Dr. Matthias von Davier (Princeton University)

Programm

• Datum: 26.05.2011
  Zeit: 15:15 - 16:45
  Ort: Raum 211b, Front building, Ludwigstr. 28

A general latent variable framework for complex assessments

The recent convergence of item response based models and factor analytic approaches has lead to developments that allow picking and choosing model features based on considerations of how to best reflect the construct of interest. Measurement has evolved from mere 'response counting' to explicitly modeling the hypothesized relationship between task performance and underlying individual differences based on substantive theory. The talk will present a family of multidimensional latent variable models and give an overview of applications of this approach to the analysis of large scale survey assessment data and data from personality assessment. Promises and limitations will be discussed as well as issues around model selection.