Tracking diversity at your conference

This is an Ask a Geek Feminist question. Questions are still being taken this round.

This one came up on the Python Diversity list:

How can we gather data on the gender balance and other aspects of diversity at our conferences without asking attendees intrusive questions? Is having numerical data not that important? But without it, if our female attendance goes from (say) 150 to 180 or to 120, we might just eyeball the crowd and think, “Not enough”, not realizing that we’re doing something important right or wrong.

Skud, Terri and I had a conversation about this in comments last year, focussing more on making it optional than on doing it without questions at all.


How do you suggest tracking the diversity of speakers? Gender can be approximated but not perfectly measured by looking at people’s first names (especially if you don’t have an ethnically diverse conference) but in general the problem we have with is that we can’t see how to do this well without a demographic questionnaire, which women especially have repeatedly said they don’t want to see because they feel like they will then attend the conference as A Representative of Womankind.


Yeah, that’s hard. Can you make the question optional, and link it to an explanation of why you’re asking it? Something like, “$conf supports diversity and is working on improving the mix of speakers at our event. To this end, we are trying to measure our progress. If you don’t mind, could you give us a few demographic details?”

If that’s still not culturally comfortable, you can get an approximation by just working off what you know. Eg. “Of the people we know, N are people of colour/from other countries/mid 20s or younger/whatever.” After the conference, you will know more of the people (esp. first-timers), and be able to adjust the figures accordingly.

We went on to discuss Australian/US/Canadian cultural differences, namely that Australians ( is an Australian conference) are used to, at best, much more limited demographic questionnaires from, for example, employers, grant funding organisations and so on than people in the US and Canada.

What do you think, folks? Do you attend events that use demographic questionnaires? How do they go down, culturally? Are they optional or compulsory? Is there a third way between that kind of measurement and educated guesses?

9 thoughts on “Tracking diversity at your conference

  1. angela

    This makes me a bit uncomfortable:

    If that’s still not culturally comfortable, you can get an approximation by just working off what you know. Eg. “Of the people we know, N are people of colour/from other countries/mid 20s or younger/whatever.”

    I have seen people (in the US) try to figure out the percentage of, say, politicians of color by looking at last names & photographs. Some of the people using this technique were academics & policy wonks, thus lending it a patina of respectability, but I think it’s really shoddy to do. As someone who is mixed-race & who has been on the receiving end of a wide range of guesses as to my background, I am sensitive to people trying to guess what I am (& they are frequently wrong). You might not know, even after the conference, unless someone’s racial background specifically comes up as a topic of discussion.

    I imagine lots of other demographic details could be as ambiguous to guess in many cases as well, so I think making estimations based on “what we know” could be wildly inaccurate.

    I appreciate the sensitivity of asking people directly but I feel like trying to make guesses, even educated ones, is potentially more problematic.

  2. Amanda

    I think optional, anonymous, open ended surveys are best. They can either be done when registering or when signing in at the conference. Open ended means having a blank line instead of check boxes with predefined options, i.e. gender___________ not gender : (b0x) female (box) male. It does make it harder to distill down into x% women y% men, but I think it’s more accurate and comfortable, especially for people in attendance on the transgender spectrum. To that end, the “visual inspection” method of “determining” the gender of folks in attendance is quite flawed (same thing for tracking racial diversity).

    1. Mary Post author

      But gender, for which there is a (problematic) binary which a lot of people advertise with their name, clothing, title, etc, is a lot easier to “head count” than other types of diversity; I think much of the problem is that this can’t be easily extended to other things.

  3. Jon Niehof

    For Anime Boston, we collected demographics at registration time, but it wasn’t tied to the individual…just did an “add one” to the (age, gender) record. This was explained in the FAQ, linked from the question in registration. (Of course, you had to trust I wasn’t lying about my code…)

    Were I writing it now (or even maintaining it now), I’d handle gender differently. We also lost the ability to do some cross-correlations (e.g. gender or age with geography) but I simply couldn’t justify storing the demographics with the registration.

  4. SarahM

    Could you request the conference attendees to indicate their preferred title (Mr./Ms./Mrs.) when registering? Most people don’t find that too annoying. There will be a good number that will prefer Dr./Prof., but you could get some demographic info from the gendered titles.

    1. Terri

      I’ve always wondered… do these surveys wind up counting the Dr./Prof. folk as male? I’m betting some do.

      Incidentally, I’m constantly irked when I can’t opt out of that Mr./Ms./Mrs./Dr. question. I prefer not to be addressed as Ms. Whatever (a quirk I inherited from my father that was amplified when I started teaching and had to hear Miss? Miss? Miss! all the time), so I try to opt out of the title to avoid that even though I don’t mind providing my gender.

  5. moose

    This is such a push-mepull-you problem. On the one hand, we want to see more women/disabled folks/non-white folks, etc. involved and part of Open Source conferences. On the other hand, you easily “other” women/disabled folks/non-white folks by trying to make some kind of identification.

    I’m working on the survey for the attendees of the Ohio LinuxFest. I’m asking for gender but it is optional (there is a “prefer not to answer”). Also, since I am making it :-), it will have a “thirdgender” option.

    On the subject of speakers, however, I’m now wondering how we can get even more diversity at OLF. We have two speakers this year who up front admitted disability [in only once case it was relevant to their talk; in both cases it is relevant to communication]. One question I ask myself is, how do I work on making less of a sea of white faces? Obviously I cannot say, “Hey, if you’re not white, submit a talk idea to…”

    (For the record, I feel like I’m tiptoeing on land mines. Please tell me if I’m being offensive here.)

    1. Mackenzie

      By the way, I am totally up for interpreting into sign language the questions from the crowd during the talk on Friday for which that may be necessary.

Comments are closed.