Abstract
The current arricle describes a novel recruitment protocol for collecting data from sexual assault and intimate partner violence survivors referred to research studies by individuals to whom they had previously disclosed. Challenges in both recruiting participants and interpreting data are described. Only 35.8% of cases had usable data for both survivors and disclosure recipients, suggesting that this referral method had limited success in recruiting matched pairs. Suggestions for modifications to improve the protocol for future research are offered. Potential advantages and drawbacks of various methods for recruiting dyads are described in order to facilitate future research on survivors’ disclosure processes, social reactions, and the influence of social reactions on survivor recovery.
Sexual assault (SA) and intimate partner violence (IPV) are major public health concerns and have been the focus of feminist scholarship. Research suggests that most survivors tell another person—usually a friend (Dworkin et al., 2019; Sylaska & Edwards, 2014)—about SA and IPV victimization. Similarly, many college students (42–51%) report being the recipient of a SA and/or IPV disclosure (Edwards & Dardis, 2020; Paul et al., 2014). Unfortunately, many survivors receive negative social reactions (e.g., disbelief, blame) from the people they tell (Ullman, 2010). These negative reactions are linked with worse mental health outcomes for survivors (Dworkin et al., 2019) and may also impact the relationship between the survivor and disclosure recipient (DR) in various ways (e.g., causing strain, codependence, role-reversal, and/or enhancing closeness; Ahrens & Aldana, 2012; Lorenz et al., 2018). Thus, research identifying antecedents of negative social reactions and strategies to prevent them or mitigate their harm is sorely needed.
Most research on SA and IPV disclosure and reactions has focused on samples comprised entirely of either survivors or of DRs. This work has provided important information about the experience of disclosing and receiving responses among survivors (see Ullman, 2010 for a review) and DRs’ experiences of hearing about and responding to survivors’ disclosures (Banyard et al., 2010; Christiansen et al., 2012; Edwards & Dardis, 2020). However, by only measuring one person's perception of support behavior provided or received, researchers only capture part of this inherently relational process. This lack of multiple perspectives may limit our understanding of the nuances and complexities involved in survivor disclosure processes and DRs’ responses to disclosures, all of which likely impact survivors’ recovery and are thus important to understand. However, recruiting dyads is inherently more methodologically complex than recruiting individual participants. To inform the methodological decision-making of researchers who wish to conduct such research, the current article describes a novel recruitment protocol used to collect data from SA and IPV survivors who were referred to the study by DRs. The challenges encountered in carrying out this protocol are discussed and situated within the broader literature on dyadic research with survivors and their informal supports (e.g., family, friends, romantic partners). In addition, most of the prior literature in this area has focused on SA, so this study expands to include a closely related form of victimization, that of IPV.
The Importance of Dyadic Research
Dyadic research represents a critically needed methodological advance in research on SA and IPV disclosure. Without research from dyadic matched pairs of survivors and DRs, it is difficult to understand similarities and differences in survivors’ and DRs’ perceptions of disclosure experiences. This dyadic research is especially important because close-relationship researchers have established the importance of perceptions in dyadic interactions and relationships. That is, the way each partner perceives an interaction can be more important than the objective reality that occurred (Rodriguez & Neighbors, 2015; Ward et al., 1997). Additionally, the way each partner perceives the exact same interaction may differ as a function of preexisting beliefs, attitudes, experiences, and expectations. These discrepancies between partners’ perceptions and attributions each makes about the other may be important for understanding these interactions and the outcomes of these interactions (e.g., survivor mental health; see Falconier & Kuhn, 2019 for a review).
Dyadic designs allow for investigation into divergent perceptions based on role (i.e., survivor, DR) and can answer questions such as, to what degree do DRs’ reports of reactions provided match survivors’ reports of reactions received? Does the amount of disagreement vary as a function of the type of reaction? Are survivors looking for certain types of supportive responses and are DRs engaging in these specific types of responsive behaviors? Do perceptions differ based on individual-level characteristics each person brings to the relationship or interaction (e.g., previous victimization, coping skills)?
In addition, increasingly, it has been recognized that disclosure is not necessarily a one-time event; rather, it may occur over multiple conversations with the same DR (Smith, 2005). Thus, in addition to exploring individual-level features that affect perceptions of support during a disclosure, a dyadic framework can also clarify dynamic relationship-centered characteristics (e.g., perceived relationship quality, strain, asymmetry). These characteristics may influence how disclosure processes unfold over time and their influence on subsequent relationship rapport and closeness.
Review of Dyadic Studies on SA and IPV Disclosure
Dyadic data collection has been an important research tool in many fields examining close relationships. Research on romantic relationships, parent–child relationships, and friendships have all successfully used these methods However, data collection with survivor-DR dyads is more challenging from a methodological and ethical standpoint compared to these other fields. Survivor-DR dyads are not socially distinct in the same way as romantic relationships and parent–child relationships. These pairs are diverse in their relationship type (e.g., family, friends, coworkers), and a pairs' status as a survivor-DR dyad is not objectively observable by outsiders (e.g., researchers). Thus, it is likely more difficult to find survivor-DR pairs than many other types of dyads, as they cannot be easily recruited simultaneously (i.e., one must recruit the other person in the dyad). This recruiting of another person raises ethical issues about protecting the survivors’ privacy and agency.
To date, there are few dyadic studies of survivor-DR pairs, and all first initiated contact with survivors before recruiting DRs. For example, Campbell (2005) interviewed 81 rape survivors seeking emergency medical care before their hospital discharge about what services they received and how they were treated by formal professional support sources (e.g., doctors, nurses, police) with whom they interacted at the hospital. Analyses showed significant agreement between survivors’ perceptions and those of legal and medical personnel about what services were and were not provided and whether these personnel engaged in “secondary victimization” behaviors (e.g., statements/actions that could be distressing for survivors). Police and doctors underestimated the impact they were having on survivors. They perceived that survivors were less distressed than survivors themselves reported they actually were. Conversely, nurses were more consistent with survivors in their assessments of survivor distress.
In another study, Davis and Brickman (1996) interviewed 128 survivors of recent sexual and non-SA (64% of which were rapes) who had reported to a victim services agency in New York City. Supportive behaviors reported by survivors and DRs included emotional support, help with daily responsibilities, and encouraging survivors to seek professional help. Unsupportive behaviors included withdrawal or hostility, blaming the survivor, and egocentric actions. Both survivors and significant others agreed that significant others engaged in supportive behavior more often than they engaged in unsupportive behavior. However, there was a discrepancy between survivors and significant others in the reported frequency with which significant others engaged in unsupportive behavior: survivors reported that this occurred more frequently than significant others.
More recently, an in-depth qualitative study of 45 survivor-DR dyads was conducted as part of a larger longitudinal study of PTSD and alcohol problems in SA survivors (Ullman et al., 2017). Survivors in the study who were interested in participating in interviews provided the researchers with contact information for a DR, whom the researchers then contacted and invited to participate in an interview (58 survivors and 45 matched DRs completed interviews). This study also showed that survivors and DRs did not always agree on their appraisals of reactions. Specifically, sometimes survivors appraised social reactions more negatively than DRs or vice versa. In other cases, survivors felt that even if a social reaction was negative, it was well-intended, thus reappraising it to be positive.
Preliminary evidence shows that dyadic matched pairs sometimes disagree on assessments of the amount of unsupportive behavior post-assault, in particular (Davis & Brickman, 1996; Lorenz et al., 2018). Thus, research recruiting dyads of victims and DRs with different recruitment referral methods are needed to evaluate these patterns further and link them to recovery outcomes in survivors and impacts on their informal network members (e.g., family, friends, romantic partners) as well as their relationship outcomes. This research is also needed to inform interventions seeking to improve survivor and supporter experiences of disclosure and support provision post-assault. Recruiting such dyads can be difficult; the current article contributes to the body of knowledge about how to collect the data we need to understand these processes. The next section reviews the recruitment methods used in previous studies, plus some using referrals to recruit one dyad member via the other.
Methodological and Ethical Issues of Dyadic Referral Studies of SA/IPV
Recruitment methods differed in the dyadic disclosure studies reviewed above. First, Campbell (2005) recruited participants through a hospital setting. When survivors presented for care, hospital staff modified their typical protocol of contacting the police and survivors’ advocates to additionally contact the research team. This required the research team to be staffed at all times with an on-call researcher. The researcher waited for an average of 4.34 h until the conclusion of the survivor's medical and legal services. Then, a nurse asked if the survivor was willing to answer some questions about the services they had received; if so, the research team collected the survivors’ data. The researchers then sought a release of information from survivors to interview all formal support sources who had worked with the survivor (e.g., doctors, nurses, police) about their interactions. Since the researchers were embedded at the hospital, in most cases, they were able to immediately conduct interviews with the professional support sources who had worked with the survivor. However, there are substantial logistical challenges in this approach given the need for on-call research staff who were able to wait multiple hours at the hospital until the survivor was available, and then subsequently conduct multiple interviews. This was an extremely difficult study, revealing why dyadic research in this population is relatively rare.
By contrast, Davis and Brickman (1996) recruited their dyads through a victim services agency in New York City from a larger sample of survivors in a study of reactions to crime. Of the 296 survivors who were interviewed for a larger study, 128 (43%) nominated a person they considered a “significant other” (i.e., a romantic partner, family member, or close friend to whom they turned about the crime) and were thus eligible for the dyadic component of the study. The researchers did not describe the degree to which significant others were unable to be reached following nomination.
In Ullman et al. (2017), after completing three survey waves, adult women survivors who had previously disclosed assault and who indicated an interest in an in-person interview were invited to participate. They were also asked to provide contact information for one friend, family member, or significant other whom they had told about their unwanted sexual experience. To ensure confidentiality and minimize coercion, survivors were asked if they would be willing to contact DRs to ask them for permission to provide their contact information to the researchers. Then, after survivors were interviewed, the researchers contacted DRs directly who were interviewed separately, which in some cases took weeks or months to accomplish given time constraints and/or scheduling conflicts. DRs could decide whether or not to be interviewed after the researchers explained the study to them. Interviews with both survivors and their matched DRs asked about the survivors’ experience disclosing and receiving support. Interviews with DRs additionally asked about the impact of hearing about SA and coping with the disclosure. Both parties were assured the research was completely voluntary and that no information reported by survivors would be revealed to the DRs and vice versa. This methodology was designed to be survivor-focused by prioritizing survivors' control.
Ultimately, this method resulted in a final sample of 45 matched pairs of survivors (n = 58) and their primary DR (n = 45), representing a 78% rate of successful recruitment of DRs. A few survivors were not interviewed because they had no DRs about whom to provide contact information. As they had to have disclosed to someone as a criterion for study participation, this issue was unexpected. Among 13 survivors without a DR interview, DRs who had agreed to be contacted ultimately did not complete interviews for various reasons (e.g., could no longer be reached, changed their mind).
It is important to note that following survivors’ preferences about the best DR to contact may have resulted in a biased sample. That is, survivors may have been more likely to refer DRs with whom they had a more positive relationship, who were presumably more likely to have responded supportively to the disclosure. DRs who responded to the initial disclosure in a manner that resulted in damage to or dissolution of the relationship presumably were less likely to have been referred to the study by the survivor. In an effort to minimize this potential source of bias, the investigators asked survivors to provide contact information for up to three DRs who were positive, negative, and mixed in their supportiveness post-assault. This led to a sample of DRs interviewed who provided a variety of reactions (e.g., negative, neutral, positive); however, this methodology still likely led to a more positive/supportive set of dyads than would have been possible had the researchers been able to interview DRs who were no longer in contact with survivors (see Lorenz et al., 2018). Given the time-consuming nature of this methodology because survivors had to be interviewed, then the DR contacts they provided were contacted one by one until one agreed to be interviewed, which sometimes took months to accomplish, and the potential limitations with respect to positive bias, when developing the next research idea, an alternative method of recruiting survivor/DR dyads was considered: recruiting survivors through DRs and gathering survey-based quantitative and qualitative (rather than solely qualitative) data.
Finally, a social network study of survivors and their DRs was conducted in sororities and described in a methodological paper by Dworkin and Allen (2017). The authors proposed surveying every member of 24 sororities about their SA victimization experiences and experiences disclosing to and receiving disclosures from other sorority members. Unlike the other studies reviewed here, the researchers attempted to recruit survivors and DRs simultaneously and independently (i.e., without a referral from survivor to DR or DR to survivor). To obtain dyadic data, the researchers originally proposed that individuals who reported SA experiences would be presented with a list of other sorority members’ names and would be asked to indicate who they had told about the experience. In addition, all sorority members would be presented with a list of other sorority members’ names and would be asked who had disclosed SA to them. The authors describe multiple practical and ethical challenges with the study methodology. In particular, the institutional review board raised “secondary subjects” concerns about asking DRs to indicate by name which sorority members had disclosed SA to them, given that these sorority members had not necessarily consented to participate in the study. As a result, the researchers removed the questions about disclosure receipt from their protocol. In addition, low levels of participation were obtained from the four sororities that agreed to participate, likely because asking sorority members multiple questions about all other sorority members resulted in a very long survey battery, and the authors decided against involving sorority leadership in recruitment efforts to minimize coercion. As a result, it was not possible to obtain dyadic data using these methods.
Present Study
In light of the challenges encountered in studies that recruit DRs via survivors, and the potential benefit of dyadic data to understanding post-assault help-seeking and recovery processes, it is important to explore other methodological strategies for obtaining dyadic data from survivors and DRs. Thus, the purpose of this article is to present the authors’ experiences recruiting matched pairs of DRs and survivors who told them about SA/IPV as part of a novel intervention evaluation to improve responses to survivors and their psychological outcomes (Edwards et al., 2020). Unlike previous studies that recruited DRs 1 via survivors, the current article describes a study where DRs invited survivors to participate in the study. This research was guided by two primary research questions: (1) How feasible is dyadic recruitment of survivors from DRs, and (2) What challenges or barriers emerge in dyadic recruitment of survivors from DRs? Methodological and ethical issues are discussed, and recommendations are made for research using survivor-DR dyads.
Method
Sample
The present dyadic study recruited SA and IPV survivors via their DRs as part of the evaluation of a new support network intervention to help DRs better respond to SA/IPV disclosures (Edwards et al., 2020). The intervention aimed to increase positive social reactions and decrease negative social reactions to SA and IPV disclosures among informal support DRs (i.e., family, friends, romantic partners). The intervention trained potential informal supports on what to say and not to say to survivors. The study was conducted with 1,268 college students from a medium-sized New England university who completed an online baseline survey and were assigned to either the treatment or control condition. This study was approved by the University's Institutional Review Board (IRB). In the current article, a novel method of recruiting survivors who disclosed to others as part of that larger study is examined.
Procedure
Six months after the intervention, participants in both conditions (henceforth called “DRs” for clarity) completed the follow-up survey online for which they received a $25 gift card. The response rate at the follow-up was 70.1% (n = 889). Participants who had received a disclosure responded to questions about features of the disclosure, their social reactions, and the survivor's perceived adjustment. If participants had received more than one disclosure, a series of questions were asked that were relevant to the larger study in order to identify to which disclosure they should refer for the remainder of the study (i.e., if any of the survivors were over 18, were a student at the university where the study took place, which disclosure happened most recently). Once a single survivor was identified, the DRs entered that person's initials to remind themselves of the person about whom they were answering questions. Thus, although DRs may have received more than one disclosure, they only answered questions about a single person.
At the end of the survey, DRs who had received one or more disclosures were offered an additional opportunity to contribute to the research. In the study described in the current article, both participants in the control (wait-list) and intervention (education and training on supporting survivors and oneself) conditions were invited to refer survivors, as the aim was to obtain victim data from victims who disclosed to DRs who did and did not receive the social support intervention. No one participating as a DR could also participate as a victim and vice versa. Participants were asked to refer the survivors who had disclosed to them in the past six months to take a survey. Step-by-step instructions were given on how to refer survivors to the survey (see Appendix A for these instructions). First, eligibility was explained (survivors had to be someone who disclosed in the past six months to the DR and had to be over the age of 18). DRs could refer as many survivors as met these criteria, so we could try to get information from victims about their experience of disclosing, one of which would be a matched dyadic pair (i.e., one would be the person about whom the DRs answered questions in their own survey). Given that the project was part of a larger study evaluating the impact of a program to reduce DRs’ negative reactions to survivors' disclosures of SA/IPV, we were interested in obtaining as much information as possible from survivors’ perspectives, so that we could evaluate the impact of the program on DRs’ behavior from the survivors’ point of view. Second, a link to the survey was provided, which the DR could pass on to survivors. In this step, talking points were provided on how to discuss the survey. These talking points included telling the survivor about the purpose of the research study, telling the victim that the researchers would not know the survivor's name, that the survivor could earn a $25 gift card, and giving the survivor contact information for questions about the study. Finally, once the DR had completed one or more referrals, they were asked to email the project staff and complete a short survey about how many people they referred. DRs received a $10 gift card for referrals regardless of whether any of the survivors subsequently completed the survey.
Survivors who completed the survey received a $25 gift card. The survey included questions about past six-month IPV and SA, psychological distress and alcohol use, and social reactions. Prior to taking the survey, survivors consented and were asked if they were 18 years or older; if so, they moved on to the survey.
Results
Feasibility of Dyadic Recruitment of Survivors from DRs
About half of DRs reported in their survey that they received at least one disclosure in the last six months (n = 481; 54.1%). Of the 481 participants who were eligible to refer one or more survivors to the survey, 217 (45.1%) said they were willing to refer someone. The survey yielded 162 responses from survivors referred by DRs. DRs reported that they identified an average of 2.2 survivors who qualified for the study and passed the survey on to an average of 1.5 survivors. However, many of these responses were not usable, as we describe in the following section. Ultimately, this approach yielded only 58 dyads with valid, usable data for both DR and survivor (35.8% of the responses from survivors). Thus, this demonstrates that dyadic recruitment of survivors via DRs presents serious challenges.
Challenges in or Barriers to Dyadic Recruitment of Survivors from DRs
Previous studies have asked survivors to provide DRs’ contact information to researchers, who then recruited the DRs (Davis & Brickman, 1996; Ullman et al., 2017), or simultaneously recruited DRs and survivors in a single setting (Campbell, 2005). The current study took a different approach: DRs sent survivors information about the research study. We did not use the researcher as an intermediary to maximize survivor privacy and autonomy. However, this design raised several challenges, which are discussed next. Some challenges were addressed while data collection was ongoing, but others could not be addressed and represent opportunities for lessons learned and future research.
Low Survivor Participation
Although 481 participants reported receiving disclosures, responses from 162 survivors were ultimately obtained (and only 58 of these were valid; see below). The challenges that may have led to low survivor participation are described below.
DR Reluctance or Inability to Refer Survivors
Data suggested that participants were not always comfortable or willing to recommend the survey to survivors. As noted above, 45.2% of eligible DRs were willing to refer someone. In the referral survey that DRs took after referring survivors (N = 52), they were asked, “If you knew eligible people and did not tell them, why didn't you tell them about the survey?” About half (n = 21) said that this was not relevant, said “NA,” or gave an unclear response (e.g., “personal reasons”). Others, however, gave insights into their reluctance. Eleven feared the negative impact on the survivor (e.g., “I didn't want it to trigger them”), and nine thought the survivor would be unwilling to take the survey (e.g., “Felt like they might not want to share their experience”). Seven said the survivor was unavailable or that they were not close enough to the survivor to ask (e.g., “We’re not close”), and five expressed that they did not know how to ask (e.g., “I was afraid”).
Possible Survivor Reluctance to Participate
Our approach precluded the collection of empirical data on why survivors chose not to participate, but it is possible that survivors who declined to participate may have had negative impressions (e.g., skepticism, discomfort) of this recruitment method for a variety of reasons. Although neither the research team nor the IRB received any reports of participant distress associated with our referral method, it is unlikely that survivors would choose these methods to report their distress (i.e., they would likely just choose not to participate). Being asked to participate in a study of one's own victimization experiences likely represents a more intrusive request than being asked to participate in a study of DRs receiving victimization disclosures. Even when the request comes from a friend, it may be experienced as an intrusive “cold call” about a private experience. The research team provided DRs with “talking points” on how to invite survivors safely and empathically to be part of the project; however, it is unlikely that this fully resolved all discomfort experienced by survivors. Survivors may find it difficult to trust others with their experience even when initiating disclosures themselves or responding to people they know who are inquiring about their well-being. This may be especially true for requests to participate in a research study. Given this context, having a request to participate come from someone they told previously about their victimization experience who is doing research on the topic, may lead survivors to feel a loss of control or even a betrayal of trust. This may be especially true given that the request pertained to research on the topic of their victimization that they were unaware of and did not have a say in initiating. Thus, despite giving DRs “talking points” on how to tell survivors about the opportunity to participate in a project, it is likely that some reasons for nonparticipation were related to negative perceptions of the project.
IRB Challenges
Working with IRBs on obtaining approvals for this type of work can be challenging based on the previously reviewed methodological challenges and safety and ethical concerns regarding research with survivors (Dworkin & Allen, 2017). For example, the IRB did not allow the researchers to collect survivor names or contact information from DRs. In addition, the IRB emphasized the need to avoid incentivizing the DRs based on survivor participation. In the current study, incentives were based on referring (as reported by the DR), and not on whether the survivor subsequently completed a survey. In addition to safety, the IRB had concerns about consent. The IRB asked that care be taken in the consent form to explain that DRs could not also take the survivor survey (and would not be incentivized for doing so). Earlier protocols did not clarify this issue, and thus, some DRs were initially incentivized for taking the victim survey. In addition, the IRB required clarity that eligibility screening for victimization experiences would only be done in the survivor survey. That is, only individuals who reported victimization could take the victim survey, but the same eligibility did not apply to the DR survey.
Lack of Researcher Control Over the Recruitment Process
Another likely contributor to low participation was the fact that the researchers did not have control over the recruitment process. Being able to ask DRs to provide the researchers with names and contact information of survivors would have made it possible for researchers to reach out to survivors directly. Although approaching survivors directly would have raised other concerns (e.g., survivor autonomy and privacy), it would have ensured consistent recruiting and participant tracking. Although talking points were provided to DRs, it is likely that they varied in how they discussed the study with potential survivors (see Appendix A).
Data Validity Issues
Another challenge with this protocol was a high rate of data with potential validity issues. As noted above, of 162 survivors, only 58 survivors had usable data (35.8%). There were three primary data validity issues: (a) participant misrepresentation (17 participants lost), (b) referred survivor denial of victimization history (63 participants lost), and (c) dyads that were unable to be matched (24 participants lost).
Participant Misrepresentation
The first challenge with respect to unusable data was that some of the survivors were not actually valid participants. Although instructions explicitly stated that DRs could not take the survivor survey, we determined, based on email addresses used to take the survey, that some DRs still completed the surveys themselves, sometimes several times, presumably to obtain the financial incentive, a $25 gift card (n = 7 of the 162 participants; 4.3%). Indeed, in accordance with the protocol for compensating survivors, these participants did receive the financial incentive before we identified that they were in fact DRs and not eligible to participate as survivors due to prior participation as DRs; this issue was clarified in later versions of the consent form so that incentives were no longer given to DRs. It is also possible that some participants who had experienced victimization during the study period did not understand the directions and completed the survivor survey. Nevertheless, any responses to the survivor survey from DRs’ email addresses were discarded. Although it might have been possible to avoid this issue by asking DRs to provide names of the survivors whom they would refer to the survey, this strategy was not pursued to avoid ethical issues around survivor privacy. Similarly, of 162 survey responses, 10 of those responses (6.1%) were duplicates (i.e., the same survivor took the survey two or more times, as identified by email address). Duplicate responses were identified and removed based on email addresses.
Survivor Lack of Acknowledgment of Victimization
DRs also forwarded the survey to some individuals who subsequently reported no victimization history (n = 63 of the 162 participants; 38.9%). Related to the challenge of participant misrepresentation, it is possible that DRs forwarded the survey to friends who had not been victimized so that their friends could receive the incentive. It is also possible that DRs did not understand the directions even though we provided step-by-step instructions. Alternatively, DRs may have indeed been told by the individual they referred about a victimization experience, but upon starting the survey, survivors were not willing to endorse victimization. Finally, it is possible that DRs defined an experience that had been disclosed to them as victimization, whereas the survivor who subsequently responded to the survey did not define their experience in the same way. Thus, some participants taking the victim survey did not report victimization, and thus, we could not use their data. In response to this problem, we modified the protocol to include a screener for victimization experiences as the first question (see Appendix B for this screening question). If individuals did not report victimization, they could not proceed with the survey.
Unmatched Dyads
Another set of challenges involved matching survivors and DRs. Two types of matching were attempted, each with unique challenges. The first type of matching involved matching each survivor to the DR who referred them so survivor and DR data could be compared. Survivors were asked to provide the name of the person who referred them. However, many survivors left the field blank or provided a name that did not match a DR in the master list (n = 24 of the 162 participants; 14.8%). Although data collection was ongoing, the protocol was adjusted to mitigate this challenge. Instead of providing DRs with a universal link to share with potential survivors, each DR was provided with their own unique recruitment survey links to share. That way, it was possible to track who referred each survivor based on the link that DRs used to complete the survey. This strategy made better use of available survey technology.
The second type of matching was more complicated. As a reminder, each DR responded to questions about one particular survivor but could refer multiple survivors to the survey. We selected this approach because maximizing survivor participation was desired in anticipation of recruitment challenges. Specifically, we aimed to conduct analyses with as many survivors as possible regarding the degree to which the intervention resulted in changes in DR responses to disclosure and, ultimately, survivor mental health. Although survivors could be matched to the DRs who referred them, this matching did not necessarily mean that the DRs answered the survey questions about that survivor. For privacy reasons, survivor names were not requested in the DR survey or survivor survey. As a result, the only marker of survivor name in the DR survey was the initials that the DRs entered to keep track of the person about whom they were answering questions. The only marker of survivor name from the survivor survey was the email address. Because the university email addresses included initials, we attempted to match survivors’ email addresses to the DRs’ entered initials. Unfortunately, this matching proved largely unsuccessful: of the 58 survivors, only four could be reliably matched to a specific DR who answered questions about that survivor. This challenge meant that while survivor-reported social reactions of participants in different treatment groups could be compared, survivors’ and DRs’ experiences and perceptions of social reactions to the disclosure could not.
Discussion
This article presented research on the feasibility and experience of collecting dyadic data on SA and IPV as part of examining the impact of a novel support network intervention on college student DRs (Edwards et al., 2020). To the authors’ knowledge, this was the first study to implement a protocol involving survivor referral by DRs and was one of the only studies to recruit dyadic partners in this area of research. Dyadic studies in the social sciences addressing different types of interpersonal relationships has proliferated; however, few dyadic studies have been conducted regarding SA/IPV disclosure and social reactions (Campbell, 2005; Davis & Brickman, 1996; Ullman et al., 2017). Research is needed to better illuminate experiences of both individuals in the dyad and their interpersonal relationship context in the aftermath of SA/IPV (Rodriguez & Neighbors, 2015; Ward et al., 1997).This study of SA/IPV disclosure dyads used a novel referral recruitment method in which survivors were recruited via DRs (Edwards et al., 2020). Only 35.8% of cases had usable dyadic data for survivors and DRs, suggesting that this referral method had limited efficacy for obtaining a significant percentage of matched pairs. One can compare this limited efficacy to previous research where victims recruited DRs; this research has yielded over twice the percentage of matched pairs (Ullman et al., 2017).
Several key challenges were encountered with this protocol. First, similar to the challenges identified in the Dworkin and Allen (2017) methodological paper on conducting social network studies of survivors and DRs, there was low survivor participation. The current study provided evidence that this was due to DR reluctance or inability to refer survivors and lack of researcher control over the recruitment process. There also may have been reluctance by survivors to participate via this recruitment method. Second, we encountered data validity issues, including DRs taking the survey as a survivor (potentially to gain the incentive) and perhaps due to survivors' lack of acknowledgment of victimization. Despite challenges encountered, a modified version of this recruitment method may add another tool for the collection of dyadic pairs to studies of violence against women. Data from dyads may improve our understanding of how to treat and intervene with survivors and their informal social networks.
Limitations
There are some limitations to the present study. First, although this study provided preliminary data on the proportion of DRs who are willing to refer survivors (just under half), this figure might be biased by nature of the sample. That is, people who participate in an intervention study on this topic might be more willing to refer survivors than participants in other studies. Second, survivors were not directly asked about their reactions to being recruited by DRs. Even if questions had been included in survivor surveys, these data would not have been obtained from survivors who chose not to participate. As noted above, future research should assess survivor reactions to this referral methodology. Third, survivor recruitment through DRs may be influenced by a positive relational bias, such that DRs are less likely to refer survivors to whom they previously responded negatively and/or with whom the relationship had deteriorated. Fourth, it is possible that some participants in the control condition could have learned about the intervention from peers. Another limitation is that the victimization screener used has unknown sensitivity and specificity. Finally, a nondiverse sample of college students was used, and the extent to which this methodology would be deemed feasible and acceptable in other populations (e.g., community samples) is unknown.
Implications
Ethical and Methodological Implications
Various ethical and methodological limitations characterize the very few studies of survivors and their informal/formal disclosure experiences, which is likely why there are so few studies in this area of research. Both Davis and Brickman’s (1996) and Campbell’s (2005) studies were successful, but both studied survivors soon after the assault at a survivor services agency and hospital, respectively. These studies present a possible ethical challenge in navigating recruitment sensitively during a time of acute crisis. However, these settings may also have the advantage of being circumstances where survivors and their support systems are easier to recruit. Hopefully, these settings also provide survivors with useful information about a variety of counseling and other support services for dealing with the assault aftermath. Given that the majority of survivors do not report to formal supports (Sabina & Ho, 2014), it is important to recruit survivors who do not present in such settings. Ullman et al.’s (2017) study was conducted with survivors assaulted in the past who had all completed three annual mail surveys and were then invited to do interviews themselves. Survivors provided DRs’ contact information to the researchers. This methodology appeared to work well with no participant complaints, but it took two years to collect interview data from all participants (including 45 DRs), and there may have been retrospective recall bias given that assaults occurred an average of 12 years ago.
Research Implications
The current study offers insights into future research that may help inform the methodological approach of research on social reactions to SA/IPV disclosure. Although preliminary research suggests that survivors and DRs may have different perceptions of the same disclosure experiences (Davis & Brickman, 1996), the extent to which these differences need to be addressed in programming is unknown. It is possible that these types of discrepancies would be more important for dyadic-focused programming rather than programming focusing on DRs or survivors alone. For example, feminist scholars have documented that victim-blaming attitudes such as rape myths are related to negative social reactions by DRs (Ahrens & Aldana, 2012; Paul & Sasson, 2013). Regardless of survivor and DR agreement on the types of responses provided, this is likely a critical component for programming with DRs.
Future research could also directly ask survivors for their opinions about this methodology and yield insights into procedures that would be preferable and maintain their privacy and autonomy. For example, it was assumed that both DRs and survivors would have been upset if DRs were asked to provide survivors’ contact information (with survivor permission, of course), and this was indeed something that the IRB would not allow. Yet, the extent to which this type of methodology would be unacceptable to DRs and survivors is unknown and in need of future research. Also, the talking points provided to DRs may need to be expanded to include the following:
I am in a research project that is trying to understand how to best help college students who have had an experience like the one you told me about. I would never tell them your name but I was asked by the research team to tell you about the project. You would tell them in a survey how other people have responded to your experiences. If you would like more information, you can contact the research team at [email address] or [phone number]. You can also see more information here on this website: [study website URL]. You will earn a $25 gift card for participating, which only includes filling out a confidential survey online. Research suggests being in research can be a meaningful experience that has personal benefits and benefits to society, too. Know that you will be anonymous to the researchers and only linked to me by your initials.
In Ullman et al. (2017), there were no IRB concerns with the methodology of having survivors provide contact information for DRs to the researchers. This was completely voluntary for survivors, who were also being asked to do interviews followed by their DRs if they agreed to participate separately. Because survivors were giving the researchers contact information from DRs with their permission, the DRs then expected the researchers to follow-up with more information about the study and at that time they decided whether or not to participate. Researchers and IRBs may need to develop guidelines for how to do such studies ethically. In particular, as noted in Dworkin and Allen (2017), researchers need to know how to address secondary subjects’ concerns. Typically, IRBs want researchers to collect data directly from participants, rather than indirectly (i.e., from others reporting on their behavior). This is especially important in human subjects research on sensitive topics. Anything beyond passing on a research opportunity may be considered by IRBs to be coercive by the DRs. However, some IRBs allow an exception for research so that researchers do not have to report as part of their researcher role (see Potter & Edwards, 2015), similar to exceptions for reporting participants' HIV status.
Policy and Prevention Implications
Results of the present study have implications for policy. First, recently adopted mandated reporting policies on college campuses may make dyadic data challenging to collect on student populations especially if the data are not anonymous (i.e., a master list is maintained over time linking participant/dyadic data). Indeed, at some institutions, researchers may have to report survivors under Title IX regulations.
With respect to prevention implications, there are several limitations of existing data recruitment methods on DRs’ responses. Most studies are largely non-dyadic, which provides only one perspective on this inherently relational topic. The few dyadic studies on this topic have been time- and labor-intensive and raise many difficult logistical and ethical challenges. In addition, they risk being potentially positively biased (survivor-to-DR method; Ullman et al., 2017), or, as in the present study, difficult to effectively match (DR-to-survivor method). Currently, prevention efforts aimed at improving social reactions to SAs (including the intervention in the present study) are based on the best available research and meta-analytic work (e.g., Dworkin et al., 2019). However, without dyadic designs, researchers may be limited in our ability to ascertain (a) the effects of particular social reactions on survivors, and (b) make recommendations to prevent negative social reactions and their impacts on survivors. That is, prevention efforts could be improved by developing feasible, practical methods to understand dyadic influences of social support more accurately. Although challenges were encountered in recruiting from DRs to survivors, it is possible that some of the changes recommended here could make such an approach more feasible.
Conclusions
In sum, this article describes referral methods of recruitment used in dyadic studies of survivors and DR's examining social reactions to disclosures of SA and IPV, including challenges inherent in this process. Dyadic studies can be used to study the topic of SA/IPV disclosure and social reactions to disclosure. Various creative methodologies are needed to obtain sufficient participation and minimize threats to validity while protecting survivor privacy, autonomy, and safety. As researchers conduct such studies, they must share successes, challenges, and lessons learned so that other researchers may build on previous work. Meta-studies (e.g., studies of/about the study process itself) should be built into research wherever possible to better understand how both survivors and DRs feel about dyadic research on this topic generally and the specific referral methods utilized. The information presented in this article can encourage ethically informed and methodologically rigorous future research using dyadic approaches to the study of social reactions to disclosures of SA and IPV.
Footnotes
Appendix A. Information for Disclosure Recipient
You are eligible to earn
Appendix B. Victimization Screening Question
Which of the following did you tell that you experienced? (Please select all that apply.)
Someone (including, but not limited to, a romantic partner) touched your genitals without your permission (but there was no penetration)
Someone (including, but not limited to, a romantic partner) used verbal coercion or threats to have sexual intercourse with you (oral, anal, vaginal)
Someone (including, but not limited to, a romantic partner) used physical force, threats of physical force,
alcohol/drugs to incapacitate you to have sexual intercourse (oral, anal, vaginal)
Your partner refused to talk to you
Your partner insulted or swore at you
Your partner said or did something to spite you
Your partner verbally threatened you
Your partner refused to allow you to hang out with friends
Your partner monitored your phone, email, or social media accounts
Your partner followed or spied on you
Your partner threw something at you
Your partner pushed, grabbed, or shoved you
Your partner slapped you
Your partner kicked, hit, or bit you
Your partner choked you
Your partner threatened to use a knife or gun on you
Acknowledgments
The authors are grateful to the participants and thank Caroline Leyva, Robert Eckstein, and Jane Stapleton for their contributions to the larger study, and 20+ research assistants and program facilitators at University of New Hampshire.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) (grant number NIH R34 #R34AA024849). Opinions, findings, and conclusions, or recommendations expressed in this article are those of the author(s) and do not necessarily reflect those of the NIAAA.
Supplemental Material
Supplemental material for this article is available online.