Men’s Vulnerability to Prisoner-on-Prisoner Sexual Violence

Abstract

The one-state case study described in this article assesses imprisoned men’s vulnerability to sexual assault by an inmate before policies were implemented to reduce sexual violence. The cases studied were substantiated in an internal hearing procedure. On average, victims were more recently incarcerated, younger, smaller, and less aggressive than their perpetrators, but many victim-perpetrator pairs deviated from this profile. The strongest predictor of victimization was a history of childhood sexual victimization. Other predictors were race, youth, build, education, and experience with incarceration.

Keywords

sexual Assault prisoner PREA

Under a legislative mandate sparked by concern with prisoner sexual assault during incarceration, the Bureau of Justice Statistics (BJS) completed the second National Inmate Survey to obtain national victimization estimates. Based on an anonymous survey completed during 2008 and 2009, with a nationally representative sample of 167 prisons, and 29,954 inmates, an estimated 1% of prisoners reported victimization through forced, nonconsensual sex with another inmate in the prior 12 months (Beck, Harrison, Berzofsky, Caspar, & Krebs, 2010). An additional 0.7% of inmates reported one or more abusive sexual contacts, such as unwanted touching of a sexual area. Men have lower estimated prisoner-on-prisoner victimization than women, but this victimization constitutes serious harm. Victimization by other male inmates is very likely to result in serious injuries (i.e., a knife or stab wound, internal injuries, being knocked unconscious) and more often to involve bribes, being drugged, offers of protection, threats of harm, gangs, and multiple perpetrators (Beck et al., 2010).

Numerous negative impacts result from prison sexual assault. They include physical injury (Beck & Harrison, 2007), spread of sexually transmitted diseases (Maruschak, 2007), and posttraumatic stress disorder (PTSD) and other psychological problems (Banbury, 2004; Dumond & Dumond, 2002; Fagan, Wennerstrom, & Miller, 1996; Struckman-Johnson & Struckman-Johnson, 2006; Struckman-Johnson, Struckman-Johnson, Rucker, Bumby, & Donaldson, 1996; Toch, 1992). Men sexually assaulted in prison experience high rates of fear, anxiety, and social disruption; they may feel a devalued sense of manhood and lowered competence and security (Dumond & Dumond, 2002, p. 73). Especially young men become highly disturbed by their own sexual arousal, which violence can precipitate (Groth, Burgess, & Holmstrom, 1977; Struckman-Johnson, 1991).

A 2007 Urban Institute report presented survey findings on institutional actions to prevent prison sexual violence. Several states had systems to identify offenders with sexual assault victimization histories or with vulnerability to assault. Typically, staff house vulnerable inmates in special areas with increased supervision and separation from likely assailants. Also, as recommended by researchers (Fagan et al., 1996, p. 60), prison employees often avoid matching people to cellmates who are likely assailants (Owen, Wells, Pollock, Muscat, & Torres, 2008a, p. 61). Such precautions were routinely used in the state we studied. Common tactics were to avoid housing men who appeared to be young, naïve, and weak with those who were larger, older, and more experienced.

Although the U.S. Prison Rape Elimination Act (PREA) may have spurred increased staff actions to assess vulnerability and refined methods to protect inmates, the Urban Institute’s 2007 report found all but California and Oregon used unvalidated vulnerability assessment instruments (Zweig, Naser, Blackmore, & Schaffer, 2007). The research we describe is a one-state case study designed to provide information to inform thinking about vulnerability. It assesses vulnerability in the state before policies were implemented to reduce sexual violence against inmates.

First, we examine differences in characteristics of male victims and their male inmate assailants to determine whether these victim–perpetrator pairs fit the commonly assumed profile—that is, the victim being younger than the perpetrator. Second, we identify individual-level predictors of male inmates’ vulnerability to inmate attacks, matching victims to nonvictims on housing security level and thus removing it as a confounding variable. Findings are relevant to understanding the likely combinations of victim–perpetrator characteristics and also to identifying male inmates vulnerable to sexual assault.

Literature Review

Researchers have focused on differences between inmate victims and perpetrators involved in the same sexual assault incident. Fagan and colleagues (1996) concluded that perpetrators of prison sexual violence tended to be older, heavier, and bigger than their victims. Additional research confirms that perpetrators are generally older than their victims but younger than the average prisoner (Chonco, 1989; Davis, 1968; Mariner, 2001). Missing from the literature is any quantitative assessment of the degree to which victim–perpetrator pairs fit a typical, multidimensional profile of victim–assailant differences.

Relevant to prediction, the BJS sponsored a nationally representative self-report survey of inmates in 2008 and 2009 that compared estimates of reported sexual victimization for groups of prison inmates on the following: sex, race/Hispanic origin, age, education, marital status, weight, and sexual history (i.e., prior victim, prior assailant; Beck et al., 2010) for a mixed group of 23,675 men and 6,279 women. For the mixed-gender sample, net of the effects of other predictors, not being heterosexual and having a history of prior sexual victimization most increased the probability of victimization during incarceration. Being mixed-race and never being married made smaller, positive contributions to the probability of victimization.

Some, but not all, studies of smaller and, in some cases, nonrepresentative samples of men have produced findings consistent with those of Beck et al. Many of these studies were conducted before PREA funding was available, and despite lack of resources for large-sample studies, they laid the groundwork for and confirmed findings by Beck et al. For example, like the BJS research, recent studies show increased vulnerability related to being nonheterosexual and prior sexual victimization. In addition, they have shown that vulnerably is associated with being young, inexperienced in prison settings, having mental health impairments, or being intellectually impaired, small in stature, and unassertive/passive (Austin, Fabelo, Gunter, & McGinnis, 2006; Beck et al., 2010; Chonco, 1989; Hensley, Koscheski, & Tewksbury, 2005; Hensley, Tewksbury, & Castle, 2003; Jenness, Maxson, Matsuda, & Macy Sumner, 2007; Warren, Jackson, Booker Loper, & Burnette, 2010; Wolff, Blitz, Shi, Siegel, & Bachman, 2007). Using official data, Austin and colleagues determined, in Texas, the most vulnerable men as white, young, having a mental health problem, and having prior sex offense convictions.

Most other recent studies used data from interviews or surveys eliciting self-reports during interviews or surveys. For example, Hensley and colleagues (2005, 2003) concluded from interviews that victims were typically white, unmarried, heterosexual, and had committed violent or property crimes. Hensley et al. (2005) confirmed Hensley et al.’s (2003) findings that victims were relative newcomers in the prison system. Somewhat limiting the strength of their inferences, however, Hensley et al. (2003) did not utilize statistical tests of mean differences to analyze their data, and Hensley et al. (2005) had a low response rate (18%). Also using interviews, but with an 86.2% response rate, Jenness et al. (2007) found that victims were usually between the ages of 18 and 25 and had mental health problems. Contradicting Hensley et al.’s (2005, 2003) conclusions, nonheterosexual and black inmates were considerably more vulnerable to prison sexual assault. With a response rate of 39%, based on a mixture of self-administered surveys and face-to-face interviews, Wolff, Shi, Blitz, and Siegel (2007) discovered that vulnerable inmates had higher prevalence of mental disorders and prior sexual victimization as well as more education than nonvictims. These cited studies identify some common predictors of vulnerability, specifically, youth, mental health problems, and prior convictions for violent or sexual offenses.

A few studies (Johnson et al., 2005; McClellan, Farabee, & Crouch, 1997; Weeks & Widom, 1998) have shown that in various samples, between 4.6% and 59% of male inmates self-report some unwanted sexual involvement as children or youth. As others have noted (e.g., O’Leary, Koenig, & Doll, 2004), researchers have not identified the mechanism through which, particularly men’s childhood, sexual victimization later influences adult revictimization. Yet there is some suggestion that vulnerability may be high for men with childhood abuse histories.

Providing a different perspective from that of quantitative research, recent qualitative studies highlight the complexity of prison sexual behavior. Based on the meanings inmates attached to sexual encounters, Fleisher and Krienert introduced the concept of a sexual behavior continuum, ranging from voluntary to coercive sexual encounters. Building on Fleisher’s work, Warren et al. identified three types of prison sexual encounters: relatively frequent consensual sex, less frequent bartered sex, and relatively rare coercive sex. Because perpetrators vary in motivation and sexual encounters have unique meanings in prison culture and context (Fleisher & Krienert, 2006; Owen et al., 2008b; Warren et al., 2010), victims are likely a heterogeneous group. Indeed, one individual could initially act as a consensual partner and later become victim or perpetrator (Warren et al., 2010). Thus, variation in the nature of sexual encounters reported as victimization may limit the degree to which there is a consistent profile of victim–perpetrator pairs or a consistent profile of vulnerable inmates.

Another often ignored complication in prior research is the failure to examine contextual predictors of vulnerability to victimization; security level of housing, for example, can mask the predictors of vulnerability to victimization. As noted earlier, most correctional institutions try to prevent victimization by varying oversight and placement of potential perpetrators and/or victims. Opportunity for sexual assault is most limited in single-cell, high-oversight facilities and is most possible in shared cells or dormitory sleeping areas with unsupervised access to common areas, such as bathrooms (Alarid, 2000, p. 401). To assess what individual-level characteristics predict victimization, the nonvictims in the study need to be in housing with security levels that match the levels for victims.

Present Study

The present study adds to knowledge about differences between assailants’ and their victims’ individual-level characteristics and about individual-level predictors that differentiate inmates who are and are not victims. It considers an 8-year record of all officially substantiated cases of male prisoner-on-prisoner sexual assaults in one Midwestern state’s correctional system. Whereas most previous studies relied on self-administered surveys or face-to-face interviews, this article examines sexual victimization cases that were officially validated through an internal hearing process. The discussion of findings considers whether results differ from studies using different methodologies. Findings replicated with both self-reported victimization and officially substantiated victimization allow for the strongest inference about differences between victims and perpetrators in the same incident and about which inmates in a correctional system are most vulnerable to sexual assault by other inmates.

Method

Substantiated Cases

Substantiated cases were heard and accepted as valid by a hearing officer employed by the correctional department. Hearing officers dismissed charges as invalid if there were technical violations during case handling or if they could not determine the incident occurred.

Three documents provided incident detail information. When inmates, their family members, or others made verbal reports to staff or wrote to an inspector or other official or staff knew of an incident, the first-contacted or knowledgeable staff person prepared an incident report to describe the alleged assault, noting the time, location, and initial account. A correctional department staff member then prepared a misconduct report that confirmed names of parties and witnesses to the incident and included a specific misconduct charge. Finally, written statements of witnesses and alleged victims and perpetrators were taken and presented at a hearing, during which records were kept of participants’ answers to questions. Victims were not required to attend hearings, but could do so if they wished. Hearings consisted of evidence presented and statements by parties involved, staff, and witnesses.

Limitations of using substantiated cases of sexual violence to identify victims and the matched group are that not all victims report incidents (Beck & Harrison, 2007; Miller, 2010; Owen et al., 2008a, pp. 69-70), some groups (i.e., homosexual/bisexual inmates) tend to report less than others (Fowler, Blackburn, Marquart, & Mullings, 2010), and staff may not appropriately categorize reported incidents as sexual violence (Eigenberg, 2000). Advantages to using substantiated cases over face-to-face interviewing or self-administered surveys are that the process reduces the count of false reports (Gaes, Wallace, Gilman, Klein-Saffran, & Suppa, 2002; Harer & Langan, 2001; Memory, Guo, Parker, & Sutton, 1999), includes people with limited reading ability (Herrick, 1991; Ryan, 1990), and removes the embarrassment and reluctance to discuss a sensitive topic with interviewers (Miller, 2010; Smith & Batiuk, 1989). Investigated incidents allow study of both victim and perpetrator characteristics, which often are unavailable in interviews and surveys without placing the victim at risk.

Sample

There were 165 officially substantiated victimizations of an inmate by another inmate in the study state between 1998 and 2006. In 13 incidents, women attacked women. In one incident, a male inmate grabbed a female inmate. Predictors of inmates’ sexual victimization may differ for men and women, since the dynamics of inmate relationships vary between men’s and women’s prisons (Fleisher, 2005; Owen et al., 2008a). The available data, however, lacked adequate cases for study of women’s victimization. This data limitation accounts for the study’s focus on men victimized by other men.

For the first stage of analysis, the unit of study is the sexual assault incident and the variables are victim and perpetrator characteristics. When a single incident involved multiple perpetrators (seven incidents), just the characteristics of one randomly selected perpetrator were considered in analysis. For the eight perpetrators involved in two different incidents, one randomly selected incident was dropped. For two victims involved in two different incidents, one randomly selected incident was dropped. One man grabbed two different men in a sexual way at the same time; only one of the victims was included in the analysis. After these adjustments, information on 140 victim–perpetrator pairs was available for analysis. The victim–perpetrator pairs that we describe were housed in both specialized units (e.g., the protective custody unit, medical areas, or the intake unit) and in general housing at the time of the incidents.

Apart from the description of how victims differed from their perpetrators, we carried out analysis to predict which men were most vulnerable to victimization. To compare victimized and nonvictimized men, a comparison group of nonvictims was matched to victims based on the restrictiveness (i.e., classification level) of their housing on the date of the assault of the matched victim. Of 140 victimized men, most (81.4% or 114 of the 140) were victimized in a general housing unit, with one of three in progressively more secure settings. To predict victimization, only men living in these three levels of general rather than specialized housing were included, since we could not rank the specialized unit security on an ordinal scale. The analysis to predict victimization, therefore, focused on settings where men were most commonly victimized. The 114 nonvictimized men were randomly selected from rosters of men housed at the same security level housing as the victimized man on the date of the incident. If a man matched as a nonvictim had an official record of sexual victimization in prison at any time, he was not included in the matched group, and an alternative was randomly selected.

Variables

Official Department of Corrections reports and case files on victims, perpetrators, and nonvictims were the source of factual information on incidents and individual characteristics.

Types of sexual assault

A review of narrative incident, misconduct, and hearing reports suggested key differences between incidents that involved threatened, attempted, or accomplished penetration (hereafter called penetration incidents) and those that did not (nonpenetration incidents). Official records provided quantitative evidence that penetration incidents were significantly more likely to involve physical force (36.7% vs. 8.8%, χ² = 16.3), a weapon (16.7% vs. 1.3%, χ² = 11.3), violence in addition to the sexual assault (43.1% vs. 9.0%, χ² = 21.5), and threats of harm (53.3% vs. 23.8%, χ² = 12.9). In each chi-square comparison, 60 penetration incidents were compared with 80 nonpenetration incidents, degrees of freedom (df) was 1, and p value was less than .01. Based on the qualitative and quantitative assessment, we developed three categories of the dependent variable: (1) assault without threat, attempt, and achievement of penetration; (2) assault with threat, attempt, or achievement of penetration; and (3) no assault (the matched group).

Most incidents (57.1% or 80 of 140) did not involve threatened, attempted, or achieved penetration. In these cases, the most common victim-reported behaviors were unwanted touching of their buttocks and being awakened by fondling. In a few cases, another inmate reported that a sleeping man was fondled. The remaining 42.9% (n = 60) of incidents involved threatened, attempted, or achieved penetration, such as anal rape or forced performance of fellatio on the perpetrator. Lending support for the identification of two categories of sexual assault, Wolff et al. (Wolff, Blitz, Shi, Bachman, & Siegel, 2006) and Beck et al. similarly differentiated abusive sexual contacts (intentional touching of specified body areas) from nonconsensual sexual acts (forced sex acts, including oral and anal sex). The dependent variable in our analysis differs from that in some prior studies that combined physical assaults with sexual assaults (Perez, Gover, Tennyson, & Santos, 2010), grouped sexual penetration with sexual touching or threats of sexual violence (Beck et al., 2010; Hensley et al., 2005, 2003; Jenness et al., 2007; Wolff, Shi, & Siegel, 2009), or only examined assaults with penetration (Wolff & Shi, 2009). However, the qualitative and quantitative review of our data suggested that the two categories are substantively different and, thus, might have different predictors.

Victim–perpetrator differences

Available records contained measures of several characteristics commonly expected to differentiate victims from their perpetrators. Race was a dichotomous variable (1 = black, 0 = non-black). Most perpetrators were white (28.6% or 40 of 140) or black (71.4% or 100 of 140), and two were categorized in official records as Hispanic. All but two people in the comparison group had official data indicating they were either black or white. We determined the various combinations of victim and perpetrator race, for instance, white victim and white perpetrator. Except for race, we computed variables reflecting victim differences from their perpetrators by using subtraction, for example, subtracting victim and perpetrator ages. Education had four categories: 8th grade or less, 9th to 11th grade, high school/GED, and more than high school. Age was calculated by subtracting the individual’s date of birth from the incident date. Total number of weeks incarcerated at the facility where the incident occurred prior to the incident indicated experience in the facility. Additional measures of familiarity with prisons were years from first incarceration to incident, and for cellmate incidents, length of time in the cell.

A count of the number of misconduct charges for aggressive acts other than sexual assault of a prisoner served as a proxy indicator of passivity; low value indicated passivity. Separate counts were taken of number of sexual misconduct charges other than sexual assault of a prisoner, other violence or threats, disobeying an order, insolence, and property misuse or destruction. Only misconduct charges before the incident were included. Reliability analysis confirmed the interconnections of the acts included in the aggressive misconduct scale (Cronbach’s alpha = .78). Other types of misconduct charges (i.e., for being out of place, possession of illegal materials, substance abuse) were unconnected to aggressive misconducts and prisoner sexual assault misconducts, so they were not considered in the analysis.

A researcher examined a photograph of each victim and perpetrator and rated the degree to which the person was “slight” in stature and not muscular (e.g., 1 = definitely slight, 2 = somewhat slight, 3 = definitely not slight). These scores were combined so a slight, weak-looking individual was rated as 2 and an individual who was definitely not slight and had muscle strength was rated as 6. Recognizing that inmates change in weight and strength, we use this measure cautiously; there was no alternative measure specific to the time of the incident.

Predictors of victimization

Already described variables that were used to determine victim–perpetrator differences were also used to identify predictors of the three-category variable: penetration victim, nonpenetration victim, and nonvictim. For the nonvictims, we used the incident date for the matched victim to calculate age and years from first incarceration at the time of the incident.

Since prior literature indicates that sex offenders are vulnerable to prison sexual victimization, additional predictors of victimization were history of juvenile assault conviction (0 = no and 1 = yes) and number of adult sexual assault convictions. To provide a full picture of the statistical effects of prior convictions, additional variables were counts of number of convictions for each crime type: property, weapons, other violence, and drug-related offenses.

The potential predictor of victimization, childhood sexual victimization, was determined from sentencing reports (0 = no and 1 = yes). Although research suggests that gang membership leads to prison sexual assaults, in the Midwestern state being studied, very few inmates were classified by prison staff as gang members (Austin & McGinnis, 2004). Staff did not identify gang conflict as a major influence on inmate behavior, and gangs were not mentioned in official records on sexual assault incidents. Thus, gang involvement was not studied. Data were unavailable for one variable we would have liked to study, that is, sexual orientation.

Analysis

Except for race, for the 140 male-on-male sexual assault incidents, one-sample t test revealed that the computed differences between characteristics of victims and their perpetrators were significantly different from zero. In addition, frequencies were examined for a categorical variable representing combinations of perpetrator and victim races.

To identify victimization predictors, t tests and chi-square tests were applied to test for differences between each type of victim (penetration vs. nonpenetration) and nonvictims. Because Kerbs and Jolley (2007) found a curvilinear relationship of age to prison sexual victimization, logistic regression was used to determine whether age and age-squared (an indicator of a curvilinear relationships) predicted victimization (DeMaris, 2004). No curvilinear relationship was confirmed, so age-squared was not considered further. The multivariate technique to compare victims and nonvictims was multinomial logistic regression.

Findings

Victims’ Differences From Their Assailants

One-sample t test of whether the difference between victims and their assailants differed from zero showed insignificance of length of time in facility and education. However, there were significant differences on four dimensions: (1) Perpetrators on average were incarcerated 9.5 years more than their victims; (2) perpetrators on average were 10.5 years older than their victims; (3) on a scale of 1 to 6, perpetrators were somewhat (.6) larger in physical stature than victims; and (4) perpetrators on average had received 10.7 more aggressive misconduct charges for incidents other than sexual assault against a prisoner while in prison than victims did.

Slightly more than half (55.7% or 78 of 140) of the perpetrator–victim pairs included a black perpetrator and a non-black victim. Also, 25.0% (35) of pairs involved two non-black inmates, 15.7% (22) of pairs involved two black inmates, and the remaining 3.6% (5) of pairs involved a non-black perpetrator and a black victim (results not shown in table, Table 1).

Table 1.

Comparing Victim and Perpetrator Differences, for 140 Matched Cases of Male-on-Male Incidents, Occurring at All Security Levels

Characteristic	Mean difference	SD	p
Educational-level difference	0.16	1.07	.086
Musculature rating difference	5.89	20.56	.001**
Age difference at incident	10.52	12.06	.000***
Difference in weeks incarcerated, at incident	8.45	64.46	.123
Difference in years with Corrections Dept, at incident	9.52	10.12	.000***
Proxy for passivity
Difference in aggressive misconducts	10.72	15.63	.000***
Difference in prior sexual victimizations	0.06	0.42	.072

Note: The t tests indicate that the mean difference between each matched pair is significantly different from zero. In other words, members of the pair significantly differ from each other.

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

There were numerous exceptions to the typical differences between perpetrators and their victims. Compared to their perpetrators, some victims were older, more educated, larger and more muscular, and had been in correctional facilities longer. In fact, Table 2 shows that only 7.1 % of the time, the victim–perpetrator pairs matched the common assumption that victims are younger, newer to the facility, less educated, physically weaker, and non-black (with black perpetrators). In 20.7% of incidents, victims were older than perpetrators (Figure 1). In 42.1% of incidents, the victim was in the facility longer than the perpetrator at the time the sexual assault occurred. In most incidents, the victim was more educated (63.6%) or larger in stature (50.7%) . It would be quite misleading to assume that differences in size, age, experience with incarceration, and education consistently characterize victim–perpetrator pairs or mark the victim as an easy target.

Table 2.

Common Stereotypes About Predicting Prison Sexual Assaults

	N	%
Assault incident met all 5 stereotypes	10	7.1
Assault incident met 4 out of 5 stereotypes	28	20.0
Assault incident met 3 out of 5 stereotypes	50	35.7
Assault incident met 2 out of 5 stereotypes	29	20.7
Assault incident met 1 out of 5 stereotypes	20	14.3
Assault incident met 0 out of 5 stereotypes	3	2.1
Total	140	100

Figure 1.

Victims’ advantages over perpetrators

Bivariate Comparisons With Nonvictims

There were several statistical differences between each victim group (penetration and nonpenetration) and nonvictims (Table 3). Compared with nonvictims, both victim groups had less education, were younger, were smaller, and had less experience in prison. They also had fewer convictions for weapons and violent crimes than nonvictims. Finally, they were more likely to be non-black, have a history of childhood sexual victimization, a history of committing juvenile sexual assault, and a greater number of convictions for sexual assault outside of prison.

Table 3.

Comparison of Two Groups of Male Victims With Comparison Group

	Victim groups
	No penetration	Penetration	Comparison group
Predictor	(M or %)	(M or %)	(M or %)
Background
Level of education	2.46**	2.26***	2.73
Black	16.2% (n = 11)****	15.2% (n = 7)****	50.8% (n = 58)
Age	27.91****	27.15****	37.23
Prior sexual victimization	14.7% (n = 10)***	15.2 (n = 7)***	1.8% (n = 2)
Musculature rating	32.79****	35.65**	41.14
Juvenile offense history
Juvenile sexual assault	8.8% (n = 6)*	13% (n = 6)**	2.6% (n = 3)
Adult offense history
Property convictions	0.94	0.94	1.40
Sexual assault convictions	0.56*	0.72***	0.33
Weapons convictions	0.12***	0.22	0.48
Violence convictions	0.24**	0.20**	0.50
Drug-related convictions	0.21	0.09*	0.33
Official record in prison
Aggressive misconducts	4.71	5.89	5.59
Weeks incarcerated	173.78****	164.26****	396.05
N	68	46	114

Note: Asterisks indicate that a group is significantly different than the comparison group.

p < .10. **p < .05. ***p < .01. ****p < .001.

Multivariate Analysis

The multinomial logistic regression analysis provided further evidence of nonspurious statistical effects on the likelihood that male inmates would be sexually assaulted by other male inmates. The chi-squared statistic (Table 4) is large and significant, showing a good fit of data to model. Table 4 also presents the odds ratio and the percent change in odds of being one of the two types of victims for each variable. The odds ratios in the results are based on analysis of the risk of victimization where no victimization (i.e., the matched group) is the comparison. A negative odds ratio indicates a lower risk of the particular type of victimization.

Table 4.

Multinomial Logistic Regression to Predict Victimization in Prisoner-on-Prisoner Sexual Assaults

	No threat, attempt, or actual penetration vs. no sexual victimization		Threat, attempt, or actual penetration vs. no sexual victimization
Predictor	e^b	100 * (e^b-1), %	e^b	100 * (e^b-1), %
Individual characteristics
Level of education	0.72	−27.7	0.53*	−47.4
Black	0.26*	−74.1	0.19*	−81.4
Age	0.90*	−9.7	0.86***	−14.4
Prior sexual victimization	19.16*	1816.2	18.93*	1792.8
Musculature rating	0.96**	−3.9	0.98**	−2.3
Juvenile offense history
Juvenile sexual assault	1.07	6.7	1.56	56.1
Adult offense history
Property convictions	0.90	−9.9	0.95	−4.6
Sexual assault convictions	1.21	20.7	1.59**	59.0
Weapons convictions	0.57	−43.5	1.09	8.5
Violence convictions	0.85	−15.4	0.84	−15.9
Drug-related convictions	0.96	−3.8	0.54	−45.8
Prison misconduct
Aggressive misconducts	1.00	−0.5	0.99	−0.8
Weeks incarcerated	1.00**	−0.1	1.00**	−0.2
Model chi-square	105.60***
df	26

Note: The base case for this model is less than high school, non-black, who do not have prior sexual victimization experience, and who do not have prior convictions on sexual assault as a juvenile and property/sexual assault/weapon/violence/drug-related offenses as an adult.

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

Nonpenetration victim group

Race, age, prior sexual victimization, and stature significantly discriminated between the nonpenetration victim group and the nonvictims. Specifically, the odds of a black man being a victim in a nonpenetration incident were 0.26 times lower than the odds for a non-black man. In other words, a black man has a 74.1% decrease in his chance of becoming a nonpenetration group victim. Similarly, being a younger age increased risk of becoming a victim of a nonpenetration incident by 9.7%, and smaller stature increased risk by 3.9%. Most striking, the odds of a man with a history of childhood sexual victimization being a victim of a nonpenetration sexual assault were 19.2 times (or 1816.2%) higher than the odds of a man without such a history. With other variables controlled, education did not significantly discriminate between men with nonpenetration victimization and nonvictims.

Most indicators of offense history did not significantly differentiate nonpenetration victims from nonvictims at the .05 or lower level, though both a juvenile sexual assault conviction and number of adult sexual assault convictions were significant (p ≤ .10) Years from first incarceration to the incident did significantly decrease men’s odds of experiencing a nonpenetration assault; with each additional year, the odds decrease by 5.2%.

Penetration victim group

Education, race, age at the time of the incident, prior sexual victimization, stature, number of adult sexual assault convictions, and time since first incarceration significantly discriminated between men victimized in an incident involving penetration and nonvictims. Parallel to findings for nonpenetration incidents, the odds of victimization are increased almost 19 times by childhood sexual victimization, representing a 1792.8% increase in the chances of such victimization. Also, each additional adult sexual assault conviction increased victimization odds almost 1.6 times ( by 59%). As for nonpenetration, the remaining independent variables modestly or slightly increased the victimization odds. However, education and both juvenile and adult history of sexual assault convictions did significantly discriminate the penetration victims from the nonvictims at the .05 or lower level of significance.

Conclusion and Discussion

Analysis of information on one state’s officially confirmed incidents indicated that compared to the perpetrators who victimized them, on average, victims had a first incarceration that began more recently, were younger, and were smaller. Suggesting a difference in passivity, compared to their attackers, victims had many fewer charges for prior aggressive misconducts. Despite these overall consistencies with common assumptions about differences in victim–perpetrator characteristics that provide perpetrators opportunities or reason to sexually assault particular inmates, the characteristics of the parties in many incidents contradicted assumed differences. Variances for the victim–perpetrator difference were large, and victim–perpetrator differences in some incidents were the opposite of the typical; for instance, some victims were older than perpetrators. Less than 10% of the victim–perpetrator pairs conformed to the combined expectations that compared to their assailants, victims were younger, newer to incarceration, less educated, smaller, and non-black and had a black perpetrator. Lending support to this finding, a study using self-report interviews (Jenness et al., 2007) documented that among California inmates, except for Asian inmates aged 18 to 25, some of every type of inmate in the random sample reported sexual assault.

Multivariate analysis to predict being a victim provided additional insight into vulnerability. Men with a history of childhood sexual assault had much elevated odds of both types of sexual victimization. In addition, consistent with prior studies (Austin et al., 2006; Hensley et al., 2005; Jenness et al., 2007), variables that significantly (but less dramatically) characterized victims of both types of attack were being non-black, young, slight, and limited in experience with incarceration. Also consistent, men with less education were vulnerable to sexual assaults involving penetration. These findings suggest that perpetrators take advantage of opportunities to overpower other inmates, but other influences come into play. Variables omitted due to data limitations (e.g., gang membership, mental health problems, sexual orientation, participation in the prison economy) no doubt would increase understanding of victimization, as would both existing and future research on the institutional context. It would also be useful to replicate the research in this same state after the introduction of PREA reforms to determine whether some groups become less vulnerable due to reforms.

Research (Heidt, Marx, & Gold, 2005, p. 533) provides considerable evidence that the confluence of childhood and adult sexual victimization produces high levels of anxiety, depression, and PTSD. A study conducted by Listwan, Colvin, Hanley, and Flannery (2010) using a sample of released prisoners found that direct and indirect victimization of multiple types that occurred during incarceration had significant correlations with posttraumatic cognitions (related to PTSD) and trauma symptoms (e.g., anxiety, dissociation) after release. Clearly, victimization and revictimization occurring under the watch of correctional administrators and staff are serious, unintended, negative effects of incarceration.

Future research would profitably consider the complex interplay of prior childhood sexual abuse, mental health impairments, revictimization of adults with childhood abuse, and vulnerability of nonheterosexual men to sexual attack both in community and in prison. We lacked data on men’s sexual orientation. However, Jenness et al. (2007) found that nonheterosexual men were at very high risk for sexual assault in prison. Nonheterosexual men in prison were at risk for sexual attack before incarceration (Morash, 2006, p. 70) and, thus, would experience the very negative effects of revictimization at uniquely high rates.

Incarceration concentrates individuals who are highly vulnerable to sexual victimization and its serious negative psychological effects with potential perpetrators with histories of committing aggressive crimes, including sexual assault. These perpetrators may themselves be child abuse victims (Eigner, 2010; Morash, Jeong, Zang, & Bush, 2010). Not all demographic groups, however, are equally likely to be subjected to the threatening circumstances of incarceration. As Jacobi (2005, p. 449) put it, “America has been on a twenty-year spree of prison building, and has filled its old and new prisons and jails with unprecedented numbers of prisoners,” who are disproportionately poor, racial and ethnic minority, and “sick” due to deficiencies in mental and physical health. Although PREA in particular and, more generally, prison reform and development efforts have resulted in increased mental health and medical services for inmates and parolees (e.g., Stahl & West, 2001), these services are far from uniformly medically or constitutionally adequate (review by Delgado & Humm-Delgado, 2009, pp. 190-193; Jacobi, 2005; Wilper et al., 2009). Extensive use of punitive incarceration has reduced resources available for childhood sexual abuse prevention and treatment of mental health impairments. Simultaneously, massive incarceration exposes large numbers of people to a cycle of revictimization, likely contributing to mental health impairments. Inmates prone to commit sexual assaults after release often do not receive treatment that might reduce this probability. The victimization of inmates promotes mental health impairments and impedes their success after leaving prison.

There is no simple, general rule of thumb to determine some mix of potential perpetrator and victim characteristics likely in a victimization incident. The massively expanded use of incarceration perpetuates failure for individuals after they leave prison. U.S. policies have produced a dramatic increase in incarceration for a widened range of crimes (Hagan, 2010; West, 2010) such that it leads the world in incarceration rates among industrialized countries. The expansive use of incarceration not only exposes a wide variety of people to sexual assault in correctional facilities; it also most exposes those individuals who will experience serious adverse psychological reactions to revictimization. Thus, it contributes to mental health impairment, which contributes to the affected individual’s return to prison after release (Cloyes, Wong, Latimer, & Abarca, 2010).

Footnotes

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

The authors received financial support from the State Department of Corrections for the research for this article. They did not receive financial support for the authorship and/or publication of this article.

References

Alarid

L. F.

(2000). Sexual assault and coercion among incarcerated women prisoners: Excerpts from prison letters. The Prison Journal, 80(4), 391-406.

Austin

Fabelo

Gunter

McGinnis

(2006). Sexual violence in the Texas prison system (No. 215774). Washington, DC: U.S. Department of Justice, Office of Justice Programs.

Austin

McGinnis

(2004). Classification of high-risk and special management prisoners: A national assessment of current practices. Washington, DC: National Institute of Corrections.

Banbury

(2004). Coercive sexual behaviour in British prisons as reported by adult ex-prisoners. Retrieved from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=522851

Beck

Harrison

(2007). Sexual victimization in state and federal prisons reported by inmates, 2007 (Bureau of Justice Statistics Special Report No. 219414; Prison Rape Elimination Act of 2003). Washington, DC: U.S. Department of Justice, Office of Justice Programs. Retrieved from http://bjsdata.ojp.usdoj.gov/content/pub/pdf/svsfpri07.pdf

Beck

Harrison

Berzofsky

Caspar

Krebs

(2010). Sexual victimization in prisons and jails reported by inmates, 2008-2009 (No. 231169; National Inmate Survey, 2008-2009). Washington, DC: Bureau of Justice Statistics.

Chonco

(1989). Sexual assaults among male inmates: A descriptive study. The Prison Journal, 69, 72-82.

Cloyes

Wong

Latimer

Abarca

(2010). Time to prison return for offenders with serious mental illness released from prison: A survival analysis. Criminal Justice and Behavior, 37(2), 175-187.

Davis

(1968). Sexual assaults in the Philadelphia prison system and sheriff’s vans. Trans-Action, 6, 8-16.

10.

Delgado

Humm-Delgado

(2009). Health and health care in the nation’s prisons. Lanham, MD: Rowman & Littlefield.

11.

DeMaris

(2004). Regression with social data: Modeling continuous and limited response variables: Wiley Series in Probability and Statistics. New York, NY: John Wiley.

12.

Dumond

(2002). The treatment of sexual assault victims. In Hensley

(Ed.), Prison sex: Practice and policy (pp. 67-88). Boulder, CO: Lynne Rienner.

13.

Eigenberg

(2000). Correctional officers and their perceptions of homosexuality, rape, and prostitution in male prisons. The Prison Journal, 80(4), 415-433.

14.

Eigner

(2010). Gender-related effects of childhood sexual abuse on males. Dissertation Abstracts International: Section B: Sciences and Engineering, 56(6), 3441.

15.

Fagan

T. J.

Wennerstrom

Miller

(1996). Sexual assault of male inmates: Prevention, identification, and intervention. Journal of Correctional Health Care, 3(1), 49-65. doi:10.1177/10783458960030010410.1177/107834589600300104

16.

Fleisher

(2005). The culture of prison sexuality and rape. Washington, DC: U.S. Department of Justice.

17.

Fleisher

Krienert

(2006). The culture of prison sexual violence (No. 216515). Washington, DC: U.S. Department of Justice, National Institute of Justice. Retrieved from http://www.ncjrs.gov/App/publications/Abstract.aspx?id=238132

18.

Fowler

S. K.

Blackburn

A. G.

Marquart

J. W.

Mullings

J. L.

(2010). Would they officially report an in-prison sexual assault? An examination of inmate perceptions. The Prison Journal, 90(2), 220-243.

19.

Gaes

G. G.

Wallace

Gilman

Klein-Saffran

Suppa

(2002). The influence of prison gang affiliation on violence and other prison misconduct. The Prison Journal, 82(3), 359-385. doi:10.1177/00328855020820030410.1177/003288550208200304

20.

Groth

Burgess

Holmstrom

(1977). Rape: Power, anger, and sexuality. American Journal of Psychiatry, 134, 1239-1243.

21.

Hagan

(2010). Who are the criminals? Princeton, NJ: Princeton University Press.

22.

Harer

M. D.

Langan

N. P.

(2001). Gender differences in predictors of prison violence: Assessing the predictive validity of a risk classification system. Crime & Delinquency, 47(4), 513-536. doi:10.1177/001112870104700400210.1177/0011128701047004002

23.

Heidt

J. M.

Marx

B. P.

Gold

S. D.

(2005). Sexual revictimization among sexual minorities: A preliminary study. Journal of Traumatic Stress, 18(5), 533-540.

24.

Hensley

Koscheski

Tewksbury

(2005). Examining the characteristics of male sexual assault targets in a southern maximum-security prison. Journal of Interpersonal Violence, 20(6), 667-679. doi:10.1177/088626050527606910.1177/0886260505276069

25.

Hensley

Tewksbury

Castle

(2003). Characteristics of prison sexual assault targets in male Oklahoma correctional facilities. Journal of Interpersonal Violence, 18(6), 595-606. doi:10.1177/088626050325113210.1177/0886260503251132

26.

Herrick

(1991). Prison literacy connection. Corrections Compendium, 16(12), 5-9.

27.

Jacobi

J. V.

(2005). Prison health, public health: Obligations and opportunities. American Journal of Law and Medicine, 31, 454-478.

28.

Jenness

Maxson

Matsuda

Macy Sumner

(2007). Violence in California correctional facilities: An empirical examination of sexual assault (The Bulletin of the Center for Evidence-Based Corrections No. 1, Vol. 2). Irvine: California Department of Corrections and Rehabilitation. Retrieved from http://ucicorrections.seweb.uci.edu/pubs#bulletins

29.

Johnson

Moss

Taylor

Williams

Carvajal

Peters

(2005). A history of drug use and childhood sexual abuse among incarcerated males in a county jail. Substance Use and Misuse, 40(2), 211-229.

30.

Kerbs

J. J.

Jolley

J. M.

(2007). Inmate-on-inmate victimization among older male prisoners. Crime & Delinquency, 53(2), 187-218. doi:10.1177/001112870629411910.1177/0011128706294119

31.

Listwan

S. J.

Colvin

Hanley

Flannery

(2010). Victimization, social support, and psychological well-being: A study of recently released prisoners. Criminal Justice and Behavior, 37, 1140-1159.

32.

Mariner

(2001). No escape: Male rape in U.S. prisons. New York, NY: Human Rights Watch.

33.

Maruschak

(2007). HIV in prisons, 2005 (Bureau of Justice Statistics Bulletin No. 218915). Washington, DC: U.S. Department of Justice. Retrieved from http://bjs.ojp.usdoj.gov/content/pub/pdf/hivp05.pdf

34.

McClellan

Farabee

Crouch

(1997). Early victimization, drug use, and criminality: A comparison of male and female prisoners. Criminal Justice and Behavior, 24(4), 455-476.

35.

Memory

J. M.

Guo

Parker

Sutton

(1999). Comparing disciplinary infraction rates of North Carolina fair sentencing and structured sentencing inmates: A natural experiment. The Prison Journal, 79(1), 45-71. doi:10.1177/003288559907900100410.1177/0032885599079001004

36.

Miller

K. L.

(2010). The darkest figure of crime: Perceptions of reasons for male inmates to not report sexual assault. Justice Quarterly, 27(5), 692-712.

37.

Morash

(2006). Understanding gender, crime, and justice. Thousand Oaks, CA: SAGE.

38.

Morash

Jeong

S. J.

Zang

N. L.

Bush

(2010). An exploratory study of the characteristics of men known to commit prisoner-on-prisoner sexual violence. The Prison Journal, 90(2), 161-178. doi:10.1177/003288551036182610.1177/0032885510361826

39.

O’Leary

A. L.

Koenig

L. J.

Doll

L. S.

(2004). Child sexual abuse and adult sexual risk: Where do we go from here? In Koenig

L. J.

Doll

L. S.

O’Leary

Pequegnat

(Eds.), From child abuse to adult sexual risk (pp. 299-309). Washington, DC: American Psychological Association.

40.

Owen

Wells

Pollock

Muscat

Torres

(2008a). Focus group methodology and findings. Vol. 2: Gendered violence and safety: A contextual approach to improving security in women’s facilities. Washington, DC: U.S. Department of Justice.

41.

Owen

Wells

Pollock

Muscat

Torres

(2008b). Improving security in women’s facilities. Vol. 1: Gendered violence and safety: A contextual approach to improving security in women’s facilities. Washington, DC: U.S. Department of Justice.

42.

Perez

Gover

Tennyson

Santos

(2010). Individual and institutional characteristics related to inmate victimization. International Journal of Offender Therapy and Comparative Criminology, 54(3), 378-394. doi:10.1177/0306624X0933524410.1177/0306624X09335244

43.

Ryan

(1990, October). Effects of literacy training on reintegration of offenders. Presented at the Freedom to Read: An International Conference on Literacy in Corrections, Ottawa, Ontario, Canada.

44.

Smith

Batiuk

(1989). Sexual victimization and inmate social interaction. The Journal of Federalism, 69(2), 29-38.

45.

Stahl

West

(2001). Growing population of mentally ill offenders redefines correctional facility design. Corrections Today, 63(5), 72-74.

46.

Struckman-Johnson

(1991). Male victims of acquaintance rape. In Parrot

Bechhofer

(Eds.), Acquaintance rape: The hidden crime (pp. 192-213). New York, NY: John Wiley.

47.

Struckman-Johnson

(2006). A comparison of sexual coercion experiences reported by men and women in prison. Journal of Interpersonal Violence, 21(12), 1591-1615. doi:10.1177/088626050629424010.1177/0886260506294240

48.

Struckman-Johnson

Rucker

Bumby

Donaldson

(1996). Sexual coercion reported by men and women in prison. Journal of Sex Research, 33(1), 67-76.

49.

Tewksbury

(1989). Measures of sexual behavior in an Ohio prison. Sociology and Social Research, 74(1), 34-39.

50.

Toch

(1992). Mosaic of despair: Human breakdowns in prison. Washington, DC: American Psychological Association.

51.

Warren

Jackson

Booker Loper

Burnette

(2010). Risk markers for sexual predation and victimization in prison. Washington, DC: U.S. Department of Justice.

52.

Weeks

Widom

(1998). Self-reports of early childhood victimization among incarcerated adult male felons. Journal of Interpersonal Violence, 13(3), 346-361.

53.

West

(2010). Prison inmates at midyear 2009 (Bureau of Justice Statistics Statistical Tables No. 230113). Washington, DC: U.S. Department of Justice, Office of Justice Programs.

54.

Wilper

Woolhandler

Boyd

J. W.

Lasser

K. E.

McCormick

Bor

D. H.

Himmelstein

(2009). The health and healthcare of US prisoners: Results of a nationwide survey. American Journal of Public Health, 99(4), 666-672.

55.

Wolff

Blitz

C. L.

Shi

Bachman

Siegel

(2006). Sexual violence inside prisons: Rates of victimization. Journal of Urban Health: Bulletin of the New York Academy of Medicine, 83, 835-848.

56.

Wolff

Blitz

C. L.

Shi

Siegel

Bachman

(2007). Physical violence inside prisons. Criminal Justice and Behavior, 34(5), 588-599. doi:10.1177/009385480629683010.1177/0093854806296830

57.

Wolff

Shi

(2009). Contextualization of physical and sexual assault in male prisons: Incidents and their aftermath. Journal of Correctional Health Care, 15, 58-69.

58.

Wolff

Shi

Blitz

C. L.

Siegel

(2007). Understanding sexual victimization inside prisons: Factors that predict risk. Criminology and Public Policy, 6(3), 535-564.

59.

Wolff

Shi

Siegel

(2009). Patterns of victimization among male and female inmates: Evidence of an enduring legacy. Violence and Victims, 24(4), 469-484.

60.

Zweig

Naser

Blackmore

Schaffer

(2007). Addressing sexual violence in prisons: A national snapshot of approaches and highlights of innovative strategies (pp. 1-187). Washington, DC: Urban Institute Justice Policy Center. Retrieved from http://www.ncjrs.gov/pdffiles1/nij/grants/216856.pdf