Well, I don't know where that discussion might have been, and I cannot say that I have a strong background in defining genders, but I have spent a lot of time designing application inputs.
My thought is to make the field free-text. I know that a lot of sites won't like how wide open that is, but in the end I think their objections are misplaced: it's easy enough to map all the synonyms of the genders you have a category for, and anyone who answers outside of those categories is someone who wasn't going to line up well with your targeted ads, cutesy avatar changes, or whatever else is driven by that gender field. What's more, if your interest is truly to do statistical modeling of your users (whether for hypothetical improvement of their experience or merely for demographic targeting) you are better off allowing as much specificity as your users can give you, because you can build clusters of them, which you can't reasonably do with that 25% Other/Unspecified.
If, for example, you wound up with a significant number of users identifying as "m2f", you could track similarities in their aggregate usage much more precisely than if they were lumped into a larger "other" category. Not that you can't have ways that your system recognizes and handles specific cases like "other", "both", "N/A", "etc", but those would no longer be the limit of what you are able to work with.
There is of course the possibility that a free text field will encourage people to enter random inputs. I don't know whether more people would do that than would randomly pick between M/F if forced to choose - when you're intentionally entering spurious info, it doesn't seem too likely that your limited choice of gender category is very statistically useful either. I suppose a Male, Female, Other (Please Specify) ____ setup might cut down on that a little, but then we're back to othering, which the free text alone will never be.
There are two cases where I think this doesn't work terribly well, one a user case and the other an application case. For the user, if they are unwilling to put down in any simple label what they consider their gender to be, this system won't work terribly well for them - I don't think gender should be a required field, or have a minimum number of characters, but if one does force that than it will be tough to accommodate those users.
For the application, I think there could be times when you want a very specific kind of quantification within gender axes. For instance, a survey related to gender issues might want to ask people to choose the closest approximation of their identified gender within the two-dimensional space you describe above. Clearly specific data mapping constraints will have more specific data input needs; in this case I think a Maleness slider and a Femaleness slider would work.
Of course plenty of sites might claim that they need that type of specific categorization for their own purposes, but most of them really just want something easy.
(no subject)
My thought is to make the field free-text. I know that a lot of sites won't like how wide open that is, but in the end I think their objections are misplaced: it's easy enough to map all the synonyms of the genders you have a category for, and anyone who answers outside of those categories is someone who wasn't going to line up well with your targeted ads, cutesy avatar changes, or whatever else is driven by that gender field. What's more, if your interest is truly to do statistical modeling of your users (whether for hypothetical improvement of their experience or merely for demographic targeting) you are better off allowing as much specificity as your users can give you, because you can build clusters of them, which you can't reasonably do with that 25% Other/Unspecified.
If, for example, you wound up with a significant number of users identifying as "m2f", you could track similarities in their aggregate usage much more precisely than if they were lumped into a larger "other" category. Not that you can't have ways that your system recognizes and handles specific cases like "other", "both", "N/A", "etc", but those would no longer be the limit of what you are able to work with.
There is of course the possibility that a free text field will encourage people to enter random inputs. I don't know whether more people would do that than would randomly pick between M/F if forced to choose - when you're intentionally entering spurious info, it doesn't seem too likely that your limited choice of gender category is very statistically useful either. I suppose a Male, Female, Other (Please Specify) ____ setup might cut down on that a little, but then we're back to othering, which the free text alone will never be.
There are two cases where I think this doesn't work terribly well, one a user case and the other an application case. For the user, if they are unwilling to put down in any simple label what they consider their gender to be, this system won't work terribly well for them - I don't think gender should be a required field, or have a minimum number of characters, but if one does force that than it will be tough to accommodate those users.
For the application, I think there could be times when you want a very specific kind of quantification within gender axes. For instance, a survey related to gender issues might want to ask people to choose the closest approximation of their identified gender within the two-dimensional space you describe above. Clearly specific data mapping constraints will have more specific data input needs; in this case I think a Maleness slider and a Femaleness slider would work.
Of course plenty of sites might claim that they need that type of specific categorization for their own purposes, but most of them really just want something easy.
--learnedax