A group of four people link arms while mid-air during sky-diving. The view of the picture looks down upon the skydivers, with a view of a coastline behind. Image by Alexander Savin, available in a CC-BY-2.0 license.

How can researchers do right by communities when collaborating on research? 

Social scientists are increasingly collaborating with communities – volunteer groups, organizations, on and offline communities –  to make progress on practice and theory parallel. From development economics to civic engagement and design, these collaborations are informing how to reduce poverty, advance justice, and strengthen democracy.

When developing studies with communities, researchers need to engage with partners in an ethical and trustworthy way. For many of us, the ethics of community collaborations are outside of our training—especially because ethics regulations focus on the rights of individuals rather than groups. When researchers treat communities well, we can build trust that enables productive collaborations and opens opportunities for more research.

In this post, we share lessons for planning and developing collaborative relationships that communities will want to return to. We’ve found that the processes we use to protect the interests of communities also help to smooth the collaboration process itself, by identifying and ironing out tension points early on. We hope that the lessons we have learned from our mistakes and successes will help you build strong, collaborative community partnerships.

Understanding Research Benefits and Risks Beyond the Common Rule

Researchers are well versed in the Common Rule when it comes to protecting individual subjects in their studies, but how do you go about protecting the communities you collaborate with? 

Risks may look very different at the community level. A field experiment to test a recruitment strategy could result in fewer new members. A control condition for testing a prosocial intervention could result in flare ups that destabilize a community event. Or the treatment you create to test the effects of rewarding volunteers for their contributions could be seen as inauthentic and decay trust in the community. In each of those cases individuals may or may not be harmed, but for groups that are sensitive to the need to build momentum and trust, a study that backfires could be a substantial set back to their efforts.   

At CAT Lab we apply the same principles as the Common Rule to minimize those risks. We take on projects where the benefits of research to a community outweigh its potential harms, in line with the principle of justice. That means putting in place systems to minimize harm – as well as to maximize benefit. For every study, we ask: Have we done enough to minimize potential harm to the community? Are the burdens and benefits equitably distributed? Does the potential benefit to the community ultimately outweigh any harm? 

With those questions in mind, we integrated a few steps into our process::

  • Working with communities to identify research questions that maximize potential benefit to those communities
  • Collaboratively building designs with community members with an eye toward minimizing risk
  • Actively seeking push back from communities on research design to uncover potential risks we may have missed in the design stage

We spell out each of those processes below.

Envisioning Research that Benefits Everyone

One of the simplest ways to optimize the benefit/risk ratio is to choose projects that everyone agrees serves the community’s goals. A study that poses risk might not make sense if the potential benefit to the community is negligible, but if it can reveal information that will lead to a sizable positive impact on the community, then a little risk may be worth it..

The community you are working with has probably spent time thinking deeply and debating which of their activities and interventions are most effective toward achieving their goals and how those interventions may be improved. Likely they have countless questions a researcher can help them answer with a well-designed study. Among those many potential answers, there is a good chance one is a contribution to scholarship and in sync with your research agenda.

Surfacing those aligned questions – that groups want answers for and that would also be a contribution to scholarship – involves conversations that could take many forms. At a minimum we recommend meeting with core community members who are in the thick of thinking about how to increase their impact – and who have the trust of the community. With that group have a conversation about their goals, challenges and the questions they have about the impactfulness of their activities and interventions.

At CAT Lab we have hosted a variety of all day – and multiple day – gatherings both on and offline to bring together group members and researchers, guiding them through discussions to reveal common research goals. While there is no magic formula to making these conversations fruitful, we have found that they can be more productive if you include a discussion about the types of research designs and methodologies you use – and the kinds of questions those designs can and cannot answer. Knowing, for example, that you can test the impact of an intervention or that you conduct qualitative interviews, will help your collaborators think through with you which of their questions you can help answer. Another key to  a successful conversation is to hold back on sharing any specific research ideas you may have until after you’ve had time to deeply listen and fully absorb the community’s questions. Keep in mind that your collaborators may be grateful for your interest and hold you, the researcher, in high esteem; as such, they may be reticent to push their ideas and instead acquiesce to one of yours even if it has negligible benefit to their goals. Encourage them to push. 

Finally, you will want to come out of the conversation with a fairly concrete research question. Write the question down and get the group’s explicit thumb’s up; being crystal clear about the question you are working together to answer will avoid confusion – and frustration – down the line. 

Collaborative design – with a ‘community liaison’

Once you’ve identified and agreed on a specific research question that has a clear benefit to the community, the next – and possibly harder – task is anticipating harms. There will be countless landmines your design could stumble into that you are unaware of: disrupting norms that communities find sacred, touching on concerns communities have about their well-being, or destabilizing fragile relationships. 

Working with a Liaison

To design a study that avoids pitfalls, you can do no better than to enlist one or two community members to be a close consultant in your design. These ‘community liaisons’ should be community members that not only are steeped in the culture of the group but also have broad-based trust among other members (that trust will brush off on your project). They should also be excited to work closely with you and, ideally, be interested in experimental design; you may demand a lot of their time so the more jazzed they are to work with you, the better. If funding for stipends or the possibility of co-authorship are possible, they can help too. 

If you are currently collaborating with a group, you are probably already working with someone, likely the person who invited you to collaborate. That person may be the ideal candidate for your liaison, or may not be. For example, if they are the lead of the project (Executive Director, President) they may not have the capacity to work closely with you. Ask your current partner who on their team or in the community – including them – may be an ideal liaison.

Collaborative design

The more you can bring your liaison into the design process, the better. We invite our liaisons to a “research summit” which is at least a one day affair to work through all aspects of the design, translating the research question into defined interventions, units of analysis and outcome measures. (Depending on the liaison’s experience with research, the summit may also involve a bit of training on study design basics.) 

The benefits of co-developing the design are countless: your liaison will help you steer clear of intervention designs that could harm the community (your main purpose), but they will also reveal mis-assumptions you have made about how their systems work, what different metrics mean, etc. 

When we work with liaisons we explicitly ask them to keep their eyes out for “gotchas”; any part of the design that might trip the study up in execution or interpretation. Many “gotchas” will be purely logistical: our Wikipedia liaisons, for example, saved us from planning recruitment efforts at times when they knew they would conflict with other major campaigns. Some will help you steer clear of transgressing community norms like, again on Wikipedia, always send messages on public “talk pages” rather than private DMs. The most important “gotchas” detect parts of the design that could hurt the community.? When we worked with Wikipedians to test the effects of Wikipedia’s “thanks” tool, which editors use to thank each other for their contributions, our liaisons warned us of the danger of sending any “thanks” that could be seen as inauthentic, and they worked with us closely to develop a treatment design that would maintain the integrity of the tool. 

If enlisting your liaison to participate in a full day research session is a non-starter, we recommend at the very least meeting with them to closely walk through a design you have already drafted, and steering them to scour for “gotchas”. Keep in mind again, they may want to be polite and defer to you, the revered researcher, so encourage them to not hold back; it could be what prevents your project from dying on the vine.

Communicating with communities – and seeking consent

Your liaison will help you navigate around most potential missteps, but they may not catch all community concerns. As part of CAT Lab’s process we seek “consent” from the community before conducting the experiment, a process that social scientists are increasingly turning to as an approach to maintain trusting relationships with the public and manage the collective risks of research. 

Just as with seeking individual consent, the community consent process gives the community an opportunity to consider the possible harms and benefits of a study before deciding whether to participate – that is, to give it the green light. Unlike individual consent, though, the process also creates a space for community members to raise concerns and risks that you and your liaison may not have seen – and allow you to adjust your plan accordingly.

Community consent looks different for each project. To devise a process you’ll first want to  identify the boundaries of the community— asking how the community defines its membership and who can give consent on behalf of the community. It could be a leadership board, the full staff or an organization, a coalition council – or even the entire community (when we work with Wikipedias we post our research plans for any Wikipedian to see on their public forums). Second, you need to learn the processes that the community normally uses for communicating and making decisions. Your liaison should help you learn who comprises (or can speak for) the community, how best to communicate with them and what “yes” looks like (e.g.consensus, majority vote, etc.)

Community consent can only happen when the community has the opportunity to weigh risks and benefits, surfacing concerns you may not have already foreseen. We advise against just sharing your plans and hoping members don’t raise any concerns. Your conversations will be much more productive and trustworthy if you encourage community members to look for fissures in your plan. By identifying and addressing those risks early, you prevent your study from crashing to a halt through problems discovered after launch. More importantly, you can proactively avoid harm.

Building Trustworthy Community Collaborations

Community collaborations provide a powerful model for advancing knowledge and contributing to society. When researchers involve communities in the work of planning studies and weighing the risks, the science is better and the trust we build can endure without burning bridges. We hope these lessons from CAT Lab’s work provide helpful guidance in your own collaborations with communities.

Not the only approach to collaboration

Above we described CAT Lab’s approach to aligning goals with communities, but ours is not the only model.

Carl Wilmsen, who has facilitated hundreds of collaborations on environmental science, distinguishes two models of collaboration, which entail different types of benefits:

  • Participation: research goals and research process and are shaped by communities
  • Reciprocity: research goals and process are directed by the researcher, but communities reap other benefits from the collaboration

As a researcher you may have a question that does not align neatly with the needs of communities. That’s where reciprocity can be important—communities may still value research if they are receiving other benefits (resources, status, enjoyment, relationships, education) from the collaboration.

Either approach has the potential to lead to successful – and mutually fruitful – collaborations. Both, however, require conversations at the outset to establish the nature of the collaboration and clarity about what each side will get out of it. It is also advisable not to combine the models, mixing collaboration with reciprocity. It may feel that you are increasing the benefit to communities if, for example, you offer funding as part of your “participatory” collaboration; that additional funding, however, may skew the power balance and inadvertently be coercive.  

If you can find agreement on a win-win outcome between researchers and communities, then you have a solid foundation for building a healthy collaboration. 

Thanks to Mark Brandt and Jan Gerrit Voelkel for reading and giving invaluable feedback on an initial draft of this post.

Image by Alexander Savin, available in a CC-BY-2.0 license.

References and Further Reading

Agans, J. P., Burrow, A. L., Kim, E. S., Garbo, C., Schroeder, M., Graf, S., & Davis, T. (2020). “You’re going to burn some bridges if you come at it the wrong way”: Reflecting on the realities of research-practice partnerships. Community Development, 51(1), 36-52.

Creighton, S. (2008). The Scholarship of Community Partner Voice. Higher Education Exchange. 

Desposato, S. (2015). Ethics and experiments: Problems and Solutions for Social Scientists and Policy Professionals. Taylor & Francis.

Desposato, S. (2014). Ethical challenges and some solutions for field experiments. San Diego: University of California. Available at www. desposato. org/ethicsfieldexperiments. pdf.

Glennerster, R. (2017). The practicalities of running randomized evaluations: Partnerships, measurement, ethics, and transparency. In Handbook of economic field experiments (Vol. 1, pp. 175-243). North-Holland.

Levine, A.S., Matias, J.N. (2021) How to Generate Research Ideas That Impact Society. Inside Higher Ed

Mortensen, C. R., & Cialdini, R. B. (2010). Full‐cycle social psychology for theory and application. Social and Personality Psychology Compass, 4(1), 53-63.

Stokes, D. E. (2011). Pasteur’s quadrant: Basic science and technological innovation. Brookings Institution Press.

West, S. E., & Pateman, R. M. (2016). Recruiting and retaining participants in citizen science: What can be learned from the volunteering literature?. Citizen Science: Theory and Practice.

Wilmsen, C. (2012). Participation, reciprocity, and empowerment in the practice of Participatory Research. University of California, Berkeley. Retrieved April 23rd, 611-629.

Zong, J., Matias, J.N. (2020) Building Collective Power to Refuse Harmful Data Systems. Citizens and Technology Lab.