Open nedjo opened 5 years ago
One option is to have a field that indicates if the data was gathered by statement or by speculation. So the person entering can say "I asked for this" or "I'm guessing about this". Depending on the context, of course, you could convert that into something like a confidence interval - I'm VERY sure or I'm NOT sure at all.
Wanted to second this. I know self-identity is ideal, but there might be room for exceptions.
For example, if an organization has a pool of potential guest-speakers, and in order to be more thoughtful about who's getting added to this list, there might be a best-effort attempt to informally surface demographics to help organizers understand their own biases. It might not make sense to expect each person added to have self-identified (when perhaps there's been no prior contact), but capturing that uncertainty feels important, so it doesn't propagate forward with assumed high certainty into more concrete/public places.
Anyhow, just wanted to add an example of a more grey area of self-identification.
@drnikki @patcon thx for the tips and ideas!
A lot of great work has gone into Open Demographics--thank you!
One context within which demographic data are frequently tracked is CRMs such as CiviCRM. Often in these systems, an individual is not entering data about themself but is being added by someone else--for example, a staffperson or volunteer responsible for membership or client engagement.
An example is the native Drupal CRM built on the Contacts module. Currently that module ships with a binary gender field. An issue is open to fix that. One option would be to replace the current implementation with one based on the Gender Field Drupal module, which implements the list of genders developed here.
But it seems like one thing to self-identify, and something quite different to have someone else apply an identifier to you. Yes, in almost all cases, the best answer to "How should I collect this?" is probably, "Don't!" Still, any thoughts or clarification about whether or to what extent the approaches developed here are applicable to data collected about others would be welcome.