Rasch Measurement Model
摘要
Rasch analysis is increasingly used for developing and evaluating interval-level scales to measure abilities. Unlike Classical Test Theory (CTT) that uses correlations between items to select items measuring a latent construct, Rasch analysis applies a statistical model that predicts the precise mathematical relationship between an item and a latent trait to be measured. Proponents of Rasch analysis argue that, compared to the typically long scales with ordinal scores that are produced by CTT methodology, application of Rasch analysis results in shorter, more reliable measures with interval-level scores. With increased application of mindfulness-based interventions, evaluation of their effectiveness requires accurate measurement of mindfulness and associated outcomes. In particular, issues with measurement precision (e.g., ordinal rather than interval scaling), item functioning, and invariance across population groups have not been sufficiently addressed using appropriate modern statistical methods. This chapter focuses on Rasch methodology and its benefits to evaluate and enhance the psychometric properties of mindfulness scales and related outcome measures. It emphasizes the value of Rasch analysis to enhance both reliability and internal validity of psychometric instruments used in mindfulness assessment.