NONLINEAR AUTOREGRESSIVE LATENT TRAJECTORY MODELS.

Area of Science:

• Statistics
• Psychometrics
• Developmental Psychology

Background:

• Autoregressive latent trajectory (ALT) models integrate latent growth curve and autoregressive models.
• Existing ALT models primarily address linear growth, limiting their application to nonlinear social processes.
• Systematic assessment of nonlinear ALT (NLALT) model identification, comparison, and interpretation is lacking.

Purpose of the Study:

• To introduce and evaluate two forms of nonlinear autoregressive latent trajectory (NLALT) models.
• To address the parameterization, identification, estimation, and testing of these NLALT models.
• To investigate potential biases when fitting alternative models to nonlinear trajectory data.

Main Methods:

• Focus on two NLALT model forms: quadratic growth and latent basis models.
• Discuss model identification, parameterization, estimation, and testing procedures.
• Conduct a simulation study and apply models to empirical weight trajectory data.

Main Results:

• NLALT models offer flexibility for nonlinear growth trajectories.
• Simulation results highlight potential biases from misspecified models.
• Empirical example demonstrates modeling of weight trajectories from birth to age 2.

Conclusions:

• NLALT models provide a robust framework for analyzing nonlinear developmental processes.
• Careful model specification and testing are crucial for accurate trajectory analysis.
• These models enhance the understanding of individual differences in growth patterns.