Why Do Domestic Abuse Organizations Engage in Gatekeeping?

Studying the interaction between politics and IPV is important for the eradication of violence against women in society. While much of this research can be done without the involvement of victims/survivors (Iyengar Reference Iyengar2009; Stevenson and Wolfers Reference Stevenson and Wolfers2006; Vigdor and Mercy Reference Vigdor and Mercy2006), occasionally their involvement is necessary to understand the individual-level impact of policies and interventions. The samples for such research rely heavily on referrals and connections garnered from DA shelters (Gelles Reference Gelles2000). Yet, DA shelters and organizations often act as impediments to this research. Why?

First, DA organizations are simply underfunded, short-staffed, and time-poor, as well as occasionally fatigued due to excessive participation in existing research (Clark Reference Clark2008).Footnote ³ Their priority with limited resources is helping survivors, and requests for research cooperation are frequent and potentially burdensome. This is the case even if researchers “only” want referrals or connections with survivors. For example, in Wisconsin, it is illegal (for good reason!) to disclose the name or contact information of a DA survivor without a signed Release of Information (ROI). Usually, an ROI provides a survivor’s consent for their information to be shared with a specific organization for a limited time period and is often used to refer survivors to other resources. However, an ROI is also necessary to share information with researchers, and to encourage clients to consent, it is likely that frontline staff must explain the purpose and consequences of the study. Advocates and other staff must, therefore, learn and communicate a nontrivial amount of information about the project. This remains the case even if researchers rely on self-referrals from survivors, which removes the need for an ROI.

The second reason why DA organizations may act as gatekeepers to researchers is their fiduciary responsibility to help survivors avoid further trauma and exploitation. While many people realize that DA includes physical and sexual violence, there is increasing awareness that it also often involves coercive control. Coercive control is a pattern of behaviors used by an abusive person to dominate and manipulate their intimate partner, and it reduces the nonabusive partner’s ability to make decisions, limits their independence, and erodes their sense of self (Ehrensaft et al. Reference Ehrensaft, Langhinrichsen-Rohling, Heyman, O’Leary and Lawrence1999; Hamberger, Larsen, and Lehrner Reference Hamberger, Larsen and Lehrner2017; Robertson and Murachver Reference Robertson and Murachver2011). Critically, the effects of coercive control have been likened to the effects of post-traumatic stress disorder (Lohmann et al. Reference Lohmann, Cowlishaw, Ney, O’Donnell and Felmingham2024) and do not end with the conclusion of an abusive relationship. Although many researchers are cognizant and mindful of these experiences, the majority are not victim advocates and likely have minimal training in trauma-informed care, which, in turn, contributes to DA organizations’ skepticism of researchers.

Moreover, even if researchers have received training, it is impossible to judge the quality of that training, as there is no standardization across organizations, states, or countries. For example, in my volunteer work with Victim Support Scotland, not only was I required to join the Protecting Vulnerable Groups (PVG) scheme—an extended criminal background check for individuals undertaking “regulated work” with children and protected adults—but I also completed VSS’s three-day foundational training plus specialist training on DA, sexual assault, hate crimes, and working with children and youth. However, when I returned to the US, although my prior training was viewed favorably, I still needed to undertake another eight weeks of training to become a volunteer advocate in Ithaca, New York, to ensure that I knew the new organization’s operating procedures and guidelines as well as how to appropriately support survivors.

In addition to researchers having minimal training, there is also always the possibility that researchers may be, at best, less-than-sensitive, and at worst, bad actors who may exploit survivors for the purposes of advancing their research agendas. Indeed, the latter are why we have ethical standards and guidelines, and given the publicity unethical work has received, without countervailing evidence that researchers are trauma informed and well trained it is understandable why DA organizations are distrustful of researchers. Therefore, to avoid exposing clients—who may have complex and long-standing mental health challenges—to these risks, it is much easier for DA organizations simply to avoid cooperating with researchers.

Finally, many DA organizations are hesitant to collaborate with researchers as there is a lack of understanding of what exactly research is and what it will be used for, which is often combined with an intense suspicion of the motives of people outside of DA advocacy. While many nonprofit fields and organizations are increasingly leveraging quantitative data for decision-making, donor engagement, and operational efficiency, this transformation—in my experience—has had minimal impact on DA organizations. Indeed, many organizations I have worked with have been outright hostile to the idea of collecting data as “proof” that programs and interventions work. My impression is that this hostility arises because those demanding proof are also simultaneously trying to reduce the already minimal funding available to DA organizations. Rather, DA organizations and advocates observe the impact on individual survivors and conclude from this powerful yet often anecdotal experience that a program “works.”Footnote ⁴

Solutions(?)

So, if a researcher encounters ambivalence or even hostility from a potential community partner, what can they do? One potential solution for accessing and recruiting a vulnerable population like domestic abuse survivors would be to try to circumvent gatekeepers and contact potential participants directly. With the advent of the internet, recruitment through social media is the equivalent of flyers, posters, or brochures. This method not only allows researchers to bypass gatekeepers and explain their research directly to survivors but also grants survivors agency over their participation. For example, Glinski (Reference Glinski2024) uses her Twitter network to recruit and partner with DA survivors to examine the different forms of economic abuse that occur after an abusive relationship. Likewise, Sandhu, Brady, and Barrett (Reference Sandhu, Brady and Barrett2023) use Facebook to connect with South Asian women who have eschewed an arranged marriage for a marriage with a partner of their choice and then experience violence including familial violence.

However, while recruitment by social media is innovative and helps to avoid gatekeepers, there are some disadvantages to accessing DA survivors in this way. First, researchers must have moderately extensive social media networks (or a cooperative community partner with an extensive network) to ensure widespread dissemination of the recruitment request. This creates a barrier for those with little to no social media/followings, as well as potential ethical issues if survivors contact researchers through for-profit, inimical platforms like “X” (formerly, Twitter). Second, a social media-based strategy requires participants to reach out based on their own initiative, which may yield a low participation rate as is the case with flyers/brochures (Pinto and Dunsiger Reference Pinto and Dunsiger2021).

Third, researchers employing paid surveys must also be concerned with the issue of screening and fraud, where potential participants may misrepresent their experiences (Cascio Reference Cascio2024), or where recruitment ads are shared with bots (Bybee et al. Reference Bybee, Cloyes, Baucom, Supiano, Mooney and Ellington2022; Pozzar et al. Reference Pozzar, Hammer, Underhill-Blazey, Wright, Tulsky, Hong, Gundersen and Berry2020; Storozuk et al. Reference Storozuk, Ashley, Delage and Maloney2020). Finally, and perhaps most importantly, gatekeepers often protect potential participants from unethical, exploitative, and disrespectful researchers (Letherby and Bywaters Reference Letherby, Bywaters., Letherby and Bywaters2007; Sandhu, Brady, and Barrett Reference Sandhu, Brady and Barrett2023), which social media recruitment may remove. Thus, while most researchers are ethical and mindful, if our aim is to minimize harm, gatekeepers are perhaps helpful for hindering the rare reprobate researcher.

Because bypassing community partners is difficult, another approach may be to attempt to mitigate the barriers that trigger gatekeeping. In some instances, like research fatigue, this will be almost unfeasible because it is the very request for collaboration that is the problem. However, for other barriers, it may be possible to create incentives that make participating in research more amenable for community partners. For example, as mentioned above, many DA organizations are underfunded and research requests often require uncompensated staff involvement and time. Therefore, compensating organizations and staff for their time conducting and participating in the research project may be helpful. One way this could be accomplished is to include a percentage of DA organization staff time in grant and funding requests under budget categories like “Consultant Services.” Indeed, according to the National Science Foundation, professional and consultant services are allowable when provided “by persons who are members of a particular profession or possess a special skill, and who are not officers or employees of the proposing organization” (National Science Foundation 2024).

Yet, like social media recruiting, in-built compensation is not a perfect solution. First, very few researchers—especially those at non-R1 institutions—have research funds significant enough to pay for staff time plus other normal (and often substantial) research expenses such as participant incentives, student hours, travel and conference costs, faculty salary, fringe rates, and indirect costs. Consequently, the ability to pay DA staff for their time may be dependent upon whether a researcher can obtain a grant for the project, and whether consultant services are an allowable budget item on the grant. However, this means that some types of research on DA will be dependent upon the availability of grant solicitations and the success of applications. Other scholars who have limited opportunities for additional funding, such as graduate students, may find the cost of researching DA impracticable.

Second, providing compensation for staff time and effort is not necessarily a guarantee of perfect collaboration. For example, I am currently part of an evaluation team analyzing the impact of Wisconsin’s Domestic Violence Housing First Pilot Program (WI-DVHF). As part of this pilot, nine organizations from across Wisconsin applied to receive approximately $200,000 each year for three years to help survivors find and maintain safe and stable housing (for more on WI-DVHF, see Sims, Barnes, and Walsh Reference Sims, Barnes and Walsh2023). Critically, every organization involved in the pilot is using some of its WI-DVHF funding to pay for internal staff and administrative positions. Additionally, as a condition of participating in the pilot, each site agreed, through a Memorandum of Understanding, that it would participate in the evaluation process. However, despite this funding for staff and formal agreement as well as significant consultative and educational efforts on our part, many of our community partners have still found the data collection process of securing ROIs and administrative data to be difficult to manage in addition to their regular responsibilities.

Next, given that a lack of research knowledge seems to contribute significantly to gatekeeping, educating DA organizations and staff about data collection, research methodology, and evidence-based policymaking could be an important component for reducing barriers. And indeed, these educational courses already exist—for example, the Jameel Poverty Action Lab (J-PAL) offers a course called Evaluating Social Programs which teaches individuals and organizations about why “rigorous impact evaluation is needed” as well as how to do evaluation through experimental means (“Evaluating Social Programs” 2024). More education about research and its importance would help both researchers and organizations. First, a greater understanding of research may help organizations to identify what types of evidence are valuable in their own work. This would allow researchers to design projects that reflect the expertise and goals of all collaborators. Second, in a world that increasingly abides by “effective giving” principles, demonstrating impact through research-informed evidence is beneficial for appealing to donors and philanthropists.

However, encouraging DA organizations and individuals to engage in this type of research education may also have disadvantages. First, many of these courses about program evaluation and research are methodologically narrow, focusing on experimental and causal inference methods to the exclusion of others, particularly qualitative methodologies. Having DA organizations undertake these types of courses, therefore, may perpetuate experimental research predominance and provide an unbalanced perspective on what constitutes good evidence. Second, these courses often remain expensive for nonprofits—for example, for a US nonprofit with an operating budget of less than $10 million, the fee for the J-PAL course is $2300 (excluding travel and accommodation expenses). While this is a significant discount on what for-profit, governmental, and academic institutions pay, it would still be prohibitive for the organizations I have worked with.