Psychometric properties of the general conspiracy belief scale using item response theory

Evaluation of the psychometric properties of conspiracy theory belief instruments has been dominated by classical approaches with limitations, especially in dependence on sample size and inaccuracies in item-level analysis. This study aims to fill this gap by applying a polytomous Item Response Theory (IRT) approach to reanalyze the General Conspiracy Belief Scale (GCBS). This study aims to re-examine the psychometric properties of the GCBS with an IRT approach to produce measurements that are more precise and independent of sample characteristics. The research design used was a quantitative replication utilizing secondary data from 2,495 students at the college level. The instrument used consisted of 15 items on a five-category Likert scale. The analysis was conducted using three polynomial IRT models, namely the Graded Response Model (GRM), Partial Credit Model (PCM), and Generalized Partial Credit Model (GPCM), with the help of R software. The results showed that the GRM model was the model that best fit the data, with most items showing high distinctiveness and providing maximum information on respondents with low to moderate levels of conspiratorial belief. Empirical marginal reliability coefficients were high, indicating that the instrument's internal consistency was perfect. This study contributes to the field by offering a more robust and nuanced psychometric evaluation of the GCBS through IRT, providing researchers with a validated framework for assessing conspiracy beliefs with higher accuracy and scale precision. However, the limitation of this study lies in the use of secondary data sourced from one particular population group, so the generalizability of the findings still needs to be further examined in a more diverse context.