Formulation and Application of the Hierarchical Generalized Random-Situation Random-Weight MIRID

Area of Science:

• Psychological measurement
• Statistical modeling

Background:

• The process-component approach is widely used in psychology.
• The model with internal restrictions on item difficulty (MIRID) is a common example.

Purpose of the Study:

• To propose a novel hierarchical generalized random-situation random-weight MIRID.
• To offer a more flexible framework for analyzing endogenous latent variables in multilevel data.

Main Methods:

• Development of a hierarchical generalized random-situation random-weight MIRID.
• Utilizing Markov Chain Monte Carlo algorithms via the WinBUGS software for parameter estimation.
• Analysis of a real dataset on guilt to demonstrate model application.

Main Results:

• The proposed model accommodates polytomous data and complex structures, including item discriminations, random situations, random weights, and heteroskedasticity.
• Successful estimation of model parameters using WinBUGS.
• Demonstration of the model's utility through a guilt dataset analysis.

Conclusions:

• The hierarchical generalized random-situation random-weight MIRID provides a flexible and powerful tool for psychological research.
• The model allows for more nuanced analysis of complex psychological phenomena within a multilevel framework.