Δαφνη Ευτυχια Αρθουρος

On a mailing list I read, someone proposed, out of idle curiosity, a survey to get a glimpse at the demographic distribution of the membership. One of the questions was the expected M-or-F one, which led to some discussion of phrasing (which I think started completely independent of my pointing out that the sample form needed at least one more option for that field).

I felt like sharing one of my posts from the middle of the discussion here (slightly edited, mostly to change email formatting to web-style). I could analyze my reasons for wanting to do so, but that'd be a whole 'nuther post worth of musing, so I'll just go right to the text...

[The previous poster in the thread] wrote:

I've been told by some of my social scientist friends that the correct term is "sex", i.e., the answer to "are you [male] or female?".

It really depends on what you're trying to measure. Quite often that would be the wrong question, and even when it's the right question, unless you attach a definition to the survey, it can be confusing.

("Confusing?" I hear folks ask ... well yeah, if you really do mean sex -- i.e. biology -- do you mean morphological sex, hormonal sex, or genetic sex? Each of those can be a valid or invalid definition depending on just what it is that you are trying to measure. And you still need at least three categories (the easy cop-out is to lump lots of things under "other"), because of various intersex phenomena. That is, even if you want to limit things to morphological birth- sex, you've got at least five categories that show up often enough to count (something like 1% of births is still a lot of people), and if you mean genetics, there's Kleinfelter syndrome (somewhere around 0.1% of births). And then you've got the questions of which sex a post-op transsexual counts as: if you're trying to figure out which bathroom they'll use and which sex acts they can perform without props, morphological sex matters; if you're screening for sex-linked genetic disease, chromosomal sex matters. If you're trying to get pregnant, both matter, and hormonal issues enter the picture. And if you're trying to decide whether someone is allowed to attend a women's music festival, then politics gets stirred into the pot, but I digress... )

Note that outside of medical contexts it's almost always morphological sex that matters when you actually mean sex at all, and more often than that it's actually gender that's meant in the first place.

I'm surprised that a social scientist would say to use sex. After all, aren't social scientists usually more concerned with whether a subject is a man or woman, rather than whether they're male or female? And man/woman is a gender thing.

Evidently sex is a pure physical distinction whereas gender refers to the psychological, behavioral, or cutltural traits associated with the sexes, so that then latter is a continuous and not a binary variable.

Well, depending on your model of gender, it can be considered a continuous one-dimensional variable, a discrete multidimensional variable, or a continuous multidimensional variable. (Okay, it can also be modelled as a discrete but non-binary linear variable, but I don't find that model at all useful.) Note that the BSRI (Bem Sex Roles Inventory, named for its creator, Sandra Bem) treats it as a two-dimensional continuous variable, allowing for (varying degrees of) both "bi-gendered" and "ungendered" in the 'middle' ground between (varying degrees of) masculine and feminine. (It does not, as I recall, distinguish between "bi-gendered" and "inter-gendered", nor really allow for "strongly female-identified butch", so even the BSRI is just a starting place. But hey, I suspect even Bem would agree that the tool is outdated now.)

So which do we mean if we want to gather statistical information about the [list] membership? Are we tallying penes, or asking how many men and women are on the list? I think we're more interested in gender than sex here, but hey, if I'm wrong I'm wrong. Could make for a cool X-rated [recurring list project] theme though, if it's really pudenda we want to count, eh? But no, I really think we want to count gender here.

Either way, there need to be at least three choices. If anyone besides me is interested in a more comprehensive list of options for statistics-gathering, we can discuss that, but just adding "other" and/or "no response" as valid options at least makes the survey possible for folks like me to fill out accurately.

        -- Glenn

PS: Why yes, I have thought about this (and read, and listened) quite a lot and consider it important. I write letters of complaint when sex/gender is a required field on a computerized form (including web forms), and have been known to pencil in a box for "other" on paper forms. The proposed [list] survey, if it ever does happen, isn't a Big Important Deal, but awareness of the issues for the next time someone here has input to the design of a form is something I do care about. It's partly a matter of personal distaste for bad science ("Oh look, your data are guaranteed to be incorrect for some subset of your respondents!"), partly a cultural/institutional invisibility issue ("How can they know whether we number enough to be taken into account if they have no way of even counting us or finding out we exist?"), and partly a matter of privacy ("If all I'm doing is registering to use a free article archive on the web, why do you even need to know my sex/gender at all?")

I've filed off the identity of the mailing list 'cause my posting the message here isn't intended as slam against the list -- which has a lot of pretty nifty people on it -- and my posting it there was intended as education more than argument (though I may have hit a more strident tone than intended). Some time after I posted it to the list, I started wondering about the "social scientists say sex instead of gender" thing. It sounds wrong to me, but for all I know, maybe that is the convention there...? Most of discussion since has been about whether people can have "gender" or only words can.