ABSTRACT
This chapter provides diagnostic information regarding different facets of rater-mediated assessments that define parts of the data that may not be functioning as intended. In order to diagnosis sources of misfit, there are a variety of tools that can be used, such as the residual analyses. The chapter focuses on residual-based approaches for examining model-data fit. This approach provides detailed diagnostic information that can be used to explore the psychometric quality of the ratings. The numerical summaries of fit indices combined with graphical displays can form a starting point for interpreting lack of model-data fit. The Wright map embodies the calibration of the persons, cues, rating scale categories, and raters on the logit scale that represents the construct. There are two major ways to present model-data fit: numerical and graphical displays. The expected ratings are based on the polytomous Rasch model with a variety of facets added as required by the particular assessment system.
