Being relatively new to this list, I dip my toe into what seems to be a somewhat fraught mailing list with some trepidation. (Read: please don’t bite this newbie).

I think we need to understand where the problems lie and therefore what problem(s) we are seeking to solve. If I understand it correctly, we are looking at the low proportion of female editors. Presumably we need to understand what is happening to women in different phases of the lifecycle, noting that not all of these phases may occur for any individual woman

• Initial recruitment – women clicking “edit” for the first time – what does/doesn’t motivate?
• Newbie phase as anonymous editor (may or may not occur)
• Newbie phase as registered user
• Active editor
• Active editor self-identifying as female (can take many forms)
• Editor taking wiki-stress break
• Blocked editor
• etc

I note that a major difficulty in working at the earlier stages of the lifecycle is that we simply do not know whether the editor is male or female until there is some self-identification. Other than the choice of a obviously-gendered user name, we often have no way of guessing the sex of the user until they are experienced enough (e.g. know about User page, etc) *and* choose to self-identity in some form.

A second and not-entirely-dependent but not-entirely-independent set of issues relates to “gender” of articles. There is data to suggest that certain topics are more of interest to women and therefore less well-developed on WP because of there being fewer women editors. Therefore, there is the possibility of slicing the problem on another axis in relation to:

• Ungendered article, by which I mean there is nothing ”gendered” about the subject matter nor any reason to think it is more likely to interest editors of either sex
• Gendered-topic article, by which I mean the subject matter has “gender” but this doesn’t necessarily alter relative editor interest
• Gender-attracting topics, which disproportionately attract editors of one sex
• Gender-controversial topics, which I draw out because this seems to be a particular battleground, by which I mean articles about feminism, women’s rights, abortion, etc and other issues which are real-world controversial topics that have definite gender issues and create major POV issues.
• etc

I note that a machine-analysis of the edits of self-identified male/female editors we can identify those articles/categories which appear to be neutral or biased in terms of editor interest. Machine-analysis can also show us which articles/categories have high levels of activity (in particular high levels of reverts and low levels of text survival and probably high levels of Talk page activity and User Talk page of editors involved) that suggest they are “controversial” (although “breaking news” can manifest similar activity patterns without being controversial in the real world) and how self-identifying editors fare during these processes (simply, do female editors exhibit different patterns of behaviour to male editors?).

And there are probably other criteria by which we can slice this issue up. I think we have to recognise this is not just “one problem” requiring “one solution”. But rather that there are potentially many scenarios where we may have a problem and, if we do have that problem, we need a solution appropriate to that lifecycle phase and that kind of article. Or to put it another way, there is a world of difference between the anonymous female editor who attempts her first edit on a living person biography, has it reverted because there is no citation, and can’t understand why her edit disappeared (noting she probably doesn’t even know that she can view the edit summary that may explain why, assuming she can figure out what the cryptic letters WP:BLP means if she did) and the experienced female editor harassed on a talk page in a “sexualised” picture-of-the-day dispute. Both situations could be the straw that breaks the camel’s back and both women might never edit again, but clearly the problem is different and the solution has to be too.

Solutions like the existing ArbCom (or Hall of Justice as proposed) are both mechanisms that depend on the editor involved being 1) sufficiently experienced to know they even exist 2) know how to engage with them and 3) are comfortable engaging with them. Despite editing WP on and off for several years, 1) I did not know of ArbCom for many years  2) I still don’t actually know how to engage with it, and 3) I am not disposed to solve my problems that way (don’t like the conflict that I rightly-or-wrongly presume is part and parcel of it). A Hall of Justice solution might work for particular scenarios (although I concur with the practicalities of staffing it with a 50% female representation) but is probably irrelevant for many others. Or to put it another way, I suspect the membership of this list is not typical of the WP editor population, nor even of the female WP editor population and we have to be careful to not just design a solution that works for us because we are probably highly atypical. I suspect that list members are 1) predominantly female 2) somewhat engaged in gender-politics 3) experienced WPians 4) feel empowered, etc. If we weren’t, we wouldn’t be on this list and wouldn’t be contributing to this conversation. We are the survivors of WP, not the ones lost to it. And we should not design solutions that just work for survivors. We have to speak and act for those who don’t survive and don’t make it to this list.

Summary. Can we break this problem up into pieces and address the pieces separately. I don’t think we get anywhere with “one size fits all”.

For example, could we run an A/B experiment where we require anonymous editors to provide an email address with their edit? Right now we have no effective means of communicating with them (writing on an IP user talk page being the ultimate exercise in futility – they don’t know it’s there). Being able to communicate with these very newest of editors might mean we can help them achieve their edit and improve their retention. It might also reduce vandalism. Yes, it makes them less anonymous and maybe there are downside with that, but unless we run that experiment, we will never know.

For example, is the ability to be anonymous or use a pseudonym more likely to allow undesirable behaviours? Do people behave better if they can be more linked to their real world identity? We could test this as a research project. Identify sets of words that disproportionately appear in “uncivil” remarks or revert stats or other possible indicators of undesirable behaviours and see if anon/pseudo users appear more likely to do these things than real-name users. If the finding was that undesirable behaviours are more likely to occur when there is no link to real world identity, maybe there’s solutions in that direction. I note that there are legitimate reasons for anonymity/pseudonymity in some circumstances but perhaps this should be the exception rather than the norm? I note that we aspire to scholarly practices in Wikipedia, yet overlook that scholarly publishing is always (in my experience at least) published under one’s real name and often has institution and email address (or other contact information) included. Frankly, in academic life, my name is my brand and my reputation. Why should WP be different?

I appreciate that the suggestions I make above might be big changes to the current culture, but if it is the culture that is the problem, it is what needs to be changed. Big problems are rarely solved by tinkering at the margins. I think the issues we have with the WP gender gap are not dissimilar to the gender gap in employment. It is usually easy to explain why you employed one person over another because most jobs have a number of selection criteria and by emphasising some over others you can often easily justify any decision retrospectively “yes, she had better qualifications, but he had more previous experience” or vice versa. This is where Hall of Justice solutions fail because they are looking at a specific situation. However, when one looks at the overall statistics of hiring staff, bias against women (or any other group) becomes harder to hide. So getting some kind of gender KPIs into the monthly WMF metrics might be another example of how we get focus on gender issue and how we can spot the macroscopic changes that occur when new software or new policies roll out. If an issue matters to an organisation, you start with metrics so people can see the problem and then you start putting changes to those metrics into annual plan of staff members, link them to bonuses, etc. While WP is very much a volunteer organisation, we might have to think hard how we reward the volunteers whose behaviours lead to improvements in the metrics (it’s easier with staff through bonuses, promotions, etc), but there must be a way (scholarships to Wikimania, the Fluffy Kitten BarnStar of Gender Metrics Improvement, etc).

Kerry, who is female, does self-identify as such, edits under her real name (User:Kerry Raymond), likes receiving Kitten WikiLove and is about to don her best asbestos suit for fear of the flames to come …