Interpreting results from Rasch analysis 2. Advanced model applications and the data-model fit assessment

Purpose: The present paper presents developments and advanced practical applications of Rasch's theory and statistical analysis to construct questionnaires for measuring a person's traits. The flaws of questionnaires providing raw scores are well known. Scores only approximate objective, linear measures. The Rasch Analysis allows you to turn raw scores into measures with an error estimate, satisfying fundamental measurement axioms (e.g., unidimensionality, linearity, generalizability). A previous companion article illustrated the most frequent graphic and numeric representations of results obtained through Rasch Analysis. A more advanced description of the method is presented here.Conclusions: Measures obtained through Rasch Analysis may foster the advancement of the scientific assessment of behaviours, perceptions, skills, attitudes, and knowledge so frequently faced in Physical and Rehabilitation Medicine, not less than in social and educational sciences. Furthermore, suggestions are given on interpreting and managing the inevitable discrepancies between observed scores and ideal measures (data-model “misfit”). Finally, twelve practical take-home messages for appraising published results are provided.