I have a dataset of 300+ students with teacher-reported data. There are 17 total teachers that reported data on the students.

I would like to run a model that tests mediation where the mediator is a teacher-reported variable (the independent and dependent variables are observed and child-report, respectively).

I know how to run this type of model in SPSS and MPlus, but I'm not sure how to structure the data to account for the nestedness of students within teachers. Is there a line of syntax that indicates the clustered nature of the data?

  • $\begingroup$ My apologies, I should have clarified. I know how to run mediation models in data that is not nested. Also, I incorrectly stated that students are nested within teachers; it's actually the case that students are nested within classrooms and classrooms are nested within teachers (three-level model). Any guidance is appreciated! $\endgroup$ Oct 7, 2015 at 4:01
  • $\begingroup$ How many teachers and classrooms do you have? A three level model with mediation is going to be potentially challenging! $\endgroup$ Oct 7, 2015 at 4:03
  • $\begingroup$ P.S. My first comment was a little brusque, on rereading. I deleted it. $\endgroup$ Oct 7, 2015 at 4:03
  • $\begingroup$ Have you come across this paper: quantpsy.org/pubs/preacher_2011.pdf $\endgroup$ Oct 7, 2015 at 4:05
  • $\begingroup$ Thanks for the great paper reference, that will be really helpful! You're absolutely right - a three-level model may be a bit too ambitious... $\endgroup$ Oct 8, 2015 at 1:11


Your Answer

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

Browse other questions tagged or ask your own question.