11
$\begingroup$

I was wondering what the policy is on questions which look like a commercial organisation is asking how to solve a business problem.

It's totally understandable that everyone needs a hand from time to time. Questions about sorting out obscure software, querying research papers, understanding some math are all fine as far as I'm concerned.

However every so often I see questions from people who are clearly employees at some company asking how to do their job. For instance, the following question seems to be asking how to build an online recommender system:

https://stats.stackexchange.com/questions/212983/manually-update-ratings-in-recomender-system

$\endgroup$
2
  • 3
    $\begingroup$ The problem with that question is how it asks for information: "I have this procedural (vice statistics) problem, and I need you to solve it." In OP's case, the answer may very well be to call recommendationEngine.update(new_data). Or it might be a policy question about how new information should be incorporated to provide the best value to the customer. The question bears SVD tags, but nothing else in the question suggests that there's any statistical or mathematical content. $\endgroup$
    – Sycorax Mod
    Commented May 18, 2016 at 14:12
  • 1
    $\begingroup$ Update: the linked question was closed on May 17 as "off-topic" with zero votes and no answers. Accordingly, it was auto-deleted by @Community ten days later, on May 27. Users with less than 10,000 rep won't be able to see it any more. (Believe me, you are not missing out on much.) $\endgroup$
    – Stephan Kolassa
    Commented May 30, 2016 at 14:52

2 Answers 2

Reset to default
28
$\begingroup$

A question should be judged based on whether it is a good statistics problem. Statistics are used in business, in academic research, in politics, in volunteer activities, and by some really weird people (ahem...) for fun and recreation.

Whether a question posted to CV concerns a specific one of these use cases should not matter. What should matter is whether it is a good statistics question.

The question you link to could just as well concern a recommender system used by a high school guidance counselor. Or by a doctor. The OP could be an academic researcher building it, a freelance software consultant, or just someone building it for fun. All this should not matter. What should matter is whether it is a good question. (I find it "unclear what you are asking.")

$\endgroup$
6
  • 7
    $\begingroup$ Agreed. Note that researchers are also "people who are clearly employees at some company asking how to do their job" with only small elasticity on what precisely is a "company". $\endgroup$
    – Nick Cox
    Commented May 17, 2016 at 11:54
  • $\begingroup$ Stephan, you make a good point. The key issue for me is where the questions would be Googleable for someone with the necessary competence. There a policy on solving routine homework questions. Why provide solutions for routine business problems? $\endgroup$
    – conjectures
    Commented May 18, 2016 at 9:12
  • $\begingroup$ Nick, to push the analogy with researchers: I'm talking about the equivalent of someone asking "Could you come up with an idea for a paper for me, and show me how to do it?" $\endgroup$
    – conjectures
    Commented May 18, 2016 at 9:13
  • 3
    $\begingroup$ I don't think that the Googleabilty of a question should be a hard criterion. Apart from that, I fully agree that routine self-study questions should be treated according to our homework help policy. But that doesn't hinge on the question coming from business. The key point is that we need people to have at least an elementary understanding of statistics, since a) we can't provide that here, and b) answering "business homework" questions may give people just enough rope to hang themselves. $\endgroup$
    – Stephan Kolassa
    Commented May 18, 2016 at 9:18
  • 2
    $\begingroup$ @conjectures Homework questions do raise related issues. If someone copies an assignment verbatim I really won't play, but I don't know how many carefully disguised questions I have answered positively. Similarly, if someone evidently wants to be given the idea for an entire project, then I won't play on those grounds alone, even though their question is usually unanswerable. But I am unhappy at the idea that there can be a policy based on whether the motives behind a question are honourable, because we can't tell that at all consistently. In essence, we exist to be exploited by others.... $\endgroup$
    – Nick Cox
    Commented May 20, 2016 at 11:42
  • 1
    $\begingroup$ @NickCox: Students'll need to have at least read & understood a homework question to carefully disguise it, so even then our policy's doing some good. $\endgroup$
    – Scortchi Mod
    Commented May 22, 2016 at 18:34
4
$\begingroup$

Given there has been no moderator comment on a policy (at this stage), this question seems like it might be on what we, as the community, think should be a good guideline or motto for answering questions.

I would like it to focus on, as @stephan kolassa has noted, good statistics AND positive sustainable answers - i.e. teaching how to fish, not handing a fish.

Sustainable answers are the ones that will discourage the selfish questions (can't be bothered learning, will just ask a forum) and encourage the learners - no matter what their statistical interest motivation is...

[maybe this question needs to be in a philosophy forum :) ]

$\endgroup$
2
  • 4
    $\begingroup$ The SE system is largely self-governing. Site policies are primarily determined by the community, not by the moderators. We respect their opinions because we have very good moderators who are knowledgeable about the subject matter & heavily involved in the site, but they aren't really a form of government. They are more like sheriffs (implementing policy) than legislatures (creating policy). $\endgroup$
    – gung - Reinstate Monica
    Commented May 22, 2016 at 16:54
  • 1
    $\begingroup$ @gung: A flattering as well as apt simile! $\endgroup$
    – Scortchi Mod
    Commented May 22, 2016 at 18:39

You must log in to answer this question.

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