Grants talk:IdeaLab/Examination of gender in biographies

On the invalidity of biography gender as a proxy of editor gender

consideration

Neat Idea, Bluerasberry. Have you thought about the problem that women may not identify as such ? --Wuerzele (talk) 22:28, 4 March 2015 (UTC)

I think this is why the project focuses on the gender of the article subject rather than the gender of the editor. And I too think this is a great idea :). Connor Behan (talk) 00:18, 5 March 2015 (UTC)

Connor Behan thank you for the oh so gentle hint. i didnt read attentively...--Wuerzele (talk) 05:00, 5 March 2015 (UTC)

I think that is a good idea, to start by measuring what can be measured. Sminthopsis84 (talk) 04:13, 5 March 2015 (UTC)

true, agree. the question is, whether merely the gender of the person in the biography allows one to infer what is termed at the end of the proposal "gender-based tension" in editing behavior. I can't confirm or deny the existence of such a thing from my experience in general, and as I reflect on my edits in biographies (mostly male personas) and writing biographies (so far all about males) in particular.--Wuerzele (talk) 05:00, 5 March 2015 (UTC)

Confused

I'm confused about how the stated solution addresses the stated problem. "3. As an approximation of the treatment of contributors of different genders, examine the differences in treatment between those contribution to biographies of people of different genders." This seems to assume that women mainly edit articles about women and men mainly edit articles about men. How about comparing the rate of edits being reverted among contributors who identify as a particular gender on their user page, say by looking at en:Category:Wikipedians by gender ? --Geniac (talk) 22:46, 4 March 2015 (UTC)

Geniac, no, i dont think this assumes men edit male bios, and women female bios. it only says differences in reviewing behavior towards biographies of different genders.

rate of reversals among self-identified gender would be adding another layer.--Wuerzele (talk) 07:25, 5 March 2015 (UTC)

It is stated that The null hypothesis in this research is that the gender of a contributor is not a factor in how their contributions to Wikimedia projects are received. But you want to test this hypothesis by looking at the gender of an article's subject. Unless you actually assume that women mainly edit articles about women and men mainly edit articles about men, you need to change either your null hypothesis or your method of testing it. Further down you make a different null hypothesis, which seems to fit your method better … --El Grafo (talk) 14:17, 5 March 2015 (UTC)

How we could change the proposal

I was talking about this as a proxy because that's how I came to the idea and because it was slightly fewer words. I suggest that someone (so probably Lane or I) comes up with a short way of phrasing that we are interested in the role of gender of article, not gender of editor, and that we are using this because it is harder to determine editor gender. Tlevine (talk) 08:24, 8 March 2015 (UTC)

It's not at all hard to determine the gender of editors. There are 12,911 males and 2,653 females and a few hundred various others listed under en:Category:Wikipedians by gender. No, it's not all editors, but still, that is a huge sample size to work with. --Geniac (talk) 10:29, 10 March 2015 (UTC)

Geniac There are more than a million biographies in English language and I expect a million more in the other languages combined. Looking at editors by gender is an interesting project also, but I think that checking for a statistically significant bias in biographies by gender would be more robust and insightful considering the 1000 fold increase in data which is available by looking at those articles. If there is no difference in treatment of contributors who edit articles about males versus females, then even the null result of "no difference" would be interesting. If this project did identify a difference, then that should encourage looking at the "gender" of articles more closely and add a layer of complexity to all subsequent analyses.

I think you are operating under the premise that if an editor seems to be of a certain gender, then others will treat them in a certain way. For this project, I was wondering what if any effects might exist in a space which there was a gender environment where all content was labeled "about a male" versus "about a non-male". I am not so ready to say that the gender gap on Wikipedia is only about the gender of contributors - there might be the existence of a bias as soon as the concept of gender is considered in any context. Blue Rasberry (talk) 14:26, 10 March 2015 (UTC)

I am operating under no such premise and I have no idea how you came to that conclusion. I have quite the opposite premise, in fact; that the gender of an editor has no effect how how other editors treat them. I am not claiming that this research would end up being pointless, just that as created, the project's goal and solution were mismatched. I see that it has since been modified, but the last sentence of the problem section still states, "This project seeks to examine the circumstances under which reviewers on Wikipedia challenge contributions of female contributors to Wikipedia", which would be an unattainable goal with the stated solution. --Geniac (talk) 00:03, 12 March 2015 (UTC)

Reflection in content

It is correct that biography gender is not a reflection of editor gender. There is no reason why male editors would not edit articles about females and vice versa (for example, I can hardly imagine an editor interested in politics who will edit en:Winston Churchill but not en:Margaret Thatcher or vice versa). What this method can measure is whether the gender gap is reflected in content, i.e. check the Foundation's message Fewer than 20% of Wikimedia contributors are women, and this disparity is reflected in our content. Thus this research will not allow to measure the gender gap, but it is still useful for measuring impact of the gender gap on our content — NickK (talk) 17:07, 9 March 2015 (UTC)

Distinguishing vandalism from real contributions

I expect that the researchers will look in detail at particular pages while they are counting the edits in various categories, so I don't mean to suggest that this needs pointing out. Undoubtedly, there are many different situations and they will be considered during the writing of the report. However, I'd like to point to this edit because it is a particularly clear case of a type of vandalism that I *think* might be more common on biography pages about (1) women (2) people of some other genders (3) anyone from an ethnic group that has few contributors in the English wikipedia (4) ... This was vandalism not reverted by ClueBot but rather by a human editor, and might perhaps appear in a count as a case of that editor repressing a "contribution" to the article. I'm really just suggesting one piece of test data to consider. Thank you for proposing this project. Sminthopsis84 (talk) 04:56, 5 March 2015 (UTC)

I'm not clear on what you are suggesting, but your comment is quite interesting regardless. I had originally thought about classifying vandalism in order to focus simply on non-vandalism edit-reversion interactions, but your ideas about differences in vandalism by article type are much more interesting! Tlevine (talk)

On factors other than gender that have something to do with reversion

confounding factors for reviewing and reverting other than gender based tension

I can think of numerous confounding factors that influence reviewing and reverting behavior, that would need to be considered, i.e. measured, to allow proper analysis and inference. Familiar to epidemiologists, this is in a nutshell the need to separate out factors that may influence the observed behavior, but have nothing to do with "gender based tension". What comes to my mind immediately is:

• length of time on wikipedia,
• professional / work category,
• degree of celebrity,
• nationality and even locality within a country,
• time of existence on WP.

A page will be reverted and reviewed less, if the bio is on a person

• in a less "competitive" profession (soil scientist vs actor, sport champion, CEO),
• less well known or only known only to a fringe audience (bicycle tool maker vs Madonna or an inventor of social networking site),
• from an obscure or "relatively uninteresting place" on earth to the masses ("developing" country, fly-over-country in the US :-) think Iowa, hinterland vs U.S. or a metropolis)
• with either a brandnew (months old, infant-) page i.e. too new to be noted, or if it has existed for many years and is stable. That said, I think the # of reverts and talk page posts need to be mathematically normalized over time, maybe also size (wordcount)- the less there is there, the less can be reviewed and quarreled about.

After writing these down I can see the trend and wonder if anyone would even call the above factors gender-based? I'd call them power-based.--Wuerzele (talk) 05:30, 5 March 2015 (UTC)

Great idea! We can test how gendered all of these other factors are. Tlevine (talk) 08:27, 8 March 2015 (UTC)

Occupation bias

I think that this research may have a potential bias based on the occupation. Articles about people with different occupations are likely to have different editing history: for example, articles about politicans (predominantly males) are more likely to be targets of POV pushing, while articles about models (predominantly female) are more likely to be targets of childish vandalism. Thus the sampling should be filtered by occupation, as comparing editing histories of en:Angela Merkel and en:Justin Bieber will not make sense — NickK (talk) 02:04, 6 March 2015 (UTC)

Another wonderful idea about related factors Tlevine (talk) 08:37, 8 March 2015 (UTC)

NickK,Tlevine ( when I wrote professional / work category I meant just that of the person 's biography.--Wuerzele (talk) 19:28, 11 March 2015 (UTC)

• This is the real problem for me. I'd be amazed if any results didn't show that the figures for eg pop stars, saints, novelists, biologists, and politicians weren't far more closely related to "occupation" than to gender. Johnbod (talk) 16:34, 20 March 2015 (UTC)

Modeling with these potentially collinear factors

Please keep the suggestions of related factors coming! People seem a bit overly worried that these will mess up the results somehow, so here I try to calm the worries.

If we completely ignore the existance of such other factors, we'll still find out how things compare by gender. Including other factors in the model would make things more interesting by

1. giving us more detail and precision
2. allowing us to see other things that are related to reversion incidents
3. assisting us in contemplating fundamental causes for differences that we find

Tlevine (talk) 08:37, 8 March 2015 (UTC)

Gender Problem

I can not hear this gender nonsense any longer. By the way, the problem of Wikipedia is not that too few women want to join in, but that no one at all wants to join. And that is because in this project, particularly in the German Wikipedia, there is too much bullying instead of cooperation. --95.116.88.109 22:41, 9 March 2015 (UTC)

Hi, this doesn't seem to be conducive to a productive environment here. Please adhere to our friendly space expectations and remember that this is not a place to have meta-discussion of the gender gap. Thanks, Keilana|^{Parlez ici} 00:41, 10 March 2015 (UTC)

agree with Keilana and wonder why IP he or she is even on this page when you "can not hear this gender nonsense any longer"!? and how does he or she know what "the problem of Wikipedia is"?

95.116.88.109 I myself dont contribute much (more) on de.WP, but because it's so fricking unreliable with its non-chalant reference rules, which may be predominantly enforced by males... Anyway, the first step is to register your IP account, the next to contribute constructively. Et voila. --Wuerzele (talk) 19:19, 11 March 2015 (UTC)

Merge with WIGI?

Hello, I see our projects, yours and the one I have proposed Wikipedia Gender Inequality Index, (based on this academic paper: [1]), as remarkably similar. I would be very keen to merge our proposals and collaborate. Essentially what I think mergeing would entail is to broaden the analysis of what WIGI already does, to include your research questions, and then apply the monthly-updating machinery that my proposal puts forth. What do you think? Maximilianklein (talk) 19:42, 10 March 2015 (UTC)

Maximilianklein I never had any interest in executing this proposal even though I typed much of it out. Tlevine and others would be managing this, so I would defer to them. Blue Rasberry (talk) 19:52, 10 March 2015 (UTC)

I was quite intrigued to see that we have vaguely similar ideas! That said, the only remarkable similarity I see between your paper and my proposed research is the use of biographies, so I think the way we might collaborate is to reuse the stuff where we acquire the appropriate data from the database dump. On the other hand, it might be worth talking about other toy research ideas. I have some very sloppy notes here. Tlevine (talk) 19:56, 10 March 2015 (UTC)

@Tlevine:, I agree that our methods are different, but I think our ends are similiar and that is important. As you point out, you make a derived measure from the database dumps using variables that are taken from Wikidata data. At that level that is also what WIGI does. The IEG that I am proposing, and perhaps it is not clear enough, is to automate taking this measure, do it every week or month, and then track the effect over time. That way we have longitudinal study to see if other projects are having an impact. Maximilianklein (talk) 00:53, 20 March 2015 (UTC)

@Maximilianklein:, I started working on this: http://dada.pink/wikipedia-biographies-edits/read-xml.py

Study the primary factor

Don't study a proxy for gender, study gender, in the abstract and particular.

As I see it, there are two major questions here, which can (and I think should) be treated separately:

1. How does the perception that Editor E is female affect how E is treated by wikipedians?

2. How do differences between men and women affect how they view and interact with Wikipedia?

For (1), get a bunch of senior editors, ideally some of the women. Have them form two separate accounts, develop user talk pages for them which identify them as male and female, and pretend to be junior editors. Try to distribute their edits among the accounts equally, and not behave notably differently between the two accounts. At the end, check for differences between community responses to their edits. Did they get reverted more as men or women? Dingsuntil (talk) 05:03, 11 March 2015 (UTC)

I think this is a wonderful proposal. Its an experiment, a prospective study as such, and completely different from this proposal which is a retrospective review. Write it up, Dingsuntil!--Wuerzele (talk) 19:22, 11 March 2015 (UTC)

Survey of women's views on Gardner's points

Has anyone done a survey on female wikipedians (or non-wikipedians) views on the significance of the points? Both the impact on them as individual editors ("The level of conflict on Wikipedia makes me less likely to edit: Strongly agree/somewhat agree/somewhat disagree/Strongly disagree") and their opinion on the total impact ("I think the level of conflict discourages women from participating") are worth exploring, although I think the former is better data. For example, I'd guess that (4) had a much bigger impact on female participation than (8), but I'm not a statistically-significant group of women. Dingsuntil (talk) 05:03, 11 March 2015 (UTC)

Dingsuntil There is no statistically-significant group of women in Wikimedia projects, so your proposal cannot be done. The project proposed here would look at 1-2 million articles that are tagged with a gender. This would test the idea that gender matters at all, because no, I do not think there is a statistically valid study examining this even though there is plenty of anecdotal evidence including focus groups and interviews.

There are maybe 300 highly active female editors on English Wikipedia. In this population, they are too frequently solicited for surveys already, and because they have been extensively petitioned to take surveys, probably most of the population is biased now because anyone eligible and willing to take a survey is likely to now be biased or fatigued by years of soliciting and questioning. Even if that were not the case an amazing turnout of 20% of all highly active female editors would just be 60 people, which is too small of a sample to draw the kinds of conclusions which could be drawn by looking at a set of 1,000,000 data points from a set of any kind. Even among moderately active women there might be ~2000 on English Wikipedia at any time, and that is too low of a useful population size for doing surveying which seeks to capture more than anecdotes. Blue Rasberry (talk) 11:27, 11 March 2015 (UTC)

Maybe that's what's really keeping women off wikipedia: too many surveys. But seriously, you don't necessarily want to survey the highly active female editors. They're already engaged, so unless you think resulting policy changes would make them less engaged, you probably shouldn't worry too much about their views on the 9 points. Ask low and moderate-activity female editors. Drive engagement there, and you boost female impact quickly. Then they recruit their girlfriends for another boost. Dingsuntil (talk) 10:03, 12 March 2015 (UTC)

I'm quite intrigued by Dingsuntil's idea, and I think it is far more feasible than Blue Rasberry makes it sound. As an example of a more feasible sampling method, we could solicit people in a public park; ask if they edit Wikipedia; ask for their gender; and if they don't edit Wikipedia, ask the sort of question that Dingsuntil proposes. I am unlikely to undertake such an undertaking on my own because I find that conventional questionnaires usually more work than they're worth.

Also, you both use the phrase "statistically significant", and I don't quite know what you mean by it. Tlevine (talk) 15:01, 11 March 2015 (UTC)

It's some stupid technical math thing. Don't worry about it. Basically, it means "We asked a lot of girls, so we're more sure." Dingsuntil (talk) 03:37, 12 March 2015 (UTC)

What about country differences – Wikipedia contribution in the most gender equal countries?

Just jumped into this, but are there any country-based statistics? Are the most gender equal countries in general (=the Nordic countries Denmark, Finland, Iceland, Norway, Sweden) also most equal when it comes to Wikipedia contribution? If so, the answers may be found in the structures in these countries. But the languages are small and I have seen no statistics on Swedish etc in the survey (the pdf files I could download). Why should Wikipedia contribution be differently than female participation in the institutions of society in general, the proportion of women in parliament and government, equality in housework and parenting, female representation on company boards, percent female entrepreneurship, percent female university graduates? Why not start searching the answers in northern Europe? Just an idea. --Caspiax (talk) 23:44, 11 March 2015 (UTC) (meditate no conflicts, peace is lovely and war is ugly)

Literature review. Already been done for editors, and confirmed.

I noticed in your background you didn't mention this 2011 paper by GroupLens research, which confirms female editors revereted more and then goes further. See hypothesis H3. Maximilianklein (talk) 00:59, 20 March 2015 (UTC)

So confused about grant applications

How do I apply for an Inspire grant? I think I need to categorize the grant page in a particular way. See source code for Grants:IdeaLab/Inspire/Ideas_by_theme.

Anyway, can someone do whatever is necessary for this to be a grant proposal?

Tlevine (talk) 08:55, 31 March 2015 (UTC)

Hi Tlevine. Sorry this is confusing. The ideas has been marked as a proposal and is included in the list of open proposals. Alex Wang (WMF) (talk) 19:41, 3 April 2015 (UTC)

Eligibility confirmed, Inspire Campaign

This Inspire Grant proposal is under review!

We've confirmed your proposal is eligible for the Inspire Campaign review. Please feel free to ask questions and make changes to this proposal as discussions continue during this community comments period.

The committee's formal review begins on 6 April 2015, and grants will be announced at the end of April. See the schedule for more details.

Questions? Contact us at grants(at)wikimedia.org.

Comments from Superzerocool

Hi @Tlevine: thanks for sharing your idea. I have two comments after read your proposal:

1. I'm sorry if I misunderstood the idea, but... How this project could improve the gender perception of the participation in Wikipedia?. The readers are a mass critic about the Wikipedia contents and the examination is a project to see how many X are male / female, but I don't see how this idea tries to reduce the gender gap in the editors.
   1. (sub-question) This project, is thinked to be implemented only in English Wikipedia?. In Spanish we have the occupation separated in both sex (ie: painter could be pintor or pintora).
2. Is possible write more about the requested budget into items like personnel, technology, etc.? (if you provide more data about this, we have a better estimation of the costs).

Regards Superzerocool (talk) 17:16, 7 April 2015 (UTC)

Overlap with WIGI project

Hi @Tlevine:. Thanks so much for your idea! I noticed that you and @Maximilianklein: discussed briefly above the overlap between this proposal and his proposal for Wikipedia Gender Inequality Index. If it's useful to have a conversation to discuss if the projects can be combined, I'm happy to set that up. Please let me know. Thanks, Alex Wang (WMF) (talk) 21:29, 7 April 2015 (UTC)

Aggregated feedback from the committee for Examination of gender in biographies

Scoring rubric	Score
(A) Impact potential Does it have the potential to increase gender diversity in Wikimedia projects, either in terms of content, contributors, or both? Does it have the potential for online impact? Can it be sustained, scaled, or adapted elsewhere after the grant ends?	5.1
(B) Community engagement Does it have a specific target community and plan to engage it often? Does it have community support?	5.1
(C) Ability to execute Can the scope be accomplished in the proposed timeframe? Is the budget realistic/efficient ? Do the participants have the necessary skills/experience?	5.7
(D) Measures of success Are there both quantitative and qualitative measures of success? Are they realistic? Can they be measured?	4.9
Additional comments from the Committee: The proposed research is interesting, but there does not appear to be any clearly articulated path toward achieving actual impact on gender diversity, regardless of the outcome of the research. If we find that women's biographies are more contentious than men's biographies, where do we go from there? There does not appear to be any community engagement or support. There is enough community engagement to justify developing a more detailed plan and timeline, and a more detailed, decreased budget. I would strongly prefer that the project not be done alone but a team created. The budget seems very high. There's very little information presented about the Project Team. There need to more vigorous measures of success in order for the project to not be an exercise in data collection. The idea is interesting and has the potential if done well to collect information that could shape Wikipedia policy and editor recruitment, but needs a stronger project team and reworked budget.

Inspire funding decision

This project has not been selected for an Inspire Grant at this time.

We love that you took the chance to creatively improve the Wikimedia movement. The committee has reviewed this proposal and not recommended it for funding, but we hope you'll continue to engage in the program. Please drop by the IdeaLab to share and refine future ideas!

Next steps:

1. Review the feedback provided on your proposal and to ask for any clarifications you need using this talk page.
2. Visit the IdeaLab to continue developing this idea and share any new ideas you may have.
3. To reapply with this project in the future, please make updates based on the feedback provided in this round before resubmitting it for review in a new round.
4. Check the Individual Engagement Grant schedule for the next open call to submit proposals or the Project and Event Grant pages if your idea is to support expenses for offline events - we look forward to helping you apply for a grant in the future.

Questions? Contact us at grantswikimedia.org