Title: Recursive Partitioning of Growth Curve Models Using LM(M) Trees

Author: Marjolein Fokkema

Affiliation: Leiden University

Abstract:
Growth curve models are a popular tool to address to address questions of
stability and change over time. Often, researchers may have a specific
interest in detecting subgroups which differ in terms of their rate of growth,
or detecting predictors of initial levels or growth over time. Several
recursive partitioning methods allow for detecting such subgroups and
predictors, like for example SEM Trees (Brandmaier et al., 2013), longRpart
(Abdollel et al., 2002) and linear (mixed-effects) model trees (L(M)M trees;
Zeileis et al., 2008; Fokkema et al, 2018). This presentation will focus on
the latter method. The application of LM(M) trees will be illustrated using a
dataset on reading trajectories of children in kindergarten. Results on a
simulation study assessing the accuracy of LM(M) trees will be presented.
Furthermore, some (dis)advantages of the different recursive partitioning
methods for growth curve models will be discussed.

References:
Abdolell, M., LeBlanc, M., Stephens, D., & Harrison, R. V. (2002). Binary
partitioning for continuous longitudinal data: categorizing a prognostic
variable. Statistics in medicine, 21(22), 3395-3409.

Brandmaier, A. M., von Oertzen, T., McArdle, J. J., & Lindenberger, U. (2013).
Structural equation model trees. Psychological Methods, 18(1), 71.

Fokkema, M., Smits, N., Zeileis, A., Hothorn, T., & Kelderman, H. (2018).
Detecting treatment-subgroup interactions in clustered data with generalized
linear mixed-effects model trees. Behavior Research Methods, 50(5), 2016-2034.

Zeileis, A., Hothorn, T., & Hornik, K. (2008). Model-based recursive
partitioning. Journal of Computational and Graphical Statistics, 17(2),
492-514.