I want to ask a question with the title Is there a canonical example demonstrating the difference between predictive and explanatory models?

I have ideas about the body of this question, with some references, but have not yet written it, although I think the title gets at the meat of it. However, I notice that we have neither an [example] tag, nor a [pedagogy] tag which makes me wonder if this kind of question might be appropriate, or if it would be too broad or too subjective a question.

PS I would welcome references which you think might be useful for me to look at on the topic of my intended question posted as comments. :)

  • 2
    $\begingroup$ It would certainly be on-topic to have a question about the difference between them and give your hopefully canonical answer but I sense you have something else in mind? $\endgroup$
    – mdewey
    Jul 28, 2019 at 10:54
  • 2
    $\begingroup$ Be aware that there are a number of existing threads on this issue, so your question may be a duplicate (eg, 1, 2, 3, 4, etc). $\endgroup$ Jul 28, 2019 at 12:48
  • 1
    $\begingroup$ When I first read the title, I figured this would be about questions seeking a tutorial, which (for better or worse) isn't something we do. Your specific question elaborated in the text is different from that. But a concern would be how could you choose which is the 'right answer' from several such posted examples? If you're not asking for code, you're not asking for a tutorial; your Q isn't a duplicate, & it's easy to see how a particular response is the correct answer, I don't see any reason such a question should be closed. $\endgroup$ Jul 28, 2019 at 12:53
  • 1
    $\begingroup$ @gung Thank you for the links. While all of these discuss prediction vs explanation, none of these provide model examples, which would be a primary motivation driving my question. $\endgroup$
    – Alexis
    Jul 28, 2019 at 15:52
  • 2
    $\begingroup$ @gung "how could you choose which is the 'right answer' from several such posted examples?" From my dictionary, 3rd definition of 'canonical': "reduced to the simplest and most significant form possible without loss of generality". There might not be such an answer. One the other hand, there might be: if I were to ask for a canonical example of a measure of data centrality, I could see acceptable answers using the mean, median or both. $\endgroup$
    – Alexis
    Jul 28, 2019 at 15:57


You must log in to answer this question.

Browse other questions tagged .