STRUCTURAL ANALYSIS OF SUBJECTIVE CATEGORICAL DATA

摘要: A general approach to the analysis of subjective categorical data is considered, in which agreement matrices two or more raters are directly expressed terms error and parameters. The method provides focused analyses ratings from several for whom have measurement distributions that may induce bias evaluation substantive questions interest. Each rater's judgment process modeled as a mixture components: an variable unique rater question well operationalizes “true” values units observation. statistical problems identification, estimation, testing such models discussed.