Wikipedia talk:Wikipedia Signpost/2017-06-09/Recent research

Discuss this story

See also the Meta-wiki talk page of this research newsletter issue

I'm more concerned about bots second-guessing and repeatedly undoing the work of humans. There's a certain class of bots that dedicate themselves to reduce the quality of non-free images. This is OK for photographs, but they often turn screenshots of software into piss-poor quality, or even an unreadable mess (recent examples: [1], [2], [3], [4]). Even when a human reverts the mistake, the bot often comes back some time later and redoes the foolishness. Diego (talk) 08:43, 9 June 2017 (UTC)[reply]
I recall having read somewhere that most people outside of Wikimedia aren't actually sure of how internal Wikipedia processes work, hence the frequent misunderstandings. I can't find the source for that now (But if you know where it is, feel free to give me a shout), but it makes me wonder just how difficult is it exactly for outside reporters to actually find our internal processes to see that we're not the Barbaric website some school teachers want me to believe. For Yasseri to not mention the BAG even once in that article is surprising. —k6ka 🍁 (Talk · Contributions) 12:24, 9 June 2017 (UTC)[reply]
- Heck, there are some 'academics' who think Wikipedia is a for-profit enterprise. There is a lot of research about Wikipedia written by people who don't know much about its internal process, and fail to even mention the existence of relevant policies/foras. Ex. I sometimes review papers on the educatonal approach, and half of them don't seem to realize the existence of the entire WP:SUP and related support framework. Perhaps even worse those papers sometimes fail to cite years of relevant literature in the field. Sigh. --_{Piotr Konieczny aka Prokonsul Piotrus| reply here} 04:27, 12 June 2017 (UTC)[reply]

It's a little disturbing after all these years that there are still academics trying breaching experiments. All the best: Rich Farmbrough, 10:24, 11 June 2017 (UTC).[reply]

@Rich Farmbrough: Hello, Rich. Which paper are you referring to? NewYorkActuary (talk) 19:44, 11 June 2017 (UTC)[reply]

Trusting Wikipedia. Vandalism attacks and content resilience: an analysis model and some empirical evidence All the best: Rich Farmbrough, 07:57, 12 June 2017 (UTC).[reply]

Jewish, Christian and Islamic in the English Wikipedia is an interesting paper. There are a few obvious but minor errors (such as "Jewish" where "Christian" is meant on P134), but one facet that attracted my attention was the statement that "Conservapedia-style" is a top collocate for "fundamentalism" - s far as I can tell the two terms only occur together on Wikipedia:Fringe theories/Noticeboard/Archive 7, which suggests that there may be some issues with the text processing pipeline.

More interesting is the failure to see the wood for the trees: as an adjectival modifier Jewish/Judaic has the highest occurrence, followed by Islamic/Muslim, and Christian in distant third place -

Jewish/Judaic	207,283
Islamic/Muslim	176,592
Christian	134,650

What is the reason for this disparity? Possible reasons that spring to mind include treating Christianity as normative, boosterism (see the still unresolved issues with the contributions of User:Jagged 85 for example), a reluctance to label people and things as "Christian" and recentism in terms of coverage of 21st century events.

All the best: Rich Farmbrough, 11:09, 11 June 2017 (UTC).[reply]

I am not fond of traditional peer review, but PLoS ONE seems to be almost like a self-publication platform. At least, I doubt that the quality of an average paper published there is better than the quality of an average conference paper. Given that even traditional peer review produces its fair share of bad papers, well... I wonder if there is any research on quality of PLoS ONE model? --_{Piotr Konieczny aka Prokonsul Piotrus| reply here} 04:23, 12 June 2017 (UTC)[reply]