Hi all,
I am trying to run a serial multiple mediation on longitudinal data (two time points) and I was wondering whether there is a package I could use in R.
Originally, I was using Mediate, which is convenient because I can use mixed-effects models that cluster my data by participants and account for the longitudinal aspect of it, but it does not allow me to include multiple mediation models in one analysis.
Then I tried Lavaan, which is convenient because I can include both mediators in the model, but it assumed heterogeneity of observations.
I was wondering whether I can specify in Lavaan that my data is longitudinal (e.g. by specifying that the data should be clustered by subject), or whether there is another package that I could use to address this.
Thanks!
Hi,
depending on the model you want to test, you can calculate the differences between the two timepoints yourself and then test the mediation directly on those differences using lavaan. There is a great paper by Montoya & Hayes (2017) on how to test mediations in a repeated-measures context using path analysis (which you can easily replicate in lavaan), you can find it here: https://www.apa.org/pubs/journals/features/met-met0000086.pdf
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling