This chapter focuses on the analysis of rating scale data within the Rasch measurement framework with an emphasis on the Rating Scale Model (RSM) and its evaluation. Rating scales are widely used in educational and psychological measurement, but their effectiveness depends on the functioning of response categories and the appropriateness of the scale structure for the intended construct. The chapter outlines the RSM’s assumptions that include equal category structure across items, ordered thresholds, and invariant measurement. It details diagnostic tools such as category probability curves, threshold estimates, and fit statistics, for assessing how well categories function. Disordered thresholds, category underuse, and overlapping curves are examined as indicators of problematic rating scale functioning. Procedures for refining rating scales are discussed that may include collapsing categories, revising category labels, and aligning rating scale design with the underlying construct. Examples illustrate the application of RSM using real datasets that show how to evaluate category performance, interpret variable maps, and ensure rating scale appropriateness. The chapter concludes that rigorous rating scale analysis can improve measurement precision, construct validity, and fairness. By integrating empirical evidence from Rasch analyses into scale design decisions, practitioners can create more effective, interpretable, and useful rating instruments.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Rating Scale Analysis for Polytomous Explanatory Rasch Models

  • George Engelhard,
  • Stefanie A. Wind

摘要

This chapter focuses on the analysis of rating scale data within the Rasch measurement framework with an emphasis on the Rating Scale Model (RSM) and its evaluation. Rating scales are widely used in educational and psychological measurement, but their effectiveness depends on the functioning of response categories and the appropriateness of the scale structure for the intended construct. The chapter outlines the RSM’s assumptions that include equal category structure across items, ordered thresholds, and invariant measurement. It details diagnostic tools such as category probability curves, threshold estimates, and fit statistics, for assessing how well categories function. Disordered thresholds, category underuse, and overlapping curves are examined as indicators of problematic rating scale functioning. Procedures for refining rating scales are discussed that may include collapsing categories, revising category labels, and aligning rating scale design with the underlying construct. Examples illustrate the application of RSM using real datasets that show how to evaluate category performance, interpret variable maps, and ensure rating scale appropriateness. The chapter concludes that rigorous rating scale analysis can improve measurement precision, construct validity, and fairness. By integrating empirical evidence from Rasch analyses into scale design decisions, practitioners can create more effective, interpretable, and useful rating instruments.