Gender Census

[ survey.gendercensus.com ]

The 13th annual international gender census, collecting information about the language we use to refer to ourselves and each other, is now open until 13th August 2026.

It’s short and easy, for most participants it takes 5 minutes or less.

After the survey is closed I’ll process the results and publish a spreadsheet of the data and a report summarising the main findings. Then anyone can use them for academic or business purposes, self-advocacy, tracking the popularity of language over time, and just feeling like we’re part of a huge and diverse community.

If you think you might have friends and followers who’d be interested, please do reblog this blog post, and share the survey URL by email or at AFK social groups or on other social networks. Every share is extremely helpful!

Survey URL: https://survey.gendercensus.com

The survey is open to anyone anywhere who speaks English and feels that the gender binary doesn’t fully describe their experience of themselves and their gender(s) or lack thereof.

Thank you so much!

[ Link to survey ]

PS: You can see some regularly-updated statistics about incoming data here, with lots of demography and graphs to peruse!

I actually just finished a 2026-specific A5 flyer, which you can find here on Google Docs!

And here is the Google Drive folder of all flyers/leaflets. :)

Thank you very much for helping with promotion, and have fun!

[ Take part in the 2026 Gender Census before 13th August 2026 ]

This is a poll for people whose gender(s) or lack thereof are NOT encompassed by the strict male/female binary, e.g. nonbinary and genderqueer people.

(You can go by other pronouns too, this poll isn't about going by name-based pronouns exclusively.)

Example: If someone's name is Shane, their pronouns might be sha/shane/shanes/shanes/shaneself. "Shane bought shaneself a long novel for shanes intercontinental flight."

Today is day 16, which is officially halfway through this year's big annual survey on identity words, pronouns, titles and family/relationship terms for people whose experience of their gender(s) or lack thereof defy or fall outside of the binary model of gender.

[ Take part here! ]

Usually at this point we'd be 80-90% of the way to the eventual total number of participants, because the responses come in very fast at the beginning and then slow down a lot after the first week, and continue to slow more gradually until the end. This year looks like this so far, with the red line being 2026:

[Those of you using screen readers may have an easier time on the original Google Sheets table of data here.]

So I would guess that we're heading for about 30,000 participants in total. This is pretty low compared to the last couple of years, so I got curious and added a few more years of data:

[Those of you using screen readers may have an easier time on the original Google Sheets table of data here.]

Yeah, it's low! The lowest since 2020. (The three lines below the 2026 lines are 2018, 2019 and 2020. All the more recent lines are way higher up.) And I think I've figured it out.

2020 is when SmartSurvey (the previous survey provider) changed their premium plan so that I would have to pay a lot more to be able to edit the "thank you" page that people see after they've submitted the form.

Before 2020, when I could edit the "thanks for taking part" page, that's where I put the "hey can you also share this survey?" social media links. So, in 2021 when SmartSurvey wouldn't let me edit it without paying extra, I put a page right after the consent checkboxes like, "before we get started, can you share the survey please?" I.e. I was asking people to share the survey right at the beginning, before they had entered any information.

I remember at the time being concerned that having the survey sharing links at the start of the survey would make people click away and I would lose participants, but it seems to have basically doubled participation for 2021 and onwards, presumably through extra sharing. Numbers got so high that it was a good idea to limit the survey to one month, even.

At the time it wasn't really possible to know for sure whether the huge jump in participation was down to the sharing page being at the start of the survey or just some other thing, like a critical mass of social media followers/mailing list subscribers. I left it there at the start because SmartSurvey didn't change that rule for several years. And this year I twitched to Trello, who do let you edit the "thank you" page, so since I was remaking the survey from scratch I figured I may as well put the sharing links at the end where people expect to see them.

Participation has been wayyyyy down. For the first few days I couldn't tell by looking at social media/emails/anon asks, because the volume is always too high to deal with and this year was no exception. But I was distracted by the aforementioned technical difficulties, and other life stuff. However, by the end of day 2 it was clear that participation was very low, and I just blamed myself for not having time to do as much promotion as usual.

A couple of days ago (survey day 14) I remembered about my suspicion several years back re: the sharing page being near the beginning causing the jump in participation. I put a "please share the survey" links page right after the consent checkboxes again just to see what happened, and I know we're in the final stretch now but participation does seem to have been bolstered. I feel like it's taking less promotional work from me to keep the numbers up.

Thus concludes my interesting discovery of the year! But while I've got your attention...

If you had to return your gender due to a manufacturing fault, please click here to participate in our annual gender language survey!

Day 15, and 25,000 responses! :)

Check out this fun graph that Google Sheets made on the public participation spreadsheet very briefly while I was updating it. :D (It's back to normal now but it was gloriously incomprehensible while it lasted.)

Nope!!! I have ADHD and I'm a bit demand avoidant, so it's gonna surprise me as much as it does you. (Context: this ask is 5 days old.)

I can't remember if I've addressed this in a report, but anyway, I'll have a bash at it now. :)

I haven't investigated this as thoroughly as I could. (I could probably look at average ages of participants from surveys as far back as those questions go, split by prior participant/new participant, and see if the prior participants' average age is going up by one year per year.)

It's not a priority though. I might play around with it later today, but no promises - I also have life things to do!

Some info that I do have is: this year about half of participants are new to the survey, and participants' average age has been increasing by about half a year per year. (That's all on the participation spreadsheet here.)

If you don't want to share your name that is 100% okay, and please only answer questions you feel comfortable answering.

Also, I don't currently plan to process the information about names without attaching it to said names. So if you are intending to not enter any names, that is fine but also it probably won't contribute much to the lists and graphs and such.

Some info that may be reassuring:

I won't be publishing the spreadsheet of responses.

I will ONLY publish a plain list of names that were entered more than once, along with how many times they were entered. So if your name is 100% unique and no one else is using it, it will not be published.

But I totally understand that you might not want to enter your name still, and that is honestly completely okay and I don't want you to feel pressured into it.

On the referrers tab of the participation spreadsheet, 16 people arrived at the survey from a flyer/leaflet, and that's right below people who arrived from Twitter (22 people).

I think it would be great and really funny if flyer/leaflet could be above Twitter. That would be so satisfying for me, because Elon Musk is a knobhead and old school hardcopy everything is the best.

So, if everyone could please print out some flyers (PDF, Google Doc) and stick them up in random unisex toilets/changing rooms, or pin them on pinboards, or stack them on tables, or put them in postage-paid envelopes from random junk mail and post them, etc. etc. that could help the cause.

Let's stick it to The Man (Elon "Knobhead" Musk).

Unrelatedly, we're at almost 25,000 responses, whaaaaat!!!

Well, I count them in several different ways (here's the write-in pronouns sheet from 2025 as an example), but for ranking for the report I usually use subject/object. So, whether or not you put a typo in the independent possessive (the fourth one) won't affect the ranking there.

However, what you're asking me is whether I can modify your entry, and I can give you my go-to answer for that!

If you entered anything you're sure is unique into a textbox in your entry, I can find it, verify that it is definitely unique and therefore yours, and then either modify it or delete it so you can resubmit.

However, I am not going to do that over Tumblr anon ask. :D You will have to email me ([email protected]), and it does mean your entry won't be anonymous to me anymore, but if it helps I have a terrible memory and I don't tell anyone any identifying info.

If you didn't enter anything unique, there's nothing I can do. I'm not going to edit a random thon entry because it might not be yours as that would be inappropriate/unethical, and it would mess with the data quality and make it less trustworthy.

Also: Thank for your efforts in this survey! It's always enjoyable to take it and it's very interesting seeing the results! We're sure it's very time and energy taking.

Why is there a limit of 5 new pronouns?

TL;DR: Only 0.3% of all participants get as far as using the fifth set of new pronouns textboxes. Google Sheets limitations means that providing five new pronoun set inputs for only 0.3% of participants is actually pretty generous! 😅

~

This is a balance between various factors, like how many pronouns people want to enter, and the cell limit for Google Sheets.

First thing, there has to be a pre-defined limit in survey design, because it's not really technically possible to tell a survey to collect new pronouns from people until they get bored of entering them.

Second thing, in 2025 9% of people checked the box to enter new pronouns (worth having entry options for extra pronouns). 25% of those people (2.2% of all participants) wanted to enter more than one set - and only 3.5% of them (0.3% of all participants, only 137 people!) wanted to use all 5 sections for new sets.

As a general rule of thumb (which is very flexible and context-dependent), if I provide several textboxes in one question, if a textbox is used by under 1% of participants I remove it. So like, if there's 10 textboxes but under 1% of participants use more than 3, I'll consider reducing the number of textboxes to 3 - or maybe 5 if that feels better.

If I apply this rule of thumb to pronouns, I could reduce it to just 2 new pronoun sets, because under 1% of participants used the third part of that section. The main reason I haven't done that is because I just... haven't really thought about it.

Third, the section for adding up to 5 new pronouns uses 44 columns on Google Sheets before processing, which is more than twice as many as any other question, and the whole survey uses over 130 columns. 44 columns for 40,000 rows of responses is nearly 1.8 million cells. The Google Sheets cell limit is 10 million cells, so that's one section using >17% of the cell allowance before I even start processing the data. (Cells for those questions count towards the limit even if the participant was never directed to that section of the survey.) Only 39,850 of the >1.7 million cells for new pronouns contained anything after processing in 2025, which is 2.3%. 98% of cells unused for a section that 9 in 10 participants never even saw!

So that you can compare them yourself, here's the survey questions and how many columns they used in Google Sheets in 2025:

Timestamp and response ID: 2

Consent checkboxes: 3

Identity checkboxes: 20

Identity textboxes: 10

Titles (with name) checkboxes: 1

Titles (with name) new title: 4

Titles (without name) checkboxes: 9

Titles (without name) new titles: 5

Pronoun checkboxes: 15

Pronoun textboxes: 44

Family words #1, checkboxes: 13

Family words #1, textboxes: 5

Family words #2, textboxes: 5

Demography and feedback: 5

This is a lot of numbers to process as a reader, I know. A whole limitation and a bunch of extra maths just because I'm using Google Sheets. But until I learn to use something other than Google Sheets to process this stuff, which may never happen, that's what I'm working with.

What's it like to feel for the pacing of the survey? (What goes into how long it's made? What thoughts go behind whether typed answers should be in a separate section or not? Really anything that has to do with the flow of the survey experience... it's really intruiging to me!)

What goes into how long it's made?

The main factor in the length of the survey is how many questions I can process the data for. Like, I get a lot of people asking in the feedback box why there are only 2 family/relationship questions, and that's mainly because if it's too much then I will find it stressful and overwhelming and it just won't get processed. It's short because that's manageable for me.

What thoughts go behind whether typed answers should be in a separate section or not?

I think this is mostly down to complexity of input for follow-up custom text entry answers.

So there's more straightforward follow-ups. That's like, check all the boxes that describe your identity. And then, if there are more, there's just 10 textboxes right underneath, which are all under one heading of "extra identities not covered above". That's simpler if it's all on one page, because people can refer back to the checkboxes if they want to and it's just 10 textboxes in a block. Super easy to scroll past, one box per identity word, and it's good to be able to see it all at once.

And then there's more complex follow-ups. That's like the pronouns question. Inputting one new pronoun set requires 5 forms plus the name of the set, whether it's gendered, and which verbs it can take. That's several questions just for one set, and people often want to (or don't want to) enter more than one set.

But, because only 1 in 10 participants want to type in a new set of pronouns, it doesn't make sense to make the other 9 in 10 people scroll through a bunch of repeated textboxes and help text and multiple choice questions to get to the next section. Similarly, many people only want to enter one or two new pronoun sets, so they should be able to skip three redundant "new pronoun" entry sections.

So, anyone wanting to enter a new pronoun has to check a box saying "a pronoun set not listed here" in the checkbox section, so that they can get taken to the first new pronoun entry section. And everyone else gets to skip a lot of pages, thus reducing the chances that they will get fed up of scrolling and clicking and just exit the whole survey.

Ultimately, whether anything is in a new section of its own is geared towards data quality by making it easier for the participant so they don't leave - while also giving everyone who wants to do so a chance to enter a bunch of extra stuff.