Wikipedia talk:WikiProject Statistics/Archive 1

From Wikipedia, the free encyclopedia


Contents

[edit] Lists

The list of statistical topics is not in its present form very sophisticated. Lists are far more versatile the categories, but this list doesn't take advantage of all of that. Look at list of mathematics articles and lists of mathematics topics (the latter is a magnificent thing!). I'll have more to say on this later. Michael Hardy (talk) 19:25, 16 March 2008 (UTC)

[edit] Proposal

I've archived the discussion of the proposal of this WikiProject that took place on Wikipedia:WikiProject Council/Proposals at Wikipedia:WikiProject Statistics/Proposal. Please do not edit that page but feel free to continue discussion here of any of the points that came up. Qwfp (talk) 23:37, 16 March 2008 (UTC)

[edit] Proposed merge of Kernel (statistics) and Kernel smoother

It has been proposed to merge the articles Kernel (statistics) and Kernel smoother. Weigh in on the discussion at Talk:Kernel (statistics)#Proposed merge of Kernel (statistics) and Kernel smoother.  --Lambiam 19:35, 17 March 2008 (UTC)

[edit] Project banner on talk pages

It's up to you guys whether you'd like to piggyback on the {{maths rating}} template or create your own. Either way, if you need help with setting up templates, customizing the assessment system, or need a bot to tag a bunch of talk pages, I should be able to lend a hand. — Carl (CBM · talk) 01:06, 17 March 2008 (UTC)

Thanks very much for your offering your help Carl. I think we might want to take you up on it at some later point but as you say, we need to discuss the options first. I have created a WikiProject Statistics talk page banner {{WPStatistics}} but so far I've only put it on a dozen or so talk pages (including some of the most viewed statistics articles), and always in addition to {{maths rating}}. Although that could be done by a bot, the number of pages in "field=probability and statistics" (230ish) is not that vast especially if several of us work on it and some human judgement is useful e.g. to decide if use of {{WikiProjectBannerShell}} would be a good idea. There may also be articles in "field=probability and statistics" on parts of probability theory field that are not relevant to statistics that we won't want to tag as being in the scope of WikiProject Statistics. And Category:Statistics would need a lot of clearing out and reorganising before we could think of tagging articles in that category or some of its sub-categories using a bot.
I was thinking of keeping the assessment side of things within the existing framework of WP:WikiProject Mathematics/Wikipedia 1.0, but then people pointed out (see discussion around the proposal archived at WP:WikiProject Statistics/Proposal) that there are articles on the non-mathematical side of statistics that would sensibly be considered to belong in statistics but not mathematics, so logically we might want a separate assessment system I guess. But I think it should be a higher priority to re-assess the articles currently in the "probability and statistics" field and make sure all the frequently viewed statistics articles are rated and then start to act on that info by starting to improve the high-importance but low-quality articles. If a few statistical but not-strictly-mathematical articles are included in the {{maths rating}} system I don't think it really matters, as long as WPM doesn't mind. Qwfp (talk) 11:43, 17 March 2008 (UTC)

I've been puting the WPStatistics above the WPMathematics tag when the article is clearly on statistics; someone else has been doing the opposite. Should there be some convention on this? Michael Hardy (talk) 19:35, 19 March 2008 (UTC)

Hmm, good question Michael. I've been putting it below but I never really thought about it — guess I was just being chronological. On second thoughts it seems to make sense that if it's definitely an article about statistics the {{WPStatistics}} tag can go first and I'll change my practice, though it doesn't seem worth going back to change the ones I've already done. But if it's more on the border (perhaps e.g. all the articles on probability distributions) then I think I'd stick with being chronological. Not sure it's too important if different people do it differently — more important that it's done, at least for frequently-viewed articles. Qwfp (talk) 20:20, 19 March 2008 (UTC)

[edit] missing statistics/statistician articles

Now that we have our own wikiproject (thanks for putting this together!), should we have maintain a list of such articles? I mean besides Wikipedia:Requested articles/Mathematics#Statistics? Thanks Btyner (talk) 02:23, 20 March 2008 (UTC)

Do you mean Wikipedia:Requested articles/Statistics? Or did you want to put it somewhere else? Michael Hardy (talk) 03:38, 20 March 2008 (UTC)
Maybe we could explicitly list least a few of them under Requests in WP:WikiProject Statistics#Article-related tasks? In particular any you think should be created but don't know and can't find out enough about to just go ahead yourself and create a stub? Qwfp (talk) 05:58, 20 March 2008 (UTC)
Yes, Wikipedia:Requested articles/Statistics sounds good to me. I'm also wondering whether it would be justifiable to set up something like User:Mathbot/Most wanted redlinks but for stats. Thanks Btyner (talk) 19:06, 21 March 2008 (UTC)

Can we set up something like the page at Wikipedia:WikiProject Mathematics/Current activity? Michael Hardy (talk) 16:50, 20 March 2008 (UTC)

Personally I can't see much advantage of Wikipedia:Requested articles/Statistics over Wikipedia:Requested articles/Mathematics#Statistics. I'm not sure statistics gets enough article requests to justify an entire subpage of Wikipedia:Requested articles to itself. Yes would be nice to have an equivalent of Wikipedia:WikiProject Mathematics/Current activity but that looks like it's taken quite a bit of work by some bot operator, so we might need to see if we can persuade a bot operator to set up the something similar for statistics. I know nothing about bot operation myself.

Anyway, I'm happy for someone else to take the lead on this — I don't have any particular role, rights or responsibilities as the project proposer any different from the rest of the members. I have no intention of spending as much time on WikiProject Statistics on an ongoing basis as I did for a few days when setting it up. I'm going on a short break over the next few days and likely to be fairly busy in real life when I get back — I'll keep an interested eye on how it's going but I certainly don't feel I own WikiProject Statistics in any way and I'm very happy for other members to be bold and take the initiative. Regards, Qwfp (talk) 21:53, 21 March 2008 (UTC)

I guess what I'm trying to get at would be a list of stats-related redlinks; for example, What about something in the style of the lists linked from Wikipedia:Missing science topics#Mathematics? Btyner (talk) 14:37, 22 March 2008 (UTC)

[edit] merge probability and probability theory?

I was just looking over the top priority articles in the project and wondered why we have an article on probability and another article on probability theory. Just from a top level view, does this seem like a good idea to others? If you think it's a good idea, what is the difference between the two? Pdbailey (talk) 21:42, 22 March 2008 (UTC)

I suspect the reason we have two separate articles is that merging them is a lot of work, since they're necessarily fairly long. I'm inclined to say that should get merged. Michael Hardy (talk) 01:05, 23 March 2008 (UTC)
I think Probability is meant to be more accessible, and also the article that deals with non-mathematical aspects such as applications, leaving the mathematically sophisticated stuff for probability theory. If they are to be merged, which I'm not convinced yet is a good idea, then we must make sure we save scary symbols like Ω for the end.  --Lambiam 14:16, 23 March 2008 (UTC)
I tend to agree it makes practical sense to continue having one article for a general audience and the other for a more advanced audience. Very similarly, there are separate Calculus and Real analysis articles, even though both articles note that they're basically the same subject covered at different levels of sophistication. (Physics does something similar in the various Mechanics articles). Best, --Shirahadasha (talk) 01:07, 24 March 2008 (UTC)
Interesting take, and I can't say that I would argue strongly against it. I have three questions (1) can we agree that the two should at least state this dichotomy explicitly? (2) do you really think that the two articles are at the levels you suggest? (3) what importance rating should they have given this separation? Pdbailey (talk) 02:40, 24 March 2008 (UTC)
Both articles in the Entropy/Introduction to entropy pair articulate the dichotomy very explicitly (For a generally accessible and less technical introduction to the topic, see Introduction to entropy). This seems a good idea. Best --Shirahadasha (talk) 22:13, 24 March 2008 (UTC)

[edit] Entropy (disambiguation)

I could use some more eyeballs on this page.

In my view, to help people get to the article they want most quickly, it is helpful to include structure in the page to group together meanings primarily related to Entropy in a thermodynamic sense, and those primarily related to Entropy in an Information Theory sense. However, because there is no provision for this is the WP:DAB guidelines, various editors specialising in disambiguation (who may know rather more about disambiguation than they do about entropy), would prefer to see all the links muddled together in a single (IMO much harder to navigate) long alphabetical list. Cf this diff: [1].

Since dab pages are supposed to help readers who do know something about the subject find the article they want, I'd greatly appreciate if members of this project could look at the two versions above, and then leave their thoughts on the talk page.

Thanks, Jheald (talk) 23:23, 23 March 2008 (UTC)

Took a while to look into this and it appears to be a tar pit of an edit war over useful vs following some interpretation of the rules. Nevertheless, I don't see what we can add since there already appears to be someone helping two editors figure this out. I think this is just a distraction to WP:Statistics. Pdbailey (talk) 02:37, 24 March 2008 (UTC)
Thanks for taking a look. But with respect, one of the most valuable aspects of a Wikiproject is to be somewhere where people with relevant knowledge can come together and seek self-defence against random rules-pushers, POV artists, and other threats.
Unfortunately, the row has become even more of (in your very apt phrase) a "tar pit", with the page currently locked down on the "wrong version" despite the efforts of various Maths editors...
One of the points at issue is whether the meaning of the word Entropy in information theory (and related links) should be considered a "primary meaning" of the word Entropy, as important as thermodynamic Entropy -- or whether it should be listed as an also-ran. It would therefore be useful to have the input of those who come to Entropy from the statistics direction, who I imagine would see the information-theory meaning of Entropy as quite as fundamental as the physics meaning, and deserving co-equal billing; to balance the input of those who may only have heard about entropy in a physical context.
I don't apologise for bringing the matter here. The only way to try to get a good outcome when something like this happens is to get enough people who do understand a subject to come to a page; otherwise everything gets railroaded by the views of those who don't. Jheald (talk) 22:04, 24 March 2008 (UTC)

[edit] Archiving discussion

I notice the new automatic archiving applied to this discussion page: (i) I think 7 days may be too short a time period (thats my interpretation of what's going on); (ii) Is there a good way of putting a link(s) to the archived stuff on the duscussion page? Melcombe (talk) 09:45, 26 March 2008 (UTC)

I half set that up last night as a bit of an experiment as it looks like this page is going to host enough discussion to need some sort of archiving, which is good news, and archiving by hand is a fairly tedious business (i've done it a couple of times). I hadn't realised that anything would actually get archived so soon. Please accept my apologies for not discussing it here first. I think you're right that 7 days is too short — lets try 28 days. I've added an archive box. In case anyone wants to change things, I've followed the instructions at User:MiszaBot/Archive_HowTo. MiszaBot archives WT:WPM and seems to do a good job. Qwfp (talk) 10:14, 26 March 2008 (UTC)
I've just undone the actual archiving and commented out the {{archive box}} so everything's back here for now, but discussions older than 28 days will get archived. Does that sound reasonable? Qwfp (talk) 10:44, 26 March 2008 (UTC)
Please don't think think that I was complaining. It's good that someone is prepared to do something. I think 28 days would be OK, but maybe start at around 2 months to allow for longer vacations? I did wonder whether it would be possible to mark certain threads as not to be archived automatically, but perhaps this would be better handled by putting the important points in the main article for the Project. As an example of the type of thing I mean, consider the stuff under "Lists" above. I have taken the liberty of copying this contribution by Michael Hardy into the main article. Melcombe (talk) 11:35, 26 March 2008 (UTC)

[edit] Other tags

I have come across this tag, {{Expert-subject|Statistics}}, which is not mentioned in the project article. Is this an advised thing to use? I have made limited use of [[Category:Statistics articles needing expert attention]] which seems not to be so blatant about complaining about an articles contents. Melcombe (talk) 12:15, 26 March 2008 (UTC)

{{expert-subject}} can be used with any WikiProject as argument. When there's a suitable WikiProject it's preferable to the generic {{expert}} tag. I wouldn't expect {{expert-subject|Statistics}} to be used by participants in WikiProject Statistics very often — it's more for use by others seeking expert statistical input and members of WikiProject Expert Request Sorting. I agree it should be mentioned on WikiProject Statistics page somewhere though.
Use of this tag places an article in Category:Statistics articles needing expert attention but an article can be placed in the category without using the tag. I'd agree that these complaining "hat" tags can be overused (see User:Shanes/Why tags are evil and its talk page). I guess the question is whether an article is inaccurate or potentially misleading and the reader needs to be warned, which is probably fairly rare.
Several of the articles in Category:Statistics articles needing expert attention got there because I went through Category:Mathematics articles needing expert attention and reallocated those that seemed clearly to fall within statistics. But I didn't consider at the time whether they deserved to be in the category or whether the tag was too blatant, so it may be useful for someone to revisit that. Qwfp (talk) 14:56, 26 March 2008 (UTC)

[edit] Conventions

I know that there is the page Notation in probability and statistics covering some stuff about notation, but is it worth developing more on other aspects of conventional usage in statistics. Where I work, we generally use capital letters for all distributions as in "Normal distribution" as oppsed to "normal" distribution, similarly for Exponential. I think that this is partly on the grounds that these are specific names for specific distributions and partly on the grounds that it makes it easier to spot where important assuptions are being stated. I note that some changes have/are being made to try to enforce the other convention. I am not really against this, but it would be good to have something somewhere, specific to statistics (or probability and statistics), that would indicate some common conventions towards which things can be moved. There must be other points also ... so some form of guideline? Melcombe (talk) 11:06, 28 March 2008 (UTC)

I'd agree that it might be useful to develop and document some conventions. There's quite a bit already at Wikipedia:WikiProject_Mathematics#Conventions and it might be more helpful to add to those pages than start a new one specifically for statistics, though I'm not sure either way.
I don't think i've come across the convention of initial capitals for names of distributions though. Wikipedia:Manual of Style (capital letters) doesn't seem to cover this explicitly but does give the general guide that "unnecessary capitalization is avoided". Another general rule is that capitals are used for words derived from proper names and the distribution wasn't named after Henry Normal! I don't much like the name "normal distribution" and sometimes prefer to use "Gaussian distribution" instead myself but there's no chance that could become a general convention. Qwfp (talk) 13:41, 28 March 2008 (UTC)

[edit] I am accused of vandalism to design of experiments

Two very very confused people have been disputing the content I put at design of experiments. I have patiently explained to them why they are wrong at talk:design of experiments. If they would just go to the library and look up the literature that is referenced in the article, they would see that it backs me up. I have a Ph.D. in statistics and I care about the subject. Anyone who cares would have read my comments on that talk page and attempted to digest what they say. But someone came along and accused me of "vandalism" to the page and reinstated the mathematically erroneous edits. Anyone who would just check the math would see the point. Can others take a look and explain to these people that I'm not just some isolated crackpot? Michael Hardy (talk) 16:13, 28 March 2008 (UTC)

[edit] Redescending M-estimator

Redescending M-estimator is very clearly in need of attention. Michael Hardy (talk) 12:55, 3 April 2008 (UTC)

It seems necessary to think about the article M-estimator as well because, while a definition of the phi-function can be found there, it is not particularly prominent. But a proper definition is needed somewhere. Is the term phi-estimator extensively used? Perhaps there needs to be an article for that, before having redescending version. The present text seems to imply that it is the phi-function that redescends, not the M-function ... the objective function for minimisation (the M- or pho- function) would flatten-off for high values.
Melcombe (talk) 15:24, 3 April 2008 (UTC)
Just a suggestion: I would support merging the redescending material into the main M estimator article. Having the redescending article seems quite too specific. Baccyak4H (Yak!) 02:16, 4 April 2008 (UTC)
I see that the article Robust statistics has quite a bit about (and links to) M-estimator... it mentions "redescending ψ functions" but doesn't (yet) link to Redescending M-estimator. Melcombe (talk) 12:16, 7 April 2008 (UTC)

[edit] Margin of error

Margin of error was promoted to featured-article status during the 2004 election campaign. Then it was demoted on 3 March 2007. Now that we're heading into another campaign, should we see if we can get it promoted again? And maybe linked to from the main page at some point int he late summer or early fall? Michael Hardy (talk) 16:35, 5 April 2008 (UTC)