User talk:Woonpton

Welcome!

"The more limited your understanding of science, the more scientists resemble masters of the occult, and the more paranormal phenomena seem likely to reflect undiscovered scientific truths." -- Wendy Kaminer

"The annals of science are littered with the names of once-celebrated scientists whose wishful thinking forced them to jump into the fringe. If their pet theories become resistant to contrary evidence, if their logic resists criticism, if their peers suspect that they have fudged results, they are expelled from the scientific community.  Pons and Fleischman were at the brink days after they went public.  Almost immediately they were told that their peak was in the wrong place.  They had to make a decision: retreat or press on despite the damaging evidence.  In the end, they leaped into the void and will never rejoin the ranks of mainstream scientists." --Charles Seife Sun in a Bottle: the strange history of fusion and the science of wishful thinking. Viking, 2008.

"The wishful thinking about fusion extends far beyond a handful of shunned individuals. Individuals ... do little damage once they are excluded from the community. The real danger comes not from these individuals but from the wishful thinking at the very core of the scientist [as human].  This, and not a threat from a handful of renegades, is what makes the dream of fusion energy so dangerous.  --also Seife

"The burden of proof, as always in science, is on those who claim extraordinary things. It is their responsibility to perform an experiment so well that it forces the scientific community to accept the results."

Thank you
Thank you. Vassyana (talk) 14:09, 8 October 2009 (UTC)


 * Where, and how, did you do it? Finell (Talk) 22:33, 8 October 2009 (UTC)


 * He must be referring to my comments on Jehochman's talk, since that's the only place I've ever had even a brief encounter with him.Woonpton (talk) 23:18, 8 October 2009 (UTC)

Hi
Back again. I was reading through the cold fusion case. How are you? This time I am really going to make an effort. The Rationalist (talk) 21:18, 9 October 2009 (UTC)


 * Hey, welcome back.Woonpton (talk) 15:22, 10 October 2009 (UTC)

Randy
I'm going to stun you further, I've never heard of Boise either! Now go and have a sit down, I realise this will have been a shock to you. :) Tim Vickers (talk) 20:41, 13 November 2009 (UTC)
 * (fanning self) Bring the smelling salts, Martha, I'm feeling faint...Woonpton (talk) 22:05, 13 November 2009 (UTC)

Randy from Boise at ANI
Would you mind putting your comments at the ANI thread about Randy from Boise in a new section? Give it a suitable title as a level 2 header. You can refer back to the previous section, and I know you edit conflicted (probably with me), but I'm trying to focus some of the discussion there into separate threads, rather than have it continue on and on. And yes, I was surprised as well. I had heard of the phrase Randy from Boise many times before, but if you search for the phrase within Wikipedia, you will find it is not that widely known. There are probably many similar phrases that others would be surprised that not everyone has heard of. Wikipedia is a very big place, after all. Carcharoth (talk) 20:46, 13 November 2009 (UTC)
 * Just ==Randy from Boise== will do the trick. It will archive separately, even if no-one replies to it. You could remove it as well. What I'm trying to avoid is a new conversation developing there, which risks the whole thread restarting. Carcharoth (talk) 21:14, 13 November 2009 (UTC)
 * As the person who started the page WP:RANDY, I don't know whether to be honored or ashamed. Short Brigade Harvester Boris (talk) 00:34, 14 November 2009 (UTC)
 * Well, since I didn't know about that page, I don't think you can take the credit or the blame for my understanding of the term, but of course I can't speak for Giano, where he picked it up from.  I did learn something from that page you started, though; I hadn't seen the original quote from Wired that the term came from; that's really great.  I've just heard the term "Randy from Boise" here and there and put together for myself who that guy is, and gone on to use the term, without knowing its etymology and historical context. So thanks, Woonpton (talk) 01:12, 14 November 2009 (UTC)
 * Just for my education, do the two words rhyme, eg (Randy, boy-se) or do you pronounce it "bO-I-se"? Tim Vickers (talk) 07:58, 14 November 2009 (UTC)
 * I've heard Boise pronounced both Boy-see and Boy-zee, stress on the first syllable (IPA: /'b$₧₩♭æ/ or /'bфӡת₤Ð/). I started to say Boy-c or Boy-z, but realized you could take the latter as Boy-zed which probably is not an accepted pronunciation. Short Brigade Harvester Boris (talk) 08:24, 14 November 2009 (UTC)
 * I'd heard of this but didn't know the history behind it either so thanks Boris. I thought that whole thread was strange but interesting in a weird kind of way. :)  Woonpton, good move removing your comment.  I really don't think we've heard the last of it though, at least that's how the ending sounded.  Take care all, -- Crohnie Gal  Talk  10:16, 14 November 2009 (UTC)

TM
What you remmoved was a direct cut and paste from the paper pg.148 which is why there was quotes around it "Direct meta-analyses showed that compared to HE, TM® did not produce significantly greater benefits on blood pressure (SBP and DBP), heart rate, TC, HDL-C, LDL-C, body weight, dietary intake, physical activity, measures of stress, anger, and self-efficacy. A subgroup analysis by study duration showed short-term significant improvement in SBP with TM®, but not over the long-term. When compared to PMR, TM® produced significantly greater benefits in SBP and DBP. When RR was compared to BF, RR did not produce significantly greater benefits on blood pressure (SBP and DBP)." Doc James (talk · contribs · email) 20:07, 11 February 2010 (UTC)


 * You mistake my edit. I removed one of the sentences listing the nonsignificant HE analyses because I didn't think they needed to be listed twice in the same paragraph, in other words I removed one of the citations to the HE analyses as being redundant.  If you'd prefer to replace the direct quote and remove the paraphrase, I have no objection, but they shouldn't both be in there.

The other sentence I removed didn't accurately reflect the subanalysis (the result of the subanalysis was nonsignificant) so I took it out. Will explain further on the article talk. Woonpton (talk) 20:15, 11 February 2010 (UTC)


 * However you left the sentence "When compared to PMR, TM® produced significantly greater benefits in SBP and DBP." Which is now out of context... So something that was baried deep in the text seems to have as great of weight as the executive summary conclusion.  Doc James  (talk · contribs · email) 20:18, 11 February 2010 (UTC)


 * Okay, I see what the problem is. If they are really referring to the HE analyses, as the context implies, then they've misrepresented their own research in the summary, since none of the subanalyses in the HE meta-analyses were significant. There was a subanalysis in a different meta-analysis (TM vs NT) but that one went the other direction: the meta-analysis was nonsignificant but when they dropped out the short term trial and did the analysis only on the two longer term studies, then there was a significant effect, but only for DBP, not SBP.  So whichever analysis they meant, the statement in the article, whether a direct quote or not, is inaccurate with respect to the actual analysis.  I agree, the PMR thing looks weird stuck on the end there.  I take it you would prefer to just reinstate the whole quoted paragraph from the results summary. Go ahead.  While I feel helpless to fix most of what's wrong with these articles, I felt fine about removing an inaccuracy, but if the inaccuracy is a direct quote from the study, then I guess there's nothing I can do.  I would guess, from reading both the abstract and this research summary, that the abstracts and summaries weren't written by the researchers themselves but by someone less familiar with the data (as a researcher who has done work for the federal government, this doesn't surprise me). Woonpton (talk) 20:39, 11 February 2010 (UTC)

Best wishes
Sorry to read at ArbCom about your recent health issues. Hope things turn out well. Fladrif (talk) 17:09, 12 March 2010 (UTC)

You're very welcome
I had previously looked, but for some reason had not, in fact, discovered until just today that the AHRQ meta-analysis had been discussed three different times before the edit-war episode in Feb 2009 when I first got involved in the substance of these articles, and that, a full year earlier, another editor had suggested perfectly appropriate language to address the meta-analysis and been run off by TimidGuy. I assumed that someone had to have brought it up before 2009 because its conclusions were widely reported in the press, but somehow I never found it in the archives until now. Now knowing that this was something already under discussion for a full year exposes the "we just wanted to discuss it first" argument for what it is. Fladrif (talk) 22:01, 18 March 2010 (UTC)

Arbitration/Requests/Case/Transcendental Meditation movement
This arbitration case has been closed and the final decision is available at the link above. The following is a summary of the remedies enacted:
 * All editors who are party to this case are instructed to read the principles, to review their own past conduct in the light of them, and if necessary to modify their future conduct to ensure full compliance with them.
 * Editors are reminded that when editing in controversial subject areas it is all the more important to comply with Wikipedia policies. In addition, editors who find it difficult to edit a particular article or topic from a neutral point of view and to adhere to other Wikipedia policies are counselled that they may sometimes need or wish to step away temporarily from that article or subject area, and to find other related but less controversial topics in which to edit.
 * Any uninvolved administrator may, in his or her own discretion, impose sanctions on any editor editing Transcendental meditation or other articles concerning Transcendental meditation and related biographies of living people, broadly defined, if, after a warning, that editor repeatedly or seriously violates the behavioural standards or editorial processes of Wikipedia in connection with these articles.
 * Uninvolved administrators are invited to monitor the articles in the area of conflict to enforce compliance by editors with, in particular, the principles outlined in this case. Enforcing administrators are instructed to focus on fresh and clear-cut matters arising after the closure of this case rather than on revisiting historical allegations.
 * From time to time, the conduct of editors within the topic may be re-appraised by any member of the Arbitration Committee and, by motion of the Arbitration Committee, further remedies may be summarily applied to specific editors who have failed to conduct themselves in an appropriate manner.
 * User:Fladrif is (i) strongly admonished for incivility, personal attacks, and assumptions of bad faith; and (ii) subject to an editing restriction for one year. Should he make any edits which are judged by an administrator to be uncivil, personal attacks, or assumptions of bad faith, he may be briefly blocked, up to a week in the event of repeated violations. After three blocks, the maximum block shall increase to one month.
 * Should any user subject to a restriction or topic ban in this case violate that restriction or ban, that user may be blocked, initially for up to one month, and then with blocks increasing in duration to a maximum of one year, with the topic ban clock restarting at the end of the block.

On behalf of the Arbitration Committee, ~ Amory ( u  •  t  •  c ) 18:31, 6 June 2010 (UTC)
 * Discuss this

Statistics
You know, at one point during the case I considered offering a free copy of How to Lie with Statistics to any member of the Committee who was seriously considering the statistical evidence provided. But then I thought it wouldn't go over very well. Anyhow, I appreciate your voice of sanity, although since you seem to have some sort of real-life expertise I feel instinctively that I should distrust you... :P MastCell Talk 22:45, 25 August 2010 (UTC)
 * And rightly so; we don't want any of that riffraff 'round here. :P Thanks, I appreciated that. Woonpton (talk)

Ping
Hi stranger, you have email. :) I hope I hear from you, -- Crohnie Gal  Talk  22:53, 28 August 2010 (UTC)

Use of the wikistalk tool for determining cooperation or collaboration among editors
[This thread was copied from William M. Connolley's talk page to give context to my answer to it (15:24 Sept 16, 2010) which was too long to post on that page but I wanted to put it somewhere, so I put it here.]

Reviewing this, Articles for deletion/Climate change exaggeration there's something very interesting about the flocking of the user accounts who voted keep here. If you run a wikistalk on the 10 or so users, you'll find a few patterns emerge. It also looks like some users have a closer working relationship than would otherwise appear. Viriditas (talk) 02:37, 8 September 2010 (UTC)


 * That debate is six months old....--*Kat* (talk) 02:42, 8 September 2010 (UTC)


 * What kind of pattern? Coke or tea? -- Petri Krohn (talk) 04:41, 8 September 2010 (UTC)


 * Here is a working link to the list I compiled. – I left out the four who had the lowes edit counts or who did not show any correlation. This left 9 so I added Mark Nutley's today's closest supporter to the list (number 10). There is a high level of correlation, even outside climate chance. The cooperation however does not seem to be politically driven. -- Petri Krohn (talk) 05:08, 8 September 2010 (UTC)
 * Um -- 1 case of 6/10 and 5 cases of 5/10? Sorry -- Random chance at work.  I went to the UT overlap for them to save time as well. 1 solitary case of 7/10 (to this page, in fact), 4 cases of 6/10.  Pretty much in line with random chance.  A lot less than found for other assortments, to be sure.  Collect (talk) 10:19, 8 September 2010 (UTC)
 * Exactly as I already said further down the page it shows nothing. I don't think there is anything particularly random about it though. If you start with editors voting in the same way on a similar article you will expect several non-random ties to come in to play. These will be so complex that any analysis based on random correlations is inherently flawed. Polargeo (talk) 10:27, 8 September 2010 (UTC)
 * Also note that I did not !vote on the article cited. Period. At all. Nada.  No connection.   Collect (talk) 10:25, 8 September 2010 (UTC)
 * I agree. Why add you into the mix? Polargeo (talk) 10:28, 8 September 2010 (UTC)
 * Petri has now run into me on (I think) three quite disparate pages, and has specifically announced a combative attitude about procedures at User_talk:Wgfinley.   Somehow I think announcing a deliberate decision to "wiki-lawyer" on that page is going to benefit him much .   Cheers. Collect (talk) 10:35, 8 September 2010 (UTC)
 * I really wish you wouldn't act like that. PK's comment is clearly intended a a humourous way of saying "discuss in detail". If everyone has to avoid anything vaguely amusing for fear of being misquoted by the Humour Police elsewhere, wiki will be the poorer William M. Connolley (talk) 10:57, 8 September 2010 (UTC)
 * I added Collect to the list because he seems to have a particular correlation with Mark Nutley. I have pumped into him and Mark Nutley in three separate context strangely related to climate change advocacy. I wanted to see if any of the others have this interest in coke, tea and caviar. The result is that they did not correlate. From this data we cannot extract signs of Organised Political Editing on Wikipedia. -- Petri Krohn (talk) 14:51, 8 September 2010 (UTC)
 * Related to Climate Change??? Um - how in heaven's name does Mass killings under Communist regimes remotely connect to climate change? Indeed, the three of us overlap on a total of two articles.    Total.   And this is some sort of major coincidence?  BTW, I do not consider Koch Industries to be especially related to climate change, and a teesny bit unrelated to mass killings as well.  Petri -- complaining about an overlap of a total of two articles is outre at best. Collect (talk) 15:06, 8 September 2010 (UTC)
 * As you see above, I am not complaining. -- Petri Krohn (talk) 15:18, 8 September 2010 (UTC)
 * Please do not describe wikistalk results using words like "cooperation" or "coordination" or phrases like "close working relationship" or "high level of correlation;" such interpretations are unwarranted. I didn't take time to count up the results, but the link Petri gave doesn't me show much of anything on a quick run through.  The vast number of overlaps are 2/10; when you have ten things, there are a lot of combinations of two, but the fact that a lot of pairs of two taken from this group of ten editors have edited articles in common doesn't mean diddly; the only thing that jumps out is that that Nyttend, Tillman and Drmies share a broad interest in geographic places in the US, although not all three of them are interested in the same places; different pairs of the three have edited different place articles. There are very few articles that even 4 or 5 of them have ever edited in common and only one I saw that 6 have edited in common (and remember, all wikistalk tells you is that these people have each edited this article one or more times in its history.  It doesn't mean that they have edited it at the same time, much less that they have edited it in a coordinated or cooperative way) and they seem to be mostly climate change related,  as you might expect from a group of editors voting on a climate-related AfD.  The degree of overlap is so small that to call it a "high level of correlation" is misleading, to say the least. Woonpton (talk) 06:14, 8 September 2010 (UTC)
 * That is not correct. The problem is with Petri Krohn's query, which failed to search for all namespaces, and left out one or two users who voted keep. The link he posted does not include these results. Viriditas (talk) 07:47, 8 September 2010 (UTC)


 * Well I don't see anything particularly astonishing in wikistalk relating to that AfDs keep voters. Just a few misguided regulars, a banned user and a sockpupetteer. Exactly what I would expect. Polargeo (talk) 08:34, 8 September 2010 (UTC)
 * After seeing that list, I edited a few of the articles on it just to add a little fat to the fire.  Fell Gleaming talk  11:04, 8 September 2010 (UTC)
 * And now a user that should be clearly topic banned then Polargeo (talk) 12:05, 8 September 2010 (UTC)
 * Why Poleargo, surely you didn't take that comment seriously? Here's a quarter; go buy yourself a sense of humor. Fell Gleaming talk  13:53, 8 September 2010 (UTC)
 * That was my sense of humour. I can't believe you missed it. Polargeo (talk) 13:59, 8 September 2010 (UTC)

I'm not used to reading wikistalk results. I can't see anything very obvious in the list (Nyttend is a new name), other than a lot of 2/10 with Nyttend and Tillman. Is that what I'm supposed to see? William M. Connolley (talk) 08:52, 8 September 2010 (UTC)
 * I don't think there is anything there at all. Polargeo (talk) 09:14, 8 September 2010 (UTC)
 * I can't say I know exactly what's there, but it does establish relationships between editors. I'm looking at this and  User:Joepnl/Vault/Climate change exaggeration, which I wasn't aware of until now. Viriditas (talk) 10:06, 8 September 2010 (UTC)
 * No still nothing. Polargeo (talk) 10:15, 8 September 2010 (UTC)

@Viriditas: I can't say I know exactly what's there, but it does establish relationships between editors. No, it does nothing of the sort. I find it rather surprising (although I guess I shouldn't be surprised any more by anything that happens on this project) that someone who is so apparently concerned about a political campaign to discredit science would be so willing to take a pseudoscientific (or maybe I should call it pseudostatistical, to be more precise) approach to establishing relationships. I suppose, then, that Viriditas endorses the use of wikistalk results introduced as evidence in the CC case to establish the existence of a coordinated pro-science "bloc" on climate change pages? Those results, while showing much more overlap than these do, were no more conclusive in showing a degree of cooperation or coordination among editors than these are, as I demonstrated on the PD talk page. And as I also pointed out on the same page, groups of SPAs who have actually been shown to have worked in concert to promote a single-purpose agenda on Wikipedia, tend to score very low on this tool, and groups of editors who are not working together, but who have been around a while and edit a lot and in a lot of different areas, tend to have a much higher degree of overlap with each other.

@Collect: Pretty much in line with random chance. As I pointed out to you elsewhere, the degree of departure from random chance is a statistical concept that must be established by a statistical test in order for the statement to have any meaning. The wikistalk tool does not provide a statistical test, so there appears to be no basis for these emphatic statements about how consistent something is with "random chance." You stated on the case pages that there was a "significant difference" between the overlap among six arbitrators (27 user talk pages in common) and the overlap among six pro-science editors on climate change articles (29 user talk pages in common). From this statement, one can assume that in your scheme of "statistical reasoning," in a group of six editors, two user talk pages in common would have to be considered a significant departure from "random chance." (this follows logically from your assertion that a difference of two--29 vs 27-- is "significant.") Extending that logic, one would have to consider that if two pages in common is a significant departure from "random chance" for a group of six editors, then one page in common among a group of ten editors (as shown on Viriditas' wikistalk link), which is much less probable by chance than a page in common among a group of six editors, could well also be a significant departure from chance. The point I'm trying to make is that this whole line of reasoning is simply without any statistical foundation, a house built on sand, and no one should draw any conclusions about degree of cooperation among editors, one way or the other, from using this tool. Woonpton (talk) 16:15, 8 September 2010 (UTC)
 * Empirical data from 300 editors taken 6 at a time establishes a pretty good basis for determining "random chance." Ask a math professor how much data is involved in 300 people taken six at a time seeking to maximize amount of overlap - and where said empirical data never got to 15 pages where all 6 people edited out of all the empirical observations (and only occurred once in all those observations), whether amounts much greater than 15 would be non-random. Collect (talk) 01:10, 9 September 2010 (UTC)
 * Well, no, I don't need to "ask a math professor" I am (or have been before I retired), a statistics professor, and you are entirely missing my point. The number of samples you analyzed to draw your conclusions becomes quite irrelevant if the conclusions are shown to be flawed, and when I showed that a group of disparate editors, obviously not editing in concert with each other, yielded an even higher number of pages in common than the group of editors that you identified as "extraordinarily cohesive," and that a group od editors known to have edited in a coordinated fashion to push a POV in a topic area scored very low on this purported measure of cooperation, the game was up.  Whatever this tool might measure, if anything, it's not cooperation or working relationships among editors.


 * I spent some time this afternoon running combinations of myself with various other editors; since I know whether I have a working relationship with these other editors,  that knowledge serves as a check to evaluate the usefulness of the tool for determining whether people have a working relationship.  What I found was interesting; I found that when I ran myself in a group of people with whom I share a scientific background and a similar view of the goals of the project and concerns about whether those goals will be achieved, but with whom I have not actually edited or even discussed specific articles, I found that the group of the four of us exhibited a high degree of "cohesiveness;" specifically there were 25 pages that the four of us had edited in common, and in "pairwise comparisons" I had 121 pages in common with Short Brigade Harvester Boris, 60 in common with MastCell, and 56 in common with Science Apologist, even though I've never edited with any of them, except for a very brief encounter with Science Apologist on What the Bleep Do We Know, in which we disagreed rather strongly. So it's definitely not cooperative or coordinated editing that this tool measures.  In this case, it seems to measure nothing but our independently similar interests and views, since for all intents and purposes we've never edited in the same topic areas.  And at the same time, I found that when I analyzed a group of people with whom I've had a fairly close working relationship editing articles together, I got a much lower number of shared pages.  And when I ran myself with people with whom I have not only never edited, but have never even encountered on discussion or policy pages, and as far as I know do not share interests or views with me (SlimVirgin, Lar) I got numbers similar to the numbers I got when I ran the group of people I did have a working relationship with. In other words, this tool does not validly serve its purported purpose of identifying working relationships among editors, since it doesn't discriminate between editors who work together and editors who don't work together.


 * Woonpton, the tool most certainly establishes relationships between editors, and this is not in any dispute. You appear to be very confused on this point.  Anyone can use this tool to create an entity-relationship model, and we used to have a tool (can't find it at the moment) that created these models based on RfA data, showing relationships by nomination.  Beyond that, you appear to enjoy engaging in fantasy and fighting with strawmen, with statements like "I suppose, then, that Viriditas endorses the use of wikistalk results introduced as evidence in the CC case to establish the existence of a coordinated pro-science "bloc" on climate change pages".  I've said nothing like that.  I've said that it establishes relationships between editors, and your reply, "it does nothing of the sort" is wrong.  Along these lines, I would encourage someone, anyone, to model the relationships of  participating editors in the current arbcom case and analyze them closely. Viriditas (talk) 21:14, 8 September 2010 (UTC)


 * No, I am not confused, in the least. This tool does not establish relationships between editors; the idea that it does is in dispute; I am disputing it, and even if all 12 million registered editors of Wikipedia were laboring under the misconception that the tool establishes relationships between editors, that wouldn't make the idea any more valid. All the tool does is indicate pages which various combinations of the group of editors in question have each edited even once, at any time in the page's history. The edits don't even have to be in the same year, much less the same discussion (if a talk page) or the same paragraph (if an article). To infer that these adjacencies establish a relationship, much less a "close working relationship" between editors is simply not warranted, and I've posted enough counterexamples to make it clear empirically that such an inference isn't warranted.


 * I'm not sure you understand what a straw man argument is; I certainly wasn't making one when I said that if you believe that wikistalk "establishes relationships" among editors, then it follows logically that you would have endorsed the wikistalk analysis that was offered in evidence on the CC case, purporting to establish statistically that "pro-science" editors were editing in a coordinated bloc. I was simply wondering whether you did or did not endorse that analysis. Woonpton (talk) 12:56, 9 September 2010 (UTC)
 * We have two users, Woonpton and Viriditas. We want to find articles where only those two editors have both edited. Here is a link to wikistalk.  Please note the drop-down list allowing the user to choose from 10 namespaces and their talk pages. (Please disregard the "summary" and "thread" namespaces for now, as that refers to the LiquidThreads MediaWiki extension currently under development.)  Default search is set to the default namespace, Main.  A check box appears to the right, allowing one to search through all namespaces.  Choose your namespace and enter up to ten users.  An explanation of the results can be found here.  As an example, I chose all namespaces and entered in two users, "Viriditas" and "Woonpton".  Here are the results.  All the tool does, is show which pages were edited by you and I.  The results are reliable, and we can compare them with another tool. Here is the same query made at Intersect Contribs.  The results are identical. Viriditas (talk) 00:26, 10 September 2010 (UTC)


 * Actually, I finally realized what happened, after running several different combinations to try to identify the source of the discrepancy. I was running the thing in two different tabs and somehow in one of the tabs the "all" button got unchecked, so it was only counting user talk pages in one tab and "all" in the other one.  So I have struck my accusation of unreliability for the tool; it does apparently reliably count what it counts, the pages that have been edited in common.   As far as I know, this thread is the first time we've ever crossed paths, so the fact that we have edited 24 pages in common says nothing at all about a working relationship between us,  which is one of the points I've been trying to make. Woonpton (talk) 07:40, 10 September 2010 (UTC)
 * The tools give us the results, but the interpretation of the relationship is up to us. Let's be precise: Based on the intersection, we've only edited one article in the main namespace together, namely Three strikes law.  Looking closer at these edits, beginning with myself, the edit history shows that I made one cleanup edit in March 2005 which brought the article into compliance with MOS:HEAD.  I made a second maintenance edit in May 2006 consisting of recent changes/watchlist vandalism patrol.  So, I made two cleanup/maintenance edits having little if any impact on the overall presentation of the content.  There's nothing interesting or controversial about these edits, and they could have been made by anyone. Now, let us look at your edit.  In May 2008, you deleted content with the edit summary, "rm unsourced (OR) graph".  You removed a graph, File:CaliforniaCrimeIndex.png that you claimed was original research.  The graph was sourced to this reference which appears to originate from data by an author named Mike Reynolds, who published the book, Three Strikes and You're Out (1996) after his daughter was murdered.  Regardless of the merits of your edits, it is clear that they were not routine cleanup or maintenance.  Therefore, the relationship between our two edits is very weak and insignificant.  If, however, I had made the same reversion of what I deemed to be OR, or made a series of edits removing content from anti-crime activist Mike Reynolds from the article, the relationship between our edits would be a bit stronger.  Looking at the editors who voted to keep at Articles for deletion/Climate change exaggeration, we see a record of strong coordination and support between each other, on many different articles, far and above routine cleanup and maintenance.  The simplest explanation for this instance of coordination is votestacking, and this observation was made by User:StuartH.  We see similar behavior on the recent arbcom case, where the same set of editors flocked to support GregJackP's disputed edits, while the overwhelming majority of uninvolved editors did not.  We also see this flocking on various AfD's and RfC's, indicating that this editing behavior cannot be classified as random, as the evidence suggests a strong working relationship over time, devoted solely to the topic of climate change, and more importantly, a working relationship that seeks to advocate for the minority POV of climate change denial. The fact that such a large number of editors are working together to promote a minority POV forces one to look closer into this phenomenon, and the relationships which emerge after analyzing the intersection of editorial contributions to common articles. Viriditas (talk) 03:20, 11 September 2010 (UTC)
 * The tools give us the results, but the interpretation of the relationship is up to us. This shows that I'm not making a dent, and what's more, that we're not even speaking the same language. As a statistician, the word "relationship" has a specific meaning to me; the wikistalk tool does not establish relationship in any sense that's meaningful to me. Broadly speaking, you could say there are three degrees of "relationship" between things.  First, there are random events, or "noise," in which there's no statistically discernable relationship between  things.  Then there are things that have been established through statistical testing to be related, associated, or correlated (in fact, the whole purpose and business of statistics is to separate relationship from randomness). Finally, determinations of causality can sometimes be made between things that are statistically associated, by the use of further careful experimental design and statistical analysis. In the quoted sentence, and indeed in this whole discussion, the three levels have been conflated to an alarming degree.


 * Let's look at my wikistalk overlap with Viriditas, since he spent so much time analyzing it. V states above that since my one edit on the three strikes law was not similar in character to his, the "relationship" established by wikistalk would have to be interpreted as a "weak" or "insignificant" relationship, but if we had made similar edits, the relationship would be stronger, even though my one edit was two years after his last of two edits to the article. Okay, I went to the supermarket today and bought Coke; by this logic, the fact that any complete stranger went to the same supermarket year before last would constitute a relationship (although a weak or "insignificant" relationship to be sure) but if on further investigation we discovered that the stranger had  bought Coke two years ago, then that would indicate a stronger relationship.   No, that's not a relationship at all, it's just two random events being misinterpreted as a relationship.  And even if a stranger went to the same supermarket at the same time I was there and bought Coke, that wouldn't be a relationship either; it's just a concurrence.  But, (keeping in mind that these are just individual cases chosen to illustrate by analogy the levels of connectedness), if my neighbor and I were planning a block party and decided that each of us would buy a case of Coke, and then we each went to the supermarket and bought the Coke, that's not only a relationship, but a causal relationship. But a computer program that counted everyone who went to the supermarket and bought Coke over a period of years would not be able to distinguish between the vast number of people coincidentally buying Coke, and me and my neighbor, or any other combination of people buying coke in a coordinated fashion, and an analyst would be making a mistake if he misinterpreted the coincidences as being associated with each other, much less as being causally related.


 * Collect at least understood that there should be some sort of statistical basis for asserting a departure from randomness before interpreting these results, when he attempted to create a sort of sampling distribution for groups of six. However, I'm puzzled by his challenge above, that I should "ask a math professor" how many ways there are to take 300 things six at a time.  If he did indeed run all the unique combinations of 300 editors taken six at a time, he would have known how many runs he'd made and could have rattled it off; it's an impressive number. I knew  that it would be a big number, 300!/6!294!, but it wasn't until last night that I sat down and calculated it, and then ran it through a computer program just to check my arithmetic; there are 962,822,846,700 of them.  If he really ran all those runs I'm impressed, but am puzzled by his assurance that he sought to "maximize the amount of overlap"--by what means?  By how he selected the sample of 300?  And if so, how did he ensure that the groups he chose to draw conclusions about were similar to the group chosen to generate the "sampling distribution"?


 * I'm also somewhat puzzled by the statement that there was only one combination in which the number of shared user talk pages was as much as 15, so he figured that anything over 15 with a group of six was significantly different from chance. The logic is good as far as it goes; by choosing a cutoff clear at the high end of the sampling distribution for the statistic, you stand a good chance of not being wrong if you say that a number higher than that is unusual (but IFF the groups of editors chosen for analysis are similar in nature to the groups of editors chosen to establish the sampling distribution).  But even if that assumption were satisfied, logistically I'm puzzled how he determined that 15 was the highest value the statistic took. The customary thing to do here would be to program your runs to compile the chosen test statistic (number of shared user talk pages among all the editors in the group) and create a sampling distribution of the statistic itself so that you would know the shape and variance of the distribution. If he didn't run a computer program that generated the distribution of the statistic, how did he establish that 15 was the highest number generated; did he look at every single one of the nearly 963 billion results to determine that?


 * At any rate, while the claim of running a large number of analyses to establish a quasi-"sampling distribution" gives a veneer of "mathiness" to the claim of "extraordinary cohesiveness" between a group of pro-science editors on the CC articles, the claim fell apart for me when I substituted non-science editors for some of the science editors in the group and got even more shared pages than Collect had found for the science editors (On the PD talk page I said 34 shared pages, but counting again I get 37, vs 29 for Collect's sample.) So I think there's a problem, a flawed assumption or premise somewhere that invalidates the conclusion. I'm not sure whether Collect has adequately demonstrated that anything over 15 shared user talk pages by six editors is (statistically) significantly different from chance (answers to the above questions could increase my degree of certainty either way, especially the implied question about how he ensured that the editors who generated the sampling statistic were similar enough to the editors chosen to draw conclusions about that one could have some confidence in the validity of the statistic.  Generally that assurance is provided by being sure that both the sample chosen to create the sampling distribution and the experimental samples are randomly selected, but in this case it would have to be a different criterion, since the experimental sample is anything but randomly chosen). But whether or not he has demonstrated that his "sampling distribution" provides a valid standard,  he has certainly failed to demonstrate that anything over 15 shared user talk pages establishes  "extraordinary cohesiveness" among these six editors, let alone a further presumption of cooperative or coordinated editing.  In other words, regardless of whether he has adequately demonstrated an "association" between the editors, he has certainly not established that his interpretation of the association, or his assignment of "cause" for the association, is correct.


 * As for Viriditas' claims above about the editors, I find nothing in his (very differently-conceptualized from Collect's)analysis of wikistalk data to support those claims; I've been over and over those results today and I just don't see anything there. There are 673 occurrences of overlap among these ten editors, including 40 pages where four of the editors overlap, 21 where five overlap, 16 where six overlap, 2 where seven overlap, 4 where eight overlap, and 1 where nine overlap (the eights and the nine are all noticeboards). Even if he had given each and every one of these 673 overlaps the same minute attention and interpretation he gave to the one mainspace overlap between him and me, there's no statistical basis to draw any conclusions from about whether these are anything but random observations, and even if if one were to accept the notion that subjective interpretations provide a valid substitute for statistics, if he were as mistaken in those interpretations as he was in this case I wouldn't give much credence to the analyses.


 * I didn't remember editing that page, at all, and was very surprised to read that I had removed a graph. It's not the sort of thing I would ordinarily do on my own; I'm not that bold an editor. So when I got some time last night,  I looked up the history of that edit.  It turns out I had been watching the NPOV noticeboard, and someone complained there about a graph they thought was misleading.  I looked at the graph and agreed that its presentation was indeed misleading; it was presented to support the idea that crime had gone down in California because of the three strikes law, but in fact, anyone with a nominal ability to read graphs could see just looking at it  that all but one of the crimes depicted in the graph had started going down before the law was passed, from two years to as long as ten years before.  I don't see anything to support Viriditas' assertion that the graph was sourced to someone named Mike Reynolds or to data supplied by Mike Reynolds; the graph was created by a Wikipedia editor and the description on the file says the graph was created to "evaluate" the claims in that blog article about the three strikes law, which you say was written by Mike Reynolds although I can't find a name on it anywhere. (The data themselves were publicly available California crime statistics, not data supplied by the author of the blog article, whoever he was.)  I would say the graph was created by the Wikipedia editor to support the claims rather than to evaluate them; if the intent had been to  evaluate the claims, the reasonable comparison would have been between CA and  states without a three strikes law; such a comparison would have shown that crime went down generally in the US during that period, in states without three strikes laws just as in states with them.  The text of the article did mention (in another section) that the reduction in crime was nationwide and not just in three-strikes states, but the text of the California section was written to suggest that crime went down in California because of the three strikes law, and this graph was drawn, and placed, to support that text, even though the graph, carefully observed, didn't even support the assertion.  There was a discussion on the article talk page which I participated in briefly after  the brief NPOV/N discussion; I said there that I thought the graph should be deleted, and someone suggested I could delete it. I said I didn't know how;  someone told me how to delete it, and I did, and have never looked at that page again til today.  I notice that the article is much more neutral now than it was that day in 2008 when I visited it, and that the graph is not there, so I assume the removal must have had consensus.  The graph was OR, or maybe more precisely SYNTH (I was new then and didn't know the difference) and didn't belong in Wikipedia. I know nothing of "anti-crime activist Mike Reynolds" and wasn't acting in opposition to him, as you seem to be suggesting above; the only place I've seen him mentioned is by you, in this thread.  My only "opposition" is to any misuse, misrepresentation, or misinterpretation of data, by anyone, and if there were editors working on that article who had an "anti-Mike Reynolds" agenda, I wouldn't know about it, or them. The implication that if there had been such editors, the similarity of my edit to one of theirs would establish, prima facie, a  relationship between us, is not credible and illustrates in a small way how misleading the overinterpretation of this tool can be.


 * Viriditas' description of what I was doing there is so far from what actually happened, it just shows how wrong a quick impression can be. No, it's not up to us to "interpret the relationship" established by wikistalk;  the wikistalk tool does not "establish relationships" and subjective interpretation of its results is not much different from reading tea leaves (and by the way, I see nothing in either Petri's or Viriditas' wikistalk data to elucidate the suggestions above that there are patterns in these data related to tea, coke, or caviar).  The establishment and interpretation of relationships rests, not on subjective interpretation, but on the strength of the data themselves and how well they stand up to the rigor of statistical testing.  There may some truth to the charge of votestacking and "close working relationship" among the editors named above, or there may not be, but wikistalk will not establish that truth, nor will selective interpretation of the results establish it, and people who point to wikistalk results as support for their suspicions about editors working as a coordinated "bloc"  are not serving the project well. Woonpton (talk) 15:24, 16 September 2010 (UTC)


 * As a side note, I'm curious what you think of this paper, which is potentially at least somewhat apposite. MastCell Talk 16:44, 16 September 2010 (UTC)


 * I wasn't able to print it out (I'm old-fashioned; in order to really absorb an article, I need a hard copy I can hold in my hand and underline and write comments in the margin and so forth) but on a quick scan from the screen, I'd say this is a very good article. It is indeed apposite, in several ways, except that the analysis is on a more sophisticated level statistically than the discussion here about the wikistalk tool.  This article is pointing out the difficulties in interpreting the results of regression-based "causal analysis" and showing how even if you subject multivariate data to this kind of "sophisticated" statistical analysis, you can't be sure that you have adequately separated influences from similarities or affinities.  The wikistalk tool of course generates nothing so sophisticated in terms of statistical analysis; it's just a counting tool, so it behooves us to be even more leery of using something primitive like this to support statements about relationships among people. Woonpton (talk) 18:52, 16 September 2010 (UTC)

Monty Hall problem
Welcome to the MHP discussion, it would be good to have a now voice in this argument which has been running for over two years now.

The article is already a FA so the argument is about rather unimportant matters of detail. I would be happy to explain what the main argument is about if you wish to join in. (Reply here) Martin Hogbin (talk) 13:58, 9 November 2010 (UTC)
 * Well, I can't think of any sane reason to jump into an argument which has been running for over two years about rather unimportant matters of detail, but thanks for the welcome. I've read enough of the interminable mediation subpage to have a pretty good idea what the points of contention are; I find them mind-numbingly tedious, trivial, pedantic, and irrelevant.  If no progress has been made in two years of discussion and two mediations, I certainly have no interest in pouring any of my time into an inevitably futile effort of trying to help.


 * As for the article being an FA, I think it's probably time for FAR. I had looked at the article a couple of years ago (I don't remember exactly when, so I can't name a specific version) and thought it an excellent article: elegant, informative, illustrated with great visual aids to understanding; it was really quite remarkable, I thought, and after that I always used that article as an example of Wikipedia at its very best. I hadn't gone back to look again but assumed it had stayed a great article, until after I started reading that mediation page recently and went back to look at the article again. I was embarrassed that I've continued to praise the article; sometime between when I saw it last and now, it has lost its soul.  It's way too long and filled with confusing and irrelevant minutiae; it's a mess.  It seems unlikely to me that a lay reader, even if they did manage to read it all the way through, would be able to make heads or tails of it. So I think it's more than a trivial argument about unimportant details; something's gone wrong with the article at its core, and it seems unlikely that anything can be done about it. Woonpton (talk) 06:21, 10 November 2010 (UTC)


 * Please pardon my intrusion. The inaccessibility of the article, that you described so well, is the very reason I stay in the fray. Against all rational explanation. Glkanter (talk) 15:42, 10 November 2010 (UTC)

Hello @Woonpton, it was nice to hear you are a statistician and your opinion about the MHP page. I'd be interested in your comments on MHP at citizendium.org, an attempt to start over with a clean slate, taking preemptive action on the bone of contention on wikipedia. See also a recent draft of a collection of "new" proofs. Richard Gill (talk) 04:46, 14 February 2011 (UTC)


 * I'd prefer not to be further drawn into this dispute, thank you. Woonpton (talk) 07:31, 14 February 2011 (UTC)


 * Very wise! I am no longer disputing, I am writing reliable sources (that's my job), since the existing ones are poor quality. Hence my sincere interest to know your reactions to new articles on MHP in other locations. In an attempt to get it right at the core. The wikipedia article is bogged down on a stupid controversy which will interest almost no readers while in the meantime life goes on and there are several new insights out there which add to the richness of the topic. Richard Gill (talk) 08:20, 14 February 2011 (UTC)
 * Trust me, you do not want to know what I think, but luckily for both of us, what I think is not relevant here at Wikipedia, where we are required to rely on existing secondary sources for our summary of a topic. We don't publish new research; if your formulations are found to be useful by other scholars in the field they will be incorporated into secondary sources (such as literature reviews or academic books devoted to the topic) by independent scholars, and when that happens, they can be incorporated into the Wikipedia article. That's how Wikipedia works.  Woonpton (talk) 18:14, 14 February 2011 (UTC)
 * Speaking of reviews, I recently had an offer from the editor of a respectable-if-not-quite-first-tier, peer-reviewed, MEDLINE-indexed journal to write a review article on the topic of my choosing. Sadly, the first 3 or 4 review topics that came to mind were those where, on Wikipedia, ignorance is currently prevailing over reason because of a lack of suitable secondary-source reviews. MastCell Talk 04:39, 15 February 2011 (UTC)
 * Well, then, get those reviews written! :-) In the case of the Monty Hall problem, it seems from the outside (I haven't edited the article or talk, but have been watching the dispute from a cautious distance, > 10 foot pole) that there are adequate secondary sources that are not being put to good use in the article, while primary sources are being endlessly debated and second-guessed as to what the source really meant,  or (more often, even) sources ignored altogether in favor of editors coming up with their own original research and synthesis; most of the disputes are about individual editors' ideas of what the Monty Hall problem is and how the solution should be argued, completely independent of sources.  It's a mess. Woonpton (talk) 05:06, 15 February 2011 (UTC)
 * Yes, it sounds pretty bad. About reviews, I think I don't have what it takes. I would never, ever feel comfortable writing a review about a topic outside my field of expertise. (I have a hard enough time feeling fully on top of my own field). But to be a truly successful charlatan dissident from scientific orthodoxy, one needs to be able to leverage one's expertise in field A to make confident pronouncements about field B, despite the fact that one knows less than nothing about field B. Most forms of arrant nonsense feature at least one individual leveraging their credentials in such a fashion. MastCell Talk 05:13, 15 February 2011 (UTC)
 * I've been thinking about your dilemma in light of an only tangentially-related phenomenon I'm seeing more and more of lately in several different topic areas, of Wikipedia editors writing and publishing papers so the papers can be used as reliable sources to advance a POV in the Wikipedia article. It makes me think of the Escher drawing with the hand drawing the hand.  At what point do we have to give up and say the whole system of academic work has broken down and the agenda-driven Wikipedia model has won over honest scholarship? Does that worry you?  It worries me, but maybe it's just time for a long walk.  Regards, Woonpton (talk) 21:50, 16 February 2011 (UTC)


 * Don't worry, I don't make my living by writing wikipedia pages, I make my living by writing reliable (indeed, highly cited!) sources in mathematics and statistics. And I do know how (and why) wikipedia works, and have no wish to change that or to break the rules. That's why I took the trouble to write up some "Truths", of similar startling nature to something like "6 is not a prime number because it is divisible by 2", and get them published in peer reviewed (scientific) literature. At least no stupid editor can now prevent a less stupid editor from using this "Truth" in the future, on the pretext that it is not written up explicitly in some "reliable source". Too bad that I'm now disqualified, by my very expertise, for editing an article in my own field. Wikipedia will have to wait ten years or so for new (to the literature, not new at all to sensible people)  insights, and for reconciliation of the squabbling school children. Richard Gill (talk) 15:32, 15 February 2011 (UTC)

I don't recall seeing the data you said you were tabulating. Did you put this someplace where I might have missed it? I think it might be quite helpful now that Elen has drafted a proposed decision. -- Rick Block (talk) 03:57, 14 March 2011 (UTC)

Comment and thanks
I was surprised to see old Monty up for Arbitration, likewise amazed by the trivia under contention, and unreasonably cheered to see your occasional good-humored comment in the mix. Thanks for that :) I hadn't seen the article when it was really good; FAR with a historical link isn't a bad idea for attracting new eyes.

As to "editors writing and publishing papers so the papers can be used as reliable sources to advance a POV in the Wikipedia article" -- this happens within academia all the time, without the bit about Wikipedia... counterfactual claims and theories can survive for a generation on the industry and focus of a group of experts who believe in them, or absent belief have hinged their careers on them. But at least it lets us offload the question of formal peer review onto a scientific publishing system that has had a few centuries to reach a productive equilibrium. –SJ + 06:53, 2 March 2011 (UTC)

Anyone home?
Are you still around, or reading this site at all? I was looking through some old discussions and smiling at some of your commentary, and I was wondering if you were gone for good or just laying low. Cheers. MastCell Talk 19:26, 3 January 2012 (UTC)
 * Ditto. -- Rick Block (talk) 04:59, 16 May 2012 (UTC)

You are missed
When you stopped editing, Wikipedia got a little less intelligent and a little less fun. The fact that Wikipedia can't retain smart, incisive, clever people like you is the reason I think the project is ultimately doomed. Anyhow, hope you're out there and doing well. Cheers. MastCell Talk 03:33, 11 September 2013 (UTC)


 * You are missed! :-)  –  SJ  +  08:22, 4 January 2014 (UTC)


 * I just had cause to look at some very old discussions, and was reminded of your contributions. You are remembered, even after a long absence, and I too miss your contributions.  And, for the record, you were right and I was wrong – Abd was a much bigger problem than I recognised at the time.  Hopefully you helped me to learn, and I am thankful.  EdChem (talk) 09:53, 16 December 2016 (UTC)