This interesting question was closed at Skeptics:

Is owning a handgun for personal protection justifiable under risk analysis?

The question involves quite a bit of statistics. Is such applied statistics on-topic on CrossValidated?


I think there's a subtle distinction to be made here: This seems like a substantive question that would require statistical analyses to be answered well, rather than a question about statistical analyses and how they can be used to answer substantive questions.

I would say that it's off-topic for Cross Validated. (Although that doesn't mean it's a bad question.)


Personally, I would like to see such questions on CrossValidated. (But I may be biased.)

However, this particular question asks 4 questions in one. So either there will be partial answers, or it needs to be split.

Also, for such issues we need to be very clear (a label like for cigarettes?) that "correlation is not causation" not to make people think that the results mean sth more (or even worse: that we have scientifically proven that weapon (not-) ownership makes someone life longer". This obvious for people here may not be for the general audience.


I agree with gung here, the question isn't suitable. The study would be interesting, and there are a few parts:

  • Come up with some good questions
  • Design your analysis
  • Get some appropriate data
  • Perform analysis
  • Communicate.

The question covers only the first part. Stats.SE covers questions related to the second and 4th part, and perhaps aspects of the 3rd and 5th part. Stats.SE would welcome questions pertaining to how best to gather data related to given questions, and how to analyse that data, but as it stands, the question is about the appropriateness of sociological questions, and that's not something related to statistics.


You must log in to answer this question.

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