2
$\begingroup$

I have discovered this site only recently, and coming from a non-mathematical or computer background, but with quite a bit of interest in statistics and scientific methodology, I've come to enjoy it very much, and spend many hours trying to learn, and challenging myself. So thank you to the many of you out there that have made CV great.

In contradistinction to other more casual sites, CrossValidated requires serious work before even posting a question, resulting in a much higher standard of content. Predictably, it is often the case that a post in CrossValidated is ranked at the top when doing a browser search.

It is in this sense that it is somewhat anticlimactic, and possibly counterproductive (?), when, having come across just the right return on a Google search within the Q&A's repository of CV, it turns out that the question is unanswered with multiple moderators' comments indicating that there's already an answer to the question on a different post.

Even though it is easy enough to just follow the link to whatever the answer is, one may claim that the fact that CV encourages specificity and detail necessarily narrows some responses at the expense of generality.

So I wonder if it would be beneficial to encourage more answers on the same or closely related topics, inviting redundancy, and to some point repetition to one, cover all possible aspects of the fields of interest in CV, and two, make up for the collimated nature of the Q&A's with some overlap in the responses.

$\endgroup$
  • 1
    $\begingroup$ Some examples might be useful for this discussion - no-one's likely to favour tolerating exact duplicates & "the same or closely related topics" is perhaps not precise enough to avoid different construals. $\endgroup$ – Scortchi - Reinstate Monica Apr 10 '15 at 9:07
  • 1
    $\begingroup$ I'm not sure I understand the question. You seem to suggest there is merit in "redundancy," but immediately thereafter you talk about how to "cover all possible aspects of the field of interest." The latter suggests something more like the opposite of redundancy, where the each answer addresses a distinct aspect of a subject. It's also unclear to me how you conceive of such "redundancy" being instantiated: we already "encourage more answers" to any given question and provide various rewards for doing so. But we try to collect them all within a common thread rather than leaving them scattered. $\endgroup$ – whuber Apr 10 '15 at 14:05
  • 1
    $\begingroup$ I can see how my wording is imprecise, and perhaps I should have held off on posting this question until I know more about the site. The cycle I often find myself in is: Google search -> Great return on CV -> Question kind of closed with comments about its having been asked before -> Follow internal hyperlink to original question -> ... but not quite the same angle... specific problem -> Now exit CV to Wikipedia or some college handout. It would be great if once in CV you'd never need to leave regardless of your level of expertise. $\endgroup$ – Antoni Parellada Apr 10 '15 at 15:09
  • 1
    $\begingroup$ @AntoniParellada: (1) If you waited until knowing more about the site you might have forgotten the challenges you faced as a new user - so this question is welcome. (2) Again, I think an example or two of a specific instance when you found yourself in this cycle would better help us consider how to improve than a general description of the issue alone. $\endgroup$ – Scortchi - Reinstate Monica Apr 10 '15 at 15:58
  • 1
    $\begingroup$ Scortchi, I'm having problems "reproducing" my findings, but I plan on illustrating the issue as soon as I get a good example. $\endgroup$ – Antoni Parellada Apr 10 '15 at 21:26
  • $\begingroup$ OK... Not exactly the perfect example, but perhaps it may help... I've been trying to understand the concept of "orthogonal contrasts" in ANOVA for the past few days, and ran into stats.stackexchange.com/questions/63379/… on a Google search. When I read the question, I thought - This is it! Except the question is not greatly formulated and there is no answer. Is 'tantalizing' a good adjective? Should CV ideally aim at having solutions for these types of fundamental questions? I don't know. $\endgroup$ – Antoni Parellada Apr 11 '15 at 4:15
  • $\begingroup$ A better example: stats.stackexchange.com/questions/145962/… $\endgroup$ – Antoni Parellada Apr 12 '15 at 12:52
  • $\begingroup$ The first example is just a q. that hadn't been answered (until you did - thank you!) rather than one flagged as a duplicate. The second has just one vote to close it as a duplicate (& three to leave it open, though you can't see those); & a couple of links to related posts (I don't think it's being suggested that they provide answers to this question). So we're still inviting answers to both these questions - I was hoping for examples of q.s that were closed as duplicates when addressing a slightly different angle from the original might have been useful. $\endgroup$ – Scortchi - Reinstate Monica Apr 14 '15 at 15:45
  • 1
    $\begingroup$ Points well taken. I will provide examples as I stumble upon them, and I hope it is understood that my perception is that this site is phenomenal thanks to the contributions. I just would like for its to be the last click on any search for answers regardless of the level of statistical/math knowledge. $\endgroup$ – Antoni Parellada Apr 14 '15 at 15:50
4
$\begingroup$

It's clear policy that exact duplicates should close (since we want to be able to find the best answer or answers to that question, not just an answer).

I think the question should be retained* so that the different ways of asking a question lead to the same answer. As far as I understand, that's the typical case.

*(I think dupes should almost never be deleted -- unless its word-for-word and has no answers; either they're so close that a separate question is no use whatever -- in which case merge it -- or they're expressed in a slightly different way that someone might hit with a search, which suggests the question has value even if it means the same thing.)

In my opinion a search that hits a closed duplicate is a success, as long as the link covers what the person wanted.

So the point of closure of duplicates -- what it's for -- is to get the best answers to that particular question in one place.

However, sometimes questions aren't quite duplicates -- a near duplicate probably contains some useful information that may help but the details may be different enough that another answer or set of answers is worth having for that specific issue.

There's also indication that we could take more care to explain why we're closing; even the newer, friendlier close messages can sometimes be confusing and even confronting, and the purpose of closing may be unclear. There's sometimes a tendency to see closing as criticism or punishment when it's actually part of how SE is supposed to work. Done right it helps people find good answers.

There's support (including founder-support) for the position of perhaps not closing when sometimes we might be inclined to. For example, there's this blog post from Jeff Atwood; see the section under "Borderline duplicates", including:

These have [...] subtle differences that may make them legitimately standalone questions

There’s often benefit to having multiple subtle variants of a question around, as people tend to ask and search using completely different words, and the better our coverage, the better odds our fellow programmers can find the answer they’re looking for.

I do think we sometimes have a tendency to "over-categorize" and lose subtle shades of distinction between questions that may help users, so we should perhaps consider whether there can be value in tolerating some posts that may seem like duplicates at first blush. As mentioned at the blog post, we can use other tools than close-as-duplicate to get people to see those other good answers - good use of tags, and links to closely related posts.

$\endgroup$
  • $\begingroup$ Excellent clarification. Thanks. $\endgroup$ – Antoni Parellada Apr 11 '15 at 6:58
6
$\begingroup$

My feeling is that questions along the same lines are going to get answers along the same lines—better to encourage distinctive questions if we want the answers to have wider coverage. We try to do that by linking to duplicates & near-duplicate questions & asking what's not covered in the answers there.

I find it counter-productive when I answer a question only to discover later on that someone else has answered what's essentially the same question already (often rather better): I could have saved time by linking to the prior question, perhaps with an explanation of how it was related if that wasn't obvious, or, if I had anything worth adding, by writing a brief complementary answer or comment; or I could have been inspired by the other answer to write something I wouldn't have thought of by myself. I also find it anticlimactic when a search turns up half a dozen posts on a topic I'm interested in but the questions & answers are all much the same.

$\endgroup$

You must log in to answer this question.

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