Item Response Theory (IRT) provides a robust framework for analysing phenomena that cannot be directly observed. In such cases, inference relies on observable behaviours, such as responses to questionnaire items or survey indicators related to the phenomenon of interest. This talk will introduce key IRT concepts, with a particular focus on extensions suited for analysing complex phenomena. Specifically, the benefits of modelling the latent trait with a discrete distribution will be
discussed, which, unlike traditional IRT models based on a continuous trait, offer a more flexible representation of heterogeneous populations. Moreover, the talk will cover the concept of multidimensionality, which allows for the representation of multiple latent traits through a multivariate latent variable approach. This is particularly useful when
analysing constructs that are inherently composed of several interrelated dimensions. The presentation will conclude with two real-world applications: assessing the risk profiles of healthcare contracting authorities in public procurement management and estimating educational poverty levels in Italy by integrating small area estimation with multidimensional IRT.
21 marzo 2025, ore 12.00
Simone Del Sarto
Department of Political Science, University of Perugia
In person: Room 24 (4th floor) building CU002 Scienze Statistiche
Webinar: https://uniroma1.zoom.us/j/83625004899?pwd=bXCtz0mp759PUh2lkqT0BUoVa0Uegg.1
ID riunione: 836 2500 4899
Passcode: 123456