Running a Mixed Model in SPSS with the mixed Command

As part of this series on Mixed Models, Mary Agyapong shared her analytic pipeline to run a mixed model in SPSS using the mixed command.

Our motivation to run this demo was a common misconception of the definition of mixed/multilevel models as a repeated-measures regression only, originating from the fact that SPSS mixed command can be run without specifying the random effect (such model would be a Generalised Estimating Equation (GEE) model, and not a mixed model).

While GEE is a valid method, it will not output the same numerical results of a mixed model run in r with lmer. Therefore, in this presentation, Mary illustrated the steps to specify the random effect in SPSS to obtain the same output in R and SPSS.

As an example, Mary used the sleepstudy dataset, the same used in the Essential Demo of a Multilevel/Mixed Model in R.

To switch on auto-generated subtitles, use the 5th icon from the right bottom corner of the player!

For Mary’s slides and further information see:

Slides

Watch the introductory talk of this series: The Hitchhiker’s Guide to Mixed Models