There are varied answers when building models for logistic regression. For selection of predictors to build multivariable model :

  1. should one choose from univariable analysis or stepwise Or
  2. first do univariable analysis, then add significant associations by stepwise approach to build multivariable model ?
  • 1
    $\begingroup$ Why not take all of the features? $\endgroup$
    – Dave
  • $\begingroup$ You pose a false dichotomy. We have many posts on model building -- and a large number of them provide methods considered to be superior to stepwise methods. We also have many posts that point out the (sharp) distinction between univariate regressions and multiple regression. They imply that neither of your strategies is appropriate. $\endgroup$
    – whuber


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.