7
$\begingroup$

A comment was posted on my answer to a question about a year ago, expressing surprise that the answer got so many upvotes given that it isn't very detailed. The comment was up-voted a few times, and my answer has subsequently been down-voted, so I guess people agree. But I wonder what's wrong with leaving some details up to the OP to figure out. I have posted plenty of detailed answers, with examples, for other questions, but in this case I think I did correctly identify what was needed (component-plus-residual or added-variable plots), and pointed to a couple of R packages that offer these. The OP apparently found the answer helpful, and had commented to that effect.

I don't really care about the reputation points. I do care about my time. I wonder how much hand-holding is expected, and I wonder about establishing an expectation where people don't have to do any digging at all to get their questions answered. Are we really here to offer silver spoons?

$\endgroup$
  • 9
    $\begingroup$ It is a fact you helped the OP there. My impression is that you could have provided more details: e.g, mentioning the functions inside the R packages, adding visual examples about that type of plots. The OP would still need to learn how to parameterize such functions, to learn the theory behind such graphs, etc. We shouldn't silver spoon when answering, but I guess the example you used is not a good one about providing guidance, without hand-holding. When studying from great (upvoted) answers here in CV I frequently need to do extra research. They help me getting started. $\endgroup$ – Andre Silva Oct 13 '15 at 15:33
  • 1
    $\begingroup$ I wouldn't want to focus on that question: it is arguable that the extra detail on R functions would actually be off-topic unless it had a clear statistical focus. The bigger issue is that minimal hints don't score many votes on internet forums. I have sometimes resorted to them myself with closing comments of the form "That is set as an exercise" (facetiously) or "You can work that out yourself using ..." but it's detailed (or especially lucid or ingenious) explanations that get the votes. $\endgroup$ – Nick Cox Oct 13 '15 at 15:44
  • 5
    $\begingroup$ In teaching the better of one's own students will respect the ideal that it's not your job to spoon-feed them. Also, within that context it is always known whether students know enough (e.g.) probability theory, calculus or matrix algebra that will let them work something out. Outside that context it's not an aim here, so far as I can see, to second-guess how much help people minimally need. Indeed, most good answers have something for several levels of readers. (I think this is a little hard to separate from whether concise answers can sometimes be more helpful than lengthy ones.) $\endgroup$ – Nick Cox Oct 13 '15 at 15:50
  • $\begingroup$ The person who made the comment is fairly prolific in r-related tags on StackOverflow. I think there is a bit of a cultural difference here between StackOverflow (programming solution) and CrossValidated (where a more general or conceptual answer might be fine). $\endgroup$ – C8H10N4O2 Oct 20 '15 at 20:23
8
$\begingroup$

You needn't pick out which functions to use from the car or effects packages. Examples of how to apply statistical methods, using any software (even if not specifically requested by the OP), can certainly be useful, but their provision isn't a necessary condition for a good answer.

Nevertheless, merely name-checking something the OP ought to look into makes for a rather meagre answer, which would typically benefit from a brief explanation of at least

  • how it relates to the question

  • what exactly it is (perhaps by giving a reference)

There's often some ambiguity in questions & answers (what is a "net effect" anyway? which of added variable & component-plus-residual plots shows it? are either of those the same as partial-regression plots?), resolving which gives readers confidence that they're not going to be sent off on a wild-goose chase, & doesn't amount to hand-holding.

All in my opinion, of course.

$\endgroup$

You must log in to answer this question.

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