I have 8 subjects. They were all tested under 2 different conditions, 5 sessions of each, with 14 trials in a session. Total data points: 1120. No data is missing. 2 dichotomous DVs were measured every trial (Did they make a response? and Did they do the task correctly?). I want to know if responding predicts correct task performance, or at least if they correlate. Would I need to use mixed-effects modeling?


1 Answer 1


Yes, you want some sort of mixed effect model but questions about SPSS code are off-topic here.

I prefer the term multi-level model; you need one because the data are not independent.

  • $\begingroup$ Apologies for the SPSS question. It's been deleted. But a bit more on the sort of stats I should be using would be helpful. Could I do a multi-level logistic regression, for example? $\endgroup$
    – Anna Casey
    Commented Feb 18, 2017 at 20:28
  • $\begingroup$ Yes, since the DVs are dichotomous, that seems right. $\endgroup$
    – Peter Flom
    Commented Feb 19, 2017 at 14:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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