Assuming that you are not studying the matter of gender self-identification
itself (i.e. the dynamics of the count of the very small number of people
who do not consider themselves either male or female), it seems to me that
a survey over time should tally (self-identified) MALE / FEMALE / DECLINE
The "Decline to Respond" category will be extremely significant -- a far
bigger threat to generation of an accurate count than the loss of the
fraction of 1% of Wikipedians who consider themselves neither here nor
there on the matter of gender.
This is, of course, not particularly "sensitive to the many categories of
non-male and non-female genders that exist in Wikipedia," which you
apparently presume to exist a priori.
"Carrite" on WP /// "Randy from Boise" on WPO
Show replies by date