14
$\begingroup$

Of course I really don't think that there are so few good statistical questions left to exhaust all questions, I have noticed that I will answer a question and find someone else saying that the question shouldn't be answered because it is the same as a previously answered question.

I am relatively new to the site and don't know a lot of the history but I find that very oftn it is happening that a question is a repeat of something from 6 months or 1 year or more in the past.

What should a new member do? When a question comes in should I be searching the site to see if it has been answered previously before I give my own answer? When asking questions should we be searching the site to make sure it is a new question before we ask it?

I would also like to know what the community members think of answering a question that is a repeat. I have done this on occasion and see no harm in it. If the answer is simple and brief isn't it better to help the OP by giving the answer rather than to ask them to search the site for the answer? I know that some moderators feel that duplicate questions should be closed immediately and the OP should be directed to the previous questions. The argument is that answering these questions puts too much repetition and clutter on the site. But isn't good service to the OP more important than the tidiness of the site. Some of the clutter and repetition could be cleaned up later. I expect that some or many of you will disagree with me. So let me hear it.

$\endgroup$
25
$\begingroup$

This is a good set of questions; perhaps many people, especially our newer users, are wondering about these things.

I need to point out that the policy about closing duplicates isn't a matter of "what some moderators feel:" it's part of the very structure of StackExchange. When you consult an encyclopedia or dictionary or even Amazon.com for information, for instance, you don't want to have to look in multiple places: you expect, and deserve, that what you're looking up will be in one place and it will be cross-referenced to any related (but different) places, so that you don't have to go skipping around to find what you're looking for.

Our site provides great tools to help the asker of questions identify similar threads. In principle, this should resolve many such issues before they ever appear, and everybody is happy. But many askers cannot recognize the similarities because they do not know the terminology or they are not knowledgeable or imaginative enough to see them. That's where the community comes in. We hope and expect that when experienced members first read a new question, they will reflect about whether they have encountered something similar and whether something similar should have been encountered by now that more than ten thousand questions have appeared. For instance, a question that asks how to do a t-test should already have been asked and answered, so a quick search for that answer is in order before going any further. Just as we ask of questioners that they do some research before posting a question, we ask of answerers that they do some research too. Think before you post!

When you find a clear duplicate question, then immediately closing the new one is the best thing you can do all around: it directs the asker to an existing solution; it prevents unnecessary duplication of threads; and it prevents unwary newcomers from investing time in formulating answers that have already appeared elsewhere. (Please note, too, that duplicate material complicates all future searches, actually making it harder for new users to ask questions and harder for experienced users to make connections among questions.) If you find a "perhaps" duplicate, then a quick discussion with the OP (using the comment mechanism) will determine whether they have a new question or not. Once again, if you ascertain that their question is a duplicate, then closing it and directing the asker to the existing thread is the fastest way to serve them and usually gives them more information then most people can hope to produce in a single quick reply.

This community has more objectives than merely dishing out answers to all and sundry. The "tidiness of the site" is intimately connected to its usability and value far into the future. Moreover, as one who has done a heck of a lot of it, I can attest that cleaning up "clutter and repetition" is time consuming, complicated, sometimes painful, and rarely gets completely done: far better is it to avoid such noise in the first place. For that reason, everyone posting answers here is constantly encouraged to provide clear, well-formulated, objectively reasoned, carefully supported answers: we want them all to be great. If users visit our site and see ragged, disorganized collections of hasty, careless, and mediocre answers, they're unlikely to return or participate. Leave that junk for yahoo (pretty well named, isn't it? -:) or answers.com or the myriad other generic Q&A forums that come and go.

In summary, if anyone has arrived here interested chiefly in participating in a question mill with little regard for quality or curating the site, then your services would be better appreciated elsewhere: Cross Validated probably holds little future for you.

$\endgroup$
  • $\begingroup$ My comment about "some moderators feel" had to do with emphasis on closing out questions. I recognize that StackExchange has directives but that doen't mean that they are not subject to interpretation. Somehow I feel that your last paragraph was directed at me. I tend to answer a lot of questions but only on subjects I think I have good knowledge. I try to give nice clean answers. I think most of my answers are high to very high quality with respect to content. Style and form is another issue. As I learn I improve the style of my answers. $\endgroup$ – Michael R. Chernick Aug 24 '12 at 20:16
  • $\begingroup$ But I don't think it is fair to portray some of my answers that are not in the best format as poor answers. Also I don't like the fact that you say and others may agree, that you encourage downvoting answers that are not neatly presented. I think that is wrong. Maybe I don't completely understand the purpose of this site yet but my impression is that we serve a costumer who is the OP and the OP should take priority over other considerations. I can now appreciate why you want to keep the site neat looking and have existing solutions easy to access. I go along with most of what you say. $\endgroup$ – Michael R. Chernick Aug 24 '12 at 20:24
  • 3
    $\begingroup$ @Michael As I tried to make explicit in the opening statement (which I admit stands a long distance from the closing paragraph :-), my answer is directed at our "newer users." It's not a personal response: those have no place here on meta and they come via other channels, as you know. It's up to you, and everyone else who may be reading this answer, to decide to what extent it might apply in any situation. $\endgroup$ – whuber Aug 24 '12 at 20:55
  • 8
    $\begingroup$ Seconding this well-thought reply, I think formatting and post quality--which often come together--play a particular role on this site, @Michael, and not only because of the broken windows rule (shamelessly copied from one of mbq's responses). The OP is surely your main interlocutor in the first instance, but thousands of users and visitors may read your reply soon or later: unlike listserv and the like, we have facilities for editing/updating our posts, that's what makes SE sites so great, among other things. $\endgroup$ – chl Aug 24 '12 at 21:13
  • 2
    $\begingroup$ @whuber I am a relatively new user. But I do think I have learned a lot about the site and have earned some respect. I am learning a lot from you and chl and mbq and others and do take your comments seriously. I will continue to work to improve my answers for everyone's benefit. $\endgroup$ – Michael R. Chernick Aug 24 '12 at 22:01
  • 8
    $\begingroup$ @Michael, this is a good question (and good answer, whuber) and I'm glad to hear your last comment. I want to add one remark regarding your comment: "my impression is that we serve a costumer who is the OP and the OP should take priority..." I think this may be the source of some past misunderstandings. I don't think the purpose of this site is to serve the OP and that should not take priority over other considerations (mods please correct me if I'm wrong). The purpose is to provide a repository of great questions with great answers that readers, including the OP, will find useful. $\endgroup$ – Macro Aug 24 '12 at 22:28
  • 2
    $\begingroup$ @Macro I know the site's goal is to produce a great repository but the process is questions and answers and as part of the process we are providing a service to the OP and I think most will agree that doing that well is important. We may not all agree that it is paramount. $\endgroup$ – Michael R. Chernick Aug 24 '12 at 23:13
  • 3
    $\begingroup$ One thing to explore here is how exactly a question must match a previous question for it to merit being closed. I, for one, think that context is often critical in giving good answers (and all too often missing from questions) and this context can make answers different, even if questions are quite similar. I'd like there to be more emphasis on good questions; I wrote a post about how to ask a statistics question that covers some of it. $\endgroup$ – Peter Flom - Reinstate Monica Aug 26 '12 at 22:12
  • 2
    $\begingroup$ Those are good points, @Peter. Typically we consider different-looking questions to be duplicates when the same answer will serve for both (which means the questions ask the same thing but perhaps with different terminology, or--quite often--they are asking trivial variations of the same homework problem), but similar-looking questions are not duplicates when they need different answers. $\endgroup$ – whuber Aug 26 '12 at 22:16
  • 2
    $\begingroup$ Also, one problem with "think before you post" is that many of the questions are from newbies, both to statistics and to the site; many have no reputation or very little. Certainly some of us both ask and answer. In fact, I think most of the highest-reputation people have done both. But the majority of questions is from new people. Yet another issue is how the site gets cross-indexed and so on. This is hard in any field; it may be esp. hard in statistics, where different nomenclature is often used for the same thing. $\endgroup$ – Peter Flom - Reinstate Monica Aug 26 '12 at 22:27
  • 1
    $\begingroup$ @Peter Are you suggesting newbies can't (or shouldn't) think? :-) The cross-indexing indeed is challenging: one hope is that community involvement in tagging questions and maintaining tag synonyms will provide a powerful mechanism to associate similar questions with each other. $\endgroup$ – whuber Aug 26 '12 at 22:37
  • 3
    $\begingroup$ :-). Newbies can think, but they may not know about searching. Often they just don't know how to search well. Most of my clients are highly intelligent people who don't like statistics. $\endgroup$ – Peter Flom - Reinstate Monica Aug 26 '12 at 22:39
9
$\begingroup$

I agree with Whuber's excellent answer, I'd just make a further point. The question isn't really about stats.stackexchange since you could ask essentially the same thing on any stack exchange site meta. This question really is asking about the stack overflow/exchange model.

When the questions talks about 'providing a service to the OP' this is something that you get on a forum, not a Q&A site. The service that a well run site like this should provide first and foremost is a great place to go to find the information you need (often via a Google search) from good answers to previously asked questions. Actually asking a question should not be the default way to use this site. Duplicate questions with 'something is better than nothing' answers clutter the place up, clog search results and encourage more duplicate/poor questions and more poorly thought out answers.

When you are used to using forums, it takes a while to get your head around the stack exchange model. These issues were hammered out in the early days of stack overflow but the eventual success of that original site shows the value of the format.

The only way to build a strong Q&A site is for the community, questioners and answerers, to commit to the format. Otherwise everyone pulls in a different direction and the whole thing becomes a mess. This does means losing members from both sides of the Q&A fence who find they don't like the format. That is a necessary growing pain along the road to building a trustable, quality over quantity repository of information.

$\endgroup$
  • 1
    $\begingroup$ Well, I've never seen a CV answer come up on 1st page of a Google search for a statistical term. So, if that's the goal, something's not working. $\endgroup$ – Peter Flom - Reinstate Monica Aug 27 '12 at 18:42
  • 2
    $\begingroup$ @PeterFlom, try searching on partial eta squared. But yes, this will happen infrequently, & topping the Google results shouldn't be our main goal. $\endgroup$ – gung - Reinstate Monica Aug 27 '12 at 19:52
  • 4
    $\begingroup$ @Peter, I did see the CV answers on the first page of the Google search. It actually happens at least a quarter of the time for me, especially on some tricky R implementation questions. (Google is not that smart though. There are some rather disappointing examples of when the Google search throws me off to my own web page -- dear Google, I want to find something new, not something I had already known three or five years ago!) $\endgroup$ – StasK Aug 31 '12 at 16:07
  • 1
    $\begingroup$ Nowadays it happens quite often ... $\endgroup$ – kjetil b halvorsen May 17 '17 at 15:58

You must log in to answer this question.

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