Predicting the Effects of Sexual Assault Research Participation: Reactions, Perceived Insight, and Help-Seeking

Abstract

This study examined effects of participating in survey research for women sexual assault survivors with other trauma histories to understand the role of study participation on reported insight and long-term help-seeking behaviors. A diverse sample of 1,863 women from a large Midwestern city participated in a three-year study on women’s experiences with sexual assault. Regression analyses were conducted to: a) examine predictors of immediate positive and negative reactions to survey participation, and b) assess the impact of the survey on reported insight and women’s long-term help-seeking behavior. Overall most women in the study had a higher positive than negative reaction to the survey (92%), with a significant proportion indicating they sought additional services as a result of participating (55%). Women with CSA, more emotion dysregulation, and more characterological self-blame had more negative reactions to the survey while those with more education and individual adaptive coping had more positive reactions. Women who said they gained insight from answering survey questions were most likely to seek additional help. This study extends the literature by examining cumulative trauma and post-assault symptoms in relation to the effects of survey participation. This is also the first study of women sexual assault survivors to find a relationship between gaining reported insight from research and subsequent help-seeking. Participating in sexual assault research may help survivors gain greater insight into their recovery, which can lead them to seek out more resources for their ongoing trauma-related problems.

Studies of sexual assault survivors have increased over the past thirty years, leading to a separate but parallel body of literature on how participating in these studies affects survivors. Researchers and Institutional Review Boards (IRBs) want to ensure that participating in survey research does not exacerbate the effects of trauma for sexual assault survivors. While past studies reveal several mixed findings (discussed below), research generally shows that most survivors participating in research experience little to no distress, that the small amount of distress experienced by some participants is not long-lasting, and that many survivors report feeling direct personal benefits (see Jaffe, DiLillo, Hoffman, Haikalis, & Dykstra, 2015; Legerski & Bunnell, 2010; Newman & Kaloupek, 2009 for reviews). However, most of these studies have focused on the immediate emotional impact of research on survivors from undergraduate students, but not diverse, community samples. There is also little longitudinal research on whether or not participating in trauma research spurs survivors to seek additional help such as counseling. We seek to address these gaps by examining predictors of reactions to survey research, gaining reported insight from participation, and help-seeking due to participating in a longitudinal community sample of diverse female sexual assault survivors.

Study Participation Effects

Past research demonstrates immediate negative reactions to participation are predicted primarily by histories of adult and childhood sexual and/or physical abuse (Edwards, Probst, Tansil, & Gidycz, 2013; Pederson, et. al., 2014), but can be mediated by severity of symptoms such as depression, anxiety, and PTSD (Massey & Widom, 2013). A more recent study showed that participants who were asked about sexual trauma had a slightly lower positive affect after two weeks compared to those who were only asked about stressful, but not traumatic, life events (Cook, Swartout, Goodnight, Hipp, & Bellis, 2015). Studies also find PTSD to be a unique predictor of negative reactions to participation (Edwards’s et. al., 2013; Newman & Kaloupek, 2004; Massey & Widom, 2013). These results show the need to further clarify the role of post-assault symptomology in how survivors react to research participation, particularly in understanding how correlates of PTSD such as emotional dysregulation predict reactions. No studies to our knowledge have examined whether or not reactions to survey participation are associated with various forms of coping such as active, individual, and maladaptive.

Some studies have also observed differences in reactions based on assault and victim characteristics. One study of college women found assault severity and attributions of self-blame to be predictors of immediate negative reactions (Edwards, Kearns, Calhoun, & Gidycz, 2009). However, another study measuring reactions to participation in trauma research found no differences in immediate reactions based on type and severity of violence and victim gender (Black, et. al., 2006). Most research does not show racial differences in immediate reactions to research participation (Black, et. al., 2006; Edwards, Kearns, Calhoun, & Gidycz, 2009) however a recent study suggests Hispanics and Blacks had more positive and negative reactions to survey participation compared to Whites and that both groups also reported experiencing less regret for taking a survey about sexual assault (Kaasa, Heaton, McAloon, & Cantor, 2016).

Benefits of Participation in Trauma Research

Many women survivors report that distress from participation does not overshadow the perceived benefits of continued participation (Edwards et al., 2013; Edwards, Sylaska, & Gidycz, 2014), such as the chance to tell their stories and feeling that they contributed to science (Cromer, Freyd, Binder, DePrince, & Becker-Blease, 2006; DePrince & Chu, 2008; DePrince & Freyd, 2004; Yeater, Miller, Rinehart, & Nason, 2012). Some of these benefits may be therapeutic, given that some survivors report that they participate as “part of their recovery process” (Ullman, 2010), in addition to other reasons (e.g., financial compensation, helping other survivors; Campbell & Adams, 2009).

Gaining reported insight may be a common and potentially important result of participating in sexual assault studies. In a study of women trauma survivors, 36% indicated they gained self-awareness, with 12% to 15% gaining awareness about trauma or drinking-related moods and behaviors (Pedersen, et. al., 2014). Gaining reported insight and validation through research may even inspire behavioral changes. Edwards, Sylaska, and Gidycz (2014) found that some women who participated in dating violence research reported the study gave them insights strong enough for some to plan behavioral changes, including ending unhealthy relationships.

Gaining insight from survey participation may encourage survivors to reach out for help. Researchers often provide lists of local resources and use non-blaming (or anti-blaming) wording in materials, instructions, and questions. Such reported insight, tangible aid, and reduced self-blame are associated with survivors disclosing and help-seeking (Carretta et al., 2015; Kilpatrick, Resnick, Ruggiero, Conoscenti, & McCauley, 2007; Patterson, Greeson, & Campbell, 2009; Starzynski, Ullman, Townsend, Long & Long, 2007).

The Current Study

The current study extends the literature by examining predictors of emotional reactions to and perceived benefits of participation in trauma research within a large, diverse community sample of survivors. Given the somewhat mixed findings discussed above, this study will examine whether physical and sexual abuse from childhood and adulthood, as well as post-assault symptoms such as depression, anxiety, PTSD, and types of coping predict negative and positive reactions to the survey. As prior research has focused largely on PTSD, we attempt to disentangle the effects of PTSD from other post-assault correlates of PTSD (increased emotional dysregulation, greater depression, less perceived control over recovery, greater self-blame, more maladaptive coping, and increased life threat and other interpersonal traumas, such as witnessing violence as a child; Frazier, 2003; Frazier, et al., 2011; Ullman, Filipas, Townsend, & Starzynski, 2007). While past literature has not examined all of these factors together, we expect them all to be associated with increased negative reactions.

Although reported insight and general positive feelings towards participation (e.g., feeling a study will contribute to science) are usually combined, we examine them separately based both on the potential theoretical impact of reported insight stated above and statistical analyses discussed below. Studies have not examined correlates of reported insight gained from survey research, yet we expect participants with a history of actively processing their experiences (i.e., using active individual coping strategies and experiencing posttraumatic growth such as new realizations or reappraisals), who use less maladaptive avoidance coping, and those who have greater ease in processing their emotions (i.e., less emotion dysregulation) to report gaining more insights. Lastly, we explore whether any particular subgroups (e.g., those with traumas or certain patterns of distress) are more likely to gain insight, but make no a priori hypotheses. We acknowledge the limitations of measuring insight with only one question, but we believe it is valuable to measure as there is very little prior research that assesses insight facilitated by survey participation.

Given that prior studies show that research participation is often a positive disclosure experience, we expect participants to report that research participation increased their chances of subsequent disclosure and help-seeking related to assault. Acknowledging a sexual assault as rape (a type of insight) is the biggest predictor of survivors seeking help from formal service providers such as counselors or other mental health providers (Walsh et al., 2015), we therefore expect survivors who report gaining insight into their experiences during the survey will be more likely to report help-seeking as a result of participation. Finally, we expect characteristics that make survivors more likely to disclose or seek help in general will also make them more likely to seek help due to research participation (for reviews of facilitators and barriers of sexual assault disclosure; see Ullman, 2007; Kennedy, et al., 2012). Such factors include having more stereotypical assaults involving strangers, violence, injury, or life threat (Orchowski & Gidycz, 2012; Walsh et al., 2015; Wolitzky-Taylor et al., 2011), a history of disclosing to more people or seeking emotional support to cope (Carretta et al., 2015), receiving more positive and fewer negative reactions to disclosure (Relyea & Ullman, 2015), and having a White racial identity as compared to having a African-American racial identity or other non-White identities (Kennedy, et. al., 2012; Starzynski, Ullman, Townsend, Long, & Long, 2007). Possibly due to greater access to resources, women with higher education engage in more mental health seeking (Ullman, 2007).

Method

Participants

The data came from a 3-year annual longitudinal survey study of 1,863 women who had an unwanted sexual experience as an adult that they told someone about. Participants ranged in ages from 18 to 71 (M = 36.51, SD = 12.54). The sample was racially and ethnically diverse: 45% African-American, 35% White, 2% Asian 7%, multiracial and 11% other, unknown, or unreported; 13% reported Latina or Hispanic ethnicity (assessed separately). While 32% of women reported having a college degree, 42% had some college education, and 26% had a high school degree or less. Less than half were employed (43%) and a majority had household annual incomes below US$30,000 (67.9%). The study response rate was 85%. Of the 1,863 women at Wave 1, the follow-up rate at Wave 2 was 72% and at Wave 3 56%. Participants who completed all three waves were slightly older (M = 37.88, SD = 12.72) compared to those who dropped out (M = 34.89, SD = 12.13). However, they did not differ by race, Hispanic ethnicity, sexual orientation, education, income, employment status, parental status, or marital status.

Procedure

Participants were recruited to fill out a 28-page paper and pencil mail-in survey asking about stressful and traumatic life events through posted fliers, local newspaper advertisements, Craigslist, and university mass emails. Fliers were posted throughout the community, at Chicago colleges and universities, community organizations (e.g., cultural centers, substance abuse agencies, domestic violence/rape crisis centers) and women-oriented businesses. Women who called were screened with the following criteria: a) unwanted sexual experience at age 14 or older, b) 18 or older at time of participation, and c) previously told at least one person about their unwanted sexual experience. Eligible participants received study materials including the survey, informed consent, stamped return envelope for their completed survey, and a list of community resources for trauma, victimization, mental health, and substance abuse. Participants were paid $25 after returning their surveys, which took approximately 45–60 minutes to complete based on pre-testing. The University’s IRB approved all procedures and documents.

Measures

Unless otherwise noted, all measures below were assessed during Wave 1.

Demographics

Income was included in earlier models, but was left out of the final models due to repeated non-significant findings and close relation to education. Race was trichotomized as Black, White, and Other with Black as the reference category due to sample sizes. Other includes women who identified as Asian, American Indian, Native Hawaiian, Pacific Islander, or who self-identified as Other. Race was not included in the cross-sectional analysis due to repeated non-significant findings but was included in the prospective analyses. Education was assessed with four ordinal categories: (less than 12^th grade, high school graduate or GED, some college, college graduate or more).

Traumatic life events

Unwanted sexual experiences were assessed with a revised version of the Sexual Experiences Survey (SES-R; Testa, VanZile-Tamsen, Livingston, & Koss, 2004). Almost all women reported sexual assault as adults (12.4% unwanted sexual contact or coercion and 86.6% attempted or completed rape). On average, participants experienced unwanted sexual experiences 14 years prior to taking the survey (SD = 12.22, Median = 11). Victimization experiences under age 14 were dichotomized with 66.1% reporting a history of child sexual abuse (CSA). History of traumas was assessed with the revised Stressful Life Events Screening Questionnaire (SLESQ-Revised; Green, Chung, Daroowalla, Kaltman, & DeBenedictis, 2006). The measure included child abuse and adult violence experiences, stalking (Logan, personal communication, 2007) and a question we added on exposure to neighborhood/community violence, “Have you ever lived in a neighborhood or community where you felt threatened or your life was in danger?”

Assault characteristics

We dichotomized answers to questions about survivors’ “most serious unwanted sexual experience” (if more than one) including whether they perceived their life was in danger at the time (with 58% saying yes), drinking prior to assault (with 31% reporting alcohol use), relationship to perpetrators (with 27% saying strangers), perpetrator use or threats of physical violence (with 78% reporting violence).

Psychological symptoms

Posttraumatic stress symptoms were measured using the 17-item Posttraumatic Stress Diagnostic Scale (PDS; Foa, 1995). Participants rated items on a scale of 0 (not at all) to 3 (almost always) regarding how often they experienced each symptom related to the assault in the past year. The items were summed to reflect the extent of posttraumatic symptoms (α= .93, M= 21.13; SD= 12.93). 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 over the past 12 months on a 5-point Likert scale ranging from 0 (never) to 5 (always). Items were averaged (α= .86, M= 2.00, SD= 0.75). Emotional dysregulation was measured with 6 items suggested by Messman-Moore (personal communication, May, 2010) from the Difficulties in Emotion Regulation scale (DERS; Gratz & Roemer, 2004) assessing difficulties with managing arousal and regulating one’s emotional state rated on a scale of 1 (almost never) to 5 (almost always). Items were averaged (α= .75; M= 2.76, SD= .94). Perceived control over recovery was measured using the seven-item Present Control subscale of the Rape Attribution Questionnaire (RAQ; Frazier, 2003). Women rated their level of perceived control over recovery in the past year on a scale from 1 (strongly disagree) to 5 (strongly agree). The scale was averaged (α = .70; M = 3.60, SD = 0.78). Characterological self-blame was also assessed with RAQ on a five-item scale from 1 (strongly disagree) to 5 (strongly agree) with items averaged (α= .76, M= 2.56, SD= 0.96).

Coping

The Brief COPE is a 28-item self-report scale of coping strategies (Carver, 1997) used to assess past-year assault-related coping on a scale from 1 (I didn’t do this at all) to 4 (I did this a lot). Based on factor analyses, we divided coping strategies into Maladaptive coping (8 items including behavioral disengagement, denial, self-blame, and substance-use (α = .86; M= 16.3, SD= 5.77), Adaptive individual coping (12 items including active coping, venting, positive reframing, planning, acceptance, and religion; α = .83, M = 29.2, SD = 7.78) and Adaptive social coping (4 items measuring instrumental and emotional coping; α = .87, M = 9.05, SD = 3.72). Posttraumatic growth was assessed with a 10-item short form of the posttraumatic growth inventory (PTGI-SF; Cann, Calhoun, Tedeschi, Taku, Vishnevsky, Triplett, & Danhauer, 2010). Items were averaged (α = .92, M = 2.15, SD = 1.10).

Social reactions and disclosure experiences

The Social Reactions Questionnaire (SRQ; Ullman, 2000) uses a 48-item scale to assess positive and negative reactions from others when disclosing assault on a Likert-type scale from 0 (never) to 4 (always), Three primary scales (Relyea & Ullman, 2015) were coded: reactions of turning against the survivor (13 items; M = .81, SD = .93, α = .92), reactions of unsupportive acknowledgment (13 items; M = 1.11, SD = .84, α = .85), and positive reactions (20 items; M = 2.23, SD = 0.95, α = .92). We asked survivors to whom they disclosed their sexual assault with over half (59%) responding they told a formal source (e.g., religious, mental health, medical, police, rape crisis) and 97% an informal source (e.g., friend). Finally, survivors reported the number of people they told. To maintain maximum variability and account for skew, numbers were categorized into a 6-item ordinal scale (28% told 1–2 people, 21% told 3, 25% told 4–5, 12% told 6–9, 9% told 10–14, and 5% told 14+).

General social support

Frequency of social contact (e.g., getting together with friends) was measured with five questions from the RAND Health Insurance Experiment (Donald & Ware, 1984). Questions were asked on scale from 1 (less than 5 times during the past 12 months) to 7 (every day). Items were averaged (α = .70, M = 3.74, SD =1.19). Perceived social support was assessed with the Social Support Questionnaire Short Form Revised (Sarason, Sarason, Shearin, & Pierce, 1987), 6 (no/yes) questions asked whether they can count on or depend on others. Items were summed (α = .84, M = 5.28, SD = 1.44).

Problem drinking and drug use

Number of alcohol problems over the past 12 months was assessed with the 25-item Michigan Alcoholism Screening Test (MAST, Selzer, 1971). Those not drinking in the past year were given a zero. Items were summed (M = 2.88, SD = 4.20, α =. 89). Past-year problem drug use was assessed with a modified, short version of the Drug Abuse Screening Test (DAST-10, McCabe, Boyd, Cranford, Morales, & Slayden, 2006), which was reliable in our sample, α= .86 (M=2.38, SD= 2.75); 35% met criteria for problem drug use.

Reactions to research participation

Seven items from the Reactions to Research Participation Questionnaire Revised (RRPQR) were rated on a 5-point Likert scale from strongly disagree to strongly agree (Newman, Willard, Sinclair, & Kaloupek, 2001). Based on similar research looking at benefits and harm, we divided items into Positive and Negative Reactions, with 3 items each. The Positive Reactions scale (M = 4.36, SD = .62, α = 0.63) included the items: I like the idea that I contributed to science; I believe this study’s results will be useful to others; and Had I known in advance what participating would be, I still would have participated. The Negative Reactions scale (M = 2.60, SD = .90, α= 0.67) included: I found the questions too personal; Completing this survey upset me more than I expected; and The research made me think about things I didn’t want to think about. One item measuring reported insight, I gained insight about my experiences through research participation, was analyzed separately based on the theoretical reasons mentioned above, an investigation of the scree plot, and the results of an exploratory factor analysis using principal axis factoring with promax rotation that showed factor loading less than .3 on other factors. Higher scores indicate more reactions and insight.

Research-inspired communication and help-seeking

To assess research-inspired communication and help-seeking, we added a question to Waves 2 and 3 asking “Did you do any of these things as a result of participating in this research?” with response categories: “Talk to someone about your unwanted sexual experience(s)”, “Seek help for problems related to an unwanted sexual experience”, “Seek help for problems related to alcohol or drug use”, and “None of the above.” Answers were dichotomized (no/yes) if each was reported at Waves 2 or 3.

Analysis Plan

First, we examined the immediate emotional impact of the survey reactions at Wave 1 including negative reactions, positive reactions, and reported insight. Preliminary analyses involved computing descriptive statistics for outcome variables and bivariate correlations among hypothesized predictors. To test our hypotheses, we ran linear regressions for positive and negative reactions and a logistic regression for reported insight (dichotomized as 1=agreed or strongly agreed to gaining insight, 0= strongly disagreed, disagreed, or neutral). We decided to categorize reported insight in this manner as violation of proportional odds prohibited use of ordinal regression and a multinomial regression showed no differences between those who marked neutral and disagree.

Second, we switched to prospective analyses, using logistic regressions to determine which Wave 1 variables predicted research-inspired behavioral effects (communication and help-seeking) across Waves 2 and 3. Because this was the first study to examine predictors of research-inspired behaviors, we performed Bonferroni corrections to protect from Type I error. If predictors had no significant correlations with any behaviors at the corrected alpha level of .00068 (.05/74), we did not include the predictor in the regressions.

Results

Similar to past studies, immediate reactions to the survey were mostly positive. Recall that we used scales to measure positive and negative reactions wherein a higher score on each scale indicates a more positive or negative reaction. In our sample, the mean score on the positive reaction scale was higher (M = 4.36) than the mean score on the negative reaction scale (M = 2.34). A large majority (92%) of participants had higher positive than negative scores on the reaction to participation scales. In terms of negative reactions, 20% of women agreed the survey upset them more than they expected and 11% found questions too personal. While 48% agreed that they experienced having to think about things they didn’t want to think about, only 4% (81 women) indicated they would not have participated had they known what the experience would be like for them. And only 2 of those 81 stated they did not wish to participate in follow-up surveys or interviews. Conversely, 91% agreed the study’s results would be useful to others and 89% liked that they contributed to science. Slightly more than half of the women (58%) reported they gained insight from the research.

Bivariate Correlations

Because of the large sample size, most correlations (any over r = .047) were significant. Therefore, we describe them using Cohen’s (1988) criteria (.10 = weak, .30 = moderate, .50 = strong). Raw correlations are not reported but are available upon request. No predictors had correlations above .66 and only two were above .60; therefore, collinearity did not appear to be a problem. At Wave 1, negative reactions to the survey had a moderate correlation with PTSD symptoms and weak associations with more depression, self-blame, maladaptive coping, turning against and unsupportive acknowledgment social reactions, life traumas, and life threat during assault, problem drug use, and less perceived social support. Both positive reactions and insight were weakly correlated with more perceived control, adaptive individual coping strategies, and positive disclosure-related social reactions. Insight was also weakly related to more interpersonal coping strategies and perceived social support. Positive reactions were weakly negatively correlated with negative reactions (r = −.15) and positively related to insight (.17). Insight and negative reactions were not significantly correlated.

We next examined correlations between Wave 1 predictors and research-inspired behaviors reported across Waves 2 and 3. Based on nonsignificant associations at the corrected .00067 level, we removed eight hypothesized predictors (perpetrator violence, drinking prior to the assault, characterological self-blame, negative and positive reactions to the Wave 1 survey, the number of people previously disclosed to, frequency of social support, and perceived social support). Research-inspired help-seeking for substance use was moderately related to Wave 1 problem drug use, alcohol problems, and prior substance abuse treatment. All of the remaining predictors had weak associations with behaviors.

Immediate Reactions to Participation

Multiple linear regressions of negative and positive reactions to the survey showed that few predictors were significant after controlling for other variables (see Table 1). In partial support of our hypotheses, immediate negative reactions were predicted by history of child sexual abuse, more emotion dysregulation, and more characterological self-blame (F (12, 630) = 10.06, p < .00, Adjusted R² = .145). More education and more adaptive individual coping predicted positive reactions (F (12, 635) = 2.32, p < .007, Adjusted R² = .024).

Table 1.

Regression for positive and negative effects of research participation

	Negative reactions			Positive reactions
	B	SE	β	B	SE	β
Education	−.030	.041	−.031	.070	.028	.109*
Stressful Life Events	.012	.012	.045	.006	.009	.033
Life Threat	.100	.073	.056	.000	.051	.000
Trauma: CSA	.171	.074	.093*	−.052	.052	−.043
PTSD	.005	.004	.081	−.003	.003	−.065
Depression	.024	.057	.021	.018	.040	.024
Emotional Dysregulation	.105	.050	.115*	.059	.035	.098+
Perceived Control	−.053	.048	−.048	.032	.034	.044
Self-blame	.134	.038	.148**	−.008	.027	−.013
Maladaptive Coping	.000	.008	.000	−.004	.006	−.040
Individual Coping	.002	.006	.019	.012	.004	.164**
Social Coping	−.011	.010	−.049	−.006	.007	−.039

Note.

⁺p<.10,

^*p < 0.05,

^**p < 0.01.

The logistic regression for reported insight showed predictors were mostly in support of our hypotheses (Table 2). Participants using more adaptive individual coping, prior posttraumatic growth, and higher education were more likely to report gaining insight from the survey. Against expectations, emotion dysregulation was positively associated with reported insight. We do not know what specific insights participants had or what parts of the survey triggered insights, but neither distress, trauma, alcohol abuse problems, nor perceived life threat were related to gaining insight.

Table 2.

Logistic Regressions for Gaining Insight and Help-Seeking due to Survey Participation

	Gaining Insight		Talking to Someone		SA Help		Substance Use Help
	OR	CI	OR	CI	OR	CI	OR	CI
Age	0.99	(0.98, 1.00)	0.98	(0.97, 1.00)	1.00	(0.98, 1.02)	1.03*	(1.01, 1.06)
Education	1.23*	(1.04, 1.45)	0.74**	(0.60, 0.90)	0.87	(0.70, 1.09)	0.69*	(0.51, 0.95)
Race: White	1.14	(0.82, 1.58)	0.94	(0.63, 1.39)	0.76	(0.47, 1.22)	0.47*	(0.23, 0.98)
Race: Other	0.86	(0.59, 1.26)	1.14	(0.71, 1.83)	1.47	(0.87, 2.48)	1.02	(0.46, 2.23)
Assault Characteristics
Stranger Assailant	0.85	(0.63, 1.16)	0.90	(0.61, 1.31)	1.21	(0.79, 1.85)	0.97	(0.53, 1.77)
Life Threat	0.87	(0.64, 1.18)	1.48*	(1.02, 2.16)	1.52t	(0.96, 2.42)	1.09	(0.55, 2.16)
Post-assault Symptoms
Depression	0.95	(0.76, 1.19)	0.92	(0.70, 1.21)	1.24	(0.90, 1.71)	0.92	(0.58, 1.48)
PTSD	1.00	(0.99, 1.02)	1.02+	(1.00, 1.04)	1.01	(0.99, 1.03)	1.00	(0.97, 1.03)
Emotional dysregulation	1.21*	(1.00, 1.47)	0.92	(0.73, 1.17)	1.09	(0.83, 1.44)	0.70	(0.46, 1.07)
Coping
Maladaptive	1.00	(0.97, 1.03)	0.99	(0.95, 1.03)	1.01	(0.97, 1.06)	1.02	(0.95, 1.09)
Individual	1.03*	(1.00, 1.05)	1.02	(0.99, 1.04)	0.99	(0.96, 1.03)	1.02	(0.97, 1.07)
Social	1.00	(0.96, 1.04)	1.01	(0.95, 1.07)	1.07t	(1.00, 1.15)	1.01	(0.91, 1.11)
Support Experiences
Turned Against	1.01	(0.96, 1.06)	1.07	(0.82, 1.4)	1.19	(0.89, 1.59)		(0.73, 1.62)
Unsupportive	1.08	(0.81, 1.45)	1.16	(0.83, 1.61)	1.01	(0.7, 1.46)	1.28	(0.75, 2.18)
Positive	1.04	(0.76, 1.44)	1.32*	(1.06, 1.65)	1.14	(0.87, 1.48)	1.22	(0.84, 1.77)
Told Formal Sources	1.00	(0.97, 1.04)	0.72+	(0.51, 1.03)	1.62*	(1.04, 2.53)	1.50	(0.80, 2.83)
Survey Insight	1.42	(1.22, 1.65)	1.26**	(1.06, 1.48)	1.41**	(1.14, 1.74)	1.03	(0.77, 1.38)
Post-traumatic Growth	0.99***	(0.98, 1.00)
Stressful Life Events	1.23	(1.04, 1.45)	1.02	(0.96, 1.09)
Trauma: CSA	1.14	(0.82, 1.58)	1.01	(0.91, 1.13)
Drug Abuse	0.86	(0.59, 1.26)					2.83**	(1.52, 5.27)
Alcohol Abuse							1.12***	(1.06, 1.19)
Substance Use Treatment							4.43***	(2.35, 8.37)
Nagelkerke R2	.08		.16		.19		.45

Note. Race variables (White and Other) use African-American as reference category.

⁺p<.10,

^*p < 0.05,

^**p < 0.01.

Prospective Analyses of Survey-Motivated Disclosure and Help-Seeking

Across the two years of follow-up surveys, 55% (n = 737) of participants reported help-seeking or talking to others as a result of participating in research: 47% (n = 629) spoke with someone about their sexual assaults, 18% (n = 241) sought help for problems related to their assaults, and 14% (n = 191) sought help for drug or alcohol problems; 35% (n = 260) did one or more of these at both waves. Contrary to expectations, most predictors for help seeking in the literature did not predict research-inspired help-seeking or disclosure across waves 2 or 3 (see Table 2). Those who reported gaining insight at Wave 1 and those who had told more formal treatment providers (e.g., religious, mental health, medical, police, rape crisis) about their assault in the past were more likely to be motivated by research participation to seek help for sexual assault-related problems. Against expectations, education was not predictive of research-inspired help-seeking. Research-inspired help-seeking for substance abuse was only predicted by Wave 1 substance problems and prior drug treatment along with demographics factors (being older, having less education, having Other racial identity relative to Black racial identity). Finally, research inspired participants to talk to others about their sexual assaults across the years of follow-up if they reported gaining insight at Wave 1, had prior positive experiences disclosing, experienced more life threat during the assault, had less education, and were younger.

Discussion

This study’s goal was to understand how participating in survey research affects sexual assault survivors. Specifically, we assessed how participants perceived the study, their immediate emotional reactions and insight, and whether survivors felt participating led them to discuss their assaults with others or seek help for problems. Overall, research participation appears to be a positive and potentially therapeutic experience as 92% reported more positive than negative reactions, and more than half (58%) reporting gaining insight from research. Over two years of participating, 55% reported the study led them to discuss their assault with someone or seek help for assault-related problems or substance use.

Reactions to survey research appeared to depend on survivors’ characteristics, history, and coping. Those with more education had more positive reactions to the survey, possibly related to their increased likelihood of mental health help-seeking and therefore having dealt with trauma adaptively (Ullman, 2007). Survivors’ preferred coping strategies were related to their reactions to participation, with individual coping related to more positive reactions. Researchers should provide options such as onsite counseling or local resources for participants who feel triggered by research questions as many will have developed active coping strategies and potentially use the resources.

Self-blame, CSA, and emotion dysregulation were associated with more negative reactions. These findings warrant future study to determine how to design instructions and questions to avoid blame. Studies could also examine how various standardized assessments affect self-blame. The relationship between CSA and negative reactions replicates previous findings (Black, Kresnow, Simon, Arias, & Shelley, 2006; Edwards, Probst, Tansil, & Gidycz, 2013; Pederson, et. al., 2014). Finally, it is possible those with emotion dysregulation are more likely to be overwhelmed by the survey questions and have more negative reactions. It is also possible that those with more emotion dysregulation experience more problems post-assault and therefore believe strongly in the importance of sexual assault research. Taken together, these results underscore the need to carefully craft questions and provide survivors with resources for mental health, as many may have residual symptomology.

Our findings indicate that participating in research can inspire participants to tell others about their assaults and seek help for problems. Although the findings should be treated with caution as they used novel single-item measures as outcomes, the findings warrant future study given the high numbers of women who reported such outcomes. Notably, gaining insight at Wave 1 was prospectively related to assault-related help-seeking and disclosure over the next two years. As hypothesized, participants with more education, histories of post-traumatic growth, and more individual coping skills reported gaining more insight. Thus, some participants may have a higher tendency for meaning/sense-making and metacognition (thinking about their own processing) while taking surveys. We had expected those with emotion regulation difficulties to have more trouble understanding emotions and thus less insight. Yet emotion dysregulation was related to having more insight. It is possible that participants gained insight into their emotion regulation difficulties. Alternatively, those who have trouble inhibiting distressing emotions during the survey may have more opportunities for thinking about and possibly gaining insight into what upsets them. This should be assessed in future research. Although we did not have hypotheses regarding other variables, insight did not appear to be related to any specific pattern of distress, substance abuse problems, coping, or assault characteristics.

We also found that participating motivated nearly half of survivors to talk to others about their assault. In addition to reported insight, participants with positive prior disclosure experiences and those who felt their life was threatened during the assault were more likely to talk to others. Contrary to the literature on help-seeking, those with more education were less likely to talk to others about their assaults. Given that the social reactions survivors receive from others impact survivors’ distress and recovery (Ullman, 2010); future qualitative research should ask why research inspires participants to talk to others, who they spoke with, and what they discussed. Our findings suggest survivors who gain insight might be likely to make new disclosures given that 57% of those who strongly agreed to gaining insight at Wave 1 disclosed to someone over the next two years, those who agreed or were neutral were less likely to disclose (49%, 50%, respectively), and those who disagreed or strongly disagreed were even less likely (40%, 42%, respectively).

Although the literature reports many predictors of help-seeking, it is perhaps not surprising that the only predictors of help-seeking for problems related to the assault as a result of survey participation were previously telling formal help providers (e.g., religious, mental health, medical, police, rape crisis) and gaining insight from completing the survey. The weak correlation of help-seeking with many variables implies that participants may be seeking help for a diversity of reasons.

Demographic factors and substance-use were the only predictors of substance-use help-seeking. We found that having Other racial identity relative to a Black racial identity predicted substance use help-seeking. These demographic findings are a departure from the literature on help-seeking that finds White women seeking help more than women of color. In our sample, women in the Other racial category identified as Asian, American Indian, Native Hawaiian, Pacific Islander, or self-identified as Other. The broader literature does not focus on women with the above racial identities. These racial differences should be examined further in qualitative research with survivors and in studies that separate substance use help-seeking from other types of help-seeking. Insight may not have predicted substance-use help-seeking as such survivors may already be aware of their substance-use problems. Thus, the survey may merely serve as a reminder to seek help.

Limitations

Our model explained a small amount of variance in participants’ negative and positive reactions to the survey. This makes sense as we would expect the main predictors of reactions to a survey to be characteristics of the survey (e.g., length, content) and we had no assessment of how participants responded to such characteristics. Also, given that we only used selected items from the RRPQR instead of the full scale, we may have had different results or explained less variance. We were also limited with regard to our question on insight as we attempted to measure a construct with only one item. There may have been an educational discrepancy in interpreting the concept of insight, with those with more education being able to more confidently answer the question. We used a volunteer sample that may have been more prepared for answering questions about trauma, and some may even have seen such participation as part of their recovery process as in prior similar studies (Ullman, 2010). Thus, they may report lower levels of negative reactions to research participation. The paper and pencil format of the survey excluded participants with literacy issues. Regarding disclosure and help-seeking, the study only included women who had previously disclosed sexual assault. Yet, this should not bias findings as most survivors disclose (Fisher, Daigle, Cullen, & Turner, 2003). The sample contained a diverse and nearly equal mix of White and Black participants (with slightly more Black women in the sample) but was limited in representations of Hispanic/Latina, Asian, and American Indian women.

Research Implications

To our knowledge, this is the first study to attempt to quantify the impact of research participation on sexual assault survivors’ help-seeking and assault-related disclosures. Such impacts were common in our sample. This study will aid researchers preparing applications for institutional review boards, who should be made aware of these findings. Colleges that are beginning to survey students about sexual assault should be prepared to increase resources to meet the likely increase in demand for services. Research studies with diverse samples such as Black women and other women of color should thoroughly prepare resource lists that are culturally relevant. This also underscores the need to continue to advance access to post-assault resources, especially for underserved populations. Among those in our sample who were likely in college (graduated high school, already in school, under age 25), 56% were inspired by research over the two years of follow-up to talk to someone or seek help for problems related to their sexual assault (20% sought help for assault-related problems and 10% for substance use problems). Future research should also more thoroughly assess reported insight by asking more questions and/or using different language so as to be more easily understood by participants. We found evidence of a connection between research-inspired reported insight and future help-seeking behavior, but we do not know what specifically may have triggered the feeling of insight or what types of insights participants felt they had. More research should, particularly of a qualitative nature, should focus on reported insight as an effect of study participation. Longitudinal research is needed to see if participants’ appraisals of research participation effects vary over time and how intervening experiences may impact these perceptions. Other factors such as mode of assessment and frequency of assessment should be examined in longitudinal studies to determine if these factors influence the duration of effects from participating in trauma research.