A parameter estimation method using an item response model considering social desirability

Abstract

In psychological research, rating scale methods such as the Likert scale are often used to measure specific constructs. However, social desirability is known to influence the reliability and validity of psychometric measures. Hattori (2007) proposed an item response model to correct for the influence of social desirability. This model simultaneously considers social desirability as an item characteristic and social desirability response tendency as an individual characteristic. This study applied Hattori’s model to real data and estimated parameters using the Markov Chain Monte Carlo method to confirm the model under actual conditions. The model’s validity was demonstrated in terms of both social desirability as an item characteristic and social desirability response tendency as an individual characteristic under actual conditions.