I am currently learning about the assumptions of logistic regression and am having a hard time wrapping my head around why independence of observations is necessary for this test. Any guidance would be appreciated.

  • 1
    $\begingroup$ Independence is not necessary, but it leads to a tractable model. I find it illuminating to consider the inverse of this question: if you think the observations are not independent, then exactly what form of independence do you have in mind? $\endgroup$
    – whuber
    Sep 20, 2020 at 21:27
  • $\begingroup$ I marked it as a duplicate of two other threads that seem to answer your question. $\endgroup$
    – Tim
    Sep 20, 2020 at 21:27


Browse other questions tagged or ask your own question.