Data science thrives on accurate data collection, yet remains an eluding prospect for gender inclusivity

Knowing one's anatomy is one thing and knowing the way it is oriented is quite another. If you know what it means you are gender aware. But unfortunately, many people do not understand the difference, ie., the difference between sex and gender. This fact is crucial for composing a good gendered-based data set.  Many people either identify themselves with a binary gender or refuse to reveal their orientation giving a huge data science challenge. A 2017 analysis found that only 0.4% of people identify with non-binary genders while it has enhanced to 1-2% a meager increase in 5 year period. However, the number of adults who expressed their gender openly is 3-5%, according to Gallup Inc. and the Pew Research Center. The problem doesn't solely lie with the people but also with the way data science surveys are designed. Some studies have clearly mentioned that gender-based data collection is rife with methodological and conceptual blunders for a gender gap to creep in.

How much flawed are the survey designs?

Most of the surveys either would have male and female as options with the occasional 'other' option. Take a look into the transgender demographics, it would be clear that it would be unfair to put them under one umbrella term. For data scientists and surveyors, it is a huge challenge as the surveys designed do not give regard to this fact nor can people come out of the veil for the consequent social stigma creating a data hole.  Alexis Dinno, a professor at OHSU-PSU School of Public Health in Portland, Ore., explains the gravity of poorly designed surveys. He says, "I'm an epidemiologist, I'm an epidemiology professor who is transgender, and I cannot tell you what the top cause of death among the transgender population is, or what the major burden of disease is, because historically we have not asked."

Inclusive Questions in Gender Based Data Collection

Is it possible at all when a majority of 56% of respondents opine other options shouldn't be included? While creating widespread awareness remains a constant responsibility, the surveyors need to learn a lesson or two for making the survey forms gender inclusive. Surveyors, researchers, and program implementors can accurately capture gender-disaggregated data by including two sets of data. This data collection technique has since long been recommended by statisticians. One set of questions can be framed around the assigned sex at birth and the other set around their personal choice., and what gender they identify with. For eg., Pew research frequently includes "Do you describe yourself as a man, a woman, or in some other way?", at times including diverse questions such as, " How do you currently describe yourself?" with options, Male, Female, Transgender, Nonbinary, Nonconforming, and None of the above, or a question like "Which of the following best represents how you think of yourself?", with options Gay or lesbian, Straight, Bisexual, Something else, I am not sure, and Refused. The surveys can get so much diverse that the options may have 'Gender not listed here', an option The General Social Survey included in 2018 with a choice for respondents to give a textual description.