7
$\begingroup$

This weekend a question got asked about exploratory analysis and p-hacking:

Is "Exploratory Data Analysis" Fundamentally At Odds With "P-Hacking"?

This is a broad topic to begin with and the question — in my opinion — is too long and lacks clarity. However, it's also an important and interesting topic, so instead of "Needs more focus" I voted to close and found threads that discuss the topic with eloquence and pointers to further resources, etc. I can only suggest one duplicate but in a comment to the question I linked to 5 (near) duplicates as it's hard to find close duplicates for a question that asks many sub-questions.

Today the question has been re-opened. This is fine by me. It's possible that I misjudged the qualities of the question.

However, I would like to ask for some practical advice on how to calibrate my understanding of what "needs more focus" and "is a duplicate" and esp. at the intersection of these two cases. Searching for duplicates takes time and often it seems easier to write a new answer than to look for great existing answers.

Related meta threads:

What is our policy on merging duplicate questions?
Many (historical) duplicate questions - how to pick a "canonical" question?

I'm not sure they address this case because the examples seem to be short questions with more obvious duplicates.

$\endgroup$
11
  • $\begingroup$ After you posted this, it has been closed again ... $\endgroup$ Jan 16, 2023 at 12:45
  • 1
    $\begingroup$ @kjetilbhalvorsen This time it's for lack of focus. This is probably more appropriate grounds for closing. I think I'm converging to the conclusion that searching for duplicates should be about finding and reading examples of effective technical writing, not so much about the close vote. I only do it for topics I find interesting anyway. $\endgroup$
    – dipetkov
    Jan 16, 2023 at 14:50
  • 2
    $\begingroup$ I appreciate your effort in sieving out the relevant posts from the CV vault: sometimes it cannot be easy to pinpoint a single post out of all those relevant ones that can be marked duplicate. But for this particular matter, the given one was more appropriate (& I voted to close). Now whether it should be closed on the grounds of lack of focus or some closed duplicate depend on the context. IMO, the post in question is too broad. In a nutshell, I would love to see a formal rule to vote to close on which basis but if it is at the confluence of two, either could suffice with proper dupe links. $\endgroup$ Jan 16, 2023 at 15:36
  • 1
    $\begingroup$ In fact, if you have many posts which have the potential to address the present queries, and one cannot narrow it down to one single dupe candidate, then plausibly the present question could lack focus (of course, this is not a proposition to go for). $\endgroup$ Jan 16, 2023 at 15:43
  • 4
    $\begingroup$ This is hardly a perfect system, but if there are a dozen question marks in a question, as there are in this case, then it's very likely that the question needs more focus. Finding duplicates for each of its 12 questions is a big task, and even if you can find 12 duplicates, the software will only allow you to attach a maximum of 5 to the question. Finding duplicates is an important and valuable contribution, but an asker should focus their question enough that it is concise and answerable in a few paragraphs. $\endgroup$
    – Sycorax Mod
    Jan 16, 2023 at 16:40
  • $\begingroup$ @Sycorax One challenge with the "lack of focus" questions is that they often receive a flurry of answers before being flagged. At that point there might be little incentive for the OP to improve their question; after all they already got a lot of engagement from the community. This could be a self-reinforcing kind of behavior. $\endgroup$
    – dipetkov
    Jan 16, 2023 at 18:24
  • 1
    $\begingroup$ I share your concerns about these imperfections of the process, especially when a question is so broad that most users on this statistics forum will be able to write a responsive answer to one of the post’s many questions. But also note that the same “time lag” exists when voting to close with "too broad" as does voting to close with a duplicate, because more votes must accumulate. If the site had larger number of active reviewers, this flaw might be mitigated, but probably not eliminated. $\endgroup$
    – Sycorax Mod
    Jan 16, 2023 at 18:59
  • 3
    $\begingroup$ I would separate the main issues as (1) making sound decisions about closure (2) giving reasonable feedback to the OP (if not a spammer or offensive). The two are linked but not identical. My impression is that being unclear is the most difficult reason for OPs to understand with common reactions (a) the question is clear to me, so why call it unclear (b) I am being as clear as I can, but my whole point is that I am confused or uncertain. So, if there are multiple grounds for closure go for something else if possible. $\endgroup$
    – Nick Cox
    Jan 17, 2023 at 14:30
  • 1
    $\begingroup$ On duplicates, I flag outliers as territory where I often think "This is just the same old question as often asked" but it can be hard to find an exact duplicate. $\endgroup$
    – Nick Cox
    Jan 17, 2023 at 14:32
  • $\begingroup$ @NickCox Your comment made me realize that I made a mistake using the word "clarity"; "focus" is much better in this case since, yes, "not clear" can mean many things. I'm wondering more about duplicates. I find existing answers useful. So I errored on the side of "this has already been answered". Wouldn't others also want to read those threads? But I understand why "lack of focus" was a better choice. $\endgroup$
    – dipetkov
    Jan 17, 2023 at 15:55
  • 3
    $\begingroup$ Another reason to close as "too broad" instead of a duplicate is iteration: when a Question contains, say, $n \ge 3$ questions, closing it as a duplicate of 1 existing question isn't a great solution because then OP edits to exclude the duplicate, leaving $n-1$ questions, whence finding a second duplicate might mean repeating this whole process, and likewise for the $n-2$, $n-3$, ... questions ad nauseam, to the frustration of all parties. $\endgroup$
    – Sycorax Mod
    Jan 17, 2023 at 16:16

1 Answer 1

4
$\begingroup$

I answer my question with a summary which incorporates advice from the comments and further reflection:

  • If a question seems too broad/includes many sub-questions, it's better to flag it as lacking focus, even if some sub-questions already have answers. This provides helpful feedback to the OP and encourages them to write questions in a better way.
  • The "is a duplicate" option is meant for close duplicates. This is helpful not only to the OP but to anyone who happens upon the question later on looking for specific information/answers and doesn't need "context".
  • Related threads are best linked to in a comment. Yes, not everyone reads comments but there is probably overlap between those who read comments and those who would be interested in reading more on the topic in question.

⭑ By happy coincidence, today someone asked about exploratory vs confirmatory analysis. They didn't find the broad thread useful even though it contains the answer to this much more specific question (somewhere, if you read everything). But I happened to know that a nice answer from @Ben discusses the exploratory vs confirmatory aspect, so I voted to close as a duplicate. Worked very well actually.

$\endgroup$
8
  • 1
    $\begingroup$ @Alexis Thank you for the comment. As the question says, I'm wondering how others make the determination whether other threads qualify as duplicates in the voting process in the case when a question is broad but related to previous questions. If I was disparaging of others, I wouldn't be asking their opinion. I'll change "exact" to "close" though not sure it's any better defined. Voting requires making a judgment call after all. $\endgroup$
    – dipetkov
    Jan 17, 2023 at 19:06
  • 1
    $\begingroup$ @Alexis My question also doesn't imply that the voting process hasn't been working so far or there is anything that should be fixed. It's only asking "why was closing as duplicate" not a good choice (that I made) in this particular case. $\endgroup$
    – dipetkov
    Jan 17, 2023 at 19:18
  • 1
    $\begingroup$ Thank you for the response and edit! I am happy now.:D $\endgroup$
    – Alexis
    Jan 17, 2023 at 20:50
  • 2
    $\begingroup$ I will say that intelligent people of good will can still have honest, well-justified differences about many issues. There isn't always a lone good response to a problem, though there may be several reasonable ones. Sometimes there's no strong reason to think that one choice was especially mistaken, even though another was selected in the end. We can try to understand the reasons, but it doesn't need to imply any great error. Even after 4 years of my time as a moderator I was still often unsure about the best choice in many cases; often better to choose something reasonable and move on. $\endgroup$
    – Glen_b
    Jan 18, 2023 at 2:12
  • $\begingroup$ @Glen_b Thank you for the advice. This isn't the first time I've been wondering about the value of looking for duplicates. Duplicates often seem to me both the best option and a complete waste of my time. As you say, if there is another reasonable option and it takes me less time to cast that vote, then it's the better choice from efficiency standpoint. $\endgroup$
    – dipetkov
    Jan 18, 2023 at 7:27
  • 3
    $\begingroup$ You may be taking a different message than the intended one. I'm all in favour of finding duplicates since (among other things) that helps later seekers of information as well (though this doesn't mean I have any disagreement with the outcome either). Rather, I meant if you make a choice and later several people conclude something else was a better choice you should not worry too much that your choice was incorrect, since you may all be choosing different responses for perfectly good reasons. $\endgroup$
    – Glen_b
    Jan 18, 2023 at 22:41
  • 4
    $\begingroup$ In short, we're all trying our best. We can disagree very easily and not necessarily have any of the choices made be wrong $\endgroup$
    – Glen_b
    Jan 18, 2023 at 22:45
  • 1
    $\begingroup$ @Glen_b I do find your comments & advice useful. I wasn't frustrated that people disagree with me but that I wasted my time to find out that people disagree with me. It has happened with flagging duplicates before. I want to clamp down on this happening (the time wasting, not the disagreement). And since I control only my own decisions, I need to be more purposeful about the process of searching for and flagging duplicates. $\endgroup$
    – dipetkov
    Jan 19, 2023 at 7:24

You must log in to answer this question.

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