Wikipedia talk:Wikipedia Signpost/2014-12-31/Recent research

← Back to Recent research

Discuss this story

I found it interesting that in 95% of the cases presented to Spanish ArbCom were dismissed in 2008. Tutelary (talk) 00:05, 3 January 2015 (UTC)

• Sichler and Prommer address an important question, unfortunately they overstate its role saying

However, none of the studies has focussed on the communication style of the collaborative network in order to answer the question why and how this female underrepresentation could be explained and ideally balanced.

Firstly I am not certain that this is a valid reading of the existing research.

However laying that aside, it is important to realise that gender gap can be categorically stated to require more to explain it than "the communication style of the collaborative network". This basically is a variation the naive response of many years ago that "men are rude, therefore women don't edit". We have established to a reasonable degree that what is perceived as a "hostile editing environment" is as discouraging to males as to females. The reasons that we have the gender gap we do are primarily that females do not start editing, and secondarily that when they do edit, they edit less, and for a shorter period. While the communication style is important, given what other research has established, it can only be expected to explain a relatively small part of the secondary problem, and hence changing it will by no means "balance" the gender gap.

Sadly the substantive part of the paper is (as the authors put it) "exemplary" (I would rather say "anecdotal", but maybe they actually meant "exemplary", though in which of its senses would be hard to divine) - and examines a couple of threads for conformance to other authors' characterisations of male and female discourse, with mixed results. It then draws some conclusions out of thin air for example: "She obviously felt offended or did not believe that the conversation would come to an amicable or at least reasonable end."

By the discussion section the paper is falling apart. It starts with a 108 word sentence that even the users of agglutinative languages might recoil at - it has been split into two paragraphs, by some sensitive copy-editor perhaps, in the middle of the phrase "prevents Wikipedia". It continues with the delightful "Gender diversity should be understood as a chance to spread up Wikipedia’s angle..." Most of the rest of the discussion is taken up with more literature review, there are attempts to propose solutions, but these are based upon arguments from someone who's biography got deleted, rather than the results of the analysis presented by the authors.

For reference the conclusion seems to be "Thus, not only women need to be trained to survive the hostile environment, maybe the competitive concept of knowledge production needs to be changed." - the "competitive concept" here seems to be the tactic of out-waiting a disputant.

The final sentence of the paper is so abominably written that it almost defies understanding.

The authors promise - that "Our findings will be analysed with regard to the impact that Wikipedia has as a source of knowledge on its users and producers considering the public discourse" is not redeemed.

How anyone can believe that analysis of two talk-page conversations can lead to robust scientifically based proposals to improve the gender gap, or indeed anything other than a list of research questions is beyond me. Even more so, given the effort that must have gone into producing a 17 page paper, merely in terms of literature review and typing, is why anyone would publish such a paper without asking a native English speaker to at least read through and pick out the solecisms.

All the best: Rich Farmbrough, 01:14, 3 January 2015 (UTC).

Anecdata and data based on small samples is very common within anthropology and ethnography, and is a perfectly valid approach to research. I'm not actually sure what the point of most of your message is, however; it seems to be railing against the injustices of runon sentences in the paper. Unless the researchers read the signpost (or frankly, even if they do), this is just...heh. "instruction and discrediting statements". Ironholds (talk) 15:54, 3 January 2015 (UTC)

@Ironholds: You may find my few lines more tiring to re-read than I found the original 17 pages, so I will summarize.

1. The authors do not appear to be familiar with existing research that is relevant to theirs.
2. They make promises that they don't fulfil
3. They pluck conclusions out of thin air
4. They base arguments on random comments from opinion pieces
5. No one proof read the paper, and the English is appalling, leading to ambiguities of meaning.

One item that may have been of interest (and of use) had it been phrased as a RQ, and applied to a significant amount of data

RQ1 To what extent do talk page contributions of gender identifiable editors on Wikipedia follow existing models of gender differentiation of on-line communication?

Questions of this subtlety are regularly answered by other researchers.

All the best: Rich Farmbrough, 21:58, 3 January 2015 (UTC).

Is reciprocity an extrinsic motivation?

Reciprocity isn't actually an extrinsic motivation: "reciprocity is significantly stronger when extrinsic motivation can be ruled out." EllenCT (talk) 08:58, 3 January 2015 (UTC)

The extrinsic/intrinsic motivation talked about in that paper is the apparent motivation for the first act. [T]he effect that the motivation of an agent has on the reciprocating behavior of another agent. All the best: Rich Farmbrough, 23:07, 3 January 2015 (UTC).

@Rich Farmbrough: do you agree with the implication in [1] that reciprocity is an extrinsic motivation? I think the senses of justice and fairness are often considered intrinsic because third parties are usually in agreement about such scenarios when they have the same information. EllenCT (talk) 00:26, 4 January 2015 (UTC)

I'm not sure. It seems to me that reciprocity can be concrete or abstract - examples from The Gift Relationship might be "I received blood and I want to give back" where the blood received is certainly extrinsic, and abstract "I (or my family) might need blood" - there is in neither case any legal obligation, though one might consider a stronger moral obligation in the first case. Whether a moral code is intrinsic or extrinsic is hard to say - certainly bushido is extrinsic. Conversely in American cases recipients of blood have been required to give (or find donors for) an equal or more often greater amount of blood. This is clearly a form of mandated reciprocity, which is extrinsic. I should, I suppose, read the full paper if I can find a copy. All the best: Rich Farmbrough, 00:45, 4 January 2015 (UTC).

Gender split in 2012 Wikipedia Editor Survey?

As Tbayer (WMF) is one of the authors of this piece, and Maximilianklein's report covers gender issues, this seems as good a place as any to ask Tilman once more what the gender split in the 2012 Editor Survey was. The survey's talk page is full of community members asking for this data, yet all inquiries directed at Tilman over the past half year have been ignored. Can we please have this data – just the simple gender split: x% male, y% female, z% other? It is now over 2 years since the survey ran, and this simple piece of data should take less than a minute to report. Thank you. Andreas JN466 11:57, 3 January 2015 (UTC)

This is an extremely important data point. If there are methodological issues, I am sure we can address them. All the best: Rich Farmbrough, 22:17, 3 January 2015 (UTC).

I agree that it is an extremely important data point, Rich, all the more since the WMF has not done an editor survey since then. So, if it is ever published, this will be the last data point on the gender gap for the foreseeable future. What do you suggest can be done to get the Foundation to release it? I am at my wits' end: I have asked about it on two mailing lists, I and others have left talk page messages for Tilman in three projects, there are enquiries from John Vandenberg and numerous other editors on the survey talk page in Meta dating back almost two years, yet the only answer I ever got was from Phoebe last month, who said, "I looked through my papers the last time you asked, and I don't think I have it. I'd send it to you if I did." This is an absurd situation: donor money paid for this survey. Yet for years now, the gender split has been shut in some drawer at WMF that nobody wants to open. --Andreas JN466 12:35, 4 January 2015 (UTC)

R

Suggested tweaks: point to the CRAN libraries rather than the GitHub ones, and note that WikipediR was released in April 2014. That's not a particularly new thing (or a particularly recent thing - if you've tried writing API client libraries before httr you'll know how much of a pain it was). Ironholds (talk) 15:52, 3 January 2015 (UTC)