PC program for comparing tracking indices in several independent groups

Area of Science:

• Biostatistics
• Longitudinal Data Analysis
• Growth Modeling

Background:

• Previous work established Cohen's kappa for tracking in one-sample longitudinal data.
• Comparing tracking stability across different independent groups was not previously addressed.
• Assessing growth pattern stability is crucial for understanding developmental trajectories.

Purpose of the Study:

• To present a statistical method for testing the equality of Cohen's kappa across multiple independent longitudinal samples.
• To enable formal comparisons of growth or adaptive pattern stability between different populations.
• To provide a method for estimating a common kappa value and its confidence interval when appropriate.

Main Methods:

• Development of a statistical test for the equality of several independent Cohen's kappa coefficients.
• Implementation of a procedure for estimating a common kappa value and constructing confidence intervals.
• Utilized a menu-driven GAUSS program for executing the described statistical methods.

Main Results:

• The paper demonstrates how to formally compare Cohen's kappa across independent longitudinal samples.
• The method allows for relative assessments of predictability in growth outcomes across diverse populations.
• A procedure for estimating a common kappa and its confidence interval is presented and validated.

Conclusions:

• The proposed method facilitates robust statistical comparisons of tracking stability in longitudinal growth patterns.
• This approach enhances the ability to assess and compare developmental predictability across different groups.
• The availability of a GAUSS program aids in the practical application of these statistical techniques.