Handling gender in statistical test data

Question

Context: imagine we're collecting data for a statistical analysis where biological sex is a factor (for example to determine optimal drugs dosage.)

Problem: when operator (the one who will insert data) is trained then a field named biological sex (or maybe even just gender) may be appropriate and enough. Available options are male, female and other (which must specified and handled separately).

The problem arises when the form must be filled by population under test (directly or with a printed form). My concerns are:

• Are gender at birth and biological sex clear enough (at least for an average English native speaker persona)?
• Is it considered a sexiest/discriminant wording for people with a different or non-binary gender identity? I need to collect only their biological sex, I'm not interested in how they identify their gender and I also want to keep wording as simple and clear as possible.

I'm currently thinking to add a separate, and unused, Gender identity (or I identify myself as) field to avoid ambiguity and to respect their desire to express their identity. Is it a viable solution? Should it be a list or free-text?

Answer 1

This is indeed a matter of phrasing. What you need to know is biological sex, not gender. Gender is personal, a sort of sense of self. Biological sex is what others have assigned you when you were born as a method of classification. Since this would be for medical purposes, use biological sex.

Gender at birth would be confusing to some, as there are people who feel like they have been a certain gender for as long as they can remember, rendering these emotionally the same. You need to know the biological sex.

The best way to approach this is to be very clear on why you need to know what sex people were assigned with at birth. If you explain that there are medical reasons why this is relevant (healthcare differences: occurrence of certain illnesses, risk factors, dosage etc.), you will steer the focus from personal gender identity towards neutral safety and health factors. It will still be a difficult subject, but you will have added an acceptable and proper reason to ask for it. Also add how you use the data and that you protect it properly.

I would not add a "I identify as X" field, because it changes the perceived focus of your survey to their gender identity. It will make it more personal than it needs to be.