Optimal two-time point longitudinal models for estimating individual-level change: Asymptotic insights and practical...

Area of Science:

• Developmental psychology
• Quantitative psychology
• Longitudinal data analysis

Background:

• Parsons and McCormick (2024) suggested two-time point models inadequately capture individual differences in linear slopes.
• Their simulation study indicated limitations in modeling individual change with limited measurement occasions.

Purpose of the Study:

• To deconfound the effects of study duration and measurement frequency on estimating individual differences in linear slopes.
• To demonstrate that appropriate study design can overcome limitations of two-time point models.

Main Methods:

• Utilizing asymptotic results to evaluate the precision of linear change models.
• Comparing the precision of various study designs, including those with different time spans and measurement intervals.
• Incorporating considerations for irregularly spaced data and missing data points.

Main Results:

• The primary factor increasing precision in estimating individual differences in linear slopes is the study's time span, not the number of measurement occasions (waves).
• Asymptotic results provide a framework for analyzing models with varying interval lengths and missing data.
• Longer study durations significantly enhance the ability to model individual change accurately.

Conclusions:

• Two-time point models can effectively capture true linear change at the individual level with careful study design.
• Maximizing study duration is crucial for precise estimation of individual differences in linear slopes.
• The findings challenge previous conclusions about the inadequacy of two-time point models for individual-level change analysis.