Top 5 tips for collecting LGBTQ+ data (by Kevin Guyan).
Start with “Do we actually need this data?”. Don’t collect sexual orientation or gender identity data simply because it is possible or because a form asks for it. Be clear about why you need the information, what you will do with it, and whether collecting it will meaningfully benefit the research or the communities involved. Where relevant data already exists, consider secondary data or working with existing datasets and community organisations rather than repeatedly asking people to disclose information.
Design for “box breakers”, no just the majority. Avoid assuming that everyone will fit neatly into predefined categories. People may identify as queer, questioning, bisexual, pansexual, non-binary, avoid labels altogether, or find that none of the available options adequately describe them. “Other”, “prefer not to say” and free-text options can be important, but the wider design of the question should also allow people to describe themselves in ways that work for them.
Remember that identify is not fixed and that data captures a moment in time. Sexual orientation and gender identity can change, and people’s willingness or ability to disclose can also change depending on their circumstances. Be transparent about what your data represents: it is an account of someone’s identify at a particular point in time and in a particular context, rather than necessarily representing their whole identity or life course. Consider using a critical disclaimer when presenting or interpreting findings.
Don’t make disclosure the price of participation. Ask yourself: “Do you need to be out to be counted?”. People may not feel safe or comfortable disclosing their sexual orientation or gender identity, particularly in certain organisational, political or social contexts. Participation should not depend on someone being willing to disclose, and researchers should consider carefully whether collecting identifiable information could create risks or harms, particularly with small samples where individuals may be easily identifiable.
Think critically about inclusion: more data isn’t automatically better data. An inclusive-looking question can still exclude people or produce misleading results. More categories don’t necessarily mean better data, and researchers need to consider who benefits, who might be harmed, what assumptions are built into the categories, and whether the question is meaningful to all participants. Inclusion should therefore be treated as something to critically evaluate, rather than an automatic good.
Key learnings from Kevin’s session based on attendee feedback:
- Will collecting the data make anything better for the communities you are working with?
- Build for the box breakers
- Do you need the data and would you ask for the same level of detail about a different protected characteristic?
- Including disclaimers around the inherent limitations of sex/gender demographic data
- Leading with ethics and safety in working out what the best approach is for different kinds of studies
- To think critically about what we are asking people to tell us, and why – The need for more research/discussion to bridge some quite radical theorising on sex and gender data vs data collection and eligibility screening practices in clinical contexts where some of this data has key implications for patient safety, treatment efficacy, adverse effects ect.