Hypothesis: women need to do more than men to “deserve” a Wikipedia page

September 4, 2024 12:49 AM

I think Wikipedia is great. I spend tons of time on it. I especially like to read about history, because it allows me to quickly jump into obscure details about anything, without the need to scout for specialized literature that might be super hard to find. But one question always creeps in the back of my mind: am I reading something as fair as it can be? How much are the editors’ biases driving my discovery process? These are testable questions! And the subject of this blog post, and a paper I recently published. from The Glass Door of Wikipedia’s Notable People
posted by chavenet (35 comments total)

It might be interesting to apply these techniques to other groups of less-privileged people, except we sort of know what the result would be.
posted by Phanx at 1:36 AM on September 4 [6 favorites]

I wonder if they can do this analysis comparing women to fictional characters from comic books and see which has an easier time getting into Wikipedia.
posted by jacquilynne at 3:33 AM on September 4 [25 favorites]

also see Jess Wade's work [guardian, wiki]
posted by HearHere at 3:50 AM on September 4 [4 favorites]

Wikipedia is great, but this is absolutely not the first time I've heard this criticism.
posted by Smedly, Butlerian jihadi at 4:00 AM on September 4 [7 favorites]

Not surprising as little as 3 years ago the site was still over representing Nazi myths. [previous metafilter]

The whole site is dependent on the editors going a good job against biases and often they as humans don’t because humans are humans.
posted by jmauro at 4:37 AM on September 4 [6 favorites]

Definitely a long-running issue. I keep hoping maybe it's getting better but, if so it's taking awhile.
posted by rmd1023 at 5:13 AM on September 4 [3 favorites]

As a longtime Wikipedian, I believe this demonstrates _another_ problem in Wikipedia's gender (and other kinds of) gap.

It's been known for years that the editor population is still mostly male, and the proportion of biographies is also heavily skewed male (though not as bad as it was 10 years ago). What this paper shows is a subtly different mechanism connected to the others: how hard it is to _create_ a new article for a notable woman.

My own anecdotal experience is that—in the Spanish Wikipedia—all a man needs to do to have their article is to play once in a soccer minor league, whereas women with decades of experience in their research fields and numerous papers and conferences are struck down as being «promotional» or a mere «resume-like» article
posted by andycyca at 6:28 AM on September 4 [24 favorites]

Unless you're the Hawk Tuah Girl, who got a Wikipedia page for answering a sex question with a surprising twelve-word sentence, only 10 weeks later. But, I guess "doing more" is a subjective term, particularly when it comes to what is important to Wikipedia editors.

If you haven't heard of her, the wikipedia page has about 50x the amount of info you actually need to know
posted by AzraelBrown at 6:31 AM on September 4 [10 favorites]

(after editing windows closed)

This is something many editors have found the hard way, and something that we knew happened, but didn't have a way to prove it even measure. So in a way this is novel merely because it proves a known experience.

(also, many new female editors face this problem during events where the local chapter tries actively to reduce the gender gap. Imagine the frustration of someone learning how to contribute to Wikipedia, excited to make their first article about a notable woman, only to have the door slammed in their faces because women are judged not as notable. The problem described in the paper is one of the causes of the other, related, gap issues of Wikipedia)
posted by andycyca at 6:34 AM on September 4 [12 favorites]

Re: "Hawk Tuah" girl - I know that viral fame can sometimes be baffling, often short-lived, etc...but seriously - THIS?

"On August 15, 2024, Welch threw out the ceremonial first pitch at a New York Mets game in Citi Field."

I know very little about pro sports, but if her garbage moment of fame is all it takes to get a promo "gig" like that...just...well, I didn't know the bar was *that* low.
posted by davidmsc at 6:42 AM on September 4 [5 favorites]

Re: "Hawk Tuah" girl - I know that viral fame can sometimes be baffling

On 3 September 2024, Welch launched her first podcast entitled Talk Tuah with Haliey Welch.

+1 for podcast title
posted by chavenet at 6:51 AM on September 4 [15 favorites]

I always get a little nervous when research attempts to quantify abstract concepts like "career success" and "achievement." With this work, I wonder how accurately the proxy variables around network connection really measure those fuzzy concepts.
posted by yellowcandy at 7:48 AM on September 4 [2 favorites]

Anecdotally it seems quite clear the bar is higher for women. I've admired how Metafilter's own Jessamyn has worked to get good quality articles on Wikipedia despite the headwinds. A key thing seems to be really solid references for notability so biased editors can't just dismiss the article. Which is a lot of work but also makes for a good article in itself in addition to one that might survive culling.

This analysis linked here is quite a subtle form of network analysis. I like it but there's a bunch of choices the authors have made in the natural language processing to estimate link weight. And the final result about significant links is a little indirect. Also curious if measuring Wikipedia-internal links is the only option. I wonder if another more direct approach might also shed light on the hypothesis that women are less likely to be counted as notable solely based on gender bias of Wikipedia editors? Just counting number of external links on articles for men and women maybe. Or comparing similar basic stats for articles that were deleted?
posted by Nelson at 8:15 AM on September 4 [4 favorites]

Sorry for slight derail, but wow - the PR page for the Talk Tuah podcast states that she is "the face of the biggest meme in the history of the internet."

Hyperbole, exaggeration, puffery, etc - all part of the PR game, I get it - but COME ON, people!
posted by davidmsc at 8:42 AM on September 4 [7 favorites]

This reminds me of how people say that mediocre men apply for a job if they have oh, 50-60% of the requirements and women won't apply unless they have 100%. I keep pointing out when people say this to me that women HAVE to be literally better than men to count against them. This Wikipedia thing? Right-o, same issue comes up again.

(Unless you make THAT joke, anyway, apparently.)
posted by jenfullmoon at 9:10 AM on September 4 [6 favorites]

It isn’t just Wikipedia. I give my students a quote at the start of every class and Brainy Quote (and pretty much every quote site I come across) is probably 90% quotes from men.
posted by Big Al 8000 at 9:41 AM on September 4 [4 favorites]

if you are looking for more quotes from "notable" women, it seems "spit on that thing" counts lol
posted by Jacqueline at 9:51 AM on September 4 [4 favorites]

Hyperbole, exaggeration, puffery, etc - all part of the PR game, I get it - but COME ON, people!

Did you use these particular words on purpose?
posted by srboisvert at 10:16 AM on September 4 [5 favorites]

Wikipedia is one of the best things. It's what the internet should be all about.

It's amazing to me that 300,000 years, 50% of the population still manages to ignore most of what the other 50% has to offer intellectually but finds a young female's comment on oral sex noteworthy. Maybe as a species if we could evolve into nebulous minds without bodies things could change.
posted by BlueHorse at 10:25 AM on September 4 [8 favorites]

It's only important because she was talking about peen.
posted by jenfullmoon at 10:59 AM on September 4 [4 favorites]

> Backboning is done to sparsify a network...

True Hacker jargon detected.

I like this approach, and feel that (with TFA's caution about whatever underlying biases are in the dataset) it usefully demonstrates and to some extent quantifies the "notability gender-gap."
posted by Aardvark Cheeselog at 11:23 AM on September 4 [1 favorite]

I always get a little nervous when research attempts to quantify abstract concepts like "career success" and "achievement."

They're trying to create reasonable proxy measures that are quantified, not to quantify 'bias'.
posted by MisantropicPainforest at 11:27 AM on September 4 [1 favorite]

@misantropicpainforest Yes. Exactly. That's what I think is a bit undercooked. I don't recall mentioning bias at all.
posted by yellowcandy at 11:33 AM on September 4 [2 favorites]

In studying history there is always a tension between "Women/minorities were widely and routinely denied education and opportunity" and "Fewer women/minorities rose to notable status, particularly in ways that required education and opportunity."

For me the smaller number of notable female writers in an era when women were seldom taught to read or write is an expected result. That doesn't change the fact that many notable women were swept under the rug and, happily, are now being rediscovered. But I think looking for exact parity in numbers is a mistake.
posted by Tell Me No Lies at 11:51 AM on September 4 [1 favorite]

But I think looking for exact parity in numbers is a mistake.

I don't really understand the network analysis here, so someone can correct me if I'm wrong, but my impression is that's not what they're doing? As I understand it, they are looking to model how notable a woman has to be before she is included compared with how notable a man has to be before he is included. It doesn't say anything about how many notable men or women actually existed in reality just what portion of them at each level of notability gets included.

If notability existed on a 100 point scale based on what percentage of people in the world had heard of someone and women needed to reach, on average, 75% of the world while men only needed to reach on average 50% of the world, that would be similar to what is being measured here. It doesn't say anything about the number of women or men who reach 50% or 75% world-wide awareness, only on how likely they are to get a wikipedia entry once they reach that level of notability.

There's a whole other bias problem in how many men vs. women reach any given level of notoriety in the first place, but I don't think that's what they're measuring?
posted by jacquilynne at 12:26 PM on September 4 [13 favorites]

For me the smaller number of notable female writers in an era when women were seldom taught to read or write is an expected result.

Yes, the article we're discussing spends a lot of time at the start acknowledging that inherent bias. It's the motivating factor for the novel network analysis that lets them come to a conclusion about Wikipedia bias independent of the historical reality. No one in the article is looking for exact parity in numbers.
posted by Nelson at 12:35 PM on September 4 [12 favorites]

superseded by a more succinct explanation by @Nelson, and yes, @jacquilynne, that is what the article is saying.
posted by drossdragon at 12:37 PM on September 4 [1 favorite]

I can't believe this discussion has derailed into complaining that there's a woman who doesn't deserve a Wikipedia page. Oh no! An undeserving woman on the internet!

What about the article that shows considerable evidence that there are many more men than women with Wikipedia pages they don't deserve?
posted by medusa at 7:18 PM on September 4 [7 favorites]

An interesting approach. It took me a few reads to understand what they are doing. They are basically counting how many mentions a person requires in an article before they get their own article, and comparing the numbers for men and women.

That's the data part. The interpretation seems a little shaky, and I'm not sure the "test" proves anything. I wonder if there are any other explanations. You could flip this around entirely - perhaps this is because those women who do have Wikipedia entries are more likely to be referenced, for whatever reason. Or perhaps un-notable women are more likely to be mentioned than un-notable men? That would have the same effect, as I understand it.
posted by Aethelwulf at 12:39 AM on September 5 [3 favorites]

Or perhaps un-notable women are more likely to be mentioned than un-notable men?

If a person of any gender is referenced in a bunch of Wikipedia articles, that person is ipso facto notable. We can certainly quibble about how many qualifies as "a bunch", but if someone keeps coming up over and over again in the literature, that is notable.
posted by adrienneleigh at 1:12 AM on September 5 [1 favorite]

That's the data part. The interpretation seems a little shaky, and I'm not sure the "test" proves anything.

What's with the scare quotes? Also who in the world is talking about proof, at all, in any sense? Studying complex social phenomena is like detective work, you look for clues and dance around the edges, not write 2+2=4 and then QED.

Or perhaps un-notable women are more likely to be mentioned than un-notable men? That would have the same effect, as I understand it.

Nope, because 'the effect' they're talking about includes the decreasing gap that occurs in March.


You could flip this around entirely - perhaps this is because those women who do have Wikipedia entries are more likely to be referenced, for whatever reason.

They're doing something more nuanced than that. From the paper:

The null model expects MF and FM edges to be equally likely.
However, that is not what we observe in the network, where F nodes are more likely
to connect to M nodes than vice versa.
posted by MisantropicPainforest at 6:40 AM on September 5 [4 favorites]

there's a woman who doesn't deserve a Wikipedia page

It wasn't about the woman herself -- it was about what Wikipedia thinks is important enough for a page. It's an example of Wikipedia crowdsourcing going "HERE'S A WOMAN, LET'S GET HER DOCUMENTED", and choosing this woman over, say, any number of scientists or politicians. She's as relevant as a lot of celebrities who have Wikipedia pages, the commentary is on the rush to get her on there versus the point of the article of how difficult it is to get women recorded in general due to biases. It's about the bias, not the woman.
posted by AzraelBrown at 7:20 AM on September 5 [4 favorites]

For me the smaller number of notable female writers in an era when women were seldom taught to read or write is an expected result. That doesn't change the fact that many notable women were swept under the rug and, happily, are now being rediscovered. But I think looking for exact parity in numbers is a mistake.

FTFA:

"The hypothesis is that women need to do more than men to “deserve” a Wikipedia page. The are a few problems with this hypothesis. For starters, we can’t really prove it by simply saying that there are way more men than women on Wikipedia. That can happen and still be fair, because Wikipedia is just working with whatever it can collect from the notoriously male-centric historiography."

Clicking on the links and reading does wonders!
posted by MisantropicPainforest at 8:04 AM on September 5 [5 favorites]

One reason that a number of minorities are under-represented on Wikipedia is the long reach of its "Articles for Deletion" system (which, yes, shares an acronym with a German far-right party, and the similarities don't completely halt there).

Mrs. Hobo got involved with Wikipedia partly as part of the Women In Red project ("red" referring to the colour of a link to a page that does not exist). And she got very good at navigating the political and bureaucratic forces behind the deletions system, simply of necessity.

I will say that the phrasing AzraelBrown used, of "choosing [a] woman over...any number of scientists or politicians", highlights the outdated scarcity-based thinking that the so-called "Deletionists" tend to take. They think of Wikipedia coverage as some kind of limited resource that must be balanced based on what they feel makes a subject "deserving". They will dress up their WP:JDLI opinions in nit-picking rules-lawyering, and exhaust people until an Editor comes by and has to make sense of the discourse before a decision lands.

The existence of a biographical page for this hypothetical person you don't find interesting has about as much effect on the page for the "scientist or politician" as our gay neighbours' marriage has on our own: nil.

Wikipedia would be immensely improved if even one third of the people who say "This page sucks: delete it!" instead went and poked through the many online archives and libraries made available to Wikipedians, and added sources or reached out to domain experts on the site to help improve the page. But marking pages for deletion starts a horrible thread of posturing and arguing (and at that point reaching out for help gets condemned as "canvassing" even though an AfD thread is not a vote), and it's one of the hotbeds of inappropriate behaviour.

Of course, pages do often need to be removed, for a variety of reasons. Topics may need to be split out or merged together, or the page may need to be rewritten. And the very best voices on that site are neither completely "Completionist" nor "Deletionist" in their approach: but their first step for responding to a bad page is to try improving it.
posted by rum-soaked space hobo at 11:00 AM on September 5 [5 favorites]

their first step for responding to a bad page is to try improving it.

Yes, for sure, but also this is also one of the most effective ways I've found to completely burn oneself out on Wikipedia, such that I've had to take a hard line of just refusing to improve any article that is up for AFD.

Consider: improving Wikipedia articles is already such a marginally rewarding / unrewarding task that the vast majority of people who could, don't. For the most part, only those of us who are wired a little strange to begin with even bother trying.* And even those mostly don't stick around for long.

Suppose an article has a 50% chance of survival if improved -- that cuts that already-tiny marginal reward of editing in half, and not many folks are going to keep going for long. Add in the fact that negative feedback has a much stronger impact than positive feedback, and the rewards of this kind of work are actually negative. Add in the fact that the actual odds of an article surviving AFD are much lower than 50%, and you have a situation that would make most humans profoundly miserable within a week.

This would be bad enough if the harms of AFD were limited to nominated articles, but -- circling back to TFA -- its harms extend much more broadly. In particular, I suspect the highly subjective and unpredictable way the notability guidelines are enforced in practice on AFD plays a major role in the outcome documented here. It would be very interesting to know if a survey of just AFD-nominated articles would show similarly biased outcomes (I suspect it would, in fact if I had to guess I suspect the effect of gender on AFD outcomes might be stronger). But even if that's not the case, even if AFD itself were firmly dedicated to gender parity -- given how deeply demoralizing the AFD process is, even a small increase in the likelihood of an article being nominated for AFD if it is about a woman would likely have a sufficiently discouraging effect over time to account for the results documented here.

And all of this harm is done (by the gleeful bullies who congregate in AFD where their behavior is socially rewarded and protected from consequences by a culture of weaponized civility) in service of a goal that is, as you note, of very limited value to the project. A hundred bad articles are less damaging than the loss of one good one.

* This topic always puts me in mind of the very first Wikipedia-related academic survey I got, way back in the mid-aughts, where the subtext of every question was "what exactly is wrong with you that you engage in this deviant activity of freely sharing information with others?" I didn't return it, and I'm glad that research on Wikipedia has gotten much better over the years.
posted by Not A Thing at 12:08 PM on September 5 [5 favorites]