Second-Order Piecewise Growth Model With Unknown Knots for Known Groups or Unobserved Classes

• In Event: Methods to Account for Cluster Dependencies and Subgroup Differences in Structural Equation Modeling

Abstract

Latent growth curve (LGC) analysis is a flexible analytical framework often used to understand individuals’ longitudinal growth. When distinct phases of growth exist, researchers can use piecewise latent growth models to estimate segmented change over time in both observed or latent variables. Additionally, group- or class-specific growth trajectories can be estimated for observed variables using piecewise latent growth mixture or multigroup models. However, current LGC methods cannot estimate group- or class-specific nonlinear growth in a latent construct. This study proposes a multigroup/multiclass second-order piecewise growth model with unknown knots, which allows for the estimation of group- or class-specific segmented change in a latent construct. Annotated Mplus code is provided to make this class of models accessible to researchers.

Authors

• Brennan Register, University of Maryland
  Presenting Author

• Gregory R. Hancock, University of Maryland
  Non-Presenting Author