useR! 2024

Diagnostic classification models are psychometric models that estimate the presence or absence of discrete fine-grained attributes. Due to the categorical nature of the latent variables, assessments using diagnostic models can provide highly reliable results with fewer items, reducing the burden on respondents. In addition, the fine-grained nature of the constructs facilitates the reporting of results that are more informative and actionable than a single overall score. The attributes can represent, for example, student proficiency on educational skills, or the presence of psychological traits or disorders, making these model useful in a variety of contexts. In this session, we will discuss the general properties of diagnostic models and describe how to analyze psychological and educational assessment data with diagnostic models using the R package measr. We will show how the R package measr, which interfaces with the popular Stan language, can be used to easily estimate diagnostic models and evaluate model performance (e.g., model fit, reliability). Finally, we’ll discuss how to draw inferences from the results to answer substantive research questions.