5

Sometimes when a question is asked there is a "simple mistake" in the subject matter rather than with Mathematica. Does voting to close "due to a simple mistake" apply only to Mathematica rather than subject matter mistakes (or misunderstandings)?

For example, consider the very recent question: P-Values For LinearModeFit Changing Due To Ordering of Features.

I posted an answer to that question but should I have made a comment about the subject matter issue, suggested a subject matter forum, and then voted to close?

2
  • 1
    I don't have a general suggestion, but for that specific example I think it's best to leave it open, because your answer could be valuable for others having the same confusion. I suppose one could argue that the whole question belongs on a different site, since the point of confusion was mathematical rather than with code, but I still think it's valuable here because I think coming to this site would be a common path for people to follow in figuring out the problem.
    – lericr
    Commented Mar 22 at 15:23
  • 1
    Trying to generalize... I think the point of "closing due to simple mistake" has to do with improving the signal-to-noise ratio. People that have "legitimate" questions might end up reading through dead end questions that superficially seem promising. I think the specific example you referenced won't degrade the signal-to-noise ratio. I also think that "simple" is a bit subjective, and that example might not be simple enough to be treated as a "simple mistake".
    – lericr
    Commented Mar 22 at 15:32

1 Answer 1

3

Generally a question about a computation involves a mix of issues related to subject matter and Mathematica functions. I think the emphasis of the simple-mistake criteria is on whether the Mathematica function is doing what it is clearly documented to do, and the OP does not recognize it as such. An OP's mistake might be because of a subject-matter misunderstanding or because they misunderstood (or did not read) the documentation. By "clearly documented" I mean that the documentation is clear, which is not always the case. The boundary dividing a subject-matter misunderstanding into simple mistakes and less close-worthy mistakes is fuzzy. Personally, I expect OP's to have the understanding of someone who took a course as an undergraduate (or in high school or in grad school, as appropriate). For stats, that's a pretty low bar, since someone might have been taught how to do ANOVA in a methods class for their science major with little explanation of the underlying statistics. I'd personally accept a slightly higher bar for stats than that, though.

In the case of the example, P-Values For LinearModelFit Changing Due To Ordering Of Features, I don't find that it is clearly documented what type sum of squares is used in ANOVA, or whether other types are implemented. Maybe it's clear to an expert, which I am not. Because of that, I find that Jim's answer significantly adds to our understanding of how Mathematica works, which is probably the most important purpose of Mma.SE. The question should not be closed.

To propose another perspective, it is probably not good for the site to assume and require that all things not-Mathematica are known by those asking questions. What level of knowledge should be expected is going to remain arguable and vague. SE's voting mechanisms are meant to keep decisions from being arbitrary, so I don't worry about the fuzziness. Personally, I'm surprised at the community's tolerance for the dozen or so questions about plotting tangent lines. Ditto for Newton's method.

1
  • 1
    "To propose another perspective, it is probably not good for the site to assume and require that all things not-Mathematica are known by those asking questions." Excellent. My biased opinion (based on that I mainly deal with Statistics issues) is that statistical knowledge (or lack thereof) is many times the issue rather than anything "wrong" with Mathematica.
    – JimB
    Commented Apr 4 at 16:40

You must log in to answer this question.

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