what should asa’s gender categories be?

Cross-posted on Social (In)queery.

The ASA is trying to respond to a request from its members to expand the options for gender on its membership form. Right now, the choices are female, male, and prefer not to answer. There is no category that acknowledges transgender members at all, but creating a new category scheme is not as easy as it might seem. For example transsexual people and transgender people have not always appreciated being lumped into the same category. Some people reject gender categories altogether and might prefer a “none” or some other less alienating gender-non-specific category.

These gender categories are important for a number of reasons. First, by having exclusive gender schemes, the ASA is not acknowledging its trans members. Second, it is a missed opportunity to collect data on the size of the trans membership in the ASA. Third, gender categories are communicative; they tell members who may not be aware that transgender sociologists work among their ranks. Finally, it is important to get the gender categories right because they are teaching sociologists what the “appropriate” categories to use are, setting an important example for us as we design survey questions , courses, departments, etc.

The ASA staff have brought the matter to the Committee on the Status of LGBT Persons in Sociology, but there wasn’t consensus there. They propose (and are planning to implement), the following scheme:

  • female
  • male
  • transgender-female
  • transgender-male
  • other: ___________
  • prefer not to answer

I am not an expert in trans issues, but this scheme sounds wrong to me in all kinds of ways. It conflates the transgender identity with the FtM* and MtF** identities, which is not a problem for some people, but others see transgender as quite different from the identities that indicate a “switch” of genders. The “other” category is one way to capture gender non-specific individuals, but it is not the most inclusive way to do so.

I looked for guidance from some trans activism website, but there is a lack of consensus on this issue there, too. We seem to be working from scratch here. With that in mind, I am going to propose three additional schemes as conversation starters.

First, the Elegant Gender Scheme:

  • female
  • male
  • transgender
  • _______________
  • prefer not to answer

Second, the Thorough Gender Scheme:


  • female
  • male


  • FtM
  • MtF
  • transgender

No gender


Prefer not to answer

Third, the Open Gender Scheme:

  • gender identity: ___________________________

I wonder if others have opinions on this issue. While I don’t imagine there is one right way to do this, I can see quite clearly how easy it is to get it wrong.

*FtM = Female-to-Male = a man who was born in a female body

**MtF = Male-to-Female = a woman who was born in a male body

38 thoughts on “what should asa’s gender categories be?”

  1. As I’m cisgendered, I’ll defer to my trans colleagues on this, but my immediate reaction is that each of these options (except the open-ended) seems to inadvertently imply that being, say, a trans woman is not fully being a woman.

    Why not have two questions? First, female/male/other/prefer not to say. Second, trans/cis/prefer not to say.

    Liked by 1 person

    1. Yes! I think so too, though I am sure that it is not intentional. I like your overlapping categories, though I wonder about two things: 1) how many cisgendered sociologists know they are so? 2) are we excluding those who are gender variant in ways that these categories don’t capture; “prefer not to say” is not the same as “prefer not to have a gender identity at all” for example.


  2. In addition to transgender people, who have a variety of preferred self-designations (including I believe a fair number who identify only with their current gender and not their past gender), there are also the folks who are hermaphroditic or physically indeterminate in a variety of ways.

    One option is perhaps to go the route race has gone on the census: check all of the boxes that apply to you. And then try to figure out which boxes you need to capture the mobilized identities among sociologists. Again drawing on the race analogy, I am well aware that there are many logically-distinct groups who are not represented in US race categories because they are not relevant mobilized constituencies in the US (the clearest example being Australian aboriginals, who are clearly NOT members of any of the recognized US categories).


    1. Perhaps we can have a question that uses the membership form to collect these mobilized identities?

      And yes, I see what you mean about intersex individuals. An overlapping scheme like Elizabeth recommends might catch less mobilized identities like this one.


  3. Tina — thank you for inviting us to weigh in!

    I would say the answer depends on why such demographic information is being collected. What does ASA hope to understand by knowing one’s gender identity and expression? How is this information used, once collected? Prior to this proposed change, how has information on ASA members’ gender been used?

    Beyond knowing how many people are trans* and gender non-conforming, are efforts being made to assess how particular policies of ASA impact scholars from these communities? Or, the climate within the discipline, and in particular departments and universities? One issue I have heard, but rarely in public discussions about the discipline, is transphobia on the job market — ranging from being denied jobs because of supposed issues of “fit” to advisors encouraging job-seekers to dress in certain gendered ways (i.e., “appropriate” for one’s sex assigned at birth). I am certain these issues extend beyond the job market.

    I am happy to see gender-neutral bathroom options at ASA conferences, and now this acknowledgement of gender diversity among ASA members. But, I hope counting trans* and GNC people is just the first step in a larger project.


    1. I can definitely understand the logic of not forcing people to lump themselves into ill-fitting categories as a matter of amity or other essentially expressive motives. However since the prospect of doing research with the data has come up, let’s acknowledge the obvious, which is that there is very little that one can do analytically with very small subpopulations. My understanding is that about 0.1%-0.5% of the general population is transgender. ASA has 14,000 members. That implies about ten to a hundred people. If we start further disagreggating into MtF vs FtM, etc, then we’re talking even smaller groups with raw n of up to a few dozen and proportion of less than a quarter percent.

      Now in general when you do an analysis of a dataset and there are categories with only a few cases you tend to drop those cases, lump them in with one of the big categories, or stick them into the “not elsewhere classified” omnibus category. Hence you have things like the vast literature on how to recode occupations into relatively parsimonious categories or scores (almost 1400 cites for Ganzeboom et al SSR 1992) or the smaller literature on how to recode religions into a small set of broad categories (over 700 cites for Steensland et al Social Forces 2000). (Note that neither Steensland et al nor Ganzeboom et al have any substantive findings, their utility is entirely methodological about how to recode small categories, which is why I chose them as examples). When you’re creating taxonomies of the rich variety of human existence everyone is a splitter, but when it comes to doing analyses everyone is a lumper. How many times have you read a study that codes race as non-Hispanic white vs else? Even studies that are about race and not just throwing it in as a control almost never, for instance, disaggregate South Asians (about 1% of the population) from the ridiculously broad category that is “API.”

      As to the present question of the ASA gender form, in practice this would mean that even if we had a highly disaggregated membership form, any analyses would end up either aggregating all the small categories into “other” or aggregating them with the closest large category (e.g. recoding MTF as female). So basically I’m highly skeptical that a disaggregated instrument would serve purposes like allowing research into issues like transphobia on the job market because you’re extremely unlikely to get enough statistical power to pursue such a question. That’s not to say that in general members of small groups aren’t entitled to fair treatment and dignity or specifically that transphobia doesn’t exist or that it ought to be acceptable, only that it’s almost impossible to study things when you’re talking about such incredibly small numbers and even more so if you split them into even small numbers.

      Realistically I’m thinking all one could really do with the data for research purposes would be as a sampling frame for qualitative interviews (and this assumes no privacy concerns) but in that case you’d be doing further data collection anyway and so could ask people to give open-ended answers about gender identity and so there’s no research (as compared to expressive) purpose in having highly disaggregated data on the ASA form itself.


      1. Yes, I agree that this is likely a very small population, and I would like to add a further barrier to using these data for research. Not only would we have small n, but in a small population that knows each other, so ethics would preclude the ASA from publishing data on our trans members due to the likelihood of their identities being known.

        I think a good use of these data would be for the ASA to know whom to contact to ask “how are we doing around trans issues?” That said, I don’t think there are any specific uses planned for the data, although we might direct a question to Roberta Spalter-Roth, who collects all the membership demographics.


    2. Eric, the ASA is responding to requests to change the gender scheme (I don’t know from whom, but I suppose it is from trans members or allies). I think they are trying to address a range of concerns of the trans community. That said, I agree with Gabriel that there is little research that can come out of these data.


  4. There are a lot of label possibilities for the small number of individuals who don’t want to check male or female. Why try bureaucratically to pre-specify all the options when the categories are contested, changing, and small? In a case like this, regardless of the uses to which you plan to put the data, I reckon you get the best data with just this:

    x Female
    x Male
    x Write in:

    Then you code the responses.

    (A further complication is that gender identities change, sometimes purposely, so that you shouldn’t just collect the data once.)


    1. I think Phil is probably right here for the reasons discussed above. You can’t do a statistical analysis on such small numbers, at best all you could do is lump together everyone who does not pick one of the big two, or exclude them from analyses of male-female gender outcomes. Finding out how people do identify and using the information to compile a list of people who want to identify themselves to the ASA for purposes of mobilizing around these issues seems like the main goal, here.

      Again drawing on the race experience, what is at stake in adding race categories to the Census is people not wanting to be “other.” That is exactly WHY the API category got added and why, after push back, that umbrella got disaggregated into a bunch of ethnic labels but not discarded.

      At this juncture, the question regarding gender is whether the mobilized constituencies would be happy enough just to have “other” recognized in a gender list. I’m speculating maybe yes, especially if you are then invited to fill in the blank.


      1. On the subject of Census ethnic aggregations, Cristina Mora at Berkeley has a book about the institutionalization of “Latino” pan-ethnicity that’s coming out in February.


    2. This is why I asked about the purpose of asking about gender. If demographic information on members of ASA is collected for (quantitative) research purposes, than, yes, that probably limits how this question can be asked. If it is about genuinely acknowledging gender diversity among ASA members, these concerns about small sample sizes are irrelevant, and offensive when used as a reason to lump “everyone else” in one category.


      1. After you ask the open-ended question you can report all the diversity in the answers you code.

        (Census asks “what is you ancestry:__________” and gets thousands of different responses which are coded into lists when it writes reports. The lack of specific alternatives offered encourages rather than discourages diverse responses.)

        Liked by 1 person

      2. It’s also interesting that the Census refuses to code some of the open-coded ancestry responses but lumps them into a NEC category. For instance, if you write-in “Jewish” or even “Ashkenazi” or “Sephardim” as ancestry the Census recodes this as “NEC” whereas it would accept “Russian” or “Yemeni” or whatever. I see this avoidance of ethno-religious categories as an excess of zeal for the establishment clause and I wouldn’t suggest that the ASA throw out even seemingly joke answers (“eg “satyr”) to a gender question, but it is interesting to think about how organizations treat open-coded data.

        Liked by 1 person

    3. Phil – “female” and “male” refer to one’s sex, not gender.

      And, I’m worried what message is sent when (cis)woman and (cis)man are provided for ease of the cisgender majority, while “everyone else” has to write in their gender identity and expression. And, if only offering “woman” and “man,” are these thought to be inclusive of transwomen and transmen (yet, still excluding other gender identities) or not?

      I emphasize again that we as cisgender people, with trans* research experience or not, should have less of a say in the matter. Ultimately, the people whom ASA is attempting to better recognize should be having the most say.


      1. (Historical aside: “male” and “female” have been used to refer to gender since before “gender” was used to refer to an attribute of people. Which isn’t to say that the usage doesn’t evolve, but the grounds for correcting somebody else’s usage like these things are natural facts rather than moving targets is weaker.)


      2. So the question becomes whether you are trying to collect data or send a message with the categories you display.

        I am only responding to a blog post that asked for our opinions. Obviously, there are experts in this who should be in charge.

        On the data quality side, when you offer many categories to a population the majority of whom has never thought seriously about the issue, you get many errors from people misinterpreting the question. (File under cis-people problems.)

        One of our students at Maryland, J Michael Ryan, has done some fascinating work with NCHs about how to do this on government health surveys. Here’s one of their papers: http://wwwn.cdc.gov/QBANK/report/Miller_NCHS_2011_NHIS%20Sexual%20Identity.pdf


      3. Hence why I have asked a few times here — what is the point of collecting this information? Or any demographic information for that matter? I don’t say that to express opposition to such efforts; rather, I want to hear a clear articulation of asking that members provide ASA with details about their background. The justification seems vague at this point.


      4. Eric, if you want an official organizational response, you will have to ask the official organization, which is not me or any of the commenters here.

        That said, I can point you to the reports that the ASA has published using the data it collects from the membership form in the past.


      5. Another example: When I was on publications committee, there was the idea that editorial boards were supposed to be diverse, and that this principle was important enough to be worthy of specific review. So we would get lists of all the editorial boards and the gender and racial identities reported on the ASA questionnaire, and review those.

        If there was no information on gender/race that was collected, it would be harder to do that sort of review. Indeed, this issue came up when I was on Publications Committee because the point has been raised that editorial boards and such ought to be diverse in terms of sexual orientation as well. To my knowledge, ASA still doesn’t collect any information about sexual orientation. Of course, that doesn’t mean that editors do not have ways of identifying gay/lesbian scholars to serve on editorial boards. It does make oversight a more complicated matter, as unless people in the room happen to know every scholar enough to know their sexual orientation, there’s not a way from looking at the list of editorial board members a way to know the representation of gay/lesbian scholars.


  5. I shared this blog post on the Facebook group for LGBT caucus. You might want to look there for responses.

    I’m not sure how many of the folks who have responded here are trans* or gender non-conforming — and, I would strongly encourage asking members of those communities directly, rather than letting the cisgender majority of ASA decide.


  6. As a trans sociologist who also studies trans issues, I am glad the ASA is moving on this issue, but I do hope careful consideration is given to the language of the change. While the ASA may indeed end up lumping folks together due to low n’s, I don’t think that should be done automatically by the question. In surveys of LGBT communities, I’ve done the following (which isn’t perfect, but none of the quantitative survey items are that I’ve seen around this topic):

    1. What is your gender? Female, Male, Other, Prefer not to answer
    2. Do you identify as transgender? Yes, No, Prefer not to answer
    3. Which of the following identities do you identify with (check all that apply): Transwoman (MtF), Transman (FtM), Genderqueer, None of the above: (write in identity here): _______________, Prefer not to answer

    Now, this doesn’t get at the issue of sex and thus doesn’t include intersex individuals, which is definitely a limitation, though individuals could write it in. I’d also strongly encourage language that does not include “Other (write in)” to avoid creating the impression that the organization is actively “othering” people who do not fit into some designed survey question.


      1. Jamie’s suggestion is also a good one in my opinion, but only if who ever is crunching the data is committed to dealing with that type of data, which can be more cumbersome than a more streamlined choose one option question. If we want to capture people’s actual identities, open ended possibilities are the only way to do that in an inclusive way.

        Liked by 1 person

  7. Judging by past reports on gender (http://www.asanet.org/research/stats/sociology_programs/membership.cfm), this information may be used just to offer descriptive statistics of ASA and section memberships by gender: “Cisgender Women” “Cisgender Men” and… ??? I do not mean to be a pest, but I strongly encourage that any request for input be framed with “here’s why we want to know, and here’s what we will do with that information.” There are some ugly moments in the history of demographic data collection. So, understandably, marginalized communities may be wary of such questions. Right now, ASA and the discipline do not have a reputation of being free of transphobic bias and discrimination – so, even greater care should be made (in my opinion) in how this change is made, how input is solicited, and how the data are ultimately used.

    Again, I hope the true experts on gender identity and expression — here, I include trans* and gender non-conforming sociologists and, to a lesser extent, all sociologists who do work on trans* communities — are being asked directly. Cross-posting this on http://socialinqueery.com/ is a good start. Also check discussions on Facebook and Twitter.

    And, some work has already been done. Look at other organizations to see if they’ve made similar changes. And, look at the best practices guides out there:






    How you ask about gender identity and expression depends on what you actually want to know, and what you plan to do with the information. So, that must be decided upon and articulated first.

    Thank you again for making this important change, and seeking feedback from ASA members.


  8. How about an amendment to Philip’s proposal:
    – Male
    – Female
    – Other
    – Please elaborate if you choose: ______________________________

    More substantively: I think this problem evokes a point David Harris made about race: simply put, there are important social effects based both on self-identification and on ascriptive classification. Effectively, asking people how they’d like to be categorized is a good source of information on the self-identification question, but a poor source of information on the ascriptive identification question. and since many important public policy questions are at least in part about ascribed, not chosen, identification, that’s important data to have.


  9. I am not quite sure what the categories should be, although I tend to agree with Phil Cohen that we should have two, and Other, and code at least at first, and then develop the categories that work empirically.

    But I’d like to suggest that we conceptually clarify if we are being asked about the member’s sex (ascribed at birth or chosen) or gender, which is a broad socially constructed continuum… from feminine to masculine to genderqueer…

    Is this about expanding Male (which is a sex/identity), Female (which is a sex/identity) and what other categories?

    Or about GENDER. I resent being asked my gender, when people really want to know my SEX (or the sex i choose to identify with and claim).

    I think it does us all a disservice for sociologists to use the word gender when we mean sex (whether it is chosen identity or ascribed at birth).


    1. Free paper ideas buried in blog comments: what percentage of the time do articles in ASR/AJS that use the word “gender” refer to it in a way that is consistent with the sex/gender distinction that every sociology major is taught, and what percentage use “gender” just as a sort of interchangeable euphemism for “sex”?


      1. “Much has been made of the supposed distinction between sex and gender. Sex is thought to be the more biological, gender the more social; the relation of each to sexuality varies. I see sexuality as fundamental to gender and as fundamentally social. Biology becomes the social meaning of biology within the system of sex inequality much as race becomes ethnicity within a system of racial inequality. Both are social and political in a system that does not rest independently on biological differences in any respect. In this light, the sex/gender distinction looks like a nature/culture distinction in the sense criticized by Sherry Ortner in ‘Is Female to Male as Nature Is to Culture?’ I use sex and gender relatively interchangeably.” –Catharine MacKinnon, Toward a Feminist Theory of the State, p. xiii.

        See also Joan Fujimura on the social construction of sex categories (http://www.jstor.org/stable/10.1086/505612), etc.

        Ever since I read this stuff I’ve lost my motivation to impose the social/biological dichotomy.


  10. I’m the author of a blog post linked above on collecting data about sex/gender (http://trans-fusion.blogspot.com/2012/02/on-sexgender-checkboxes.html).

    I’m also intersex, and have gender transitioned from the sex I was assigned at birth.

    My main input here would be that very few people identify their gender as “transgender.” It’s not a gender identity, it’s a description of one’s relationship to one’s identified gender. Many trans* people who identify with a binary sex find the suggestion that we identify ourselves as “transgender” rather than male or female to serve the needs of a data gatherer quite offensive. The proposed “transgender-female” and “transgender-male” alternatives are likely to be found almost as offensive by binary gender transitioners, because the cis gender counterparts are not so labeled. In essence, the set of terms ASA proposes will be experienced as reading “woman, man, fake woman, fake man, other.” I understand that the concern is that writing “cis woman, cis man, trans* woman, trans* man, other (please identify_______)” will confuse cis people who are unaware of the term that describes them, but the solution of dropping the modifier is terrible.

    Speaking as an intersex person, the experiences and needs of people who are intersex by birth and of nonintersex trans* people are quite different. Thus I’d strongly suggest disaggregating these.

    The alternative I suggest is three-staged:

    Gender identity: Male __, Female __, Alternate Self-identification (please write in) ______________.

    Do you have an intersex condition? Yes__, No__.

    Are you trans gender? Yes__, No__.

    This avoids concerns that cis gender people aren’t aware of the term cis gender and won’t check it off, and of the conflation of categories under “other” (genderqueer people, binary gender transitioners, and intersex individuals). Allowing people to write in their own gender identification under other also strikes me as vital at this time, because of the huge proliferation of gender identity terms in the trans* community in recent years. For many younger trans* people in particular, having a very specific and complex gender identity term is experienced as empowering and of deep personal import. There’s no way to compile a list of all such terms, but not allowing people to state them will be experienced as hurtful.

    A note on intersexuality: there is a deep schism in the community between those who identify as “intersex” (now a political term) and those who identify as “people with DSDs,” that acronym being that of the current term, Disorders of Sex Development,” now in use in the medical profession. In the organizations I now am active in, the phrase “intersex/DSD” is generally used so as to include both the politically active and the medicalized, more closeted segments of the community. However, most people who are not born sex-variant are unfamiliar with the term “DSD,” so I just employ “intersex” in the demographic question I propose.

    I’ve also chosen the term “trans gender” in my third question as a sort of middle ground between the term most often used by cis academics today (“transgender”), and the term most often used by trans* community activists (“trans*”, with the asterisk serving as a wildcard character for the many groups in the community: transsexuals, genderqueer individuals, agender-identified people, etc. etc. etc.).

    Identity terms are fraught territory. Researchers with good intentions who are not aware of the concerns of trans* and intersex people often wind up alienating the people they think they are assisting, and I hope the ASA doesn’t make this mistake.


      1. I second that. What a great discussion this has been, both here and over at Social (In)queery.

        I think the common ground in these schemes is that one question is not enough to capture this, and especially not one question of mutually exclusive categories. Intersexroadshow makes the reasons for this very clear.

        I am going to bring this discussion to the ASA staff’s attention, and I am asking a few of the folks who have participated here to chat with the ASA staff and try to come up with a plan (it may well be the one proposed here, but it’s not for me to decide). I will certainly keep everyone updated on ASA’s decision.


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: