Differences in Women’s Substance-Related Sexual Assaults: Force, Impairment, and Combined Assault Types

Abstract

This study furthers previous research on sexual assaults (SAs) involving substances and/or force by examining effects of perpetrator behaviors of alcohol and/or drug impairment level (none, impaired, incapacitated) and/or force during SA in relationship to various assault and recovery outcomes. A diverse sample of 632 women from a large Midwestern city participated in a study on women’s experiences with SA. Of this sample of substance-involved SAs, 37.3% (n=236) reported a forcible-only unimpaired assault, 50.6% (n=320) reported a combined impairment/incapacitation and force assault, and 12% (n=76) reported an impaired/incapacitated only assault. Multivariate analyses of covariance (MANCOVAs) and chi-square analyses compared assault types as defined by combined alcohol and/or drug impairment level and/or force to determine how these assaults differed in demographics, other assault characteristics, and post-assault experiences. Assault types differed on several demographic, assault, and post-assault factors with most differences showing that the combined assault type was related to worse outcomes than forcible-type assaults, including greater re-experiencing, avoidance, and numbing PTSD symptoms. Implications for clinical intervention include recognizing that assaults involving substance use and force are traumatic and warrant individualized treatment.

Alcohol-involved sexual assault (SA) is common with up to 50% of assaults committed in the context of drinking by victim, offender, and/or both parties (see Lorenz & Ullman, 2016a for a review). Alcohol-involved assaults, where victims were drinking at the time of the assault, have greater perpetrator violence, less physical and verbal resistance by survivors, and greater injury (Abbey, BeShears, Clinton-Sherrod, & McAuslan, 2004; Littleton et al., 2009; Ullman & Najdowski, 2010). Also, survivors of more violent alcohol-involved SAs appear to receive more negative social reactions when disclosing than survivors of less violent assaults (Ullman, 2000), suggesting the intersection of alcohol and violence during rapes impacts how survivors are responded to by others. Generally, SAs involving alcohol involve greater self-blame, avoidance coping, and negative social reactions (Lorenz & Ullman, 2016b; Macy, Nurius, & Norris, 2006; Ullman & Filipas, 2001; Ullman & Najdowski, 2011). Additionally, SAs involving alcohol/drug facilitation are also associated with negative post-assault mental health outcomes, including PTSD and depressive symptoms (Zinzow et al., 2010; Zinzow et al., 2012). Research is needed to better understand whether findings showing these differences of substance-involved versus non-substance-involved assaults differ when considering the co-occurrence of perpetrator use of force during the assault and alcohol/drug impairment level. Furthermore, PTSD and depressive symptoms commonly found following SA need to be examined further to inform clinical practice with victims of various assault types.

Most past studies have examined how assaults and their aftermath differ according to pre-assault alcohol use without looking simultaneously at other possible perpetrator behaviors during the attack (Abbey et al., 2004; Ullman & Najdowski, 2010; Zinzow, Resnick, Amstadter, McCauley, Ruggiero, & Kilpatrick, 2010). For example, few studies of victims have looked at perpetrator substance use and how this may impact victims’ SA experiences from victims’ perspectives. In addition, only looking at use or non-use of substances is too simplistic as the amount of substances consumed and resulting impairment and/or incapacitation due to substance use may impact assault outcomes (see Littleton, Grills-Taquechel, & Axsom, 2009; McConnell, Messman-Moore, Gratz, & DiLillo, 2017 for exceptions). For example, Jaffe et al. (2017) reported that community residing women SA victims who were intoxicated had more severe PTSD symptoms controlling for assault severity, particularly reexperiencing symptoms. In addition, alcohol-involved assaults are related to greater withdrawal from others and drinking to cope with distress, both of which are forms of avoidance coping (Lorenz & Ullman, 2016b).

Less is known about voluntary drug use at the time of the assault, or the co-occurring use of alcohol and drugs at the time of the assault, and how this may influence post-assault impacts. Drug-facilitated SA is a more common topic in the literature, though the rates of drug-facilitated SA are much lower than rates of voluntary drinking at the time of the assault (Kilpatrick, Resnick, Ruggiero, Conoscenti, & McCauley, 2007; Lasky, Fisher, & Swan, 2018). Some research has shown that voluntary marijuana use was a predictor of subsequent SA victimization, while binge drinking was not (Martino, Collins, & Ellickson, 2004).

Substance-Involved Assaults Compared to Non-Substance-Involved Assaults

Given the prevalence of substance-involved SA, it is important to review how these types of assaults differ from non-substance involved assaults. While both assaults are traumatic, substance-involved SA victims can experience greater PTSD symptoms and self-blame than non-substance-involved assaults (Peter-Hagene & Ullman, 2018; Ullman & Najdowski, 2011). This is of particular concern given that victims of substance-involved SA are also less likely to seek formal help (e.g. police, rape crisis center, medical; Walsh et al., 2016). Post-assault substance misuse post-assault (e.g., of alcohol, marijuana, other illicit drugs) is also more likely for victims of substance-involved assaults (McCauley, Ruggiero, Resnick & Kilpatrick, 2010). Victims of substance-involved assaults are also less likely to acknowledge their rape/assault experiences afterward, which could account for delayed help-seeking (Littleton, Radecki, & Berenson, 2008; Walsh et al., 2016). Perpetrators of substance-involved assaults also differ from perpetrators of non-substance-involved assaults in being more likely to adhere to stereotypes about drinking women and hold other alcohol-related expectancies (e.g. alcohol enhances sex drive; Pegram et al., 2018). Beyond the differences in assault characteristics and post-assault outcomes, there are some similarities between substance-involved SAs and non-substance-involved SAs. For example, risk factors for these different types of assaults have shown that early childhood sexual abuse (CSA) is associated with both forcible-type assaults and incapacitated/drug-facilitated assaults (Walsh, DiLillo, Klanecky, & McChargue, 2013). Additionally, CSA leads to problems of substance misuse, which could translate to substance-involved revictimization (Wilsnack, Vogeltanz, Klassen, & Harris, 1997). Given a dearth of studies, research is needed to see if there are differences amongst different types of substance-involved SAs beyond the force-substance binary.

Force, Impairment, and Incapacitation

To date, few studies have examined combinations of behaviors that typically occur in SAs, which further complicates understanding differences in post-assault outcomes as reviewed above (Littleton et al., 2009; McConnell et al., 2017). One study of college SA victims used latent profile analysis which showed that survivors of assaults characterized by alcohol and violence were uniquely vulnerable, showing less resistance to assault and greater self-blame (Macy, Nurius, & Norris, 2007a,b). This is important because assaults also vary in the degree to which perpetrators use various types of coercive tactics including physical force or verbal coercion and the associated consequences of such tactics (Abbey et al., 2004).

Previous rape type comparison studies

In the two studies done to date, interesting findings were revealed. Littleton et al. (2009) examined a sample of 340 college student alcohol-related SAs that were categorized into three groups: impaired (trouble walking, trouble sleeping; asleep, moving limbs, n=138), incapacitated (unconscious, n=72), and not impaired (n=130). Impaired and incapacitated groups showed less self-blame than non-impaired victims and more disclosure for those incapacitated and no differences in maladaptive coping. Groups did not differ on physical force used, receipt of negative social reactions upon disclosure, PTSD symptoms, anxiety, depression, or posttraumatic cognitions. McConnell et al. (2017) studied a community sample of 161 completed rapes broken down into forcible-only rape (n=48), impaired rape (n=56), combined rape (impaired and forcible; n=29); and incapacitated (n=28) rape types. Combined type rapes were associated with significantly more severe PTSD symptoms than forcible-only and impaired type rapes. Women in the combined group reported significantly higher levels of intrusive and hyperarousal symptoms than those in the forcible-only and impaired groups, as well as significantly higher levels of avoidance/numbing symptoms than those in the impaired group. Women in the combined type rape group reported significantly higher PTSD scores than those in the forcible-only and impaired groups. Women in the incapacitated group did not differ significantly from women in the forcible-only, impaired, and combined groups on any of the three symptom clusters.

While yielding important findings, these two prior studies looking at combinations of tactics and level of alcohol impairment had small sample sizes for rape type groups compared and limited outcomes examined in relation to combined rape tactics groups. Littleton et al.’s (2009) study was also limited to a college sample, and McConnell et al.’s (2017) community sample size was small overall, and therefore also limited. Additionally, they focused solely on alcohol impairment (or did not specify that they looked at more than just alcohol impairment) at the time of assault and did not indicate any overlap between impairment and incapacitation, which is not easily delineated (Ullman, O’Callaghan, & Lorenz, 2019). Furthermore, other disclosure outcomes (e.g. extent of disclosure), prior victimization experiences (e.g. CSA experiences) and other perpetrator behaviors during the assault have also been missing from comparisons of these assault type groups, but are important to examine to further understand other factors unique to substance-involved SA, and differences between types of substance-involved SA. Understanding the differences of extent of disclosure, for example, could provide insight into differing social reactions that substance-involved SA survivors may receive from others, which may be related to how impaired they were during the assault. The current study will address each of these gaps in existing knowledge of substance-involved SAs.

Current Study

The present study seeks to both go beyond studies examining attack behaviors separately and the limited two past studies (Littleton et al., 2009; McConnell et al., 2017) by looking at combinations of perpetrator behaviors (e.g. use of force) to examine how combined assault types based on force and/or alcohol/drug impairment level relate to various demographic, assault, and recovery outcomes. In order to better understand SA recovery, assault types need to be conceptualized and explored in a complex way and then compared for their associations with important variables of relevance to theory, clinical practice, and prevention efforts. We also sought to extend the literature by looking at other outcomes not addressed in previous literature (e.g. child sexual abuse (CSA), extent of disclosure) to further understand how the assault types in the current study differ, especially given that CSA is associated with substance-involved revictimization and post-assault substance misuse (Classen, Palesh, & Aggarwal, 2005; Wilsnack et al., 1997). Additionally, knowledge of whether other aspects of the perpetrator’s behaviors (e.g. perpetrator substance use at time of assault) vary by assault type can help to target preventive interventions and possibly clinical treatment strategies to survivors. Perpetrator substance use is important to analyze given that perpetrator alcohol/drug use at the time of the assault can increase the likelihood that they will also use violence during the assault (Busch-Armendariz, DiNitto, Bell, & Bohman, 2010). Three primary research questions drove the current study. 1) How do different victim demographics vary in different types of substance-involved SA?; 2) How do assault characteristics differ across substance-involved SA types; and 3) How do post-assault outcomes differ across substance involved SA types?

Similar to McConnell et al., (2017), we expected that women in the combined assault type group would have more severe assault experiences, more self-blame, worse coping, more negative social reactions and greater psychological symptom outcomes than those in the forcible-only and not impaired or impaired/incapacitated only groups. In line with past research, we expected differences in PTSD symptoms among groups, with the combined assault type group more likely to experience more severe PTSD symptoms than the other two groups. Three specific hypotheses were tested in the current study. First, other perpetrator behaviors (e.g., perpetrator substance use) will show increased risk for the combined impaired/incapacitated and force group than the forcible-only, unimpaired group and the impaired/incapacitated only group. Second, unlike previous research, all post-assault symptoms, particularly PTSD symptoms, will be worse for the combined impaired/incapacitated group than the forcible-only, unimpaired group and the impaired/incapacitated only group. Third, greater self-blame and negative social reactions will be reported for women in both the combined impaired/incapacitated only group and the impaired/incapacitated only group than the forcible-only, unimpaired group.

Method

Sample

A volunteer community sample of 1,863 women from the Chicago, Illinois area participated in the full three-wave longitudinal study from the years 2010–2013. There were 2,192 surveys sent out to women calling about the survey in response to recruitment of which 1,863 returned completed usable responses for an 85% response rate at Wave 1. No women calling about the survey were screened out by our criteria; however, some did not complete/return the survey of those who we sent it out to who requested it be mailed to them.

The final sample for the present study consisted of 632 women who indicated having an unwanted sexual experience and reported voluntary alcohol and/or drug use and some level of impairment at the time of the assault. The sample was then categorized into the assault type variable based on responses to the force variable and the level of alcohol/drug impairment variable. While the final sample reported an average of 6 unwanted sexual experiences and the majority of the sample reported more than one unwanted sexual experience (n=488), participants provided details regarding their most serious SA, which comprised the final sample and all variables assessed in this study. The final sample was ethnically diverse; about half were White, one-third African American, and the remainder another race/ethnic groups (see Table 1 for demographics and assault characteristics). The sample was well-educated, 46%) were employed, and a third were in school (32%). Most women had low household incomes of less than $30,000.

Table 1.

Descriptives for survivors of substance-involved sexual assault.

Variable	N (%)	Range	M (SD)
Age		18–69	34.14 (11.46)
Education
Less than high school	58 (9%)
High school/GED	87 (14%)
Some college	259 (42%)
Graduated college/beyond	219 (35%)
Race
White	323 (53%)
Black	194 (32%)
Other	98 (16%)
Income
$10,000 or less	228 (38%)
$10,001–$20,000	119 (20%)
$20,001–$30,000	66 (11%)
$30,001–$40,000	57 (9%)
$40,001–$50,000	36 (6%)
Over $50,000	101 (17%)
Perceived life threat
Yes	285 (47%)
No	320 (53%)
Perpetrator using intoxicants
Doesn’t know	154 (25%)
None	29 (5%)
Alcohol	232 (37%)
Drugs	46 (7%)
Both alcohol and drugs	167 (27%)
Childhood sexual abuse (CSA)
Any CSA	347 (61%)
No CSA	221 (39%)
Number of perpetrators
One person	493 (79%)
Two people	68 (11%)
Three or more people	62 (10%)
Relationship of perpetrator
Stranger	185 (29%)
Non-stranger	445 (71%)
Highest level of abuse (adult)		1–6	5.33 (1.45)
Fondling of sexual parts	31 (5%)
Kissing	30 (5%)
Performed oral sex on you	19 (3%)
Performed oral sex on them	28 (5%)
Anal sex	20 (3%)
Vaginal sex	483 (79%)
Highest level of resistance		0–7	4.36 (2.19)
Stay still or freeze	88 (16%)
Reason/plead	67 (12%)
Cry or sob	85 (15%)
Scream for help	6 (1%)
Run away	14 (3%)
Physically struggle	217 (39%)
Physically fight	78 (14%)
Extent of drinking detail at disclosure
Said briefly	73 (22%)
Discussed a little	76 (23%)
Talked in a general way	63 (19%)
Talked about in detail	118 (36%)

Procedure

We recruited female participants via weekly advertisements in local free newspapers, on Craigslist, and through university mass mailings. Fliers were posted in the community, at area colleges and universities, and at agencies catering to community members generally and gender violence victims specifically (e.g., cultural, domestic, and rape-crisis centers, substance abuse clinics). Interested participants called the research office and were screened for eligibility using the following criteria: (a) unwanted sexual experience at age of 14 or older, (b) 18 or older at the time of participation, and (c) previously told someone about their unwanted sexual experience. Women were informed about the amount ($25) of compensation. Eligible participants received surveys, consent forms, community resources lists for victims, and stamped return envelopes. Participants were paid upon receipt of their completed surveys. The Wave 1 response rate was 85%. The university’s institutional review board approved all study procedures and documents.

Measures

Sexual Experiences Survey (SES).

Sexual victimization in adolescence/adulthood (i.e., 14 or older, the age of consent in the state where the measure was created) was assessed using Testa, VanZile-Tamsen, Livingston, and Koss’s (2004) revised version of the Sexual Experiences Survey. Testa et al.’s revised version (SES-Revised) assesses various forms of SA, including unwanted sexual contact (e.g. unwanted fondling, kissing), verbally coerced intercourse (e.g., gave into sexual contact due to pressure/arguments), attempted rape (e.g., attempted unwanted sexual acts with threat or force), and completed rape resulting from force/threat of force or incapacitation. This version of the SES was reliable in the present sample, Cronbach’s α = .77. We coded highest severity of sexual victimization ranging from 1 (least severe) to 6 (most severe), see Table 1. Most of assaults were completed rapes (87%). CSA before age 14 was assessed with the SES-R as a 5-level ordinal variable (0 = no victimization, 1 = sexual contact, 2 = sexual coercion, 3 = attempted rape, and 4 = completed rape) that was dichotomized to assess differences between assault types (0 = no CSA; 1 = any CSA).

Assault characteristics.

A level of violence variable was dichotomized into a two-level variable indicating that no force versus any force was used during the assault (e.g. verbal threats, physical violence, weapon used) and used to create the assault type groups. Highest level of resistance used by survivors during the assault computed from several items (stay still/freeze, reason/plead, cry/sob, scream for help, run away, physically struggle/push away/hit/scratch, or physically fight/kick/punch/use a weapon/martial arts), with summary scores indicating highest level of resistance from 1 (least) to 7 (most).

In addition to the objective measures of violence and resistance, we examined the negative peritraumatic response of whether victims thought their lives were in danger during assault (yes or no). We asked each woman to indicate her relationship to the perpetrator by checking one of the following options: stranger (23%), acquaintance or casual first date (50%), romantic partner (14%), relative (2%). For assaults with multiple perpetrators with different relationships to the victim, we created a separate category (11%), which was dichotomized to compare stranger versus non-stranger assaults. Women indicated whether they had engaged in drinking only (61%), drug use only (18%), or both drug use and drinking at the time of the assault (21%). Women indicated consuming an average of 4.34 drinks at the time of the assault. Women also indicated whether the perpetrator had engaged in drinking only, drug use only, both drug use and drinking, neither drug use or drinking, or the women did not know if the perpetrator engaged in drug use or drinking. Finally, women’s level of impairment was computed from several items following Littleton et al. (2009): (1) not impaired including those who used alcohol and/or drugs but did not indicate yes to any impairment or incapacitation questions, (2) impaired including: those who used alcohol and/or drugs and indicated “difficulty speaking”, “difficulty moving limbs”, and/or “difficulty walking”, (3) incapacitated including: those who used alcohol and/or drugs and indicated “I was asleep” and/or “I was unconscious”. Level of impairment was coded as how “out of it” they were from 1 (I was not out of it) to 6 (I was unconscious (blacked out)), M = 3.31, SD = 2.09.

Assault type.

Victims reporting either using a) alcohol at the time of assault, b) drugs, or c) both alcohol and drugs were further broken down into three assault type categories. Based on their responses to the impairment question and the force question, they were either put into the forcible-only unimpaired group (they experienced force and indicated drinking, drug use, or both at the time of assault, but self-reported that they did not feel “out of it”), the combined impaired/incapacitated and force group (they reported force and some level of being impaired), or the impaired/incapacitation-only (no force reported, but reported some level of impairment based on their responses to that 6-level question “how ‘out of it’ were you at the time of the assault?”). Incapacitation and impairment could not be coded as two separate groups as in McConnell et al. 2017, because there were 11 impaired only cases, and it is not always possible to distinguish impairment and incapacitation (Ullman, O’Callaghan, & Lorenz 2019).

Center of Epidemiologic Studies Depression Scale (CESD-7).

Depressive symptoms were assessed using a seven-item version of the Center of Epidemiologic Studies Depression Scale (CESD-7) modified by Mirowsky and Ross (1990). Participants rated their symptoms (e.g., I felt lonely) over the past 12 months on a 5-point Likert scale ranging from 0 (never) to 5 (always). Items were averaged (α = .84, M= 2.01, SD= 0.69).

Posttraumatic Stress Diagnostic Scale (PDS).

PTSD symptoms were assessed with the Posttraumatic Stress Diagnostic Scale (PDS; Foa, 1995), a standardized 17-item instrument based on DSM–IV–TR (APA, 2000) criteria. On a scale ranging from 0 (not at all) to 3 (almost always), women rated how often they experienced each symptom (i.e., re-experiencing/intrusion, avoidance/numbing, hyperarousal; e.g. “Having upsetting thoughts or images about this experience that came into your head when you didn’t want them to”) during the past 12 months, in relation to their most serious SA. The PDS has acceptable test–retest reliability for a PTSD diagnosis in assault survivors over two weeks (α = .74; Foa, Cashman, Jaycox, & Perry, 1997) and was reliable in our sample (α = .92). The 17 items were summed to assess posttraumatic symptoms (M = 20.72, SD = 12.57); and 68% qualified for the PTSD diagnosis.

Social Reactions Questionnaire (SRQ).

In order to assess the extent to which women who were drinking at the time of the assault disclosed that they were drinking, a 4-level ordinal variable (1 = said briefly, 2 = discussed a little, 3 = talked in general way, and 4 = talked about in detail) was used to assess the extent drinking was discussed in women’s disclosure. The Social Reactions Questionnaire (SRQ; Ullman, 2000) measured how often victims received each of 48 social reactions since the assault (0, never, to 4, always). The SRQ consists of seven subscales detailing various types of negative (controlling, blaming, egocentric responses, distracting the victim, or treating the victim differently) and positive (emotional support, tangible support) reactions to assault disclosure. Overall, women received more positive (M = 2.17, SD = 0.94) than negative (M = 0.97, SD = 0.75) reactions. We further separated negative reactions into acknowledgment-without-support reactions (e.g., acknowledging the assault happened, but failing to give adequate support, misplaced efforts to control the victim’s decisions) and turning against reactions (e.g., blaming the victim, not believing her story) based on the confirmatory factor analyses reported elsewhere (Relyea & Ullman, 2015a). The turning-against scale consisted of 13 items, M = .86, SD = .90, α = .92. The acknowledgment-without-support reactions scale consisted of 13 items as well, M = 1.06, SD = .76, α = .83.

Rape Attribution Questionnaire.

Two 5-item subscales of the Rape Attribution Questionnaire (Frazier, 2003), a valid and reliable self-report measure of SA victims’ attributions about why the assault occurred, assessed behavioral (e.g., “I should have resisted more”) and characterological (e.g., “I am a careless person”) self-blame on scales ranging from 1 (strongly disagree) to 5 (strongly agree). Both scales were reliable in the current sample: behavioral self-blame, Cronbach’s α = .81 (M = 3.66, SD = 1.00); characterological self-blame, Cronbach’s α’s.75 (M = 2.67, SD = 0.93).

Brief COPE Scale.

Participants completed the Brief COPE, a 28-item self-report scale of coping strategies (Carver, 1997). Strategies used in the past 12 months to cope with the assault were assessed on a scale ranging from 1 (I didn’t do this at all) to 4 (I did this a lot). Based on a factor analysis, Maladaptive Coping was computed as the average of responses to eight items, including the behavioral disengagement, denial, self-blame, and substance-use subscales (M = 17.01, SD = 5.89, α = .83). As Carver suggested, we performed our own factor analysis to create adaptive coping scales. Principal axis factoring with Promax rotation yielded two separate positive coping scales: Individual and Social Coping. Adaptive Individual Coping includes 12 items assessing adaptive, active forms of individual coping (M = 28.43, SD = 7.69, α = .83) such as: “I thought hard about what steps to take.” Adaptive Social Coping includes four items assessing active, adaptive, interpersonal forms of coping (M = 8.86, SD = 3.69, α = .89) such as “I tried to get advice or help from other people about what to do.” Participants completed a 5-item scale of drinking to cope (Cooper, Frone, Marcia, & Mudar, 1995) with assault in the past year assessed from 0 (I didn’t do this at all) to 3 (I did this a lot), M = 1.38, SD = .96, α = .91.

Results

Three types of analyses were conducted to determine differences between assault groups among several variables. Chi-square analyses were conducted among categorical variables in comparison with assault type to evaluate possible differences (see Table 2). For continuous variables, a one-way analysis of co-variance (ANCOVA; Table 2) or a multivariate analysis of co-variance (MANCOVA) was used to test for group differences among variables (see Table 3).

Table 2.

Significant chi-square tests of categorical study variables, ANCOVA test of age, and assault type.

	Assault Type
Variable	Forcible-only unimpaired (n=236)	Combined impaired/incapacitated and force (n=320)	Impaired/incapacitated only (n=76)	Test statistic (χ2)	(df, N)
Race
Black	.404	.253	.301	12.05**	(2, 615)
White	.435	.580	.575	14.08**	(2, 615)
Perceived Life Threat
Yes	.626	.428	.155	52.96**	(2, 605)
No
Perpetrator intoxicant use				24.29*	(8, 628)
Doesn’t know	.280	.224	.227
None	.059	.041	.027
Alcohol	.373	.344	.467
Drugs	.106	.054	.053
Both alcohol and drugs	.182	.338	.227
CSA				7.18*	(2, 568)
Any CSA	.660	.606	.478
No CSA	.340	.394	.522
Extent of drinking detail at disclosure				12.29+	(6, 330)
Said briefly	.303	.208	.122
Discussed a little	.247	.240	.163
Talked in a general way	.202	.177	.224
Talked about in detail	.247	.375	.490
	M (SD)	M (SD)	M (SD)	F
Age	37.83a (11.75)	32.18b (10.91)	33.48b (11.62)*	13.14**	(2, 613)

Note. Standard deviations appear in parentheses below means. Means with differing subscripts within rows are significantly different at the p < .05 based on Bonferroni-adjusted pairwise comparisons.

^**p<0.01,

^*p<0.05, + p<0.1

Table 3.

Differences in continuous study variables and assault type.

	Assault Type
Variable	Forcible-only unimpaired (n=236) M (SD)	Combined impaired/incapacitated and force (n=320) M (SD)	Impaired/incapacitated only (n=76)M (SD)	F	df
Level of resistance	4.55a (2.16)	4.39b(2.19)	3.25a,b(2.17)	5.04**	2, 528
Level of sexual abuse (adult)	5.17 (1.61)	5.46 (1.30)	5.36 (1.50)	2.55	2, 528
Total PTSD scores	18.83a (12.80)	21.95a (12.15)	19.78 (12.55)	4.80**	2, 565
Depressive symptoms	1.98 (.71)	2.06 (.70)	1.87 (.73)	2.60	2, 565
Re-experiencing symptoms	4.44a (3.71)	5.37a (3.91)	4.94 (3.99)	5.24**	2, 568
Arousal symptoms	6.50 (4.73)	6.90 (4.31)	5.79 (4.40)	1.86	2. 568
Avoidance symptoms	3.16a (2.23)	3.59a (2.08)	3.43 (2.05)	3.17*	2, 568
Numbing symptoms	4.84a (4.23)	6.10a (4.29)	5.62 (4.36)	5.52**	2, 568
Characterological self-blame	2.66 (.95)	2.69 (.92)	2.49 (.90)	1.33	2, 607
Behavioral self-blame	3.54a (1.05)	3.78a (.94)	3.59 (1.03)	3.90*	2, 607
Maladaptive coping	16.48a (5.78)	17.75a,b (6.02)	15.65b (5.27)	5.14**	2, 473
Positive individual coping	28.51 (7.36)	27.90 (7.59)	27.02 (7.00)	.626	2, 473
Positive interpersonal coping	8.64 (3.74)	8.89 (3.64)	8.56 (3.90)	.332	2, 473
Drinking to cope	1.35 (1.02)	1.38 (.89)	1.18 (.92)	1.18	2, 473
Positive social reactions	2.20 (.99)	2.12 (.91)	2.18 (.85)	.176	2, 523
Turning against social reactions	.85 (.91)	.86 (.91)	.79 (.74)	.167	2, 523
Acknowledgement without support reactions	1.10 (.81)	1.03 (.74)	.95 (.68)	.547	2, 523

Note. Within each row, means with different subscripts are significantly different at p < .05, Bonferroni-corrected. PTSD= post-traumatic stress disorder.

^*p < .05.

^**p < .01,

^***p < .001.

Demographics

Of the 632 assault survivors included in the present study, 37.3% (n=236) reported a forcible-only unimpaired assault, 50.6% (n=320) reported a combined impaired/incapacitated and force assault, and 12% (n=76) reported an impaired/incapacitated only assault. Race was dummy-coded into three variables for white, Black (African American), and other. Chi-square tests were run to determine if there was a relationship between race and assault type. Survivors in the combined impaired/incapacitated and force group and the impaired/incapacitated only group were more likely to be White than those in the forcible-only unimpaired group. Similarly, women in the forcible-only unimpaired group were more likely to be Black than those in the combined impaired/incapacitated and force group and impaired/incapacitated only group. Other races showed no differences by assault type. No relationship was found between education and assault type or income and assault type. An ANCOVA (controlling for race) was conducted to examine between-group differences in age among the assault types. Age was significantly different between groups, F(2, 613) = 13.14, p<.001. Bonferroni-adjusted pairwise comparisons revealed that women in the forcible-only unimpaired group were significantly older than: combined impaired/incapacitated and force and impaired/incapacitation only groups.

Assault characteristics

Chi-square analyses of several assault characteristic variables were conducted to evaluate differences across assault type groups. Of the five assault characteristic variables, three were significant and two were not significant. Perceived life threat was significantly related to assault type. The forcible-only unimpaired group was more likely to perceive that their life was in danger than both the combined impaired/incapacitated and force group and the impaired/incapacitated only group, and the combined impaired/incapacitated and force group was more likely to perceive that their life was in danger than the impaired/incapacitated only group. Perpetrator intoxicant use at the time of assault was significantly related to assault type. The impaired/incapacitated only group was more likely to report that perpetrator alcohol use only than the combined impaired/incapacitated and force group and the forcible-only unimpaired group. Additionally, the combined impaired/incapacitated and force group was more likely to report the perpetrator was using both alcohol and drugs at the time of assault than the forcible-only unimpaired group and the impaired/incapacitated only group. Childhood sexual abuse (CSA) was significantly related to assault type and higher for women in the forcible-only unimpaired group and combined impaired/incapacitated and force group than in the impaired/incapacitated only group. Number of perpetrators and victim-perpetrator relationship were unrelated to assault type.

To examine between-group differences in assault characteristics, a MANCOVA (controlling for race) was conducted on level of resistance and level of sexual abuse (not CSA) reported at the time of the assault. A significant multivariate effect was found, F(4, 1054) = 3.78, p=.005, Wilks’ Lambda = .972. Significant univariate F tests showed resistance level differed by assault type (p=.007), but level of sexual abuse did not (p=.079). Follow-up Bonferroni-adjusted post-hoc tests showed forcible-only unimpaired and combined impaired/incapacitated and force groups reported higher resistance levels than impaired/incapacitation only group.

Post-assault sequelae and social reactions

To examine between-group differences in psychological symptoms between assault types, a MANCOVA (controlling for race) was conducted on total PTSD scores and depressive symptoms. A significant multivariate effect was found, F(4, 1128) = 3.20, p=.013, Wilks’ Lambda = .978. Significant univariate F tests showed that total PTSD scores (p=.009), differed significantly by assault type, but depressive symptoms (p=.075) did not, though a marginal trend revealed more depressive symptoms for the combined impaired/incapacitated and force group than the other two groups. Follow-up Bonferroni-adjusted post-hoc comparisons showed that the combined impaired/incapacitated and force group reported a higher PTSD total score than the forcible-only unimpaired group. To further evaluate which PTSD symptom clusters varied among rape types, a MANCOVA (controlling for race). A significant multivariate effect was found, F(8, 1130) = 3.20, p=.001, Wilks’ Lambda = .956. Significant univariate F tests showed that re-experiencing symptoms (p=.006), avoidance symptoms (p=.043), and numbing symptoms (p=.004), but not arousal symptoms (p=.156), differed by assault type. Follow-up Bonferroni-adjusted post-hoc tests showed more re-experiencing, avoidance, and numbing symptoms in the combined impaired/incapacitated and force group than the forcible-only unimpaired group.

To assess differences in self-blame post-assault, a MANCOVA (controlling for race) was conducted on characterological and behavioral self-blame. A significant multivariate effect was found, F(4, 1212) = 2.79, p=.025, Wilks’ Lambda = .982. Statistically significant univariate F tests showed that behavioral self-blame differed among assault type groups (p=.021), but characterological self-blame did not (p=.264). Follow-up Bonferroni-adjusted post-hoc tests showed more behavioral self-blame for the combined impaired/incapacitated and force group reported than the forcible-only unimpaired group.

In order to assess differences in coping post-assault, a MANCOVA (controlling for race) was conducted on drinking to cope, maladaptive coping, positive individual coping, and positive interpersonal coping. The multivariate test was marginally significant, F(8, 940) = 1.76, p=.081, Wilks’ Lambda = .971 with significant univariate F tests showing that maladaptive coping differed among assault type groups (p=.006). Follow-up Bonferroni-adjusted post-hoc tests showed that the combined impaired/incapacitated and force group had higher maladaptive coping scores than both the forcible-only unimpaired and the impaired/incapacitation only groups.

Additionally, similar to Littleton and colleagues (2009), a MANCOVA (controlling for race)¹ was conducted on social reactions to assault disclosure (positive reactions, turning against reactions, acknowledgement without support reactions), but was not significant, F(6, 1042) = .347, p=.912, Wilks’ Lambda = .996. For additional disclosure variables, a chi-square analysis examined differences in the extent that women disclosed drinking at the time of assault between assault type groups. Extent of disclosing drinking was marginally significant between assault type groups. The impaired/incapacitated only group was more likely to disclose drinking in great detail than both the combined impaired/incapacitated and force group and the forcible-only unimpaired group, and the combined impaired/incapacitated and force group were more likely to disclose drinking in great detail than the forcible-only unimpaired group. The forcible-only unimpaired group was more likely to briefly disclose drinking than both the combined impaired/incapacitated and force group and the impaired/incapacitated only group.

Discussion

This study goes beyond the two past studies comparing rape types based on level of impairment due to alcohol and/or drugs and perpetrator use of force in relation to various demographics, assault characteristics, and recovery outcomes. Our study had a large, diverse community sample of 632 SA survivors, and our combined assault type was much higher (50.6%) than the 18% in other samples, which could be due to the nature of our sample or suggest that this type of assault is more common than previously found (McConnell et al., 2017; Zinzow et al., 2010). However, we only had 12% reporting an impaired/incapacitated only assault versus 17.4% in McConnell et al. 2017) and 21% in Littleton et al., 2009). Our sample had 37.3% with a forcible-only unimpaired rape, whereas McConnell had 29.8% with such assaults, which probably reflects our community sample that experienced higher severity on several assault and psychological outcomes.

Older victims and African American victims were more concentrated in the forcible-only unimpaired group, consistent with Zinzow et al. (2010). Higher level of CSA was more likely in both the forcible-only unimpaired group and the combined impaired/incapacitated and force group than the impaired/incapacitated only group, which could indicate different pathways to SA revictimization for alcohol/drug-related assaults and/or force-involved assaults, which warrant further research. Both the forcible-only unimpaired group and the combined impaired/incapacitated and force group reported higher levels of resistance than the impaired/incapacitation only group, which is unsurprising given how these groups are defined. What is surprising, however, is that there were no significant differences between the combined impaired/incapacitated and force group and the forcible-only unimpaired group, especially given that some women in the combined impaired/incapacitated and force group reported being incapacitated at the time of the assault. Further research should work to tease apart these different assault characteristics, and how they may impact recovery and post-assault responses.

For other perpetrator behaviors, we found that perpetrator alcohol-use only at the time of the assault was more likely in the impaired/incapacitated only group than both the combined impaired/incapacitated and force group and the forcible-only unimpaired group; however, the combined impaired/incapacitated and force group was more likely to report the perpetrator using both alcohol and drugs at the time of the assault than the forcible-only unimpaired group and the impaired/incapacitated only group. This is the first study to look at perpetrator intoxicant use at the time of the assault in the context of comparing assault type groups, so this factor should be analyzed in future research including its impact on other assault characteristics and post-assault outcomes, as it may reveal a higher risk profile associated within combined assault types.

Perceived life threat was greater for forcible-only unimpaired victims than combined impaired/incapacitated and force and impaired/incapacitated only groups, and the impaired/incapacitated and force group reported greater peritraumatic fear than the impaired/incapacitated only group, as in McConnell et al.’s (2017) study. In our study, characterological self-blame did not differ by assault type, but behavioral self-blame was higher for the combined impaired/incapacitated and force group than the forcible-only, unimpaired group. This differs from McConnell et al. (2017) who found no self-blame differences among rape type groups but is similar to Littleton et al. (2009) who found that nonimpaired victims had less self-blame than impaired or incapacitated victims, though that study did not include the combination of force and substance use, which is novel in the current study. This reflects our hypothesis that the combined impaired/incapacitated and force group would report worse post-assault outcomes, and future research should investigate if the greater self-blame relates solely to the alcohol/drug use, or if perpetrator behaviors also have an impact. Although the multivariate test was not significant, we did find differences in maladaptive coping, unlike McConnell et al. (2017). Maladaptive coping scores were significantly higher for the combined impaired/incapacitated and force group than both the forcible-only, unimpaired group and the impaired/incapacitated only group, which could mean that recovery outcomes for assaults involving force and alcohol/drug use require more clinical intervention and attention.

The finding that combined impaired/incapacitated and force group victims had more PTSD symptoms than the other groups may be important and is consistent with our hypotheses and other findings in the literature showing these assaults are traumatic (Jaffe et al., 2016) as well as related to higher PTSD symptoms for combined impaired/incapacitated and force rape types (McConnell et al., 2017). Like Littleton et al. (2009) we found no differences in depressive symptoms between groups, though our findings were marginally significant and warrant further investigation. Reexperiencing, numbing, and avoidance symptoms were higher for the combined impaired/incapacitated and force assault types than the forcible only unimpaired group, which is consistent with McConnell et al. (2017). However, McConnell et al. (2017) and Zinzow et al. (2010) found combined rape types with higher PTSD symptoms than other groups, except for numbing/avoidance symptoms which did not differ, so further research is needed to better understand PTSD symptom profiles for different assault types.

In terms of reactions to disclosure, there were no significant differences between groups, like Littleton et al. (2009), although incapacitated victims in that study felt more stigma than others and McConnell et al. (2017) did not report on social reactions in their study. Victims in the incapacitated/impaired only group and the combined group were more likely to discuss drinking in great detail, whereas the forcible-only unimpaired group just mentioned drinking briefly. Other studies have not looked at this variable, but disclosing for impaired/incapacitated victims, regardless of whether or not they report force at the time of the assault, may be riskier and lead to more social reactions of all types than assaults with force but no impairment that may better fit traditional rape stereotypes. Future research should investigate different pathways to disclosure between assault type groups in order to understand specific disclosure patterns.

Limitations

This study had several limitations including a retrospective design in which assaults varied in how long ago they occurred. Future research should replicate this study with victims closer to when their assaults occurred to address this limitation. In addition to the retrospective design, this study relied on self-report data from a sample of victims only, which could impact the generalizability of this study. Furthermore, the fact that the majority of the sample reported completed rapes also limits the generalizability of this study. Results regarding perpetrator behaviors at the time of the assault are also limited given that those data come from the survivor perspective and not the perpetrators. Future research should examine perpetrator behaviors at the time of the assault from the perpetrator perspective.

Assault type groups were larger than past studies, but we could not differentiate impaired from incapacitated victims, as too many cases were overlapping in these categories, though making this distinction is not always representative of a survivor’s experience (Ullman, O’Callaghan, & Lorenz, 2019). Further, our sample size of the impaired/incapacitated only group was low compared to the other two groups, which probably resulted in less statistical significance than expected. This is also likely because we had a more severe assault severity in this study with most women having completed rapes and some level of force used. We also did not include a no-alcohol group, but instead looked at how levels of alcohol’s/drug’s impact (impaired, incapacitated, forcible-only unimpaired) differed between groups all of whom had victim pre-assault drinking and/or drug use. Given that risk factors for and consequences of alcohol related versus non-alcohol related assaults are different (Lorenz & Ullman, 2016a), it is more important to focus on how alcohol related assaults vary in their impacts based on victim appraisals and the co-occurrence of force, than just comparing groups with or without alcohol and/or force, following McConnell et al’s (2017) recommendation. We used Testa et al.’s (2004) revised SES measure which used the original age 14 cutoff used by Koss et al., (1987) to demarcate child sexual abuse from SA in adolescence/adulthood. This is a limitation as this cutoff does not necessarily capture the distinction of child sexual abuse versus adolescent/adult SA from clinical, developmental, or legal perspectives. Finally, while yielding no significant results, our abuse severity variable was skewed, mostly because of the number of completed rapes in the sample. Future research should work to capture a range of assault severity amongst these assault type groups.

Implications for Research and Clinical Intervention

This study is vital as the third study of the role of alcohol/drug impairment and force in outcomes of SAs. Such descriptive research can help to better characterize these very different types of SA and their impacts, so that clinical and prevention strategies can take account of how such assaults vary and develop unique approaches to deal with them. Like previous research, combined assaults have serious consequences in comparison to forcible-only unimpaired assaults as well as impaired/incapacitated assaults. Our study, however, was unique compared with prior research in examining both pre-assault drug and alcohol use for perpetrators and victims. Future research should investigate if the use of a specific intoxicant (or the combination of intoxicants) affects any of the assault, recovery, and disclosure outcomes we analyzed. Qualitative data are clearly needed to better understand how survivors’ experiences in these distinct assault type groups differ and what they need for their recovery process. Additionally, qualitative data that further investigate differences and similarities between being impaired, being incapacitated, and how use of force or threat of force may be used in these types is needed to better understand intoxicant-involved SAs and their consequences for victims. Differences in CSA among assault type groups imply that victimization histories may influence SA risk later in life and could be related to the type of substance-involved assault one experiences. Future research should explore these pathways to see if CSA is a predictor of different substance-involved assault types and associated post-assault outcomes.

Our sample provides implications for both literature and intervention on self-blame in victims, as well as understanding how PTSD symptoms manifest in different types of intoxicant-involved SAs. Unlike previous research, the combined impaired/incapacitated and force group reported higher behavioral self-blame than the forcible-only unimpaired group. The differences in self-blame show that further intervention efforts should work to break down rape myths associated with victims who may be drinking and/or using drugs at the time of the assault. Intervention efforts should not be limited to challenging myths about alcohol-involved assault, but also about voluntary drug use prior to assault. Future researchers should examine how the combination of both being impaired/incapacitated and type of intoxicant (alcohol, drugs, or both) may impact victims’ self-blame in addition to their acknowledgement of the rape, which our study did not address, and how this may impact other post-assault outcomes.

Previous researchers have theorized that victims of incapacitated/impaired only rapes would experience less PTSD symptoms than combined and forcible-only groups, and that combined rape type groups will experience more PTSD symptoms than other rape tactics groups (McConnell et al., 2017; Zinzow, 2010). Our results have also reflected these assumptions; however, this could be because our combined group did contain women who were incapacitated during the assault, which has not been looked at in previous studies. In McConnell et al. (2017), only impaired rapes were combined with force, and our data challenge the assumption that incapacitated rapes cannot include force as well and call attention to the need for further discussion of recovery options for victims who experience combined impaired/incapacitated and force assaults. Additionally, coping differences have not been shown in previous research, but our study found differences in maladaptive coping among assault types that, while marginal, could indicate the need to tailor clinical treatment and interventions for victims who experience a combined impaired/incapacitated and force assault. For example, therapeutic treatments should not focus on reducing self-blame that is more common in assaults where victims voluntarily used alcohol and/or drugs, but also acknowledge that they may have experienced severe physical trauma, and work to address both of these in tandem. Future research should investigate what these maladaptive coping strategies are, and why they may result more from combined impaired/incapacitated and force assault types. It may be more challenging to help victims recover from such assaults, so more data are needed to understand victims’ experiences and how recovery occurs and can be facilitated by therapists and in various treatment settings.