Modeling (in)congruence between dependent variables: The directional and nondirectional difference (DNDD) framework

Area of Science:

• Psychometrics
• Statistical Modeling
• Quantitative Psychology

Background:

• Incongruence between dependent variables is a common challenge in statistical analysis.
• Existing methods for modeling differences have limitations in capturing nuanced relationships.

Purpose of the Study:

• To introduce and validate a new approach, Directional and Nondirectional Difference (DNDD), for modeling incongruence.
• To compare the DNDD approach with existing methods for analyzing differences between variables.

Main Methods:

• Decomposition of incongruence into orthogonal directional and nondirectional components.
• Monte Carlo simulation to examine circumstances of difference emergence.
• Application of the DNDD approach to a field dataset.

Main Results:

• The DNDD approach provides richer insights into the antecedents of incongruence compared to arithmetic, absolute, or squared differences.
• Simulation results illustrate the distinct information provided by each difference component.
• The field dataset example demonstrates the practical utility of the DNDD method.

Conclusions:

• The DNDD approach offers a more comprehensive framework for understanding variable incongruence.
• Proposed extensions include modeling with a known target value and using multilevel analysis.
• The DNDD method has potential for various practical applications in psychological research.