|
|
Time-varying Effect in Multilevel Gro... |
|
Message/Author |
|
|
I have three waves of nested assessment data (L1 = time, L2 = school, L3 = district). The model will include some L2 and L3 covariates. I am particularly interested in the relationship between achievement (L1) and some time-varying L1 predictors. I hypothesize that the coefficient for those relationships will change across the three time points. I had considered using a Time-Varying Effects Model (TVEM; https://www.methodology.psu.edu/ra/tvem/) or a piecewise regression model (e.g., Leroux, 2019) but I don't think three waves of data will be sufficient for those models. Any suggestions on how to capture the change in the L1 regression coefficient over time within Mplus? I noted the mention of estimating a time x feature interaction using dummy codes for time in the recent Hamaker & Muthèn (2020) article. Is that a possible solution in my situation? If so, can you point me to any articles/examples of that method being used in practice? Thank you in advance for your time. References: Hamaker, E. L., & Muthén, B. (2020). The fixed versus random effects debate and how it relates to centering in multilevel modeling. Psychological Methods, 25(3), 365–379. Leroux, A. J. (2019). Student mobility in multilevel growth modeling: A multiple membership piecewise growth model. The Journal of Experimental Education, 87(3), 430-448. |
|
|
You can simply do this as a two-level model where the variables measured at the 3 waves are spread out in wide format so that level-1 is now school and level-2 is district. Have a look at UG ex 6.12. |
|
Back to top |
|
|