Dynamic Response Strategies: Accounting for Response Process Heterogeneity in IRTree Decision Nodes

Area of Science:

• Psychometrics
• Cognitive Psychology
• Statistical Modeling

Background:

• Self-reported trait measurements require controlling for response style effects for valid interpretation.
• Traditional psychometric models assume constant trait and response style influences across questionnaire items.
• Respondent motivation decline can lead to heuristic, response-style-driven responding over time.

Purpose of the Study:

• To propose two dynamic IRTree models accounting for item position-dependent trait and response style effects.
• To model systematic continuous changes and random fluctuations in response strategies.
• To provide a more accurate cognitive model for analyzing changes in response strategies over items.

Main Methods:

• Development of two dynamic Item Response Theory (IRTree) models.
• Simulation analyses to evaluate model accuracy in capturing dynamic response processes.
• Application to an empirical dataset to demonstrate benefits over traditional IRTree models.

Main Results:

• The proposed dynamic IRTree models accurately capture dynamic response process trajectories.
• Models reliably detect the absence of dynamics, indicating constant response strategies.
• The continuous dynamic model offers a parsimonious cognitive model for strategy changes.
• The extended model with random fluctuations shows high flexibility and fits data closely.

Conclusions:

• Dynamic IRTree models offer significant advantages over traditional models for analyzing response strategies.
• These models provide a more nuanced understanding of how response processes evolve within a questionnaire.
• The proposed models enhance the validity and interpretation of self-reported trait measurements.