LGBTQ issues are worth researching yet rarely included in studies. People have asked why. Is it an oversight or lack of interest? From a methodological standpoint, it is much more difficult. The issue comes down to sample size and definitions.
On almost any research project, I request demographic information. Depending on the research needs, questions may include location, ethnicity/race, education, and income. Yes, gender (or sex) is often included. When I ask these questions, the answer may be “I am HUMAN.;” or “why does this matter?”
Demographic questions fulfill two purposes. First, it helps the researcher to gauge the representativeness of the sample. If the survey comes back with 85% female and 90% white, problems need correction. Second, demographic subgroups have different needs, interests, and buying patterns. Including these measures allow demographic subgroups to speak for themselves.
The problem is one of representativeness. Often members of the minority community complain that they are responsible for explaining the experiences of their whole community. The same problem exists in research. Does a sample identifying as LGBTQ represent the opinions and needs of the entire group?
There are many ways to determine the necessary sample size (or sub-sample). The population size, sampling methodology, measure sensitivity, and other issues are factors. However, let’s consider a target sample of 200 individuals representing a group. The Human Rights Campaign used US Census data to indicate that the LGBTQ community is 8-10% of the population (depending on how you define it). To reach 200 LGBTQ participants, you would need a target population of at least 2000. A sample size of 2000 is possible, but there are further considerations:
It may be challenging to get candid responses to the LGBTQ question without extreme sensitivity. Generalizability may be an issue if there is a difference between participants willing and unwilling to identify as LGBTQ. As with any demographic subgroup, it is unfair to represent the whole with a flawed sample.
There is an ethical issue involving collecting data you cannot use. For example, in a recent survey, I saw the gender question asked with the possible choices of male, female, non-binary, transgender to male, transgender to female, and other. The researcher should ask if it is possible to collect enough responses that the two transgender questions are useful. If not, it is information you must protect for no positive effect.
A Facebook friend posted this advertisement. A t-shirt with the words “There are more than two genders,” yet the shirt is only sold for men or women. It seems to be an apparent inconsistency. However, this is a great example of definitional issues – the difference between gender self-realization and the manufacture of the shirt for body shape.
The main definitional issue is inconsistencies within the LGBTQ community. Would it be reasonable to place all LGBTQ people into one category on many issues? For example, are the experiences of gay men and women the same? In what way are trans people different from gay? How do you divide the group into categories that provide reasonable subgroups? For some issues, an LGBTQ subgroup may be reasonable (even if imperfect).
The conclusion of this post is not to discourage research. Yet research must be built with logic and sensitivity so that the participants’ help is honored with valid results. As much as we want to include a group, if we cannot do it well, it should be avoided.