Global Wikimedia survey: the number of women users drops to an all-time low of 9%

The 2018 global Wikimedia survey came out in October and I nearly missed it.

In other years, Wikipediocracy would have been on it, but they seem to have even fewer women than usual these days, and the only two people who cared – Andreas Kolbe, because he did care, and Greg Kohs, because he wanted to embarrass the WMF over their sloppy statistical methods and extract revenge for some backroom deal over paid editing that fell through – are gone.

So what does the survey say?

Only 9% of the users are women. Compare this to 2010, when, according to Kolbe’s old piece in Wikipediocracy,  “only 12.64% of contributors are female.” The report mumbles around about “we are not sure if this is due to a change in the sampling strategy”.  But responsiveness or even methodology does not seem to be the WMF goal here. On the talk page someone has pinged EGalvez (WMF) with the question

“Among high activity editors, the lowest response rate was 25% from Portuguese Wikipedia while the highest was 25% from German Wikipedia.” Presumably one of those percentages is wrong? Thanks.”

The question was never answered and the apparent error remains on the page.

cat gifSo the answer to participation of women remains: “Some work is needed to learn what effect this sampling change had, if any.” Well, they have the raw data, when are they going to rework it to give a meaningful comparison?  IMO it has already been done, and the results are not something we will ever see made public.  Instead we have ever-shifting goalposts.  Participation of women is now rebranded as “diversity” and “inclusiveness”, which is in turn rebranded as age, educational level, and geographical area.

What else?  Results are posted by “teams”, that is, the WMF group requesting data points on various stuff. So after clicking through the general report and reading it, you then find out you have to click on 11 different boxes to see the results for different internal WMF requestors.  Here is the Support & Safety page: Community Engagement Insights/2018 Report/Support & Safety.

And what else?

  • “Dutch Wikipedia was lower in measurements of collaborative intent, awareness of self and others, feedback and recognition, individual commitment to diversity, and inclusive interactions.” LOL, no surprise there, just take a look at the crude, misogynist language they use over on Sucks.
  • “Compared with editors, program and affiliate organizers perceive that their communities place more value in diversity”. Hm, maybe that has something to do with the “Proposed changes to the Wikimedia Foundation Bylaws“.
  • “Sentiments towards a sense of belonging and having an inclusive culture in the Wikimedia community were two of the lower scores among the diversity and inclusion measures. …the average response for women in the survey was 8% lower than men…” No shit, Sherlock.  If you can find any women.
  • “Although 54% of Wikimedians on the projects agree they are “freely able to express my thoughts without being attacked on Wikipedia”, women reported statistically significant lower scores than men.” That pretty much means that 46% do NOT feel they are freely able to express their thoughts.”  But which thoughts?  Don’t forget these are the same people who think a good use of their time is deliberating whether an F-bomb is sanctionable.
  • “Among developers, there was 20% average increase from 2017 for all measures of collaboration and engagement”.  Hmm, might this be a result of the Technical Code of Conduct.
  • “Across Wikimedia audiences, an average of 22% felt unsafe or uncomfortable in any online or offline space in the last 12 months. In 2017, we asked the same question, but did not set a time limit of 12 months. So while it cannot be stated that harassment has decreased, we can say with some confidence that it has not worsened.”  What? First, where is the survey from last year, and what were the numbers.  Second, what happened to all the people who were harassed before? Did they by any chance just leave?  This is like measuring unemployment statistics by counting the number of people getting unemployment compensation.  You aren’t measuring the number of people who have run out of benefits, or who have stopped looking for work out of discouragement.
  • “Experience of harassment has not declined since 2017 and appears to remain steady“.  So all those people they keep hiring to solve the harassment problem have not made any headway.
  • Finally this one from “Support & Safety” (isn’t it Trust and Safety?):  “We still do not understand the reasons why people feel unsafe.”  Well, yeah they did kinda tell you:

Except the problem is that you asked the people being harassed why they were being harassed. How can anyone be expected to speculate on someone else’s motives? You should have asked the people doing the harassing. “Have you ever done x, y, or z, if so, who and why.”  And instead of this multiple-guess format, how about a space for open-ended answers.

18 thoughts on “Global Wikimedia survey: the number of women users drops to an all-time low of 9%

  1. The 2018 report was mentioned on Wikipediocracy a few times in passing, including once by myself (link below), but I’m afraid it’s not because we don’t care; it’s because those numbers are weighted (apparently to emphasize non-English contributors) and the actual statistical change over the past year or two is practically nil, as they admit themselves:

    Obviously our focus is on the English Wikipedia almost exclusively (though that isn’t to say we ignore statistics and events on non-English Wikimedia sites completely, and we certainly don’t do it deliberately). Because they’re smaller and have fewer Americans, they tend to be easier to manage and administer – so in most cases they do less social/cultural damage and there are fewer reasons to criticize them.

    As for our own diversity issues, since most of our members find us via Wikipedia, our statistics tend to reflect theirs. I’d love to see more women participating, but this is the reality, I’m afraid. Still, the percentage of Wikipediocracy members who identify as female isn’t significantly lower than it was in 2012, but in any event, we don’t really concern ourselves with it all that much. There’s no way to really know what any given member’s gender is without asking for some sort of IRL identification to register (something we’re definitely not going to do).

  2. “Numbers are weighted”, okay I think I see what you mean, the 9% is actually across all the projects. They used to report all the language wikis separately, for example I seem to remember Sweden had a very representative group of women, but now there is only a rough percentage for broad geographical regions. A comment on the talk page:

    “After the Harassment consultation 2015, I’m once again surprised there is no data (well, at least not presented here) which would illustrate differences between individual editions of Wikipedia. The health indicators are well-chosen (collaboration, engagement, inclusiveness, harassment, etc.), but the data is presented at a high level, which is probably sufficient to indicate global trends, but can’t detect hotspots and is therefore not actionable at all. E.g. if Foobar-language Wikipedia receives dramatically lower health scores than the average, wouldn’t that be a concrete, actionable finding? Absent that, one can only conclude things are – averaged over the entire globe – a little better or little worse than before, and that’s pretty much it.”

    Looking at the methodology page, it looks like this guy had all the data on his laptop on a g-doc and just dumped what he wanted in the report. None of the talk page questions are answered, and the page isn’t linked to the previous surveys, although someone did come by and put some kind of (non-helpful) category on it. It looks like just a PR exercise.

    The “team” thing is weird too, some groups you can sense some frustration in their “next steps” statement, that they weren’t able to get useful data points. Others seem like it’s twisting their arm to get them to even submit some questions for the survey, and they didn’t bother to do any follow-up. Obviously the survey thing has been taken away from Support and Safety/Community engagement and given to a different and perhaps more circumlocutory group so now the continuity is lost, for instance why are there now 3 age divisions, didn’t there used to be 4? And what’s with removing and replacing the non-native speakers on enwiki? So every year some detail is changed so that you are comparing apples and oranges and cannot make meaningful conclusions. But I suppose it’s better than the year they put up a powerpoint with a graph asking whether diversity was a priority and the dudes answering the survey just said no.

    And why is Legal in the survey? They used it to ask about COI, but where are the diversity issues – maybe if the survey is only 9% women they should not be using it to determine their workflow and priorities. But what do you expect from a group that holds its policy meeting during the WikiWomen luncheon, obviously there is no overlap between the two groups.

    An interesting thing from the Communications group “A higher proportion of males used Wikimedia projects pages. Female editors reported a higher use of mailing lists, social media, the Wikimedia Foundation blog, and conferences.” So women prefer off-wiki or face-to-face communication, anything on-wiki is obviously hostile territory, but how much financial support is available for conferences? Participation by women ends up limited to those hand-picked by the Foundation, but this is an elite rather than grassroots group, probably on the basis of their history of getting other institutions to part with funds for Wikipedian-in-Residence or grant money.

    The way GorillaWarfare has been treated on Wikipediocracy is shameful, especially when you consider how carefully the pedestals for the other arbitrators have been burnished.

  3. Well, that was quick.

    This account is globally locked. The latest lock log entry is provided below for reference:

    00:23, 9 February 2019 ERoss (WMF) (talk | contribs) changed status for global account “User:EGalvez (WMF)@global”: set locked; unset (none) (No longer works for WMF)

    Some badass staffer is working on a weekend.

    But back to what Somey said back then, “about how the WMF has been trying to pull the ol’ switcheroo on the gender-gap conversation, redefining it as a problem of biographies about women”.

    “Apparently there also hasn’t been any change in the percentage of people who claim they’ve ‘experienced harassment’ on WMF projects, so maybe the ol’ Trust and Safety Manager isn’t quite getting it done either.”


    But hey, we got Valerie now, any day now she might be making her first edit.

  4. Please consider that Graaf Statler is a, to say it politely, somewhat confused person. Luckily, he has been blocked indefinitely, so we don’t really care about him. He seems to think that wikipediasucks is quite an important forum, but I don’t see that.

    1. Yes and I see Graaf is complaining about me over there again. There are a few people at Sucks who do not work well with other groups, but perhaps they have found their niche.

      The Dutch Wikipedia did not do well in the global Wikimedia report. In fact they were lower than all the other projects. But if the problem was just one individual, maybe they can recover,

      Since this blog is about gender, I have also detailed some of their gender failings in a new post. People like Graaf who do not understand gender problems need education and information, but where are the articles in the Dutch Wikipedia?

      You may be interested in what I wrote about Sucks earlier:
      Sucks narrows the gender gap
      Dutch perverts and gropers: #ZegHet

      1. I would say: let the Graaf be the Graaf, he’s absolutely hopeless….. It’s nice to see he found a place at sucks, where he can think he’s quite the man.

        1. I think it is not quite that simple, there is also the question of abuse.

          My Evil Twin will not let Graaf comment here at all. She says he is abusive, and that abuse hurts people–not only the people who are abused, but also the people who witness the abuse.

          When people find out I volunteer for Wikipedia, they want to give me books about abuse. I meant to write about this, but it is a very complicated subject, and I will have to look at the book again.

          The first important thing is to stop the abuse, and to detach from the abuse. Because when the abuse is happening, all you can think about is the abuse, you are in reaction mode, to protect yourself, and you forget your identity. So when the abuse finally stops, first you are thinking about Graaf, then you are thinking about not-Graaf. You have to catch your breath, and get to a point where it is not about Graaf, where you can think about yourselves and your own goals.

          If I have a chance I will try to write something about abuse, because as I said before, it is very complicated, to even understand what abuse is, and then what to do about the abuse.

  5. “Community Engagement” doesn’t mean what most would think it meant. It’s not about the WMF engaging with the editing community, it’s about finding ways to increase the editing community’s level of “engagement” with their product (the online encyclopedia). For example, try running an experimental bot that puts barnstars on their talk pages to see whether that increases their level of “engagement”. I still pine for the day when “engagement” to them actually means to engage with the folks who make their jobs possible.

    Last night I watched the DVD documentatry “Citizen Jane” about Jane Jacobs’ 1960s battles with Robert Moses over his plans and implementation of “urbal renewal” in New York City. See

    I couldn’t help but make the analogy with Wikipedia.Think of Jimmy Wales as Robert Moses. Greenwich Village was the tight-knit professional editors of old-school encyclopedias meeting in person in their offices to hash out the contents of the encyclopedia. A safe place to work with lots of eyeballs overseeing each other, analogous to people looking out their second- and third-story windows on the crowds walking, working and playing on the sidewalks below. Wales/Moses tears down the old encyclopedias and builds highrises where editors work in total isolation without ever making physical contact with each other. This turns out to be a mind-numbing environment for many and they take it out on the buidings/encyclopedia by vandalising it. In contrast to their former lowrise, intimate work environment, now they look out their windows and see few people outside; the lack of eyeballs means that many dark corners aren’t very safe anymore and those adventurous enough to edit in them risk getting mugged. Eventually most of the highrise public housing projects were blown up and torn down, the same fate many hope eventually befalls Wikipedia while they watch it from their crow’s nests.

    Maybe the answer really is that the infrastructure in which Wikipedia editors work is fundamentally broken, and rather than try fixing the people to make them behave nicer in this environment, a new, more editor-friendly infrastructure needs to replace the existing system. But, I don’t think the Foundation has any more idea about what an optimal infrastructure looks like than Robert Moses did with regard to New York City.

  6. They know exactly how to fix the infrastructure but they can’t do it because consensus. In a top-down organization it would have been done years ago, just like YouTube Twitter, etc. and they would already be working on the next problem.

    On WP you have everyone in the inner city throwing rocks at each other, vandals and whatnot, including arbcom, and you have an increasing number of users moving out to the suburbs, away from the people throwing rocks – the programs and affiliates, who if you noticed, were included in the survey this year, and will be included in the strategy process, such as it is. It is the Al Capone solution: the local wikis are corrupt so they will turn to federalism — you may recall that they got Al Capone for federal tax evasion, not for corruption. The only problem is the suburbs are expensive and not everyone can afford to move out and/or pay the dues, which includes paying for your own entry fees and accommodation at in-person events and conferences, which women and academics must have for safety, or obtaining grants and racking up a wiki-resume that looks good enough to get free entry and plane tickets. There is also an unspoken limit to how many people can participate on this level and a corresponding competition for resources. But it seems the more lackluster the reward, the more want to compete for it.

    The answer of course is that the suburbs are soulless and the inner city must once again be reclaimed, and made safe, (and to complete the urban planning analogy), probably with unscrupulous real estate developers and gated neighborhoods (more affiliates).

  7. Yes, one of the major operations built in the suburbs is Wiki Ed. I keep seeing their work because, well, the instructors seem incapable of teaching their students even the most basic of the editing guidelines so when the students move their class assignments to mainspace inevitably several of them trigger maintenance flags that I routinely patrol. I’ll give an example here.

    As stated on the user page of the article creator
    this article comes from the University of Washington’s COM 482 Interpersonal Media – Online Communities: “This course combines an in-depth look into several decades of research into online communities and computer-mediated communication with exercises that aim to give students experience applying this research to the evaluation of, and hands-on participation in, online communities.”

    Curiously (to me anyway), Meganjoss has chosen to write about an Australia-based online women’s clothing company, “Hello Molly” See the edit history to see my corrections to basic rookie mistakes and ommissions.

    This article has been twice-before deleted, first in September 2015 because “Article about a real person, which does not credibly indicate the importance or significance of the subject. Probably related to Wikipedia:Long-term abuse/Orangemoody” and again in November 2018 because “Unambiguous advertising or promotion: see WP:YFA, WP:Paid WP:RS, notability guidelines.” I’ve restored all the deleted history so anyone can review it.

    The cynic in me is wondering whether Orangemoody has discovered that if they submit their “paid” editing products via the WMF-approved Wiki Ed program, they will have a better chance of sticking in the encyclopedia. And they even get to exploit the labor of university students to improve their bottom lines.

    Say it isn’t so.

  8. Not so much “suburbs” as capitol city of neighboring and so-far non belligerent country that could probably eat a couple of our larger affiliates for breakfast.

    Orangemoody I had to look up. Some people who were unsuccessfully trying to have articles created about them were approached by someone who offered to write them for pay. The going price was about $400 per article, and the articles were not all that good, but better than nothing for someone who doesn’t know how to edit themselves. Once the deal had been struck and the article was online, some refused to pay, and some of those articles were subsequently nominated for deletion, and since the notability was mostly shaky to begin with, the result of the added scrutiny was to make the articles undefendable. There is a nice IP map of orangemoody/moodies done here by Our James, which unfortunately has no scale and no labeling, so we are not even sure if we are looking at a city or a whole country.

    The articles were about British firms and presumably the IPs were all in England and the payment was in British pounds. But has this Orangemoody decided to move to Seattle, pay who knows how much tuition to enroll in some communications course so they can write some article stub, then get paid in the U.S. with Australian currency? I would say unlikely.

    The professor seems to be following some canned program provided by the Wiki Ed, so if they are triggering maintenance flags (I assume something like cats and internal links) and you say they do it routinely, then the person to talk to would be a Wiki Ed person and not the hapless educator or their superiors who may already be suspicious of Wikipedia, and rightly so. I would also think an article like this could be left alone until after the end of the course, or the student won’t learn what they were supposed to learn, and their grade could be negatively affected by circumstances outside their control. But it the students are triggering these flags, it might be better to have the page patrolled by someone more experienced before the bitey junior wannabe patrollers can arrive. Also it is very hard to keep up with these constant changes, ACTRIAL autoconfirm, etc, so the training program might be out of date. I would also say they are being taught to create these things in sandbox so the instructor can check it before moving out, so I think there is no way to know that this article had been created previously (maybe that is one of the flags?). I don’t find the topic odd, it was one of the things that Sue Garner predicted about attracting women to the project that women would cover more traditionally female topics like fashion and hair styles.This is also the time of year when there are deep deep discounts on clothing and other merchandise, that started with Black Friday, and just kept getting cheaper, so I right now I am doing more research on clothing sales than on Wikipedia and gender (don’t ask me about the black Steve Madden combat boots).

    There are also reasons for an article being recreated other than an Orangemoody hiding under the bed. An article might have been notable the first time and deleted in error, or it might have become more well known than it was before. Or it still might not be notable but people are just interested in the brand and don’t understand notability. I think the German Wikipedia is less deletionist in this regard, and they even welcome participation by people employed by the firm, the COI freakout is more of an English Wikipedia thing. That said, I would not think it worth my while to create any article previously deleted on any language wiki, even if notable it would not be worth the drama.

      1. Yes, I assumed she was just a fan who had bought some clothes from them
        But other editors might not have made the same assumption, despite WP: “assume good faith”

        Their canned program could still use some tweaks to make the transition from user/draft space to live articles more seamless. I’d rather see the paid professors or Wiki Ed staff ensure that they don’t use level 1 headings or neglect to include a references section and categorize the article, and use sentence case rather than title case for section headers… rather than take advantage of the goodwill of unpaid me to clean up after them.

  9. “After searching through articles that needed to be created on Wikipedia, I began to look at clothing sites…”

    Hmm, maybe the instructors suggested searching Wikipedia:Requested articles for ideas. A on-time-editing IP dropped the suggestion tro create “Hello Molly” at Wikipedia:Requested articles/Arts and entertainment/Fashion on 14 September 2015.

    That could be “Orangemoody” or one of their competitors. Clever of them. It only took 3+ years for a volunteer to drop by and take up the suggestion. Or maybe it just took that long to get enough celebrity models to wear the brand to make it happen.

  10. So this Orangemoody moves to Serbia, makes an IP request for an article, four years later the article is created in Seattle, so Orangemooody moves back to England, or maybe Sydney, and picks up the $400. How diabolically clever. What a business plan.

    Or alternatively whoever said “this article must have been written by Orangemoody because Orangemoody re-created deleted articles and this article has been deleted”, maybe they are an idiot and you shouldn’t listen to them. The talk page for the list of articles created by Orangemoody is full of such speculation.

    Here is the instruction page, choosing an article should be about week 3, but articles about fashion companies is not one of the options. I don’t think a reasonable instructor would try to cramp the style of a female student by telling them outright not to do fashion if that’s what they chose, most instructors would have that much sense. I don’t think its notable, though, at least by English Wikipedia standards. Or is it? Can a famous person convey notability to a fashion house by wearing their dress? Because that’s all those sources are, all based on promotion. But if you look at something like the Givenchy article, that’s exactly how that article is structured. So what is notability for fashion? Is this a women’s issue? (don’t ask me, I’m in the middle of a boot obsession, currently studying AR 670-1 military standards for women boots [$130-$175] and why they are all the color of sand these days). Or maybe Harley Davidson [$145-170]…. or Timberland [@ $170.00] …. as long as I don’t buy any….

    Incidentally, there’s a completely ridiculous controversy Hello Molly has supposedly gotten involved with the upstairs tenant, fortunately it’s the Daily Mail, so it won’t make it into Wikipedia.

  11. Love playing devil’s advocate and pitching you softballs that you hit out of the park. Other than the idea that the IP was in Serbia (I think IP is about as reliable as caller ID is for telling you the location of the caller) I pretty much agree with you. (ooh, on the one hand she’s very cultured and knowlegable about religion and so many other cultural topics, on the other she’s into fast cars, military boots, and some techie stuff.. why I’m a big fan)

  12. Ah, so I did detect a hint of reductio ad absurdum in that last remark.

    The boots are All Salvidrim’s Fault, since he mentioned Quebec politician Catherine Dorion’s green Doc Martens I have not been able to get them out of my mind. Although for a while I was furiously googlng “buttery leggings”, which I think is what you are supposed to wear with the boots. It turns out they are a premium brand called LuLaRoe, and everyone is falling all over each other to produce knockoffs of tights, tunics, trapeze dresses, or whatever the latest terminology. After looking at a number of satisfied customers – who are, shall we say, not at their ideal weight – posting pic of themselves wearing this apparel I have retreated to brushed denim and flannel, at least for the duration of the cold snap.

    I too appreciate my readers and the odd bits of information they turn up about devving and adminning and so forth. This sockpuppet investigation stuff is starting to sound very unreliable… although I know you can’t use a VPN or proxy to log in to some types of websites (Wikipediocraxy, cough cough), I’m not sure whether WP is one of them.

Leave a Reply to genderdesk Cancel reply

Fill in your details below or click an icon to log in: Logo

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

Google photo

You are commenting using your Google 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