A Social Desirability Item Response Theory Model: Retrieve-Deceive-Transfer.

Area of Science:

• Psychometrics
• Statistical Modeling
• Survey Methodology

Background:

• Traditional item response theory (IRT) models may not fully capture response biases.
• Understanding and correcting for respondent deception is crucial for accurate opinion measurement.

Purpose of the Study:

• To develop a novel IRT model that accounts for respondent overreporting and underreporting.
• To simulate deception and transfer mechanisms within a psychometric framework.

Main Methods:

• Incorporation of a deception term and a transfer term into a multidimensional rating scale model.
• Utilizing a Bayesian framework with Markov chain Monte Carlo (MCMC) algorithms for parameter estimation.
• Validation through simulation studies and an online experimental study.

Main Results:

• The proposed IRT model accurately recovers parameters in simulation studies.
• Demonstrated the practical applicability of the methodology in a real-world online experiment.

Conclusions:

• The developed IRT model provides a robust approach to address response biases in surveys.
• This methodology enhances the validity of opinion measurement by accounting for deceptive response patterns.