tips for making your survey/questionnaire/census/psychological test more inclusive!

if you don’t understand any of the terms used in this post, please feel free to ask me what they mean! (but i also recommend doing a quick Google search first)

i’ve been taking part in a lot of university-run studies lately! a lot of them include demographic information and i’ve noticed that it can often be super hard to describe yourself in the context of three or four multiple-choice questions, which in my opinion leaves out a lot of relevant information and provides less accurate results. i want to talk about some of the things i’d love to see in more surveys, etc.

gender options: these are super tricky. first of all, i do want to point out that there are studies in which gender is not relevant, and in those cases, having a mandatory question to which the only two answers are “male” and “female” is only serving to make things more difficult for your participants. if you do include a choice like that, “gender” is always preferable to “sex” or even “designated sex” or “legal sex”.

if you can, ALWAYS include a third option — it can be “other” or “did not wish to report” or even a “fill-in-your-own” box. if you are not able to provide a third option for some reason (i guess this could happen, for example if you have to present your findings to a conservative organization that may reject them based on that data), there are alternatives!

for example, something as simple as changing the question from “What is your gender?” to “Which of these describes you more accurately?” or “Which of these do you identify with more?” can make non-binary test subjects feel more receptive to having to make the choice.

never list your gender options as “male”, “female” and “transgender”! most transgender people consider themselves to be male or female, whichever is their gender identity, and forcing them into an “other” category is harmful and alienating. if you’re interested in collecting data on how many participants are transgender, you should make it a separate question! a simple “Are you transgender?” does just fine.

when considering whether or not a third gender option is “necessary” for your survey, remember that the entire point of such a census is to obtain firsthand results that are as accurate as possible. in the interest of accuracy, please provide a nonbinary option, even if you are testing a relatively small group or in areas you consider to have a low genderqueer population! (accuracy aside, you are also providing a positive example to other testers who see your work — and probably making the day of any nonbinary person who takes your test and unexpectedly finds themselves included.)

sexuality options: as with the gender options above, in most studies, sexuality is not relevant. don’t feel obligated to include a question about sexuality just to be more inclusive, but DO NOT fall into the trap of assuming your test participants are heterosexual just because the question of homosexuality is not relevant to your subject. if you are offering a test that mentions the participant’s spouse, and this test is only available to females, it’s always best to say, for example, “how much would they agree with your opinion” or “how much would he or she agree with your opinion” and not “how much would he agree with your opinion”.

in the same vein, don’t assume that your test participants are legally able to marry any partners they may have. don’t assume their relationship carries a risk of pregnancy. don’t assume that your participants are monogamous (i saw some EXCELLENT survey language the other day: “If you have multiple partners, please pick one for the following questions, and answer all questions with the same partner in mind.”) don’t even assume that your participants have/want/will ever have/have ever had a relationship; that’s bad testing!

when you do include a sexuality question, my advice is to always include at least “heterosexual”, “homosexual”, “bisexual” and “asexual”. however, it is best to have all of these, in addition to an “other” option and/or the option to write in your sexuality.

if you are running a test relating to someone’s partner, and the gender of their partner is a factor, it is better to ask what their partner’s gender is than to rely on what they reported as their sexuality. obviously, the “bisexual”, “asexual” and “other” options leave it open to the possibility of an any-gender partner. there are also definitely cases in which someone will report as heterosexual while dating a same-gender partner or homosexual while dating an opposite-gender partner. if you don’t want to do this, you should use more specific language: if your study is testing something specific in gay male relationships, ask if your participant is a man dating another man, etc.

never list your sexuality options as “heterosexual”, “homosexual”, “bisexual” and “transgender”. people who are transgender also have a sexuality! it is a bad idea to force your transgender participants to choose between reporting their sexuality or reporting their transgender status. as mentioned above, if you want to know if there are transgender participants in your study, it is best to have it be a separate question.

as a final aside, i want to talk about surveys that deal specifically with testing a negative or positive bias towards transgender people, people with disabilities, people with mental illnesses, etc. if you are running a study dealing with social perceptions against a marginalized group, i recommend asking the participant if they are a member of that group to improve your data. don’t underestimate how valuable it is to know whether the marginalized group in question seems to have a high or low opinion of itself, whether it considers itself oppressed in society, or anything else you intended to ask of a privileged group! remember, the marginalized group your test revolves around does not just exist hypothetically.

if you read this post and are using it to improve your testing methods, thank you so much! if you read this post and felt that something was omitted, it means that i either am not educated enough on the subject to provide guidance or that i did not notice any bias was present. i recommend writing your own set of guidelines to supplement any gaps in this one.

if you notice research that you feel could be improved (with some of the guidelines mentioned in this post, or with your own ideas for inclusiveness), please make a difference by contacting the group running the survey and letting them know! usually, such an error is present due to a lack of knowledge and not any actual bias. real scientists are interested in obtaining data that is accurate and relevant, not data that confirms their existing beliefs.