12
$\begingroup$

I have been an active participant in the site for a while now, and I never really figured out how to deal with issues that keep recurring (nor have I figured out "Community" topics). I would appreciate some pointers. Is there a place for statistics FAQs and their ever-growing answers? Thanks.

| |
$\endgroup$
  • $\begingroup$ The FAQs are the frequently answered or asked questions, which means most. A negative principle here is not being Wikipedia, but only for other reasons are the better answers being edited to approach something definitive. The more solid questions are usually better than W., especially when they capture a range of views on contested or complicated issues. I'd phrase this as centred on how do we persuade new and/or low-reputation visitors to read the help and use the resources (except that I don't know any data on visits to existing pages that resulted in happy visitors and no new question). $\endgroup$ – Nick Cox Mar 25 '14 at 16:16
  • 2
    $\begingroup$ I wish I had an answer to the part about FAQs, because I spend too much time every day looking up duplicates to new questions. The Community Wiki question is addressed on Meta SO at meta.stackexchange.com/questions/11740/…. $\endgroup$ – whuber Mar 25 '14 at 16:30
  • 1
    $\begingroup$ We don't seem to have a [faq] tag like they do on SO (eg, r-faq). I wonder if we should? Note that SO's are language specific, whereas one here would end up filled w/ heterogeneous material on r, hypothesis testing, regression, clustering, pca, etc. A real question is whether it would actually help make material better organized & more accessible. $\endgroup$ – gung - Reinstate Monica Mar 25 '14 at 16:36
  • $\begingroup$ I'm a bit disappointed to learn this. We need a permanent place for FAQs and need to have a single place to see what is there, so that we refer "new" questions to that place instead of duplicating answers. I don't think a question being frequently asked helps organize the site, because it can be frequently asked in duplicate and under other guises. $\endgroup$ – Frank Harrell Mar 25 '14 at 17:13
  • 2
    $\begingroup$ This is becoming reminiscent of the recent thread meta.stats.stackexchange.com/questions/1982/… which pivoted around the obvious "Why no X? Because no one wrote one". Somebody so inclined could write a FAQ that fit our format; whether it was labelled as such would be secondary. $\endgroup$ – Nick Cox Mar 25 '14 at 21:01
  • $\begingroup$ How would it be formatted Nick, as a new question titled "Statistics FAQ" that is heavily monitored and curated by the community? $\endgroup$ – Frank Harrell Mar 25 '14 at 21:07
  • $\begingroup$ I'd like to see some suggestions from the moderators here, but anybody posting and then answering a question as if it were an FAQ would presumably get edits, comments and votes on merit. A random example could be "FAQ: Why are stepwise model selection procedures often criticised? Do they have no merits?". Nor would FAQ flavour rule out other answers. $\endgroup$ – Nick Cox Mar 26 '14 at 1:27
  • 1
    $\begingroup$ I guess the challenge is that we might want all the FAQs to be seen on a single page so would need to know how to format sub-questions. (?) $\endgroup$ – Frank Harrell Mar 26 '14 at 12:38
  • 2
    $\begingroup$ I see little point in debating the detail about a category that doesn't exist yet. If someone is able and willing to write something of FAQ character or style they have a chance to set precedents for how to do it, and in turn others might do something differently. Let standardization or tolerance of diversity arise as a question of its own accord. $\endgroup$ – Nick Cox Mar 27 '14 at 1:06
  • 1
    $\begingroup$ Nick let's at least figure out if we want one big question that keeps having individual FAQs added to it, or a separate topic for each FAQ with a title like FAQ: Does x imply y? And what is the role of "Community" vs. individual postings? Sorry there's some things I've never taken the time to understand. $\endgroup$ – Frank Harrell Mar 27 '14 at 12:05
5
$\begingroup$

A very similar question came up around a year ago, although I'm unaware of any material change in the site in response to that. I will suggest cardinal's suggestions are still just as relevant and highlight other communities' behavior in response to similar problems.

I think the faq tag on the main site is the simplest way forward. I imagine you have some answers of your own in mind about evaluating regression models: you could add a faq tag to the question yourself and then in the future make sure to link back to it whenever a similar question returns.

An interesting project may be to examine the network of answers and see which ones are linked to most frequently. These might be good candidates for applying the faq tag.

| |
$\endgroup$
  • 1
    $\begingroup$ I focus on actionable behavior (as opposed to debating the need for such an faq thread) as no action came out of previous discussion. I suspect any of the suggestions will have a net positive impact on the community - it is just necessary for some key individuals to take the steps to get the ball rolling. $\endgroup$ – Andy W Mar 28 '14 at 1:46
  • 4
    $\begingroup$ Sounds good to me. When I have time in a couple of weeks I'll start a controversial one to see how the process works, e.g. FAQ: How do you interpret a large $P$-value? $\endgroup$ – Frank Harrell Mar 28 '14 at 16:06
  • $\begingroup$ @Frank: It looks like you never got around writing this large p-value FAQ. Are you still thinking of doing it? $\endgroup$ – amoeba Dec 12 '14 at 14:20
  • $\begingroup$ Yes, just haven't had time. $\endgroup$ – Frank Harrell Dec 12 '14 at 14:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .