What Mistrust Can Teach Us: Research and Racial Justice

Longstanding bias and mistreatment have suppressed Black progress and eroded community trust. Here are some ways we can start rebuilding that trust.

Stacey Houston, II

Published in

Meta Research

11 min readJul 27, 2021

by Stacey Houston and Randy Illum

As a result of numerous highly visible examples of disregard for Black lives, the national spotlight on racial injustice may have been brighter over the past year or so than ever before in recent history. To be clear, racial and ethnic minorities in this country have always felt, called attention to, and fought against their mistreatment. But over this past year, many more people, organizations, and companies — our own included — have rushed to take action, to question our commitment to racial justice, and to understand how they can contribute.

While this effort is honorable (though overdue), the same injustices that triggered our desire to take action can also impede our ability to take effective action. We’re hindered by the assumptions we make about the best way to support the communities most impacted. We have taken important steps toward beginning to understand what we can do to better serve the Black community, but have we made room to consider whether Black people’s voices are truly included in the work we’re conducting? How do we interrogate our commitment to racial justice, and who takes part in that introspection? The answers to these questions will have a direct impact on the progress we make.

Questioning our course

At Facebook, as at many other companies, we rely on our research to guide our product decisions and priorities, including our work related to racial justice. Last summer, a few FB employees published an article highlighting the strides the company has taken in using research to build products that improve the lives of Black people. And, to be fair, the progress has been somewhat inspiring. But — you knew there would be a “but” — we must continue to question how the inequities baked into our standard processes run counter to our desire to do better.

Any attempt to serve Black people more justly must start with an understanding of history — specifically, the role of tech companies and employees in the Black community’s mistrust of research. Rooted in historical mistreatment and abuse by researchers, this mistrust is a barrier to progress toward racial justice. To be clear, we believe this barrier is tech companies’ problem to solve, not our communities’. Toward that goal, we hope this article poses some useful questions and encourages us all to be more critical of how we go about correcting our course.

Historical mistrust in research participation — The Tuskegee Syphilis Study

Running from 1932 to 1972, the Tuskegee Syphilis Study is just one infamous example in a long history of racist mistreatment in American research. Told they were being treated by the U.S. Public Health Service and CDC for syphilis, a group of Black men in Alabama were actually receiving a placebo. The study led to the deaths of 128 people and harmed many others, deepening Black communities’ mistrust in research institutions. Today, even when the government mandates the inclusion of women and racial/ethnic minorities in government-funded research, Black people participate at much lower rates than white people. Research with Black people suggests that mistrust in who (e.g., the institution, company, researcher) is conducting the research is one of the strongest barriers to participation.

The legacy of the syphilis study and countless other examples of racism, mistreatment, and abuse is evident in Black people’s continuing mistrust in medical institutions. What’s important for us to understand as employees of tech companies is that the lingering collective memory of mistreatment in research — coupled with disparities created by organizations running it — likely influences Black people’s willingness to engage in any research.

Brand perceptions and trust deficits

Black people have been vocal about their lack of trust in big tech. A recent study estimated that only 37% of Black Americans trust large corporations to do what’s right in response to racial injustice and systemic racism. Rooted in a history of incidents like the Freedman’s Savings Bank failure, this lack of trust clearly extends to social media companies, as highlighted by the pervasive belief among the Black community that “shadowbanning” is common practice.

In addition to intensifying the barriers noted above, the lack of trust in tech companies likely contributes to mistrust of our research process and larger workforce. Employees of tech companies need to be aware that Black people’s mistrust in research may stem not only from a history of abuse but also from mistrust of the company or the industry as a whole, compounding their hesitation to participate or speak openly with us.

Many companies have started down the path of trying to do right by the Black community. For Facebook, Inc., this has included creating the Instagram Equity team and Racial Justice Research Council, hiring people across teams to work specifically on issues of inequity, and conducting research focused on Black users. Ideally, these steps will improve the experiences of Black people and build trust. As we attempt to make progress through these efforts, we must continuously ask ourselves at least a few questions:

In what areas, and to what degree, do we have the authority and the legitimacy to engage in the space of racial justice?
What problems should we solve?
What should we build?
Have we built the right foundations, including trust, to ask the questions posed above?

How bias blocks progress

Even when we are able to overcome the initial barrier of mistrust and conduct research with Black users, they may still be unlikely to be forthcoming or to provide accurate information about their experiences, opinions, and feelings. Research suggests that Black people are less likely to provide accurate information about their experiences, opinions, and feelings. This issue, commonly known as response bias, doesn’t apply only to Black people, but it does present a major obstacle toward understanding and meeting their needs.

One particular form of bias we try to mitigate is social desirability bias, in which research participants might be compelled to say what they believe will satisfy or please the researcher. The desire to put forth the most favorable view of oneself might lead to untruthful answers. For example, a participant who didn’t vote in the most recent election might say they did.

Years of research has suggested that the person asking the question — or believed to be asking the question — plays a huge role in the response to interview and survey questions. Race is important in the consideration of social desirability bias. For example, Black respondents may put on a performance when interviewed by a white interviewer, “donning the Black mask” instead of speaking frankly, which they’d be more likely to do in the presence of a Black interviewer.

Social desirability bias is often exacerbated by interviewer effects, which can go in many directions. One could make the case, for example, that a Black researcher may elicit an over-sharing of racialized experiences. But when a majority white cast of researchers (e.g. the research organization at most large tech companies) suddenly begins to interact with a majority-Black population of users — as is arguably happening more over the past year or so than ever before — the propensity to give socially desirable responses is worthy of consideration.

The race-of-interviewer effect: a mini thought experiment

Imagine two employees are undertaking a new study to understand how your product can better support Black people during the pandemic. As the two researchers begin to outline the study design and generate hypotheses, they start having conversations about the concerns outlined above. And so, as part of their larger effort to understand how COVID has changed the way Black people use their products, they decide to investigate whether the race of the interviewer changes participants’ responses. One of the interviewers in this scenario presents as Black while the other presents as white, and they divide 50 interviews between them.

One plausible outcome is that you’d find that the race of the interviewer makes a substantial difference in participant responses. For example, a large majority of participants might explicitly discuss their race or the unique challenges they, their families, or their friends have faced while using their products with the Black interviewer. On the other hand, it would be surprising if the participants discussed the same things with the white interviewer.

This example, while made up, likely reflects a reality that many of us have failed to consider as we’ve raced to understand Black people’s needs in our products. Researchers need to envision the collective research and conversations that have transpired across companies wherein Black people have either acquiesced, given less than their version of the truth, or opted out of research altogether due to lack of trust in the intentions of our work.

Any difference in participants’ willingness to express what could be central components of their racial identities, especially as those identities intersect with their experience using our products, reflects a history of long-lasting neglect that stands to be repeated if we do not actively work against it.

What might we do?

The first things we must do are to acknowledge that the barriers presented here are the results of historical and persistent mistreatment, and to reflect on how that truth might inform the best way forward. These first steps will help us move away from simple or singular solutions toward a system of solutions that address both the surface level and the more deeply embedded source of challenges. With that goal in mind, here are a few places to start.

What companies can do

Increase visibility of our Black researchers and employees

People sometimes mistrust institutions, organizations, or companies in part because of the perception that the people who work within them don’t look like them and/or don’t have their interests in mind. As an anecdotal example, in each of the branded studies I (Stacey) have conducted, at least one participant has made some version of this comment: “I didn’t even know Black people worked at Instagram.” To me, this implies a bit of relief that someone is listening to and working on the issues they’ve discussed in the interview. Highlighting that there are a group of people here who look like them, and who may even share similar experiences, could help build trust that people’s experiences, insights, and recommendations won’t fall into a void.

Hire more Black researchers and employees (in senior roles)

While increasing visibility might be a partial solution, it can also have severe negative consequences, including tokenism and overburdening. Tokenism and gaslighting could occur if we tout a relatively small number of employees solely for the purposes of brand perception. Overburdening could occur if we elevate low numbers of employees and expect them to be solely responsible for representing an entire community’s thoughts, feelings, and opinions. We need to hire more employees of color — and dedicate resources to develop and retain them — so that they can interact with our users and drive questions that stem from and reflect some of the lived experiences of their communities. Additionally, we need to make sure that Black employees are well-represented in leadership so that prioritized problems are more reflective of the entire community we serve.

Provide training and support on working with underserved communities

Provide training on impression management (e.g., seeking to be respected and seen as competent) and stereotype threat (the pressure to disconfirm and to avoid being judged by negative stereotypes), each of which can influence response bias. There are also resources and trainings to better understand how racism is baked into quantitative data collection and methods that companies could offer to their researchers and data scientists. Companies should package these resources and other related resources (e.g. from internal and external experts) to offer training so that employees can build more empathy into the ways we learn about community needs.

What individual employees can do

Engage in training and support on working with underserved communities

There are ways to demonstrate appreciation and acknowledgment of sensitive topics like racialized experiences when we engage with underserved communities. Community-based participatory research, for example, can help reduce barriers to trust. Built around the idea of doing research with — not on or about — communities, such approaches can maximize impact and minimize harm when executed well.

While companies can offer training, it’s the responsibility of employees to learn from the wealth of research on the effects of phenomena such as impression management (e.g., seeking to be respected and seen as competent) and stereotype threat (the pressure to disconfirm and to avoid being judged by negative stereotypes). Be mindful of who is doing the outreach to participants and whether the outreach message is one that signals trust. These and other well-researched concepts offer guidance on how we might build better understanding across lines of social division.

Be an ally, be humble

When working with underserved communities, it’s paramount to listen first. Researchers and other “experts” often have a set of hypotheses and assumptions about what underserved communities need. But those needs often differ from what outside observers expect (one need only look at the history of Urban Planning to see concrete examples of this). A major step toward being an ally is to bring the communities we want to serve in as true stakeholders, or better yet, have them bring us in so that we can be partners in the process. Listen, be aware of your implicit biases, and work to understand the needs of those we want to better serve.

Work with more contractors and vendors who specialize in working with (and who are a part of) underserved communities

By working with more contractors and vendors who work with specific underserved communities or populations, we can bring in voices that would otherwise be overlooked while reducing the burden put on employees who represent these communities. Employees of color at our companies cannot fully reflect the heterogeneity of experiences of the communities on our platform. Working with vendors or contractors who are part of the communities we intend to better serve also keeps any monetary gains within those communities, demonstrating our commitment while shifting power and building trust.

While we still have a long way to go on this front, we’re encouraged by recent progress. Since launching supplier diversity efforts in 2016, Facebook has spent more than $1.7 billion with US companies certified as minority-, women-, veteran-, LGBTQ-, or disabled-owned. Learn more and connect with us about these efforts on our Supplier Diversity Website.

Bending the arc of justice

Like many across the industry, we’ve focused this article on racial justice for Black people in the US, and focused even more narrowly on research. But the problem is transcendent. This is a call for introspection not just across functions, levels, and teams, but also across industries. We need to deeply interrogate how the historical and continuous mistreatment of underserved communities intersects with our identity as a people, as employees, and as companies to influence whether and how we travel the path of equity and justice.