Mastodon demographic survey, May 2017

Mastodon demographic survey, May 2017

@cassolotl@cybre.space

I ran a survey of Mastodon users, because that's a thing I like to do in communities I like and am passionate about.

It ran for one week, from 22nd to 28th May 2017. There were 544 participants, mostly acquired through boosts of one promotional toot that got 193 boosts at the time of writing. (I also posted on Reddit, Twitter and Tumblr - but they didn't get much attention.)

I have only self-taught/amateur experience in running surveys, so I'll point out the flaws as best I can.

Spreadsheet of results here.

Participation

It dipped a bit from 25-27th May because I was away from home, and not promoting the survey so much.

Age

For this question, you literally just wrote your age into a text box as a whole number. 18 people left this question blank.


I also did a graph of people by age group:

Over two thirds (67%) are 21-40. This is pretty representative of most similar social networks; mayyyybe Mastodon skews a little older - the median age was 31. The youngest participant was 15, and the oldest was 70.

Gender and sexuality

Gender

The options were:

  • Man
  • Nonbinary or otherwise binary-defying/transcending/rejecting/redefining
  • Woman
  • Other [text box]

I tried to keep this simple and inclusive, and I think I did okay. Not many people used the text box, which I think is a good sign that people felt included.

Before you look at the graph, be aware that people were able to select more than one gender - but not many people did. (So that's why it's not a pie chart.)

This may be biased by the fact that I'm nonbinary, so I might be attracting more nonbinary and trans followers and therefore more nonbinary and trans survey-completers... but it also might not. I don't really know.

Thanks to a limitation of Google Sheets outlined below, I can't count the "other [text box]" responses except by manually trawling through every response in the spreadsheet and manually counting them, which does not sound like fun so I'm not going to do it. I mean, there might be a way to do it automagically, but I don't know about it. If you know a way and you download the spreadsheet and wiggle the data and post it online, let me know - I'll link to it here. :)

Trans/cis stuff

The options were:

  • I'm transgender
  • I'm cisgender
  • Other [text box]

This was another where you could select more than one, and again, most people selected only one.

I added this question about 67 responses in. This caused some outrage. I do get that, but also, I'm doing this for fun and I do what I want. So I figure it's actually okay to give you the graph and the statistics, as long as I tell you that 65 or so responses are blank because that question wasn't there yet. You're adults, you can handle it.

This may be biased by the fact that I'm trans, so I might be attracting more trans followers and therefore more trans survey-completers... but it also might not. I don't really know.

Sexuality

Again, I added this about 67 participants in. I'm sure you can cope with that, though, right?

The options were:

  • I'm a lesbian
  • I'm straight
  • I'm queer
  • I'm gay
  • I'm asexual
  • I'm bisexual
  • Other [text box]

This was choose-as-many-as-you-like, and I was surprised by how many people took advantage of that. A fair few people chose "queer" and something else, for example.

In the graph, I included everything typed into the "other" box that got more than one response.

Note that a bunch of the "blank" options were probably because I added the question about 67 respondents in. 81 people have no sexuality information in their response.

A note on bigots

One person just wrote "simple guy" in the other box, which is... awful, really. A bunch of people wrote "normal" in the gender, cis/trans and sexuality questions, which is equally awful. Ew. Don't do that, people. It tells us nothing about your identity except that you are awful. Trans and queer people are normal.

Country of residence

33 countries were represented. The top 3 were USA, France, and the UK.

This is undoubtedly affected by the fact that the survey was mostly promoted in English, on largely English-dominant instances.

Disability

This one was one-choice-only, so yay, pie chart!

Race/ethnicity/ancestry/etc.

This is extremely difficult to report on, because I just provided a blank text box. I counted the at-a-glance most common words and put them in a graph, but if someone else can do better with the data please feel free and I'll link to it here.


A couple of people got really angry that I was even asking this at all, calling it useless or offensive. But, y'know, I think having confirmation that we are overall more white than Twitter is some very useful information, if you know what I mean.

Occupation

This was a choose-as-many-as-you-like question. The options were:

  • Job of some kind
  • In education of some kind
  • Neither job nor education presently
  • Volunteer work
  • Other [text box]

I feel it should be noted that a bunch of people used the "other" text box to say that they were a full-time parent, and I also feel like I should have included that as an option in the first place.

Coders/devs/programmers

This was a vague and (for some people) confusing question, and that's because I'm most definitely none of the above and have no idea what I'm on about. But I wanted to ask anyway, because I get the feeling that a lot of folks here are on the "code it yourself" path, and I wanted some numbers about that. I know, I know, they're very unreliable numbers, pinch of salt, etc.

So the majority of us are coders or something maybe perhaps, I guess?

Accessing Mastodon

How do we use Mastodon? Here were the options:

  • Laptop or desktop browser
  • Some kind of app on your mobile device
  • Phone browser
  • Tablet browser
  • Carrier pigeon, owl, etc.

I like to drop silly answer options in there sometimes just to get across to people that this is really not a serious survey at all and please stop sending me messages about how biased and unrepresentative it is. So, big up to the ~7% of participants accessing Mastodon from Hogwarts, you rebels. 💖

Mastodon in relation to other apps

Oh god(s), this question.

About 50 participants in, I decided that the answer "I use it the same as before I joined Mastodon" was too confusing/misleading, so I changed it to "My use of it is unaffected by Mastodon". And then people whinged at me about how very very bad that is. Yes, yes, this is very bad. Be aware that I've done that. I split the results of this question in the spreadsheet and graphs to reflect this.

Anyway, this was a fun question! I made two graphs, but this was my favourite:

You can see it in more detail on the Google Sheet, but basically:

  • Most people had previously used Twitter, and most people had not used any of the other things.
  • Mastodon has in particular caused people to use Duolingo and Patreon more.
  • Of all the apps/sites/things listed, the one that people were least likely to have tried was another incarnation of OStatus (such as GNU Social, Friendica, Quitter, etc.)
  • People are using Twitter and Facebook less because of Mastodon, and people are most likely to have replaced Twitter, Facebook and another OStatus thing entirely with Mastodon.
  • Probably other insights can be made - shout if you think there is something else I should include.

Bias

I used Google Forms, because I am used to it and it's easy and user-friendly and I know how to work with Google Sheets. I knew in advance and was reminded many times through the survey's duration that this would bias the results - only people who were willing to give personal information to Google Forms were taking part, which means the many anti-Google people of Mastodon were not taking part. I'm comfortable with my decision, because I doubt I would've had the energy to do this survey using different software, and survey is better than no survey. Plus, I'm doing it for myself, for fun.

There's also bias just because the survey is in English. The vast majority of Mastodon do not speak English as a first language, and I didn't translate the promotional toot until partway through the survey, and even then it was only translated into German and French. So, that limits the survey pool a lot. Similarly, there was probably bias because I posted about the survey most days, and I did so in my waking hours - timezones significantly shifted from mine might have missed out.

And finally, there is bias because I did the survey and it was promoted entirely by people (directly or indirectly) connected to me posting and boosting. That means the respondents are limited to a certain degree-of-separation. I'm on cybre.space so I'm thinking... witches.town, awoo.space, cybre.space, and connected instances?

Limitations

Google Forms is great, but it's unhelpful in one way that really bugs me: If someone checks "other" and leaves the text box blank, this is not reflected in the spreadsheet of results. If anyone checked only "other" and then left the text box blank, it will look to me like they skipped the question entirely.

Summary

This was fun! Thank you to everyone who took part.

This was a survey of people who can read English, are willing to use Google Forms, and are directly or indirectly connected to me.

Basically, even in my queer circle we are overwhelmingly white, male, cis, straight, not disabled, and into code.

Thanks

Big thank you to @luluberlu@witches.town for translating the toot into French, and to @maunzikation@computerfairi.es for translating it into German! :) Also thanks to everyone who replied reminding me of biases and such. I was aware of them all already but if I hadn't been it would have been extremely useful and important! And finally, thank you to everyone who boosted the toot to get it as far out of my personal circle as possible.