Over the couple of years, I’ve spent a lot of time researching, writing, and taking surveys, and I’ve been keeping an eye on how these surveys handle questions about identity. Demographic questions are frequently considered an afterthought – questions that everybody is used to answering these days about their gender, race, age, and more. But when these questions are not written thoughtfully, the impact can actually cause respondents – your employees, customers, or other stakeholders whose opinions you care about, or you wouldn’t be surveying – to feel excluded by your words.

So how can you write inclusive demographics questions and avoid this outcome? Here are 6 things I’ve learned:

1. Be expansive in your options (when it makes sense).
2. Don’t “other” people.
3. Allow multi-select.
4. Don’t ask unnecessary questions.
5. Be transparent about how you’ll use the data.
6. Allow people to skip questions.

Be expansive in your options (when it makes sense).

The problem

Being asked to responds to open-ended questions is one barrier to survey response rate. Yet, many demographic questions include a limited list of options for survey takers to select from and otherwise ask the individual to fill in the blank. For example, a question on sexuality may ask people to select from straight, lesbian, gay, and other. (We’ll talk about the impact of the “other” option later on.) This excludes many identities such as bisexual, pansexual, asexual, queer, questioning, etc. and puts the burden on people taking the survey to type out this information when they’re not provided a choice to select that represents their identity.

The solution

Do your research to learn about the less commonly included identities in questions like these – and then include them in your survey! You won’t be able to list every possible identity in each category, but providing a longer list serves two purposes: it reduces the barrier of entry to your survey for many people who can now check a box instead of filling in the blank, and people who still don’t see their identity listed will respond to the effort you made to be more inclusive. An excellent demonstration of this is a survey I recently took that had every subregion of the globe listed as an option for a question on race/ethnicity.

Note that you probably don’t want to ask for very detailed responses to questions that don’t require a high level of detail for your decision making process or that feel very personal. We wanted to know whether any employees live with health conditions, but did not need to know specifically which. Instead, we asked whether employees live with a disability, a chronic illness, a mental illness, a sensory impairment, or neurodivergence. These categories balanced the information we needed with a respect for the level of information we believed employees would be comfortable disclosing (even anonymously) at work.

Don’t “other” people.

The problem

Even when you include a broad scope of options for your demographic questions, you’re unlikely to list every option that every respondent might identify as. That’s what open-ended responses are good for! Just be considerate about how you phrase this option. Many surveys list this option as “Other.” This phrasing is exclusionary; you are literally othering people who don’t fit into your predefined list of options.

The solution

Consider phrasing the open-ended options as “A [blank] not listed above” (e.g., “A gender not listed above”). Now, you’ve acknowledged that you recognize your answer choices can never encompass every possible gender/sexuality/race/etc. and that you are open to seeing however your respondents might self-identify.

Allow multi-select.

The problem

Many demographic surveys ask people to put themselves in a box: are you Caucasian or Black or Asian? Single-select (also known as radio button) answer choices limit people to categories that often are not representative of how they actually identify. What should people who are biracial select in this scenario? Single-select demographic questions ignore the diversity of identities that your employees and customers potentially hold.

The solution

That better choice is multi-select (checkbox) options, so you allow each individual to choose any of the options that fit how they self-identify. Combined with expansive answer choices as described above, people taking your survey will now be able to select all of the answer choices that fit their identity. On a recent survey, we asked employees to “Please select all that represent your current identity” when asking about gender and sexuality. The addition of the word “current” acknowledges that many aspects of identity are fluid, and your survey only represents a single point in time.

Don’t ask unnecessary questions.

The problem

You should have a plan for your data from the beginning of the survey development process. What changes will you make based on the answers you get? Surveys set expectations. It can be easy to get carried away when brainstorming a survey with all of the things you’re curious about knowing from your audience. But if you ask questions that you’re not prepared to take action on, you can foster discontent if survey respondents don’t feel like their input is being taken seriously.

The solution

If you don’t have any plans for using the data from any given question, then don’t ask. For a recent survey I was involved in writing, we debated a question (which ultimately did end up on the survey) asking about employees’ current living situation. Because we’re in a pandemic and a large proportion of employees are working from home, this question helps us understand how to acutely support employees who are living alone or caring for elderly family members, among other living circumstances. If it weren’t a pandemic, we may have opted not to ask this.

Don’t get this confused with limiting your answer choices, though. As discussed above, if you are going to ask a question, it’s important that your answer options are inclusive.

Be transparent about how you’ll use the data.

The problem

Because demographic questions are personal, people need to feel comfortable with how you are going to use their data. There are many reasons that someone may choose not to answer a question, but being asked to disclose potentially sensitive information for no apparent reason is one. This is especially true when demographic questions are part of a survey where understanding personal information is not the primary goal. You already know your rationale for asking each question on your survey – that also needs to be communicated to your audience.

The solution

Carefully craft an introduction to your survey so that the people taking your survey understand why you are asking particular questions and how you’ll handle their data. This can be done in an intro within the survey itself or in the body of the email you send to distribute the survey. You don’t need to explain in detail your rationale for each individual question, but instead your overall reasoning for asking demographic questions in the survey. Understanding this will make your survey respondents feel more comfortable with giving you the requested information.

Allow people to skip questions.

The problem

Your goal should be making everyone taking your survey feel comfortable while taking it. Although you’ve taken all of the tips above into account when crafting your survey, still not everybody will want to answer demographic questions – even if you offer a “prefer not to answer” choice. That’s expected, and okay! People can still provide valuable insights into the other questions you have for them, if you give them the opportunity.

The solution

Allowing survey takers to skip questions, especially those asking personal information, is imperative. Set your survey up in such a way that answering demographic questions isn’t mandatory for continuing with or submitting the survey.

With these tips in mind, you’re well on your way to crafting an inclusive survey that will collect the responses you’re hoping for!