Accommodating Continuous Time Metrics within the Discrete-time Latent Change Score Model Using Definition Variables.

Area of Science:

• Statistics
• Developmental Psychology
• Epidemiology

Background:

• Longitudinal models traditionally use a single time metric to assess change.
• The COVID-19 pandemic highlighted the need to model changes across distinct phases while accounting for age.
• Existing methods may not precisely capture complex temporal dynamics.

Purpose of the Study:

• To extend the discrete-time latent change score modeling framework.
• To precisely model wave-to-wave changes by incorporating continuous time metrics.
• To account for age and pandemic phases simultaneously in longitudinal analyses.

Main Methods:

• Proposed an extension to discrete-time latent change score modeling.
• Included continuous time metrics by regressing out initial age.
• Utilized definition variables instead of age bins.
• Applied the model to adolescent marijuana expectation data.
• Conducted a simulation study to compare the approach with existing models.

Main Results:

• The proposed model offers a more precise way to analyze longitudinal data compared to traditional methods.
• Simulations demonstrated the advantages of incorporating continuous time and definition variables.
• The model effectively captures changes influenced by pandemic phases and age.

Conclusions:

• The extended latent change score model provides a robust framework for analyzing complex developmental changes.
• This approach enhances the precision of longitudinal modeling, especially during dynamic periods like a pandemic.
• The findings have implications for understanding adolescent development and substance use trajectories.