Comparison of IRT, CFA, and CTT

  CTT provides a method of developing and evaluating a test or measure, however, respondent and item characteristics are inseparable and contribute to measurement error. CFA is a subset of Structural Equation Modeling (SEM) and is focused on the specification of a measurement model for a construct. CFA allows the researcher to specify the relationships among the items and the measure or measures, as well as the relationships among the constructs being measured. IRT allows for a clear evaluation and separation of the items that make up a measure and the respondents that provided data for a specific measure. While these approaches may appear to be vastly different, they share many similarities and often result in very similar conclusions being drawn about a psychological measure. Each of these approaches also has specific advantages and disadvantage