User talk:Jimbo Wales

"Venezuela government controls almost all the newspapers and television stations"

Hello Jimmy. I'm watching your "State of the wiki" talk for Wikimania 2015. There you said that "Venezuela government controls almost all the newspapers and television stations". Where did you get that idea?

That is a typical hoax repeated in many places. Here I gathered some info into two tables, using data from Wikipedia articles.

National newspapers in Venezuela Newspaper Owner Diario 2001 Bloque De Armas Correo del Orinoco State-owned Diario VEA State-owned El Nacional Miguel Henrique Otero El Nuevo País Rafael Poleo El Universal Epalisticia S.L. Tal Cual La Mosca Analfabeta C.A. Últimas Noticias Cadena Capriles

National television networks in Venezuela Television network Owner Venevisión Gustavo Cisneros Venezolana de Televisión State-owned TVes State-owned Televen Corporación TELEVEN, S.A ViVe State-owned Globovisión Juan Domingo Cordero Vale TV Arzobispado de Caracas Meridiano Televisión Bloque De Armas TeleSUR Public company

I hope that this data helps to clarify this topic. Regards. emijrp (talk) 14:45, 5 September 2015 (UTC)

1. "The national airwaves are now almost entirely dominated by the government and its obligatory announcements, called cadenas." - Reporters without Borders
2. "In a 2015 report by the Institute for Press and Society (IPYS), over 25 media organizations had changed in ownerships between 2010 and 2015 with the new owners having "a direct relationship" to local governments and the national government that were linked to Chavismo.[21]" - see Media_of_Venezuela which links to news about report here.
3. "Under new ownership, the network purged its newsroom and stopped airing live speeches by opposition leader" - Washington Post

Control of the media need not take the form of direct government intervention - threats leading to self-censorship and purchase by government proxies is sufficient. I stand my my assessment, which is in agreement with every serious human rights organization I have seen write about the issue.--Jimbo Wales (talk) 13:42, 6 September 2015 (UTC)

@Emijrp: ¡Por vida! _Fortuna ^{Imperatrix Mundi} 13:51, 6 September 2015 (UTC)

Venezuela's situation is definitely a bad one - a cautionary tale that shows that even when freedom of the press is badly abused, abolishing it is a worse cure than the disease. Without suspending RCTV's license and persuading some of the other channels above to take a more 'neutral' position, Chavez still might not have faced another coup attempt; but by pursuing censorship online and off, he overthrew himself - the society he might have wanted was overthrown by one of excessive security and a failure to contemplate or compromise that has left them vulnerable in the oil downturn. For those abroad who wanted to turn back the Pink Tide, a bad leftism is far more useful than the complete overthrow of Chavez' government - indeed, one scarcely hears of the accomplishments of Uruguay, which has surged far ahead of the U.S. on that first press index, or of Bolivia or any other country that has shown that leftist politics can work, and work with democracy and human rights. That said, is the vague connection of the owner of Globovision with government interests really as close as that of Roger Ailes, who is CEO of Fox News? What about MSNBC, given the chumminess between Microsoft and the NSA? Even when you consider Ted Turner, definitely an independent media voice ... what does someone whose land has a higher GDP than Belize have to do with the legendary General Will? He may be independent of government, but what owner of large media is independent of the prerogatives and interests of wealth? I think it may be better to take a more remote, bottom-up view: what happens to you if you use a watt or two to broadcast your video on TV in your block, without getting it cleared by Somebody Better Than You? Then ask what country's media is free. Wnt (talk) 15:15, 6 September 2015 (UTC)

In all countries we should care a lot about journalistic freedom, and about what non-journalistic influences may be impacting what we see and hear and read. Having said that, I think it's unwise to make a comparison that would suggest that things in Venezuela are probably routine - they most certainly are not.--Jimbo Wales (talk) 15:38, 6 September 2015 (UTC)

That's not really my intent - I don't want to minimize censorship in Venezuela. The resignations at Globovision over their change in editorial policy are certainly a clue that something is particularly wrong there. Still, biased pro-government reports can occur anywhere when emboldened by a lack of competing voices to debunk false claims. Censorship is like Ebola - some people may just have a fever, while others are vomiting blood, and that matters for their prognosis ... but it is always the same virus, waiting for its next victim. I think that (because of the EM frequency licensing structure that Venezuela and the US share) the Internet offers the best alternative to biased, censored, or just "owned" news. For example, Wikipedia, a site where people can write and contribute directly, usually does far better at putting complex political issues in context than 99% of media reports. So it concerns me most when censorship is targeted at ordinary people on the Internet. Wnt (talk) 21:29, 6 September 2015 (UTC)

If you assess the situation of Venezuela media using Reporters Without Borders (see #Criticisms of RWB) and The Washington Post... Well, then obviously you will have that impression. In the same fashion that you check who controls the Venezuelan newspapers/televisions, you should check who controls that NGOs and newspapers, and their interests.

I wonder if the world was worried about freedom of expression in Venezuela before Chavez/Maduro, when all the media was private in the hands of businessmen and there were no public ones. Obviously not, because the super-neutral mass media and NGOs didn't report about it.

For a real problem for freedom of expression, see Concentration of media ownership. Here is a table with the most linked mass media as references in Wikipedia:

Mass media in references [1]

Mass media	Links	Country	Owner(s)
BBC	386,665	United Kingdom	Statutory corporation
The New York Times	211,769	United States	The New York Times Company
The Guardian	107,524	United Kingdom	Guardian Media Group, Scott Trust Limited
The Daily Telegraph	55,543	United Kingdom	Telegraph Media Group, Press Holdings, Barclay Brothers
Washington Post	44,674	United States	Nash Holdings LLC, Jeff Bezos
The Independent	38,983	United Kingdom	Independent Print Ltd, Evgeny & Alexander Lebedev
Los Angeles Times	30,894	United States	Tribune Publishing
Time	30,791	United States	Time Inc.
ABC	29,968	Australia	Statutory corporation
Reuters	28,290	United Kingdom	Thomson Reuters, The Woodbridge Company, Thomson family
USA Today	27,024	United States	Gannett Company
Daily Mail	26,251	United Kingdom	Daily Mail and General Trust

Almost all them owned by rich people, big companies, holdings, banks, etc. I don't see any owned by the poor, do you? Where is their freedom of expression? --emijrp (talk) 16:38, 6 September 2015 (UTC)

The funny thing is, as an exercise in debating, I could take that exact data and make effectively the inverse argument: that the English Wikipedia has a left-wing bias. The American right-wing (I'm American, so this is my frame of reference) pretty much considers all of those except maybe the Telegraph and the Mail to be leftie socialist rags, especially the top three in that table. And indeed, it's not hard to find right-wingers online complaining about how Wikipedia is a liberal propaganda machine. --71.119.131.184 (talk) 20:22, 6 September 2015 (UTC)

@Emjirp: That's a good point about the rich being overly represented compared to the poor. However, the media's consisting of mostly 1%-controlled corporate outlets such as the above influences society as a whole, with Wikipedia simply using the most reliable sources available. --Rubbish computer 22:32, 7 September 2015 (UTC)

It's a little more nuanced than that - we actually have a choice, and can make the right one or the wrong one. A biased media is not strictly a censored media, which is to say, there are exceptions that slip through. An editor can pick through these back-page stories like Noam Chomsky in order to get a much clearer overall picture of what is going on. The problem is that recently there has been an upsurge of editors insisting that it is not enough that a fact merely be sourced reliably - they want it excluded unless many or most sources happen to mention it. Which is to say, they don't want a comprehensive encyclopedic resource, but a consensus summary that accurately reflects both the facts and the bias of the majority of the sources, without including inconvenient facts that broaden the context. When people see this kind of thing they need to push back hard against it, or yes, Wikipedia might as well be run by some big company. Wnt (talk) 12:01, 8 September 2015 (UTC)

Getting better?

Is Wikipedia getting better? --Bob K31416 (talk) 15:43, 6 September 2015 (UTC)

@Bob K31416: That depends on how you classify "better". It is growing. Rubbish computer 19:29, 6 September 2015 (UTC)

There was a study before showing it gradually getting more neutral in terms of wording, but still being biased to the left. Unfortunately I can't remember where. Rubbish computer 19:30, 6 September 2015 (UTC)

I agree at it would depend on what you mean by better since there are several ways to interpret better (including more non Western content, Quicker response to vandalism and POV Pushers, improving the software to make editing easier, etc).--65.94.253.185 (talk) 03:05, 7 September 2015 (UTC)

"It is growing." Well, then it obviously must be better, mustn't it?. Because everyone knows that bigger is better. Q.E.D.

However, if instead of looking at the number of articles it contains and creates (many of which are of mediocre quality) you instead look at the number of active contributing editors, then Wikipedia has actually been shrinking since 2007, and continues to do so. Of course, no one involved in Wikipedia wants to hear that - especially old guard "well-entrenched" editors who wish to perpetuate, as part of their own Peter Pan denial of changing reality, the period just prior to circa 2007 when Wikipedia was in its heyday and had become a web institution. However, since its 2007 zenith at over 50,000 active editors, Wikipedia editorial participation has been steadily declining due to it becoming a bureaucratic behemoth with no structure or leadership and a high level of policy creep. By its very nature Wikipedia eschews any sort of central planning and conventional expertise. Thus it considers contributions by subject matter experts in real life to be a conflict of interest since they get paid for their expert knowledge, and that's not "the Wikiway" - which is amateur volunteerism. Consequently, Wikipedia articles are created mostly by unemployed management consultants, dilettante telephone sanitizers, and a Scouse hairdresser called Rita who mostly contributes on her two half days off work.

Over the intervening years since its 2007 peak Wikipedia's culture, which was always feisty and argumentative (but ultimately in a good, constructive way), has now become top heavily bureaucratic and highly confrontational (in an obstructionist and persnickety unproductive way). The rules and guidelines for contributing to the project (which used to be just the "five pillars" of policy guidance) have now reached labyrinthine levels that long ago crossed "Teal Deer" thresholds, becoming internally inconsistent and self-contradictory in the process. This, in turn, only creates more opportunities for daily acrimony and disputes to occur, thereby requiring an ever-increasing volunteer work force of officious and sometimes abusive admin panjandrums to police it. Jimmy Wales has been dismissing suggestions that the project will get worse for years now (despite hard evidence to the contrary), but is on record as stating that he believes the project cannot significantly improve without an influx of new editors who have different interests and emphases (not to mention gender!). Yet Wikipedia's complete intransigence - or perhaps its inept incompetence (e.g., the "Visual Editor" debacle); it's actually a lot of both - at abating the ever-increasing levels of acrimonious confrontation and bureaucracy is not only failing to attract his desired new blood, but is preventing what new editors that do venture to dip their foot into the Wikipedian waters from also staying very long, in addition to driving away long established "old guard" editors as well.

On the flip side of the coin, with Wikipedia receiving more than ten billion page views every month that keep it in the top ten of the most used websites in the world, and with the project still creating lots of new articles and pages, there are many Pollyannaish Wikipedians that feel everything is simply fine and dandy and generally tickety-boo. Wikipedia has continually grown from day one and will continue to do so in the foreseeable future. For instance, it hasn't even started to scratch the surface yet when it comes to documenting all the Finnish and Czech ice hockey players, so still plenty of work to do right there. Ask any Wikipedia editor and he'll tell you it's "a work in progress" with still no end in sight where that Borgesian day is eventually arrived at when the encyclopedia will have finally documented and defined everything that has ever existed in the world. No doubt the editors of Encarta felt as equally confident and bullish. If Wikipedia is bigger and brighter today then it can only be even bigger and brighter still tomorrow. Indeed it can, but one does suspect that such ostrich Wikipedians may have never read this particular article. — not really here discuss 05:56, 7 September 2015 (UTC)

Wow, that'a really, really long comment there. I doubt that many people will read all of it. I prefer to be concise. The best available metrics show that the number of active editors is actually increasing modestly, rather than decreasing. Read a recent Signpost article about the data. Cullen³²⁸ Let's discuss it 06:31, 7 September 2015 (UTC)

There ya go, old fella. I've now broken my post up into four separate messages so that even people with IQs less than ambient room temperature, or recovering from cataract surgery, like yourself will now be able to follow it. I also made sure I did it in as many edits as I could in order to maximize my edit count. That way I will soon be a Senior Editor, which will hopefully allow me to wander around Wikipedia with a gold star stuck on my forehead like an over-achieving preschooler too. Good call. — not really here discuss 08:15, 7 September 2015 (UTC)

Thank you for the gracious remarks. They reflect well on you, I'm sure. Cullen³²⁸ Let's discuss it 21:13, 7 September 2015 (UTC)

My favorite way of checking this is to "click random article" on 10 articles, and go back and look at them a year ago, 5 years ago, 10 years ago. Every time I have tried, it's unambiguous: Wikipedia is getting better by this test.--Jimbo Wales (talk) 08:28, 7 September 2015 (UTC)

So just checking that ten random articles all get longer in any chosen five year span conclusively proves Wikipedia is getting better? Wow, that's really scientific! Do you perhaps have any sources you could cite that verify that such a methodology is remotely meaningful of anything? How do your random checks prove the quality of the writing is improving; or that it's becoming more NPOV; or more factually correct; or better sourced? Does you random check methodology conclusively show that the percentage of women editing Wikipedia is significantly increasing over every five year span? Or that more articles are being written about Africa than, say, Finnish ice hockey players? BTW, by "citable sources" I mean reliable secondary sources from independent publishers ... not a Signpost article produced by Wikipedia as an Orwellian morale-raiser for the troops. — not really here discuss 09:22, 7 September 2015 (UTC)

Perhaps your could read what Jimbo actually wrote? The word "longer" or any synonym doesn't appear in his post. --NeilN ^{talk to me} 16:01, 7 September 2015 (UTC)

Perhaps YOU should read what Jimbo actually wrote. And Jimbo too, because he states below that he "suggested checking the quality." Show me where the word "quality" appears in his two line post? He merely states, "Every time I have tried, it's unambiguous". Unambiguous WRT what? Length, quality, truth, format, sourcing, dates, wombats? It's a meaningless statement. He may have intended "quality" - and I'm sure he did - but it's not what he wrote. He cannot expect readers to magically divine his intent. Incorrectly assuming what another person says goes right to the heart of WP:AGF. It applies just as much to expecting someone to correctly divine what you really meant to say (but didn't) as it does to expecting them not to read something entirely different into what you did say.

His persnickety reaction to my post may have been justified if I had chosen to divine "wombats" given the context of the discussion, but not for choosing any of the others meanings I listed as they are all pertinent. I went with "length" because that is the ONLY objective criterion on my list; all the other criteria require subjective analysis and assessment and are thus POV, therefore they could not possibly be considered unambiguous. — not really here discuss 06:59, 8 September 2015 (UTC)

Gee, Gamergate much? That's a lot of questions, and virtually every single one of them contains an invalid premise or is asking me as if I made claims that I did not make. Let me answer your questions, all of them, and then you can go away and never ever post on my talk page again unless you take that chip off your shoulder.

1. "So just checking that ten random articles all get longer in any chosen five year span conclusively proves Wikipedia is getting better?" - Checking the length is not what I suggested. I suggested checking the quality. 'any chosen five year period' - not 'any' - the relevant 5 year period, the one ending now. Does checking 10 articles "conclusively prove" anything? Of course not. I said that it's my favorite way - it is something that I do from time to time, and I encourage others to do it.
2. "Do you perhaps have any sources..." No, I don't. I made the method up out of thin air. But it's a good idea, and you should try it sometime. I would actually love it if we had a tool to allow lots of people to do it and track the results across thousands of articles over many years.
3. "How do your random checks prove.. (various things)" - try it and you'll see what I mean. All those things are true.
4. " Or that more articles are being written about Africa than, say, Finnish ice hockey players?" - This particular method is focussed on the quality of individual articles and will completely miss problems with balance across various fields. A different test would be required to deal with that. Again, I told about a favorite way to check on the quality - it is not the only way nor even a comprehensive way. I never claimed it was, so your hostility is unwarranted.
5. As to the rest of your comments - they contain little content but they do reveal your agenda, so thanks for including that.--Jimbo Wales (talk) 16:05, 7 September 2015 (UTC)

Of course you didn't say 'longer', but since you gave no methodology for assessing 'quality', he was teasing you with 'length'. Given that the random method generally gets you 10 articles that you would never find in a standard reference work, a better method is to start with random articles taken from a standard reference work, and see if Wikipedia through time is approaching the quality of the standard work, using an appropriate understanding of 'quality'.

Another method, if you are a specialist in some subject, is to watch the progress about articles in that subject. As you know, I know a little about this guy, and this recent edit was just plain weird. On that measure, it's not getting better at all. And that's despite my occasional attempts at improvement. Peter Damian (talk) 17:19, 7 September 2015 (UTC)

Apparently it's not getting better. I have had many differences with Fram, but he is on the mark here. Peter Damian (talk) 09:30, 7 September 2015 (UTC)

@Me? I'm not really here: I know bigger isn't better, I was simply stating a fact. Rubbish computer 11:49, 7 September 2015 (UTC)

@Me? I'm not really here: Do you have a source for the number of active contributors continuing to decrease? --Rubbish computer 12:06, 7 September 2015 (UTC)

I have a number of sources (a dozen or so off the top of my head, possibly more) that support most of the points I made in my original post. If I was posting that text into an article they would, of course, have been cited at the appropriate places. But this is a Talk page and I made my comment extempore, not with those sources directly to hand, some of which I have not read in a long while, so I will have to go Google them in order to locate them. There is a possibility that some no longer exist (as some might go back as far as 2009). I will post each source as and when I find it and append it as a bullet underneath this reply. This might take awhile. However, before even starting that process, I first wish to address Jimbo's response(s). — not really here discuss 20:03, 7 September 2015 (UTC)

OK, here is the first reference source for some of what I initially posted. It can be summed up by the quote: "So long as an illiterate drug addict can override the work of a Harvard professor, Wikipedia will never be an authoritative reference." This is the source / basis for my comments at the end of my first paragraph (although obviously I cannot source my own humor) and it is clearly pertinent to the sort of posting interchange that just occurred with user JBL (which is why I found it first). Peter Damian also appears to be having a problem with this obvious flaw regarding how Wikipedia works.

• http://www.dba-oracle.com/t_credibility_wikipedia.htm

— not really here discuss 03:59, 8 September 2015 (UTC)

Here is a source reference that covers much of the material in my first post here (it is one of many and no doubt there will be much overlap). It addresses your initial question as to whether I could cite sources for "the number of active contributors continuing to decrease". It states categorically that they decreased by a third from the zenith contributor rate sometime in 2007 until ... well until it was published in October 2013. It doesn't prove the editor count hasn't increased again during the last two years, but I would have thought that highly unlikely, if only based on the sample behavior of people posting on this section of Jimbo's Talk page alone. My comments re paid feminism advocacy that upset Jimbo is based mostly on things I've read elsewhere (hopefully I can find those links too). This source actually contradicts that statement, claiming that articles on women's literature and feminism have actually declined in favor of computer games, which appears to be the exact opposite. I had no idea what "Gamergate" was until Jimbo stuck that label on me in order to try and suggest I had a hidden agenda, and so I had to go read the Wikipedia article on that topic in order to find out about it (to be honest I didn't expect to find an article on it). My source reading for that comment came from an Ivy League university source, if I remember correctly, but it may still have been an indirect reference to "Gamergate" without calling it that as such. I'm next going to redact that part of that comment until I can find, re-read and re-assess my sources supporting (or not) its validity. It is not at all critical to many of the other points I made, but it appears to have touched a nerve, and I apologize to Jim for that. However, there's no hidden agenda here.

• http://www.dailymail.co.uk/sciencetech/article-2477432/Is-decline-Wikipedia-A-staff-quit-site-thanks-dumbed-software-auto-delete-tools.html

— not really here discuss 05:35, 8 September 2015 (UTC)

While the Daily Mail article above mostly addresses the fact that long time editors are leaving because they are not willing to limbo dance in response to the new initiatives taken by Wikipedia in 2007, this study addresses the parallel lack of retention of new blood since 2007. However, it is purely a statistical analysis. Nobody appears to have talked to any exiting newbies to find out first hand why they left. Their reason for leaving is mostly speculative based on statistical analysis of new user accounts. But it does confirm statements I made in my initial post (and have repeated since) re linear falling newbie retention.

• http://www-users.cs.umn.edu/~halfak/publications/The_Rise_and_Decline/

— not really here discuss 06:20, 8 September 2015 (UTC)

This WSJ sourced article from 2009 even comes with a video for the ADD crowd that now appear to dominate Wikipedia. Although from 2009, nothing appears to have changed from what was already trending even six years ago. Are Jimbo and the folk at WMF simply just covering their eyes and ears and ignoring these long term trends or are they not able to do anything about them? I don't believe they are completely unaware of them since they keep coming up for discussion at the annual Wikimania meetings. A couple of notable quotes included in this particular source:

- "Wikipedia is becoming a more hostile environment, contends Mr. Ortega, a project manager at Libresoft, a research group at the Universidad Rey Juan Carlos in Madrid. Many people are getting burnt out when they have to debate about the contents of certain articles again and again."

- "He argued that Wikipedia needed to focus less on the total number of articles and more on 'smarter metrics' such as article quality."

• http://cuppacafe.com/author/the-perfessor/page/204 (the second discussion item, not the first one on climate change)

— not really here discuss 06:49, 8 September 2015 (UTC)

Using Jimbo's method, I pushed the random article button in the side menu 10 times. I only tried the 5 year part of Jimbo's time ranges of 1, 5, and 10 years, which would have been more thorough. Below are the links to the diffs from 5 years ago to now. In cases where the page was created less than 5 years ago, I gave the current version, which is essentially the diff from it's nonexistence 5 years ago to now.

1. https://en.wikipedia.org/w/index.php?title=Kim_Hee-sun&type=revision&diff=678916354&oldid=383599363
2. https://en.wikipedia.org/w/index.php?title=Dunleavy&type=revision&diff=652077569&oldid=369884357
3. https://en.wikipedia.org/w/index.php?title=Iniquity_%28band%29&type=revision&diff=662721521&oldid=378161666
4. https://en.wikipedia.org/w/index.php?title=Chamaemelum_nobile&type=revision&diff=672437156&oldid=380965667
5. https://en.wikipedia.org/w/index.php?title=Walter_G._Alexander&type=revision&diff=679298867&oldid=372115853
6. https://en.wikipedia.org/w/index.php?title=Uilenburg_%28Amsterdam%29&type=revision&diff=545941955&oldid=379618447
7. https://en.wikipedia.org/w/index.php?title=Tadahito&oldid=536153308
8. https://en.wikipedia.org/w/index.php?title=Niemi&type=revision&diff=540632279&oldid=372534500
9. https://en.wikipedia.org/w/index.php?title=Simon_Petrie&oldid=655260314
10. https://en.wikipedia.org/w/index.php?title=Eoxin_E4&oldid=670415197

--Bob K31416 (talk) 14:32, 7 September 2015 (UTC)

Great. I don't have time right now to study all those... how did we do in your random set?--Jimbo Wales (talk) 16:07, 7 September 2015 (UTC)

Here's some comments for each page. (I use "page" because some are disambiguation pages.)

1. Improved. Changed from small article with no figure and almost no cites to moderate sized article with 36 cites and figure.

2. Improved. A disambiguation page that grew with more wikilinks to articles and a discussion of item that was disambiguated and with a figure added for the discussion.

3. Unreferenced and about the same.

4. Uncertain but probably improved. It would take study. Reflist increased from 2 to 6, which is a good sign.

5. About the same with a few lines added.

6. A stub about the same.

7. Stub created about 3 years ago.

8. A disambiguation page that is about the same.

9. A small article created a little less than 5 years ago.

10. A stub created 8 months ago.

Overall, it looks like an improvement to me. (Just an aside, but most of the 10 pages were about people.) --Bob K31416 (talk) 21:03, 7 September 2015 (UTC)

I would be careful about inferring from this result anything about the rest of the encyclopedia, considering the limited subject matter of these pages and other factors. --Bob K31416 (talk) 21:31, 7 September 2015 (UTC)

You have just spent upto 7 hours (based on the difference in timestamps between posting your initial ten diffs and then posting back with your cursory analysis of them) discovering for yourself first hand the wisdom that Peter posted at 17:19, just under three hours after you first posted your diffs, and which was also the basis of my, "Wow, that's really scientific!" comment made at 9:22, over five hours before you even embarked on this experiment. You just stated in your last two posts:

• "Just an aside, but most of the 10 pages were about people."
• "I would be careful about inferring from this result anything about the rest of the encyclopedia, considering the limited subject matter of these pages and other factors."

What you have just said, in effect, is: "Yeah it appears in a touchy-feely sort of way that some have improved, while a few others have remained about the same, so 'overall, it looks like an improvement' (your exact words)" ... and then having thought about your analysis some more, you felt obligated to post further to add words to the effect: "But this doesn't really mean anything, given how skewed the sample was."

Not only was the random sample skewed, it was a totally indeterminate and meaningless sample rate for an encyclopedia that now boasts however many million articles. If you know anything about statistics and probability sampling rates then you would have realized, without embarking on your recent exercise to indulge Jimbo, that in order to make any useful inferences about a sampled population, choosing an insignificant sampling rate is basically counter-productive. It is simply going to fool you into believing that your sampled results mean much more than they really do. That is, it's not going to be predictive of anything meaningful, no matter to what kind of population (in this particular case, all currently extant Wikipedia articles) that sampling is applied, because the likely error rate is going to be way too large.

Even if we allow Jimbo the grace of a much larger than normal margin of error in his choice of sample size (since it is meant to be a quick sanity check, frequently applied, rather than a one-shot prediction of who is going to win the upcoming election), sampling only ten random articles would only have had some merit if the encyclopedia was orders of magnitude smaller than it currently is. All the necessary formulae are in that linked article should you wish to perform the math yourself. In layman's terms, the exercise you just undertook is the equivalent of trying to predict the outcome of a general election in the United States by asking only half a dozen voters how they voted as they left the polling booth. Thus it was an exercise in futility before you even began it. Which is the conclusion you came to yourself the more you thought about it afterwards.

It was also what I meant with my "that's really scientific" remark, but I can hardly expect you or anyone else to infer all of the above from that single remark. However, that observation came out in that curt manner due to some other numpty having previous played the WP:TL;DR card who I was also trying to satisfy with my post. IMO the "Teal Deer" contribution to WP guidance is the biggest cause of confrontation on Wikipedia (and thus editors leaving) because anyone trying to have intelligent open discourse in order to achieve consensus can be simply closed down by someone else, who cannot provide a convincing counter-argument for anything themselves, by their repeatedly using it to try and silence any arguments they disagree with by simply declaring the more constructive contribution to be longer than a Tweet. No wonder most differences of opinion on Wikipedia never reach consensus as they are meant to, but end up in AN/I instead. — not really here discuss 02:06, 8 September 2015 (UTC)

I also forgot to add in my comment above, for such a sampling exercise to be meaningful, in addition to the sample being large enough you usually also require a control group reference population in order to make any sense of what your sampled data is telling you. I don't wish to turn this discussion into a Math 101 course so you will have to work that one out for yourself. In this particular case, Wikipedia is your friend (but that is not always the case). However, exactly that point had already been made by Peter, as I stated above, before you ever reached the conclusion you did yourself. Please go read Peter's post again. A standard reference work, such as the Encyclopedia Britannica, that Peter refers to would be the control group or reference population in this instance. Peter explains in his post that unless you choose some sort of reference point for apples-to-apples comparison then your random sample is going to be fairly useless - such as disambiguation pages incorrectly treated as articles or bios of people or heavy metal bands of dubious notability. If Jimbo had instead randomly selected ten articles from the EB (all of which are written and peer-reviewed by subject-matter experts) and asked, "OK, how does Wikipedia treat these same ten topics?" and then compared the Wikipedia articles with the corresponding EB ones based on some well-defined and mutually agreed upon definition of what constitutes "quality" then we might actually have the makings of a useful metric. All Jimbo has done is given you a pseudo-metric that looks and feels like it is more scientific / mathematical than it really is. It's the Wikipedia equivalent of a proof that all triangles are isosceles. — not really here discuss 02:42, 8 September 2015 (UTC)

In contrast with my earlier comment below, I would like to note that this exhibits real style. I mean, we've got nearly 50Kb of edits here in the past week (plus tens of Kb on other talk pages) and in the middle one finds sighing laments about WP:TL;DR. The care and craftsmanship applied is almost touching. --JBL (talk) 02:47, 8 September 2015 (UTC)

Also of course the choice of a username echoing WP:NOTHERE is all class. --JBL (talk) 02:50, 8 September 2015 (UTC)

Thanks for violating WP:AGF by making your puerile assumptions about the origins of my user name simply to post a personal attack. My user name has absolutely nothing to do with WP:NOTHERE (bit unfortunate that) but is instead a nodding reference to Paul Lake, plus also the fact that I left Wikipedia due to encounters with people like you but would infrequently post anonymously now and then to correct things. Interestingly, what I discovered (purely by accident) while posting anonymously is that others paid much more attention to just the edits being made (simply because they had nothing else to go on in order to make entirely invalid and almost libelous assumptions about the person that made them like you just did) such that they never led to any confrontation and were rarely challenged. I only re-registered with a user id. because I had to change my static IP address of some fourteen years or so standing, by which time I had a lengthy Talk page associated with it, and it just seemed easier to redirect to a user name than another IP address. I was reluctant to do that (hence my new user name) for the reasons you just validated. — not really here discuss 03:30, 8 September 2015 (UTC)

Making the Wales method work

The wall of text by the editor who is not really here makes exactly one valid point. A sample size of 10 is not big enough to draw serious conclusions. Big deal, everybody knows that already. So the question arises "How could your make this method work to draw serious conclusions?"

• 1. you need to understand that this is just one way to evaluate the improvement of the encyclopedia. It could answer the question "Is the average article getting better?", but ignores completely how many articles there are.
• 2. It would be nice to have some sort of measurement of article quality. Many measures are presumably related to article quality, e.g. size, number of editors editing in the last month or year, number of footnotes, number of tags on the page, and perhaps even the number of see-also links, or illustrations. But these don't really get to the heart of the matter. I wouldn't completely trust the stub, start, C, B, A, FA ratings. They generally seem out of date and inconsistently applied, and probably change in meaning over time. A subjective assessment of quality might be applied (say 1-10) but care would have to be taken to make sure different reviewers rate consistently.
• 3. A sample size of 400 articles should be able to do the job, if you want to test for a 5% change in one of the variables. (Notice this doesn't depend on the number of articles in the encyclopedia, whether it's 10,000 or 10,000,000 articles)
• 4. Since you want to measure the quality of articles, ditch the disambig pages, but leave in every other type of article, e.g. lists
• 5. I can't see any reason that the random article function, which is actually pseudorandom, shouldn't be good enough. It wouldn't be good enough if for some reason it selected e.g. newer articles, or larger article more frequently. Anybody have any info on the random article function?
• 6. One fly in the ointment is that deleted articles would not be sampled, so the "average article" from 1 or 5 years ago would be biased. Presumably, if our editors believed the encyclopedia was better off without the article, then the bias would work against finding improvement. The missing deleted articles were bad, and now that they're gone the encyclopedia is improved. I doubt the %'age of deleted article is high enough to effect any results however, and you never can tell for sure whether our editors delete good articles.

So it is definitely possible to make this method work, with just a couple quibbles as is usual. I'd suggest doing it over time, say 25 article each week. Then you'd have a large enough sample size to draw conclusions every 3 months, and then 4 samples per year to see how things change over time. Anybody interested? Smallbones_(smalltalk) 04:27, 8 September 2015 (UTC)

Very interested. One twist I'd like to add. Presumably we care a lot about improvements in quality in articles that people actually read, as well as improvements in quality in articles that are more obscure. It might be useful if the random selection of articles were weighted to article popularity.--Jimbo Wales (talk) 06:32, 8 September 2015 (UTC)

By Friday I'll set up a user page/sandbox, outlining my ideas. I'll bounce the basics off a few people before then. It would be a lot of work just doing one article, but I think it could be designed that it could be done in 30 minutes per article. The subjective quality rating would be the hard part, probably a rubric (rating guide) would have to be developed. The idea of selecting the most popular articles is good, but I only have some vague ideas now on how to do it. I'd have to see at least 5 qualified reviewers sign up before I'd commit to this. It's not a 1 person job. Perhaps call it WP:Random article. One thing I'd insist on, Jimmy Wales could not be a reviewer - people might think he is biased. Sorry Jimmy, you don't get to (have to) do the hard work of rating, but your suggestions on designing the rubric, work flow, etc. would be appreciated. Smallbones_(smalltalk) 11:24, 8 September 2015 (UTC)

Regarding weighting for popularity, one way is to exclude articles with fewer than a certain number of page views in the last 90 days. For the list of 10 pages previously given, here's what their page views were in the last 90 days, along with a very brief description of each page.

1. 25,686 – person

2. 594 – mostly people disambiguation page (dab)

3. 526 – musical band

4. 4,525 – plant

5. 345 – person

6. 866 – city, stub

7. 67 – a given name of people, stub

8. 507 – people dab

9. 321 – person

10. 460 – chemical, stub

--Bob K31416 (talk) 15:18, 8 September 2015 (UTC)