The skeptics SE gets a lot of questions related to interpretation of crime statistics in one way or another. Each of the ones that I can see refers to a specific statement, but the same kind of issues come up over and over again.

I was considering an ask-and-answer along these lines:

Q: What factors should I consider when evaluating a claim about crime statistics?

A: Here are some of the most important considerations:

  • How are the relevant crimes defined? If the definition is different between time periods or jurisdictions, this will affect data. [insert examples here]
  • How is the reference period determined? For example, if a serial killer has murdered 200 people over 20 years, but the murders are only discovered in 2015 and the killer is only convicted in 2017, when do those murders appear in the time series?
  • etc. etc. (cherry-picking of time series or crime types, per-capita vs. raw numbers, reporting rates, assumptions of causality, ...)

Would a meta-question & self-answer of this kind be appropriate here? I haven't seen any others of this kind on skeptics, and I'm not sure whether that's due to a community rule that I've missed, or just because people aren't posting them.


You must log in to answer this question.

Browse other questions tagged .