Wikipedia:WikiProject Statistics/Proposal

A copy of two subsections of the 16 March 2008 version of WikiProject Council/Proposals

Projects

Statistics

Please see below at #Probability and statistics under "Task forces/Work groups" where discussion of a proposal for a probability and statistics task force/work group of WikiProject Mathematics has clear support but has expanded to discuss whether it would be more appropriate to create a separate WikiProject Statistics. Qwfp (talk) 14:21, 14 March 2008 (UTC)

Task forces/Work groups

Probability and statistics

Description

A subproject of Wikiproject Mathematics to improve coverage and content of articles on probability theory and statistics, the history of probability and statistics, and biographies of probabalists and statisticians. One step towards article improvement will be to revisit the assessment of the probability and statistics articles at WP:WikiProject Mathematics/Wikipedia 1.0/Probability and statistics and consider rating more articles. Also improve navigation by reorganizing sub-categories within category:probability and category:statistics, reviewing existing navigation templates such as {{statistics}} or constructing new ones, and perhaps constructing a Portal.

Interested Wikipedians (please add your name)

1. Qwfp (talk) 10:21, 12 March 2008 (UTC)
2. Salix alba (talk) 18:43, 12 March 2008 (UTC)
3. Pdbailey (talk) 00:31, 13 March 2008 (UTC)
4. Baccyak4H (Yak!) 02:31, 13 March 2008 (UTC)
5. Tayste (talk - contrib) 05:09, 13 March 2008 (UTC)
6. Petergans (talk) 07:55, 13 March 2008 (UTC)
7. Johnbibby (talk) 15:37, 13 March 2008 (UTC) (I'm happy to be involved in thus but am unlikely to have more than 1 hour per week)
8. Michael Hardy (talk) 20:12, 13 March 2008 (UTC)
9. Shirahadasha (talk) 13:03, 14 March 2008 (UTC)
10. Melcombe (talk) 16:59, 14 March 2008 (UTC)
11. Btyner (talk) 02:20, 15 March 2008 (UTC)

Comments

If we can get enough people interested, this could be a worthwhile project, prob and stats do seem to lag behind most other maths articles, but seem to have high hit counts. --Salix alba (talk) 18:43, 12 March 2008 (UTC)

This is a badly needed subcategory, being a part of mathmatics is not good for these articles, and having a core group will help a lot. I hope this works. Pdbailey (talk) 00:31, 13 March 2008 (UTC)

I would see my role as providing comments from a scientist's viewpoint. Petergans (talk) 07:55, 13 March 2008 (UTC)

-- -- --

I agree with some views expressed here and elsewhere that either probability and statistics or statistics alone could be a stand-alone WikiProject, and that might be more logical in some respects. I've suggested starting off as a sub-project of Wikiproject Mathematics mainly to reduce the administrative overhead, as explained at WikiProject Council/Guide/Task forces. In particular, WPM has an excellent assessment framework and it would be quite a bit of work to separate out probability and/or statistics, with no clear advantage. Also it avoids starting off with too many admin and re-organisation tasks to do that might drain our initial enthusiasm (I know they would quickly drain mine if they got left to me as project creator). I'd rather keep creation of a separate WikiProject as a longer-term option.

Possibly WikiProject Statistics could be a redirect to the sub-page for this sub-project for the time being, WikiProject Council permitting? For one thing, that would mean that articles in need of statistical editing expertise could be tagged with {{expert-subject|Statistics}} to get our attention. It would also increase our visibility a bit. WikiProject Probability already exists but has been all but dormant for 18 months, so that could have a variant on the {{inactive}} tag at the top with a prominent link to this sub-project.

Creating separate sub-projects for probability and for statistics within WPM would conflict with the current structure of their assessment scheme so would again involve tedious admin work to split it. Also it's often difficult to draw the line between probability and statistics and I don't think it would be useful to try — for example many of the articles in Category:Probability are probability distributions, many of which are of as much or more interest to statisticians than probabalists.

Qwfp (talk) 13:00, 13 March 2008 (UTC)

Permission for creating a redirect is not necessary, but I think I can say that you have that permission anyway. From the looks of things, right now you have enough people to start a task force/work group on the subject. If I can be of any help in setting up the group, let me know. John Carter (talk) 13:26, 13 March 2008 (UTC)

Thanks John. I'll leave it until the weekend to set this up as I've just posted brief notes at Talk:Statistics and Talk:Probability theory and a couple of other places so I'd like to give people a chance to notice them and comment here, for or against. I thought it would probably be ok to create that redirect, but it's good to be given the nod. To be honest I'm amazed anyone not interested in joining the sub-project read that far down my comment — maybe the redlink caught your eye. I try only to create ones that I think have a chance of turning blue! Qwfp (talk) 14:02, 13 March 2008 (UTC)

The {{Maths rating}} template has a field=probability and statistics flag. It should be a fairly easy job to change what's displayed to indicate the prob/stats project/sub-project in the template when this field is set. That should save a fair bit of work. --Salix alba (talk) 13:44, 13 March 2008 (UTC)

I'd lean toward WikiProject Statistics as a separate project. Michael Hardy (talk) 20:13, 13 March 2008 (UTC)

I agree - there's a lot more to Statistics than just the maths. Tayste (talk - contrib) 06:02, 14 March 2008 (UTC)

Hmm, I'm feeling much less sure about this now, I might even be starting to come round... I admit I only studied the Task Force Guide which it now seems to me concentrates on the advantages of these over Projects and says little on the disadvantages. Looking belatedly at the WikiProject Guide I can see I overlooked an important advantage to full Project status : a Project has its own banner to put on Talk pages and that is the most effective way to steadily recruit new members, which is essential to long-term viability. I know that as one of the substantial minority of statisticians without a maths degree I dithered over whether to join WPM whereas I would have joined a WikiProject Statistics without hesitation. I wasn't around at the time, but a bit of wikiarcheology has made me suspect that WP:WikiProject Probability's sadly comatose state may be partly attributable to the fact that its banner was created too late in the day and put on precisely one talk page (can that really be right? Is there some hidden history behind that I haven't dug up?).

Also as I've only joined one WikiProject I hadn't appreciated the degree of overlap and even nesting of Wikiprojects that's common. However I still think it would be best and by far simplest to stick with WikiProject Mathematics' established and comprehensive assessment procedure. Neither do I fancy adding a banner by hand to 234 talk pages, so if we decide to go down that route I'll leave that to one or more volunteers with experience in using editing tools that could make that less tedious.

I'd welcome a bit more reasoned debate (as opposed to !voting) as I'm now genuinely undecided. I'll post another brief note on WT:WPM to alert them to the expanded scope of this discussion.

Another thing: if we do go for a full project, "WikiProject Statistics" or "WikiProject Probability and Statistics" (or "WikiProject Statistics and Probability")?

Final thought (for now) re Tayste's specific comment above - are there articles any of you can suggest that you would tag as belonging to WikiProject Statistics but not WPM ?

Qwfp (talk) 08:34, 14 March 2008 (UTC)

One type of thing here would be the frequentist-Bayesian comparison stuff and similarly the stuff which is either strongly pro-frequentist or pro-Bayesian which goes beyond simply setting out the underlying maths of these approachs (some of this comment may relate to parts of existing articles that would benefit from having such stuff removed). Other possibilities would be some aspects of the design of experiments, possibly the whole of the topic of design of questionaires and some parts of doing population surveys. Melcombe (talk) 16:59, 14 March 2008 (UTC)

...and perhaps I can add : graphical methods of statistical analysis; exploratory data analysis; good design of graphical results. Melcombe (talk) 17:19, 14 March 2008 (UTC)

I was thinking of some of the more practical aspects of statistics, e.g. Misuse of statistics. There are some aspects of statistics that overlap as much with other disciplines as they do with maths, e.g. how it's used in various social/soft sciences. Another example is statistical computing (e.g. numerical optimisation) which is equally a computer science area. Tayste (talk - contrib) 21:36, 14 March 2008 (UTC)

I would strongly agree with a separate WikiProject, not a branch of WP:MATH. Mathematics works by deducing specific conclusions from general axioms. Science works by inducing general laws from empirical observations. Statistics as a field has some elements of both. Judging statistics topics by mathematical criteria would be inappropriate because topics of critical importance in statistics as a profession are often scientific, not mathematical, in nature. --Shirahadasha (talk) 13:03, 14 March 2008 (UTC)

I would suggest WikiProject Statistics. Pure probability articles can be addressed by both WikiProjects. --Shirahadasha (talk) 13:04, 14 March 2008 (UTC)

It is not generally true that math works by deducing theorems from axioms. That's a caricature. Some people prefer to put the finished product in that form, but if you look at how the finished product got finished in the first place, you see something different. And the question of what is a good axiom system, worthy of attention, and what are good definitions, occupies the attention of all mathematicians and is not a problem solved by deducing the answers from axioms.

On the other question: what is there in statistics that is not just math, how 'bout this: think of the various ways of making the Behrens-Fisher problem into a precisely defined mathematical problem, and then think of the philosophical reasons adduced for preferring one or another of them. The latter are not mathematics. Michael Hardy (talk) 20:44, 14 March 2008 (UTC)

To answer a question above, I would definitely tag Data Monitoring Committees with WikiProject Statistics but not WikiProject Mathematics. A few other examples that come to mind would be Statistical ethics, Loss function, bias (statistics), American Statistical Association, Study statistician, Non-inferiority trial, and W. Edwards Deming. Best, --Shirahadasha (talk) 06:06, 16 March 2008 (UTC)

Discussion here closed: Consensus favoured creation of WikiProject Statistics so I have now created it. Please feel more than welcome to continue the discussion on its talk page. Many thanks for everyone's input, Qwfp (talk) 16:25, 16 March 2008 (UTC)