2019 editor gender survey – the final solution

Yesterday, Midsize “Somey” Jake commented, “Surely by 2109, all of this gender-related conflict will be behind us?”

So that got me thinking.  What can we infer from the statistics.  We had all kinds of surveys from the WMF, so why not put some of them to use.

I was going to do the calculations on the back of a napkin, like some blogs do with their breakfast calculation posts, but unfortunately I have spilled some coffee and the napkin is now a bit damp.

So I have delegated the task to my Evil Twin, who has outsourced the task to one of the other twins who runs the Desk’s statistics department.   I’m not really sure who did the analysis, since yesterday Andrew Brundle of AB Website Design Bangkok came by acting all mean and dropped a bunch of c-bombs in the comments, and the other twin is already really shy, but now she is so utterly terrified she will not even reveal her pseudonym.

But here is the calculation.

The answer is that women will disappear from Wikipedia completely by 2028, unless they are quickly replaced by men impersonating women.

I like it when people show their work.  Then you can use their research to see if the results can be replicated. In this case the research was based on published WMF data, see Global Wikimedia survey: the number of women users drops to an all-time low of 9%. The drawing was done in 3D Paint with a 5 pixel marker, which I believe is the default, and which I believe produces an annoying popup on the desktop, so this may be the last time we see this methodology.

In case anyone wants to snatch it for nefarious purposes, I have cleaned it up a little, and added the copyright info. 😉

Women in Wikipedia 2019 1

Always happy to help the folk over at Wikipediocracy whenever they have a burning question about gender.

Advertisements

2 thoughts on “2019 editor gender survey – the final solution

  1. Okay, I’ll gladly yield to your superior long-form comeback capability…

    However, you’ve reminded me that there’s another point we could be missing, which is that since the WMF keeps changing their counting methodology, the whole issue could become a form of long-term gaslighting. Even now, it’s hard to get journalists and researchers to understand that on a site that encourages people to be anonymous and expose as few personal details as possible, you can’t get these statistics without begging people to volunteer them – and that when they do volunteer them you can’t trust them, and that since the respondents are self-selecting you can’t get meaningful numbers even if you decide to trust them anyway. Now try getting them to understand that by changing their counting methodology each time, the WMF practically guarantees different results, at which point they can use the differences to argue that the issue shouldn’t even be brought up at all.

    And FWIW, it did occur to me that the WMF folks could just falsify the results any way they like. It would be risky because some disgruntled ex- (or even current) employee might spill the beans on them, but I guess I wouldn’t totally put it past them.

  2. Falsify? no that’s not how statistics works. Here is how it works. The nine percent was for the whole projects, all the Wikipedias together. The 13% was for en.wiki, but it was originally more like 11% I forget, or maybe more like 8%, done by an outside agency, but the numbers were fudged with a particular number from a paper by a WP insider because it was claimed that research of percentage of non-responses from a phone survey could be used to upwards inflate the number of women for an online survey. Like right, if someone has your name and phone number and can call you, you will behave exactly the same way to them as if you were anonymous online. Of course they never went back to the original outside agency to ask if it was a legitimate massaging of their results. Everything is transparent yes, but who has time to click all the sources and read all the footnotes, not to mention understand and evaluate what they are reading. All these surveys are measuring something different, and therefore cannot be compared directly. They never do the same survey twice. I have heard it said that if you do not know the results in advance, you had best not do the research.

    I don’t have any insider information on this, it’s just possible that this time they want genuinely reliable data.

    I have been saying for a long time they are about to make major changes, and the changes will be painful, but they must do so for their own corporate survival. So my gut feeling is either they want to have something to justify their changes, or they want to be able to monitor short term changes in data for their own purposes, perhaps to help them adjust mid-course.

    And your own “long-form comeback capability” is much superior to mine, with a potential for a much larger reach, but you rarely use it.

    My graph of course is a sort of joke, interpolating a data trend from only two data points.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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