Do we know the demographics of metafilter users? August 17, 2022 12:27 AM   Subscribe

Reading the recent user survey summary, I see a lot of talk about the identity of the user base. Do we actually know what the demographics are of MetaFilter users? Has there ever been an attempt to survey the users' identities? It would be interesting to know how people think of themselves in terms of nationality, politics, gender, race, sexuality, class, etc. I often read assertions in comments about the identities of the user base, but I am wondering if they correspond to the user base in reality.
posted by mortaddams to MetaFilter-Related at 12:27 AM (33 comments total) 6 users marked this as a favorite

Hi, sorry for the delay posting this! As far as I know, we do not have data like this, and we've been awfully careful to *not* collect the sort of information that is often used to monetize groups / communities / users. Loup or Jessamyn may have more information about any sort of demographic info we do have.
posted by taz (staff) at 12:33 AM on August 17 [10 favorites]


Many years ago there was a survey that produced an average age of 34 for Mefites, and so much lying and joking that that data and all other data was interesting but statistically useless.
posted by Cranberry at 2:08 AM on August 17


MetaFilter: interesting but statistically useless.
posted by bryon at 3:36 AM on August 17 [26 favorites]


I think that the upcoming Steering Committee will be eager to maintain the momentum which the Transition Team started with their survey.

Personally, if I were part of implementing a means by which we can survey the membership more often, I would start by researching best practices for how to securely gather, temporarily store, and use anonymized PII. That would allow us to answer these demographic questions less intrusively.
posted by thoroughburro at 5:06 AM on August 17 [4 favorites]


I just want to say that I appreciate Metafilter's commitment to respecting anonymity in the past. I hope that as the steering committee takes over that this policy, of not collecting demographic data whenever possible, will continue.
posted by JDHarper at 6:24 AM on August 17 [17 favorites]


Highly intelligent and highly opinionated.

That is all I need to know.
posted by JohnnyGunn at 7:00 AM on August 17 [4 favorites]


Back in MetaFilter's peak activity years, I wrote my PhD dissertation about this place, including surveys over time with demographic data. It's lengthy and flawed, for various reasons, but there are some charts that are fairly representative. Things are wildly different now.

I just discovered that the link to PDF is broken (site is defunct), but you can download it here.

Perhaps the mods can update the first link in the How do you pronounce MeFi MetaTalk thread please?
posted by iamkimiam at 7:15 AM on August 17 [18 favorites]


I don't mind providing my demographic information. I used to be a demoralized Democrat, but now I'm more of a Demogorgon. But I've always been effing graphic.
posted by It's Raining Florence Henderson at 8:51 AM on August 17 [4 favorites]


Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups, assuming the goal is to acknowledge, celebrate, and increase diversity of a community. Individual anonymity goes without saying (I hope) but aggregate and intersectional stats are key for a DEI effort.

(He says, as a member of what's probably the most overrepresented demographic category in history.)
posted by supercres at 10:15 AM on August 17 [7 favorites]


Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups

How about assuming the best of our community? By all means make a case for collecting demographic info, but you can do that respectfully without attribution bias.
posted by tangerine at 11:43 AM on August 17 [7 favorites]


I would start by researching best practices for how to securely gather, temporarily store, and use anonymized PII.

It is worth noting that this is a pretty complex question, and depends on what one considers to be an adequate level of security, and what one considers to be PII. Knowing the general culture and politics of MetaFilter, it seems likely that any discussion on this question is going to arrive at a general consensus that only fairly strong answers to these questions are going to be considered acceptable. I'd be concerned that solving this problem to a level that the community is willing to accept could balloon very quickly beyond what MetaFilter's limited resources are capable of.

That said, it's possible that MetaFilter might be able to contract this problem out using tools similar to what academic researchers use. One such tool is called RedCap, which is used to conduct, among other things, medical survey research, with a level of security and privacy that is capable of meeting FDA standards. As MetaFilter is not a non-profit, it would not be eligible to license the RedCap software directly, nor would that probably be desirable to do, but it seems that other organizations can contract to run RedCap studies either from Vanderbilt University or a third-party called RedCap Cloud. I have no idea what the cost of such options to MetaFilter would be, nor whether MetaFilter would be eligible to contract with either of these services, but it is probably cheaper than trying to build something in-house that meets the level of privacy standard that I predict most Mefites would demand.
posted by biogeo at 11:48 AM on August 17 [9 favorites]


I think it might be possible to use projects like Presidio to avoid expensive licenses or SaaS.
posted by thoroughburro at 11:53 AM on August 17


Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups, assuming the goal is to acknowledge, celebrate, and increase diversity of a community. Individual anonymity goes without saying (I hope) but aggregate and intersectional stats are key for a DEI effort.

I will note that at least for the Transition Team-driven study, there were notes from a number of people concerned that they could be identified by demographic questions, as well as significant privacy concerns generally related to identity. There were also some difficulties deciding exactly how to parse various identity categories effectively (e.g. with respect to some gender, national, racial, and ethnic identities) and how narrowly vs broadly they should be parsed. We opted not to ask questions about identity not because we were not interested in the results (profoundly untrue) but because we wanted to focus cleanly on community issues while minimizing privacy concerns and miscategorization concerns. Finally, the main direct use of collecting demographic information we could articulate involved marketing MeFi to online advertisers in the hyper-specific mode currently popular on the Internet, and that was an income stream that seemed immediately like a bad idea anyway. So we opted instead to focus on questions that seemed more immediately important to the vision that users have for the site, regardless of who precisely those users are.

When you're trying to celebrate diversity, it does pay to ask what kinds of diversity you're interested in measuring. We had multiple people on the Transition Team who have experience trying to fill out surveys that ask demographic diversity questions that have no answer for you, or which make clear that your answer is not among those imagined by the survey designer. It seemed important to us not to highlight that inability to imagine for users during that first pass attempt to get a feel for what users think the community is, should be, and needs to do.
posted by sciatrix at 12:38 PM on August 17 [30 favorites]


Yeah, what sciatrix so eloquently said!

It did seem a bit odd to me at first, but in the end it worked out better because we were focusing on what people do with the site as opposed to who they are. The data was more pointed, IMO.
posted by Brandon Blatcher at 12:43 PM on August 17 [3 favorites]


Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups

There are also other explanations.

1. The most recent user survey was put together by a group of volunteers who came together to help MetaFilter be better and they were trying to do many things at once. The makeup of the Transition Team is almost entirely not people in overrepresented groups.

2. Our userbase has privacy and other concerns about sharing this information as sciatrix stated so well.

3. Historically this hasn't been done done since 2004. Here's the data from 2004 (relevant MeTa). It wasn't official and was very self-selecting. The previous administrations were, shall we say, somewhat conflict averse and so someone would float the idea of doing something where more demographic data would be collected, a whole bunch of nitpickers would poke holes in it (Have you met MeFites?) and it would never happen.

4. We're not fucking facebook. We have a very limited amount of resources, people are trying their best. I respect that this is not good enough or quick enough for everyone and I'm not saying mistakes may not have been made. Anyone who wants to imply that the absence of stats collecting is so that MeFi LLC can remain head-in-sand about the diversity problems here is warmly invited to help collect that data.
posted by jessamyn (staff) at 2:03 PM on August 17 [45 favorites]


And analyze it. One of the most common responses to the concern about not providing enough options to allow people to accurately self identify is to provide a "fill in the blank here" option. Which, from extensive personal experience, is generally not super compatible with quick, clean analysis.

In the TT survey we chose to emphasize non-directed questions at the expense of easy analysis in order to avoid biasing answers from folks. Even with just ten questions, our volunteer survey coders did a truly phenomenal amount of work getting the analysis completed and done on time. Adding more open-ended questions without a clear goal for the resulting data did not seem like a good use of volunteered, gifted time.
posted by sciatrix at 2:13 PM on August 17 [18 favorites]


"This is important information to gather" is not in any way a direct attack on the recent member survey, and it's weird and offputting that it's being taken as such.

Demographics are important to gather.
And it makes sense that the recent survey did not ask for them, given privacy concerns.
And therefore it would make sense to gather them in a different way, in a way that is voluntary and also does not raise privacy concerns.
posted by lapis at 8:12 PM on August 17 [4 favorites]


lapis, I don't think anyone is treating that idea as a direct attack. However, supercres said that not doing so is a "head-in-the-sand tactic," which is an attack in that it ascribes malicious, or at best negligent, intent, and I think it's understandable that some of the people who have worked hard on collecting the data that we do have would want to push back on that.

The tendency to respond to something about site governance that one disagrees with not simply by stating disagreement, or asking for clarification, but by immediately jumping to an accusation of ill intent, bad faith, negligence, or incompetence, this tendency is unfortunately pretty deeply ingrained in site culture on MetaTalk. If MetaFilter is going to successfully move to some form of community-governance model, we can't afford to keep doing it. Otherwise we're going to burn through all the good will and energy of the people willing to volunteer their time to help run the place, and within a year no one's going to want to step up to be on a steering committee or work on specialized projects.

I don't know if supercres intended their comment to come across as an attack, but if not the phrase "head-in-the-sand tactic" was poorly chosen. For that matter, I think describing each other's words as "weird and offputting" is probably not great for site health, either. Disagreement is fine and healthy, but that kind of othering and minimizing language leaves no room for good faith engagement as peers. None of us are going to communicate perfectly, but I think all of us need to step up our game in how we talk about these kinds of issues if MetaFilter is going to survive as a more democratically-governed community.
posted by biogeo at 9:33 PM on August 17 [20 favorites]


"Head in sand" tactic is accurate for the idea that demographics don't matter, and it was not directed at any particular initiative here or lobbed as accusation about anyone. It was a statement about not having demographic data. If such a statement is treated like an attack, that's a problem. Users need to be able to say, "Hey, we need X to help with diversity and inclusion efforts here" and not get treated as if they're attacking people.

This is not a situation where both sides have equal power. A status quo in which we ignore demographics is an unequal status quo. A status quo in which we say demographics don't matter is an unequal status quo. Voices defending that inequality are speaking on behalf of inequity, and it needs to be ok to say that.
posted by lapis at 10:24 PM on August 17 [3 favorites]


To me what is problematic is not so much the head in sand part but the combination with tactic of overrepresented groups (Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups,).
I am very hesitant to post this comment but i feel quite strongly that labelling this lack a tactic, by whomever really misses the point of how fraught the entire issue of collecting demographic data is.

And to the practical side: let's assume the whole issue of security was solved, how to implement a survey that will do justice to the very very different cultural practices around collecting ethnic (race) related data outside the English speaking world (rethoric question, i don't think it can be done). Asking for ethnicity and race is for example a BIG no go where i live.

Personally, i am very happy that Metafilter as a company does not collect my demographic data beyond what is in the profile.

(English is my second language, if any wording is offensive, this is not intentional)
posted by 15L06 at 11:18 PM on August 17 [9 favorites]


Not collecting statistics on diversity is a head-in-the-sand tactic by overrepresented groups

Add the word "often" or even "usually" to that line, and I'll be able to agree.
posted by Too-Ticky at 12:34 AM on August 18 [2 favorites]


To explain why i think it is impossible, perhaps i should add that the reasons for the different approach to collecting ethnic and or race data are historical and cultural. If a European organisation or entity requested my ethnic data or asked my race i would immediately recoil and refuse strongly.
But - partly through reading Metafilter - i do know that collecting such information is considered normal elsewhere.
posted by 15L06 at 12:34 AM on August 18 [4 favorites]

Add the word "often" or even "usually" to that line, and I'll be able to agree.
I think this is sort of implied? Saying "X is a tactic used by Y" does not mean that the only reason anyone would ever do X is because they are Y.
posted by wesleyac at 3:53 AM on August 18 [1 favorite]


Sort of, yes. But why not be clear about it?

If it means "Not collecting statistics on diversity is always, including in this case, a head-in-the-sand tactic by overrepresented groups", then I'll have to disagree; I don't believe it's being used as such a tactic in this case.
If it means "Not collecting statistics on diversity is often a head-in-the-sand tactic by overrepresented groups", then I can agree.

As a non-native speaker of English, I like my implications explicit.
posted by Too-Ticky at 4:08 AM on August 18 [3 favorites]


Also, if I may ask: what is the function of the question mark in "I think this is sort of implied?"
Is it a question? If not, what is it that that question mark is doing there?
I see that a lot on here, possibly more and more often, and I don't understand it. But I assume it has a meaning or function.
posted by Too-Ticky at 4:10 AM on August 18 [1 favorite]

But why not be clear about it?
I think it would have been better to be explicit about it! I just wanted to point out that I read that statement differently than it seems like other people did, in the hopes that it would help other people see that perspective.
Also, if I may ask: what is the function of the question mark in "I think this is sort of implied?"
Is it a question? If not, what is it that that question mark is doing there?
I see that a lot on here, possibly more and more often, and I don't understand it. But I assume it has a meaning or function.
It was meant to acknowledge that language is a subjective thing, and I don't believe that it's incorrect to interpret that statement differently than I did. That's sort of redundant with "I think", but I find it's better to overdo it than underdo it when it comes to indicating that I'm uncertain, since I've heard that I've done the reverse of that too much in the past.

It's the second usage here, approximately.
posted by wesleyac at 4:21 AM on August 18 [2 favorites]


wesleyac: It's the second usage here, approximately.

Thank you! That link explains a lot.
posted by Too-Ticky at 4:46 AM on August 18


supercres: "(He says, as a member of what's probably the most overrepresented demographic category in history.)"

People who dislike any song by Train but especially "Hey Soul Sister"?
posted by signal at 9:35 AM on August 18 [4 favorites]


LOL, is deconstruction of a vapid throwaway phrase considered a pile on attack?

I noted the head-in-ground comment, thought it was slightly pejorative but not especially mean or accurate. Made a point. Then returned to this thread this morning with spiced tea and whew, mifi in a nutshell.

Considered doing an analysis of the deconstruction mini-thread but naw.


(side question, just what does proportional voting mean at Mefi?)
posted by sammyo at 6:54 AM on August 19


Internet: "We will leech and squeeze every bit of information possible out of you."
Metafilter: "We will not do that."
Mefite: "Therefore Metafilter supports cis white male supremacy."

Never change, Metafilter. Never change.
posted by Pyrogenesis at 12:51 AM on August 20 [5 favorites]


Voluntary surveys exist. There are consensual, transparent ways to get information. I understand there are limitations, but please don't assume that everyone who's saying gathering information is important is talking about data mining.
posted by lapis at 10:50 AM on August 20 [4 favorites]


Would an anonymous demographic survey whose participants are self-selecting be a good enough resource for analysis in this case where we are talking about a website community?

I’m not being coy or playing gotcha. Because a survey monkey survey about demographics doesn’t seem like that much volunteer effort, even if it’s imperfect.

But if I’m being too naive about it, please tell me.
posted by kimberussell at 6:09 AM on August 21 [2 favorites]


I think it would be a start. I know there are various best practices on this kind of thing, though, and I'm not familiar enough with them to know what they are. But research and guidance exists.
posted by lapis at 10:24 AM on August 21 [3 favorites]


« Older On not repurposing abortion rights terminology   |   MetaFilter Usage Statistics Newer »

You are not logged in, either login or create an account to post comments