I aim to assess the effects of difficulty (continuous variable) and trial type (0/1) on whether a subject has been correct in a logistic regression model. However, I have also measured subjects impulsivity scores, which is a continuous variable. For impulsivity I only have one estimate per subject, for the other variables I have trial type dependent estimates.

So far I have estimated a logistic regression model per subject, normalized all coefficients by the vector norm of the weight vector per subject and correlated the regression coefficients with the impulsivity scores on a group level. I wondered whether this is a reasonable approach? I could also estimate a single mixed effect model and incorporate subject as a random factor. However, there is again no way to incorporate the continuous variable impulsivity in that model, right? How would you proceed with that problem?


1 Answer 1


In a mixed effects logistic regression you can include both categorical and continuous covariates that may or may not change within a subject.

  • $\begingroup$ So you would add a vector of num trials*num_subjects that just repeats the impulsivity score trial times per subject? $\endgroup$
    – Carlos
    Aug 20, 2019 at 8:27
  • $\begingroup$ It would be the same as including the sex of the subject, i.e., it remains the same for all measurements of a particular subject. $\endgroup$ Aug 20, 2019 at 10:05
  • $\begingroup$ Yes. Thanks for your help! I will try that. $\endgroup$
    – Carlos
    Aug 20, 2019 at 11:55

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.