On the Origin of Species: The evolution of diagnostic modelling within the psychometric taxonomy
State of the Art talk by Jonathan Templin, Department of Research, Evaluation Measurement and Statistics, University of Georgia, USA.
Chair: Terry Ackerman, Tuesday 21st July, 12.20 - 13.05, Palmeston Lecture Theatre, Fisher Building.
Diagnostic models have seen an increased research focus in recent years, predominantly used in the measurement of cognitive abilities. As latent variable models that utilize categorical latent variables to classify individuals, diagnostic models link the diagnostic status of an individual with observable outcomes such as item responses or reported behaviours. The use of diagnostic models can provide highly informative feedback as the current diagnostic status of an individual, feedback which may streamline the process of remediation (in educational contexts) or recovery (in psychiatric contexts).
In this talk, I present the diagnostic modelling paradigm from a state-of-the-art perspective linking existing diagnostic models with more commonly used psychometric methods. To further place diagnostic models in the psychometric taxonomy, I highlight the statistical properties of diagnostic models, yielding insights into the types of situations where diagnostic models are most effectively applied. The talk culminates with a review of ongoing research and unresolved issues in diagnostic modelling, providing my thoughts as to the current and future direction of the field.