Identity Inclusion in Market Research

About a 4 min. read

Authors
James Kelley, Ph.D
Analytics Director

As a market researcher, political scientist, teacher—and human—I spend time thinking about concepts related to identity. I personally identify as a white, straight and cis male—in other words, I am not a person who has been under-represented or disenfranchised because of my race, sexual orientation, gender identity or other characteristics. But, I still have an obligation to understand that other individuals—including potential participants in our research—have been under-represented, misunderstood, and marginalized. Thoughtful consideration of identity offers opportunities to avoid bias, to better understand your audience (specifically how they view themselves and how they want to be identified), and can help your clients better communicate with their consumers. This is critical to good market research, and it’s simply the right thing to do.

Here are a few key—but non-exhaustive—ways market researchers should include identity in their work:

1. Create a safe space: Participating in qualitative research or taking a quantitative survey instrument is a daunting task. Everyone can relate to the “short” survey that ends up being ~20-minutes. Imagine reading the first few questions of a survey that asks you to divulge your identity, and your identity is not represented. You may feel unseen and/or unheard.

To help minimize these negative experiences, consider providing opportunities for respondents to self-identify using open-ended questions. While this requires coding responses after field, it’s one easy way to ensure everyone has a chance to identify themselves accurately. If you must use a closed set of alternatives, here are some recommendations:

  • Race and ethnic identity: Consider including longer, more inclusive lists of ethnic/racial identities, allowing respondents to select as many options as they identify with, and ask about LatinX identity separately (don’t assume any of these identities are mutually exclusive).
  • Gender identity: Move away from the binary, biological terminology (male, female). Instead, consider using more identity-related terms (such as man, woman rather than male, female), providing a non-binary option, and allow a safe space for people to self-identify (“Another gender identity” with an open end as optional).
  • Sexual orientation: Asking questions about sexual orientation (not preference!) is inherently personal and is only appropriate if you have a compelling hypothesis. If you’re going to ask about sexual orientation remember to be as inclusive to the LBGTQIA+ community as possible (and understand that terminology around identities is evolving), and ensure there is sufficient representation among identities. Vanderbilt University has a thoughtful set of guidelines, which you can access here.

2. Be comfortable with some discomfort: Decisions you make related to asking questions about identity might result in backlash. Feedback related to your survey will either appear in open-ended responses, especially for gender and sexual orientation, or via email. Consider these responses to come from two groups: allies and antagonists.

  • Allies will provide feedback, indicating that you’ve missed a group identity, aren’t using the correct term, or need to revise the way you ask a question. These folks are genuinely trying to help and often point out ways you can improve your questions. I file these under “learn and adapt.”
  • Antagonists will generally indicate things like “There are only two genders”, include profanity and/or include language about identity politics. These individuals are fundamentally activated by inclusion and there’s nothing you can do about it. I file these under “ignore.”

3. Ask identity-related questions at the end of the survey. Unless your research is focused on identity and you need specific representation, or if you need to click balance on these dimensions, save personal questions for the end. Inclusive identity questions early in a survey could activate participants and bias your results. Additionally, including demographics at the end of a survey is increasingly common in social science research.

4. Don’t waste data. Now, it’s time to analyze the results. As researchers, we’re inclined to save banner space by including the modal groups for analysis and “othering” those who don’t fit into the larger groups. What’s the point in gathering identity data if you’re not going to use it? Consider the following:

  • Include all identity groups with a readable sample size (n=30 per the central limit theorem) as banners in your analysis. Where there’s not 30 respondents, consider grouping individuals together, but only if there are theoretical reasons why these groups might be similar. Do not assume homogeneity among non-modal groups, as this is both “othering” by definition and is a loss for meaningful insights related to identity.
  • Take the time to read the data by identity subgroup. See where subgroups are different from others, understand their relationship to your client or subject matter, postulate why these differences exist, and think about ways to unpack causal mechanisms and generate new questions for future surveys. One way to be a good researcher and an ally is to treat identity-related data with the care and the attention it deserves.

As market researchers we are not expected to be identity experts. However, we are expected to uncover insights that help to understand our consumers—not just the modal group. Refreshing the way we frame identity questions and using the data to uncover meaningful insights is one way we can be better humans and better researchers.