Wikipedia:Wikipedia Signpost/2019-04-30/Recent research

Recent research

Female scholars underrepresented; whitepaper on Wikidata and libraries; undo patterns reveal editor hierarchy

By Lane Rasberry, Thomas Niebler and Tilman Bayer

A monthly overview of recent academic research about Wikipedia and other Wikimedia projects, also published as the Wikimedia Research Newsletter.

"Female scholars need to achieve more for equal public recognition"

Reviewed by Thomas Niebler

The underrepresentation of women in science, not only STEM fields, is one of several problems the research community is dealing with. One aspect of this is the public recognition of female researchers, for example on Wikipedia. Women have to achieve more than men to be equally recognized in public. One specific example is given by Donna Strickland, whose Wikipedia page was first deleted in 2014 due to copyright violation, was declined by Articles for Creation for failing AfC notability rules (see Signpost coverage) and has been recreated only after she had won the Nobel Prize in Physics in 2018.

In their work,^[1] Schellekens et al. supported this claim with a systematic analysis of the interdependencies of the existence of a researcher's Wikipedia article, their gender and their h-index, i.e. a measure of recognition of a researcher's scientific works. Using a logistic regression approach, they found that "regardless of [the] field of study, being male significantly increases the chance of being recognised and featured on Wikipedia". They state that a "male economist has to achieve an h-index of 11 for a similar probability of public recognition as a female economist with an h-index of 19", while "similar patterns are observed for Physics and Philosophy". As a conclusion, they discuss several factors that have to be worked on in order to increase public visibility of female researchers without them needing to win a Nobel prize first.^{[supp 1]}

"ARL White Paper on Wikidata: Opportunities and Recommendations"

Reviewed by Lane Rasberry

On 18 April, in a blog post the Association of Research Libraries published "ARL White Paper on Wikidata: Opportunities and Recommendations"^[2], a white paper describing how research libraries can collaborate with Wikidata and its community of contributors to enrich library resources. Recommendations in the paper include that library leadership should "Give staff time to experiment and contribute to Wikidata" and "Expand capacity with Wikimedians in Residence or fellowships". Any Wikimedia community member seeking to establish an institutional partnership with a library, STEM organization, cultural partner, university, government agency, or knowledge center of any sort would do well to share this document to introduce the case for ongoing Wikimedia engagement in Wikidata, Wikipedia, Commons, and the rest of the Wikimedia platform.

Conferences and events

See the research events page on Meta-wiki for upcoming conferences and events, including submission deadlines, and the page of the monthly Wikimedia Research Showcase for videos and slides of past presentations.

Other recent publications

Compiled by Tilman Bayer

Other recent publications that could not be covered in time for this issue include the items listed below. Contributions, whether reviewing or summarizing newly published research, are always welcome.

"Wikipedia access and contribution: Language choice in multilingual communities. A case study"

From the abstract:^[3]

Seventy-seven first-year audiovisual communication students [at Pompeu Fabra University ] made contributions to Wikipedia as part of the assessed work in the first year course titled “Digital Culture.” Before and after writing Wikipedia articles, the students responded to two questionnaires that enquired about their language-related habits when using the site and about their language choice for contributing to it [with the vast majority stating Catalan and/or Spanish as their main languages]. ...
Students favor the English edition of Wikipedia when consulting it despite the fact that this is the language they assess themselves as being less proficient at in reading. More generally, our research shows that multilingual Wikipedia users move seamlessly from one language edition to another, thus refuting the cliché that relates minority languages with exclusively local and self-referential topics.

Undo patterns reveal "dominance order" among editors

From the abstract:^[4]

... In this article we claim that dominance (respectively deference) is revealed by users undoing (respectively redoing) edits of others. We propose methods to turn the history of Wikipedia pages into a dynamic multiplex network resulting from three types of interaction events: dyadic dominance, dyadic deference, and third-party assigned dominance ties. We analyze [...] a sample of page histories comprising 12,719 revisions by 7,657 unique users. On the dyad level we analyze whether two users tend to agree on a dominance order among them or whether dominated users tend to fight back. On the neighborhood level we analyze various degree effects including whether dominant users tend to dominate in the future and whether subordinate users tend to get dominated. On the triad level we analyze whether users have a preference for transitive closure over cyclic closure of dominance ties. These dynamic patterns shed light on the micro processes that can foster or impede the emergence of a global linear hierarchy."

See also our coverage of other publications by the same two authors.

"Wikipedia graph mining: dynamic structure of collective memory"

From the abstract:^[5]

... we propose a distributed graph-based event extraction model, inspired by the Hebbian learning theory. [...] We focus on data-streams with underlying graph structure and perform several large-scale experiments on the Wikipedia visitor activity data [i.e. pageviews ]. We show that the presented model is scalable regarding time-series length and graph density, providing a distributed implementation of the proposed algorithm. We extract dynamical patterns of collective activity and demonstrate that they correspond to meaningful clusters of associated events, reflected in the Wikipedia articles. We also illustrate evolutionary dynamics of the graphs over time to highlight changing nature of visitors' interests.

See also blog post, code and dataset.

"(De)constructing public opinion. The collective creation of articles on political developments in the Wikipedia in Spanish"

From the English abstract:^[6]

In this article we will see how the interaction of different users, both active and anonymous users who make vandalism, influence the reconstruction of public opinion from the cases of two articles in Wikipedia in Spanish.

"Suggesting Specific Segments as Link Targets in Wikipedia"

From the abstract:^[7]

... there are a number of links in Wikipedia pointing to corresponding segments, because the whole article is too general and it is hard for readers to obtain the intention of the link. We propose a method to automatically predict whether a link target is a specific segment and provide which segment is most relevant. [...] Through evaluations on Wikipedia articles, our method performs better result than existing methods."

"DOI Links on Wikipedia. Analyses of English, Japanese, and Chinese Wikipedias"

From the abstract:^[8]

In this paper, we analyzed Digital Object Identifier (DOI) links among English, Japanese, and Chinese Wikipedias [...]. Most of the DOI links in these Wikipedias were revealed to be CrossRef DOIs. The second most-referenced in jawiki were JaLC DOIs, whereas those in zhwiki were ISTIC DOIs. JaLC DOIs were uniquely referenced in jawiki, and ISTIC DOIs tend to be referenced in zhwiki. In terms of DOI prefixes, Elsevier BV was the largest registrant in all languages. Nature Publishing Group and Wiley-Blackwell were also commonly referenced. [...] overlapping analysis showed that jawiki and zhwiki share the DOI links with enwiki at a similar high rate. The analysis of revision histories showed that the DOI links had been added to enwiki before they were included in jawiki and zhwiki — indicating that the majority of DOI links in jawiki and zhwiki were added by translating from enwiki."

References

^ Schellekens, Menno H.; Holstege, Floris; Yasseri, Taha (2019-04-12). "Female scholars need to achieve more for equal public recognition". arXiv:1904.06310 [cs.DL].
^ "ARL White Paper on Wikidata: Opportunities and Recommendations". Association of Research Libraries. 2019-04-18.
^ Soler-Adillon, Joan; Freixa, Pere (2017-12-15). "Wikipedia access and contribution: Language choice in multilingual communities. A case study". Anàlisi. 0 (57): 63–80. doi:10.5565/rev/analisi.3109. ISSN 2340-5236.
^ Lerner, Jürgen; Lomi, Alessandro (2016-11-30). "Dominance, Deference, and Hierarchy Formation in Wikipedia Edit-Networks". In Cherifi, Hocine; Gaito, Sabrina; Quattrociocchi, Walter; Sala, Alessandra (eds.). Complex Networks & Their Applications V. International Workshop on Complex Networks and their Applications. Studies in Computational Intelligence. Springer International Publishing. pp. 95–107. doi:10.1007/978-3-319-50901-3_8. ISBN 9783319509006.
^ Miz, Volodymyr; Benzi, Kirell; Ricaud, Benjamin; Vandergheynst, Pierre (2017-10-01). "Wikipedia graph mining: dynamic structure of collective memory". arXiv:1710.00398 [cs.IR].
^ Azcárraga, Luis Álvarez (2018-06-14). "(De)construyendo la opinión pública. La creación colectiva de artículos sobre acontecimientos políticos en la Wikipedia en español". Caleidosocopio - Revista Semestral de Ciencias Sociales y Humanidades. 21 (37): 91–120. doi:10.33064/37crscsh904. ISSN 2395-9576. ("(De)constructing public opinion. The collective creation of articles on political developments in the Wikipedia in Spanish", paper in Spanish, with English abstract)
^ Wang, Renzhi; Iwaihara, Mizuho (2016-12-07). "Suggesting Specific Segments as Link Targets in Wikipedia". Digital Libraries: Knowledge, Information, and Data in an Open Access Society. International Conference on Asian Digital Libraries. Springer, Cham. pp. 394–405. doi:10.1007/978-3-319-49304-6_42.
^ Kikkawa, Jiro; Takaku, Masao; Yoshikane, Fuyuki (2016-12-07). "DOI Links on Wikipedia". Digital Libraries: Knowledge, Information, and Data in an Open Access Society. International Conference on Asian Digital Libraries. Springer, Cham. pp. 369–380. doi:10.1007/978-3-319-49304-6_40.

Supplementary references and notes:

^ One notable attempt to diminish this recognition imbalance on Wikipedia is done by Jess Wade, by writing a short article about another female STEM researcher each day.

← Previous "Recent research"

Next "Recent research" →

In this issue

30 April 2019 (all comments)

News and notes

In the media

Discussion report

Discuss this story

These comments are automatically transcluded from this article's talk page. To follow comments, add the page to your watchlist. If your comment has not appeared here, you can try purging the cache.

The quantitative data analysis in the paper by Schellekens, Holstege and Yasseri is pretty stark: whether you look at physics, economics or philosophy, a man active in one of those fields is considerably more likely - as much as twice as likely - to have a Wikipedia article as a woman. There is a clear selection effect or systemic bias at work. It is not just that there are fewer women active in these fields, or that women are less prominent: even for people of equivalent accomplishment, a man is significantly more likely to have a Wikipedia article than a woman. (That said, we didn't have an article on George Smith (chemist) until he won the Nobel prize last year, and we still don't have an article on the Stanford professor Tony Heinz, the only remaining redlinked former president of The Optical Society.)
However, it is not correct to say that an article on Donna Strickland was "first deleted in 2014 due to lack of relevance". An article was deleted in 2014 as a copyvio. As far as I am aware, no one created a new article until a WP:AFC draft article in March 2018, which was rejected in May 2018 for lack of independent sourcing (see Draft:Donna Strickland). No one said she was not notable or lacked "relevance", whatever that means, let alone deleted an article for that reason. The Signpost from last year includes several articles discussing what happened at some length - for example - so I am surprised to see the events being misrepresented in this report.
And now we have the debate around the article on Clarice Phelps, with its intersection of ethnicity and gender, which has reached its second DRV... 213.205.240.174 (talk) 16:01, 1 May 2019 (UTC)[reply]
@213.205.240.174, Bluerasberry, ThomasNiebler, and HaeB: - (sorry for multi-ping, didn't know who wrote what) - "relevance" is indeed an odd choice of words. The original version was indeed deleted for copyright, and the link the deletion writing sits on goes to the declined draft, not the deleted article, linking to both the draft and the article's log history would seem reasonable. I disagree on the specific Strickland case (premised on AfC rules), but yes, this very interesting article was indeed a clear statement that it's not merely a paucity of sources that's leading to under-representation, the fault lies here, too. Nosebagbear (talk) 22:31, 1 May 2019 (UTC)[reply]

@Nosebagbear: "Relevance" does seem incorrect. You seem to have thought this through. We are a newspaper but still also a wiki - would you be comfortable correcting the text? Blue Rasberry (talk) 23:21, 1 May 2019 (UTC)[reply]

@Bluerasberry: - hi, I've made some changes in what I believe was a neutral fashion - please let me know if I've either failed with that or there are any copyediting errors. Nosebagbear (talk) 09:49, 2 May 2019 (UTC)[reply]

@Nosebagbear: I checked what you did. This is great! Wikipedia editors know that we do not judge "relevancy". We judge "notability" by the community-designed definition. Thanks for the changes. Blue Rasberry (talk) 10:39, 2 May 2019 (UTC)[reply]

I've seen "wp:undue" used all the time as a form of judging relevance on Wikipedia. AugusteBlanqui (talk) 11:59, 2 May 2019 (UTC)[reply]

Thanks Nosebagbear! I have also added a link to the Signpost's earlier coverage for fuller background.

We usually avoid changing the content of Signpost articles after publication, but corrections of this sort are a worthwhile exception.

Re "didn't know who wrote what" - there is a separate byline for each review (right below the section heading, I agree it's a bit confusing with the overall bylines on top), in this case Thomas authored it. But I should have caught this too, as the editor of this section / the research newsletter. Regards, HaeB (talk) 18:18, 4 May 2019 (UTC)[reply]

Yes I spotted that in my most recent read and self-trouted myself! Nosebagbear (talk) 19:21, 4 May 2019 (UTC)[reply]

"Self-trouted". That's a phrase I wish I'd heard before. Nicely done. SchreiberBike | ⌨ 19:50, 4 May 2019 (UTC)[reply]

I'm late to the party, but could "was declined by Articles for Creation for failing AfC notability rules" be changed to "was declined by Articles for Creation for failing to demonstrate notability"? This clarifies that, despite what about a million journalists have reported, no-one ever decided that Strickland was not notable enough for an article. The problem was that a particular draft did not adequately demonstrate notability. Adrian J. Hunter^{(talk•contribs)} 05:37, 10 May 2019 (UTC)[reply]

@Adrian J. Hunter: - my word choice did attempt to clarify that point - it doesn't do it particularly well (unless you already are a Wikipedian) but I'm not sure yours would clarify the point either Nosebagbear (talk) 08:19, 10 May 2019 (UTC)[reply]

Love this column! The "Female scholars" study is really interesting. One methodological aspect I'd like to know more about is how they determined the existence of a Wikipedia article for a given researcher. Based on section C of page 3, it sounds like they may have just used the Wikipedia API to check for the existence of an article (or redirect) for "Jane Smith" or whatever the name was. If so, couldn't their results be caused (or at least confounded) by the fact that Wikipedia has more articles about men as a whole, and therefore a male scholar with a low h-index is more likely to get a false positive match for someone with the same name who does have an article? This would also be consistent with their finding regarding the negative interaction effect, which "suggests that Wikipedia's bias towards men is strongest among scientists with relatively low indexes". They mention manually checking the article-lookup results for 30 people and finding no errors, but I imagine an error rate of less than ~1 in 30 could still have a significant effect with a sample size of 15k. I'd love to get my hands on the data and poke around. One way you could try to address this would be to check the existence of an article for the given name, and verify that it's categorized somewhere under the umbrella of Category:Scholars and academics? Or they could grep the article's text for the name of the scholar's institution (which they have from Google Scholar). Maybe they already did something like that? Though I feel they probably would have mentioned it. Colin M (talk) 18:11, 6 May 2019 (UTC)[reply]

What do you think of The Signpost? Share your feedback.

Home

About