9
$\begingroup$

At times I see that someone will provide an answer that, while not directly addressing all of the points in a question, will nonetheless give a good insight on something that is either partially solving the problem or is tangentially related in some way. I feel somewhat torn on these answers. On one hand, I think the commentary they provide can be very useful (for example a question about regression assumptions where an elaborate answer is only given for the linearity assumption), but I also wonder if it should be upvoted if it doesn't address the complete problem. What is the correct way to consider these answers?

For the record, I don't mean answers where somebody answers the full question and another person will chime in with additional commentary that supplements it. I mean more for standalone answers that do not build off of others.

$\endgroup$
2
  • 5
    $\begingroup$ The latter point is very important. I think it's fine to upvote the latter. What's in question is whether upvoting a single, partial answer might make it less likely that others come along to provide a fuller answer. $\endgroup$
    – gung - Reinstate Monica
    Nov 10 at 12:18
  • 6
    $\begingroup$ Just some thoughts: (1) A "partial answer" suggests the question is not a good one for the site because it raises multiple issues. Maybe the best route is to encourage the OP to narrow the question. (2) Sometimes you can encourage the respondent to flesh out a partial answer into a full one. By withholding an upvote at that time, you retain the power to reward a later improvement to the answer. (You can also award a bounty later, too.) $\endgroup$
    – whuber Mod
    Nov 10 at 15:15

3 Answers 3

Reset to default
5
$\begingroup$

The help-page mentions this (emphasis mine):

Make sure your answer provides that – or at least a viable alternative. Your answer can say “don’t do that,” but it should also say “try this instead.” Any answer that fully addresses at least part of the question is helpful and can get the asker going in the right direction. State any limitations, assumptions or simplifications in your answer. Brevity is acceptable, but fuller explanations are better.

So, there is no doubt they are indeed acceptable answers. But upvoting them depends on the specific post and the overall scenario in concern.

Context is important and relevant here. whuber has raised a good point. The help-page again goes on in the same vein:

Save yourself some frustration and avoid trying to answer questions which... ...are unclear or lacking specific details that can uniquely identify the problem.

If someone has answered such ambiguous query (same in case of a broad question) and the bone of the post is seeking more information, check if it could have been possible only with leaving a comment and then may ask the answerer to do the same. However, if there is also some good advice (in the spirit of the first quoted para above) that has the potential to help OP, albeit in that ambiguous set of information, then that should be upvoted.

Nevertheless, answers are not monoliths - they are prone to be altered once OP has provided enough details for the responder to update their post. So, the best approach would be to wait (you can follow the post), both for OP's clarification as well as the subsequent update of the answer, if it is needed.

$\endgroup$
5
$\begingroup$

A very good question that did get me thinking.

I think whuber's comment is absolutely key:

A "partial answer" suggests the question is not a good one for the site because it raises multiple issues.

So here is what I will do in the case of good and partial answers: upvote them. Because they are good. Yes, that will likely reduce the probability of later full answers. I will consider this a feature, not a bug of upvoting the partial answer: if the asker is unsatisfied with the (good!) partial answer, they should post a new, more focused, follow-up question.

Essentially, I believe that rewarding good partial answers will have an educational impact on the asker that will overall improve the quality of this site.

$\endgroup$
5
$\begingroup$

A partial answer should be both confident and diffident. It should be confident because you think you have something worth saying that is (much) more than a comment, but it is not a full answer. It should be diffident for the same reason.

Manifestly, whatever can be said as a comment should be a comment, starting with "I agree with X" (sometimes worth saying) or "I have the same problem" (not usually worth saying). I agree also that really poor or confused questions should not be answered. Votes to close, downvoting, flagging, and commenting on what's poor or confused are among the tools provided, depending on how and why it is not a good question.

It's a good idea to flag the diffidence early on, to explain why you're posting a partial answer. Ideally you will then fend off possible comments (or thoughts) to the effect that you didn't answer the question completely, or even at all.

I can address part of the question here, namely ...

There are already excellent answers from @User1865345 and @Stephan Kolassa, but here I want to add a further perspective.

No-one has answered yet, but here are some thoughts to start a discussion.

What I have to say won't fit into comments because .... (For example, you may be posting a graph, or more mathematics or code than is easy to read in comments.)

This divides opinion, I guess, but I think it can be in order to try to widen or shift a thread because the question as posted obscures or ignores a deeper or more worthwhile question, or because there is an interesting or useful question alongside. Naturally the OP has control over whether they accept any answer, and they or others might riposte that your answer is irrelevant or marginal to the question. That's part of the risk.

Posting a partial answer carries some obligation to keep an eye on the thread to see if someone else posts a more complete and/or better answer. If that happens, there can be various options, including

  • leaving your partial answer, because it hasn't been superseded or it remains your view

  • taking what is different and summarizing it as a comment on someone else's answer

  • occasionally adding material as an edit to someone else's answer (best done only if you think the poster "knows" you from your contributions or direct exchanges and won't mind)

  • just deleting your answer.

I think tolerance for (and appreciation of) partial answers on CV goes along with a statistical culture that recognises and even celebrates the possibility of different attitudes and styles, particularly in data analysis, but also in statistical philosophy and theory. People new to statistics are often very surprised or disappointed by this, but it is part of what they may need to appreciate. CV is the site on Stack Exchange I know and like best; on Stack Overflow I often see very limited tolerance for comments or discussion, even when relevant to the question and likely to be helpful to readers.

Self-exemplifying: I didn't comment on how to vote. I just picked up on ideas so far and ideas nearby.

$\endgroup$
2
  • 2
    $\begingroup$ CV is the site on Stack Exchange I know and like best; on Stack Overflow I often see very limited tolerance for comments or discussion, even when relevant to the question and likely to be helpful to readers.---Couldn't agree more with both sentiments. I think what often prevents me from posting an answer is the diffidence part, but perhaps if I am cognizant that it is conditional upon the entire discussion then it isn't so bad, so long as it is done in good faith. $\endgroup$
    – Shawn Hemelstrand
    Nov 11 at 12:31
  • 3
    $\begingroup$ @Shawn Hemelstrand Raw experience: I often comment on Stack Overflow about the importance of giving a reproducible example and other habits that are similarly good practice. Sometimes that is an extra comment in my answer. I have just had that edited out of several answers as unnecessary information by another user, supported by a moderator. Hmph! (Biter bit, some might think.) Moderators here are always sensible and open to discussion. $\endgroup$
    – Nick Cox
    Nov 13 at 16:35

You must log in to answer this question.

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