75 percent in the long term to 89 percent in the short term (exponentiation of 0.559 and 0.635, respectively, from Model 2 of Table 3). On the other hand, the results also point to the significant role of major depression in sexual victimization, a risk that has generally been neglected, though is theoretically well-motivated, at least with respect to the general case of stress generation. To date, much of the literature on sexual violence and its antecedents has focused on a different set of risk factors. Most notably, alcohol use is frequently implicated in date rape, especially on college campuses, and remains a cornerstone of popular discussions of the circumstances surrounding sexual violence. Yet, the results of this study suggest that, at least from the standpoint of a broader set of psychiatric disorders, major depression is as important a risk factor, consistent with stress generation theory. The results also indicate that the consequences of major depression for sexual violence are not explained by the fact that women suffering from major depression might also suffer from alcohol abuse.

Although the results suggest the effects of rape on major depression may be inflated in models that do not adjust for the sequencing of rape and depression, the results also point to the especially long at-risk period associated with rape, another critical (and often overlooked) dimension of timing. Perhaps the most conservative test of a relationship between sexual violence and major depression is to examine the risk of another episode of major depression in Wave 2. A lag in excess of 10 years is conservative in the sense that it is well outside the window normally examined in studies of rape. Prior studies have shown that the effects of sexual violence on the symptoms of depression can fade over a window even as short as two years (e.g., Frank and Stewart 1984). Such a test is also conservative in the sense that many early life risk factors appear to affect later onset only through their relationship with Wave 1 major depression, such as the presence of a psychiatric disorder in either of one’s parents. In the models, Wave 1 depression is allowed to a have a direct effect on Wave 2 depression. Yet, even employing a conservative empirical design, the present study finds that rape is still associated with the risk of depression years after the event.

The results are consistent with stress generation theory, but they encourage additional research to uncover the social and psychological processes that link major depression to sexual violence. The results point to several general possibilities. One possibility, directly relevant to stress generation theory, is that those who suffer from major depression are more likely to encounter risky situations and/or assess risks inaccurately. Studies of stress generation have produced evidence consistent with this idea (see Liu and Alloy 2010 for a review), though this argument would seem to apply better to alcohol abuse than major depression and, in the current study, both are significantly related to sexual violence. At the very least, major depression is not plainly associated with the same type of risk-taking as alcohol abuse and, in some specific ways, would seem to generate a lower risk for sexual violence. For instance, those who suffer from major depression in adolescence tend to marry sooner than those who do not (Gotlib, Lewinsohn, and Seeley 1998). Furthermore, young people with major depression report lower levels of sexual activity, including lower levels of sexual interest and arousal (Deumic et al. 2016). A closer look at how depression affects the interpersonal environment with respect to sexual violence would be useful. A somewhat different possibility, discussed earlier, is that those who suffer from major depression are subjected to predation and targeted for their apparent vulnerability. The analyses presented here do not prove the case for predation, and predation is not outside the scope of stress generation theory, though future research on the relationship between psychiatric disorders and sexual violence would benefit from considering the perceptions and actions of both the victim and the perpetrator rather than focusing only on those of the victim (see Hammen 2006 for a discussion of the transactive nature of stress generation).

Limitations

This study has several limitations. For one, reports of sexual violence depend on a respondent’s understanding of what rape is. Although it is unlikely that respondents are reporting incidents that did not occur, it is probable that some respondents who were raped failed to report their experiences as such (see National Research Council 2014 on undercounting rape in the National Crime Victimization Survey). The prevalence of rape found in the NCS is somewhat lower than what has been reported in surveys that assess sexual violence in more encompassing ways, though it is not entirely inconsistent with other studies. The National Intimate Partner and Sexual Violence Survey, for instance, found that 19.3 percent of women reported having been raped during her lifetime, but also that approximately 43.9 percent of women reported other forms of sexual violence (Breiding 2014). It is possible that more NCS respondents would have reported sexual violence had they been cued as to the kind of experiences that qualify rather than simply asked to report “rape” in the context of a list of other traumatic life events. The NCS also likely misses some women who have been raped by virtue of focusing on non-institutionalized adults. The NCS does not include homeless women and women in shelters, though both populations report higher rates than the general population (Fisher et al. 1995).

In addition, the current study partly depends on respondents accurately recalling when an episode of depression occurred, something that may be more difficult to date precisely, especially relative to reporting on the timing of a traumatic event. The NCS did not approach this issue casually and, in fact, greatly improved upon previous studies (e.g., Simon and VonKorff 1995). In the NCS, respondents were first asked if they remembered the exact age at which a syndrome emerged. If they could not, they were prompted, using progressive increments across an age range, starting with “Was it before you first started school?” and so on, after which they were assigned the upper bound of the implied range (Kessler, Berglund, et al. 2005). This technique greatly improves the accuracy of age-of-onset reports (Knäuper et al. 2006). The NCS technique is not perfect, however, and there is likely some inaccurate assignment of respondents to the before/after conditions this study relies on. Longitudinal data assessed over a longer window of the life course would help to overcome this problem, though given the early onset of some cases of major depression, as well as the early onset of sexual violence, data collection would need to start early in the life course to eliminate reliance on retrospection entirely.

A related problem pertains to cohort differences in the recall of events. The understanding of sexual violence has shifted over time, as has the recognition of psychiatric disorders (Chasteen 2001). There are significant inter-cohort differences in the prevalence of psychiatric disorders, with a generally higher prevalence among more recent cohorts (Kessler, Berglund, et al. 2005). It is unclear, though, what role cohort differences play in the adequacy of the reports used in this study. For instance, some evidence suggests that age-of-onset reports, when ascertained in a crude fashion, tend to be about 10 years before the interview, regardless of the age of respondents at the time of the interview (Simon and VonKorff 1995). Assuming recollections of sexual violence do not suffer from the same problem, this implies sexual violence will appear to precede the onset of psychiatric disorders more in older cohorts than younger ones. The present study used two waves of data, overcoming this issue at least in part, though distortions in recall are minimized best, again, in closely spaced waves of longitudinal data. Closely spaced waves would allow researchers to discern accurately whether the depression was active when the victimization occurred.

Conclusion

When considering sex differences in major depression, it is important to think not only about sex differences in general stress but also the specific sources of that stress. Sexual violence is common among women and plays an important role in the development of major depression. The present study shows that the pathway from sexual violence to major depression may be overstated in previous studies and also shows that such a pathway persists long after the event occurred. At the same time, the study reveals that major depression increases the risk of sexual violence. Two-way relationships of this sort can bias causal estimates and require thinking seriously about timing and sequence, though they are nonetheless a critical aspect of the relationship between sexual violence and mental health. Exploring these relationships further could shed light both on the determinants of major depression and the risk factors for sexual violence, literatures that are best considered together given their overlapping themes of gender, violence, and vulnerability.

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.

References

Abbey

Antonia

Ross

Lisa Thomson

McDuffie

Donna

. 1994. “Alcohol’s Role in Sexual Assault.” Pp. 97–123 in Addictive Behaviors in Women, edited by Watson

R. R.

Totowa, NJ: Humana Press.

Abbey

Antonia

Zawacki

Tina

Buck

Philip O.

Clinton

Monique

McAuslan

Pam

. 2001. Alcohol and Sexual Assault. National Institute on Alcohol Abuse and Alcoholism. Retrieved May 31, 2019 (https://pubs.niaaa.nih.gov/publications/arh25-1/43-51.htm).

American Psychiatric Association. 1987. Diagnostic and Statistical Manual of Mental Disorders. 3rd ed., Rev. Washington, DC: Author.

American Psychiatric Association. 1994. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: Author.

Armstrong

Elizabeth A.

Gleckman-Krut

Miriam

Johnson

Lanora

. 2018. “Silence, Power, and Inequality: An Intersectional Approach to Sexual Violence.” Annual Review of Sociology 44:99–122.

Breiding

Matthew J.

2014. “Prevalence and Characteristics of Sexual Violence, Stalking, and Intimate Partner Violence Victimization—National Intimate Partner and Sexual Violence Survey, United States, 2011.” Morbidity and Mortality Weekly Report 63:1–18.

Briere

John

Jordan

Carol E.

2004. “Violence against Women: Outcome Complexity and Implications for Assessment and Treatment.” Journal of Interpersonal Violence 19:1252–76.

Brownmiller

Susan

. 1976. Against Our Will: Men, Women and Rape. New York: Bantam Books.

Campbell

Rebecca

Dworkin

Emily

Cabral

Giannina

. 2009. “An Ecological Model of the Impact of Sexual Assault on Women’s Mental Health.” Trauma, Violence, & Abuse 10:225–46.

10.

Chasteen

Amy L.

2001. “Constructing Rape: Feminism, Change, and Women’s Everyday Understandings of Sexual Assault.” Sociological Spectrum 21:101–39.

11.

Cherlin

Andrew J.

Hurt

Tera R.

Burton

Linda M.

Purvin

Diane M.

2004. “The Influence of Physical and Sexual Abuse on Marriage and Cohabitation.” American Sociological Review 69:768–89.

12.

Clum

Gretchen A.

Calhoun

Karen S.

Kimerling

Rachel

. 2000. “Associations among Symptoms of Depression and Posttraumatic Stress Disorder and Self-Reported Health in Sexually Assaulted Women.” The Journal of Nervous and Mental Disease 188:671–78.

13.

Collins

James J.

Messerschmidt

Pamela M.

1993. “Epidemiology of Alcohol-Related Violence.” Alcohol Health and Research World 17:93–100.

14.

Connolly

Nicole P.

Eberhart

Nicole K.

Hammen

Constance L.

Brennan

Patricia A.

2010. “Specificity of Stress Generation: A Comparison of Adolescents with Depressive, Anxiety, and Comorbid Diagnoses.” International Journal of Cognitive Therapy 3:368–79.

15.

Conway

Christopher C.

Hammen

Constance

Brennan

Patricia A.

2012. “Expanding Stress Generation Theory: Test of a Transdiagnostic Model.” Journal of Abnormal Psychology 121:754–66.

16.

Cuevas

Carlos A.

Finkelhor

David

Clifford

Cynthia

Ormrod

Richard K.

Turner

Heather A.

2010. “Psychological Distress as a Risk Factor for Re-Victimization in Children.” Child Abuse & Neglect 34:235–43.

17.

Deumic

Emira

Butcher

Brandon D.

Clayton

Anita D.

Dindo

Lilian N.

Burns

Trudy L.

Calarge

Chadi A.

2016. “Sexual Functioning in Adolescents with Major Depressive Disorder.” The Journal of Clinical Psychiatry 77:957–62.

18.

Devries

Karen M.

Mak

Joelle Y.

Bacchus

Loraine J.

Child

Jennifer C.

Falder

Gail

Max

Petzold

Jill

Astbury

Watts

Charlotte H.

2013. “Intimate Partner Violence and Incident Depressive Symptoms and Suicide Attempts: A Systematic Review of Longitudinal Studies.” PLoS Medicine 10:e1001439.

19.

Dworkin

Emily R.

Menon

Suvarna V.

Bystrynski

Jonathan

Allen

Nicole E.

2017. “Sexual Assault Victimization and Psychopathology: A Review and Meta-Analysis.” Clinical Psychology Review 56:65–81.

20.

Ehrensaft

Miriam K.

Moffitt

Terrie E.

Caspi

Avshalom

. 2006. “Is Domestic Violence Followed by an Increased Risk of Psychiatric Disorders among Women but Not among Men? A Longitudinal Cohort Study.” American Journal of Psychiatry 163:885–92.

21.

Fedina

Lisa

Holmes

Jennifer Lynne

Backes

Bethany L.

2018. “Campus Sexual Assault: A Systematic Review of Prevalence Research from 2000 to 2015.” Trauma, Violence, & Abuse 19:76–93.

22.

Fisher

Barbara

Hovell

Mel

Hofstetter

C. Richard

Hough

Richard

. 1995. “Risks Associated with Long-Term Homelessness among Women: Battery, Rape, and HIV Infection.” International Journal of Health Services 25:351–69.

23.

Frank

Ellen

Stewart

Barbara Duffy

. 1984. “Depressive Symptoms in Rape Victims: A Revisit.” Journal of Affective Disorders 7:77–85.

24.

Freeman

Daniel H.

Temple

Jeff R.

2010. “Social Factors Associated with History of Sexual Assault among Ethnically Diverse Adolescents.” Journal of Family Violence 25:349–56.

25.

Golding

Jacqueline M.

1994. “Sexual Assault History and Physical Health in Randomly Selected Los Angeles Women.” Health Psychology 13:130–38.

26.

Gotlib

Ian H.

Lewinsohn

Peter M.

Seeley

John R.

1998. “Consequences of Depression During Adolescence: Marital Status and Marital Functioning in Early Adulthood.” Journal of Abnormal Psychology 107:686–90.

27.

Grant

Bridget F.

Harford

Thomas C.

1995. “Comorbidity between DSM-IV Alcohol Use Disorders and Major Depression: Results of a National Survey.” Drug and Alcohol Dependence 39:197–206.

28.

Hammen

Constance

. 1991. “Generation of Stress in the Course of Unipolar Depression.” Journal of Abnormal Psychology 100:555–61.

29.

Hammen

Constance

. 1999. “The Emergence of an Interpersonal Approach to Depression.” Pp. 21–35 in The Interactional Nature of Depression: Advances in Interpersonal Approaches, edited by Joiner

Coyne

J. C.

Washington, DC: American Psychological Association.

30.

Hammen

Constance

. 2006. “Stress Generation in Depression: Reflections on Origins, Research, and Future Directions.” Journal of Clinical Psychology 62:1065–82.

31.

Herman

Judith Lewis

. 1992. “Complex PTSD: A Syndrome in Survivors of Prolonged and Repeated Trauma.” Journal of Traumatic Stress 5:377–91.

32.

Jordan

Carol E.

Campbell

Rebecca

Follingstad

Diane

. 2010. “Violence and Women’s Mental Health: The Impact of Physical, Sexual, and Psychological Aggression.” Annual Review of Clinical Psychology 6:607–28.

33.

Kendler

Kenneth S.

Heath

Andrew C.

Neale

Michael C.

Kessler

Ronald C.

Eaves

Lindon J.

1993. “Alcoholism and Major Depression in Women: A Twin Study of the Causes of Comorbidity.” Archives of General Psychiatry 50:690–98.

34.

Kessler

Ronald C.

2008. “National Comorbidity Survey: Baseline (NCS-1), 1990-1992 [ICPSR 06693.V6].” Ann Arbor: Inter-University Consortium for Political and Social Research [distributor].

35.

Kessler

Ronald C.

2015. “National Comorbidity Survey: Reinterview (NCS-2), 2001-2002 [ICPSR 35067.V2].” Ann Arbor: Inter-University Consortium for Political and Social Research [distributor].

36.

Kessler

Ronald C.

Berglund

Patricia

Demler

Olga

Jin

Robert

Merikangas

Kathleen R.

Walters

Ellen E.

2005. “Lifetime Prevalence and Age-of-Onset Distributions of DSM-IV Disorders in the National Comorbidity Survey Replication.” Archives of General Psychiatry 62:593–602.

37.

Kessler

Ronald C.

Chiu

Wai Tat

Demler

Olga

Walters

Ellen E.

2005. “Prevalence, Severity, and Comorbidity of 12-Month DSM-IV Disorders in the National Comorbidity Survey Replication.” Archives of General Psychiatry 62:617–27.

38.

Kessler

Ronald C.

Üstün

T. Bedirhan

. 2004. “The World Mental Health Survey Initiative Version of the World Health Organization Composite International Diagnostic Interview.” International Journal of Methods in Psychiatric Research 13:93–121.

39.

Khalifeh

Hind

Dean

Kimberlie

. 2010. “Gender and Violence against People with Severe Mental Illness.” International Review of Psychiatry 22:535–46.

40.

Kilpatrick

Dean G.

Ruggiero

Kenneth J.

Acierno

Ron

Saunders

Benjamin E.

Resnick

Heidi S.

Best

Connie L.

2003. “Violence and Risk of PTSD, Major Depression, Substance Abuse/Dependence, and Comorbidity: Results from the National Survey of Adolescents.” Journal of Consulting and Clinical Psychology 71:692–700.

41.

Kilpatrick

Dean G.

Saunders

Benjamin E.

Veronen

Lois J.

Best

Connie L.

Von

Judith M.

1987. “Criminal Victimization: Lifetime Prevalence, Reporting to Police, and Psychological Impact.” Crime & Delinquency 33:479–89.

42.

Knäuper

Bärbel

Charles

F. Cannell

Schwarz

Norbert

Martha

L. Bruce

Ronald

C. Kessler

. 2006. “Improving Accuracy of Major Depression Age-of-Onset Reports in the Us National Comorbidity Survey.” International Journal of Methods in Psychiatric Research 8:39–48.

43.

Koss

Mary P.

Figueredo

Aurelio José

. 2004. “Cognitive Mediation of Rape’s Mental Health Impact: Constructive Replication of a Cross-Sectional Model in Longitudinal Data.” Psychology of Women Quarterly 28:273–86.

44.

Koss

Mary P.

Harvey

Mary R.

1991. The Rape Victim: Clinical and Community Interventions. 2nd ed. Thousand Oaks, CA: Sage Publications.

45.

Krause

Elizabeth D.

Kaltman

Stacey

Goodman

Lisa

Dutton

Mary Ann

. 2006. “Role of Distinct PTSD Symptoms in Intimate Partner Reabuse: A Prospective Study.” Journal of Traumatic Stress 19:507–16.

46.

Lehrer

Jocelyn A.

Buka

Stephen

Gortmaker

Steven

Shrier

Lydia A.

2006. “Depressive Symptomatology as a Predictor of Exposure to Intimate Partner Violence among Us Female Adolescents and Young Adults.” Archives of Pediatrics and Adolescent Medicine 160:270–76.

47.

Littleton

Heather

Ullman

Sarah E.

2013. “PTSD Symptomatology and Hazardous Drinking as Risk Factors for Sexual Assault Revictimization: Examination in European American and African American Women.” Journal of Traumatic Stress 26:345–53.

48.

Liu

Richard T.

Alloy

Lauren B.

2010. “Stress Generation in Depression: A Systematic Review of the Empirical Literature and Recommendations for Future Study.” Clinical Psychology Review 30:582–93.

49.

Logan

T. K.

Walker

Robert

Jordan

Carol E.

Leukefeld

Carl G.

2006. Women and Victimization: Contributing Factors, Interventions, and Implications. Washington, DC: American Psychological Association.

50.

McCart

Michael

Zajac

Kristyn

Kofler

Michael

Smith

Daniel

Saunders

Benjamin

Kilpatrick

Dean

. 2012. “Longitudinal Examination of PTSD Symptoms and Problematic Alcohol Use as Risk Factors for Adolescent Victimization.” Journal of Clinical Child & Adolescent Psychology 41:822–36.

51.

Mellins

Claude A.

Walsh

Kate

Sarvet

Aaron L.

Wall

Melanie

Gilbert

Louisa

Santelli

John S.

Thompson

Martie

Wilson

Patrick A.

Khan

Shamus

Benson

Stephanie

Bah

Karimata

Kaufman

Kathy A.

Reardon

Leigh

Hirsch

Jennifer S.

2017. “Sexual Assault Incidents among College Undergraduates: Prevalence and Factors Associated with Risk.” PLoS ONE 12:e0186471.

52.

Miron

Lynsey R.

Orcutt

Holly K.

2014. “Pathways from Childhood Abuse to Prospective Revictimization: Depression, Sex to Reduce Negative Affect, and Forecasted Sexual Behavior.” Child Abuse & Neglect 38:1848–59.

53.

National Research Council. 1996. Understanding Violence against Women. Edited by Crowell

N. A.

Burgess

A. W.

Washington, DC: The National Academies Press.

54.

National Research Council. 2014. Estimating the Incidence of Rape and Sexual Assault. Edited by Kruttschnitt

Kalsbeek

W. D.

House

C. C.

Washington, DC: The National Academies Press.

55.

Orcutt

Holly K.

Erickson

Darin J.

Wolfe

Jessica

. 2002. “A Prospective Analysis of Trauma Exposure: The Mediating Role of PTSD Symptomatology.” Journal of Traumatic Stress 15:259–66.

56.

Risser

Heather J.

Hetzel-Riggin

Melanie D.

Thomsen

Cynthia J.

McCanne

Thomas R.

2006. “PTSD as a Mediator of Sexual Revictimization: The Role of Reexperiencing, Avoidance, and Arousal Symptoms.” Journal of Traumatic Stress 19:687–98.

57.

Robins

Lee N.

Wing

John

Wittchen

Hans Ulrich

Helzer

John E.

Babor

Thomas F.

Burke

Jay

Farmer

Anne

Jablenski

Assen

Pickens

Roy

Regier

Darrel A.

Sartorius

Norman

Towle

Leland H.

1988. “The Composite International Diagnostic Interview: An Epidemiologic Instrument Suitable for Use in Conjunction with Different Diagnostic Systems and in Different Cultures.” Archives of General Psychiatry 45:1069–77.

58.

Schnittker

Jason

. 2000. “Gender and Reactions to Psychological Problems: An Examination of Social Tolerance and Perceived Dangerousness.” Journal of Health and Social Behavior 41:224–40.

59.

Simon

Gregory E.

VonKorff

Michael

. 1995. “Recall of Psychiatric History in Cross-Sectional Surveys: Implications for Epidemiologic Research.” Epidemiologic Reviews 17:221–27.

60.

Skrondal

Anders

Rabe-Hesketh

Sophia

. 2004. Generalized Latent Variable Modeling: Multilevel, Longitudinal, and Structural Equation Models. Boca Raton: Chapman & Hall/CRC.

61.

Teplin

Linda A.

McClelland

Gary M.

Abram

Karen M.

Weiner

Dana A.

2005. “Crime Victimization in Adults with Severe Mental Illness: Comparison with the National Crime Victimization Survey.” Archives of General Psychiatry 62:911–21.

62.

Tjaden

Patricia Godeke

Thoennes

Nancy

. 1998. Stalking in America: Findings from the National Violence against Women Survey. Washington, DC: National Institute of Justice, Office of Justice Programs, U.S. Department of Justice.

63.

Walsh

Christine

MacMillan

Harriet

Jamieson

Ellen

. 2002. “The Relationship between Parental Psychiatric Disorder and Child Physical and Sexual Abuse: Findings from the Ontario Health Supplement.” Child Abuse & Neglect 26:11–22.

64.

Weissman

Myrna M.

Wickramaratne

Priya

Nomura

Yoko

Warner

Virginia

Pilowsky

Daniel

Verdeli

Helen

. 2006. “Offspring of Depressed Parents: 20 Years Later.” American Journal of Psychiatry 163:1001–1008.

65.

Wheeler

Jessica L.

2015. “A Review of National Crime Victim Victimization Findings on Rape and Sexual Assault.” Presented at the 2015 Online Symposium on Intimate Partner Violence, Council on Contemporary Families, University of Texas, Austin, April 20. Retrieved May 31, 2019 (https://contemporaryfamilies.org/crime-victimization-brief-report/).

66.

Wolpert

Lewis

. 2001. “Stigma of Depression—A Personal View.” British Medical Bulletin 57:221–24.