Retired UvA econometrician publishes in leading journal Psychometrika
A team of long-retired researchers has succeeded in publishing a methodological paper in the leading journal Psychometrika. The article introduces a new and much faster method for estimating complex statistical models that are widely used in psychology, sociology and economics.

Econometrician and University Professor Bernard van Praag (85) was, among other roles, professor at Erasmus University Rotterdam, director of the SEO Economic Research institute (1992-2001), and one of the founders of the Tinbergen Institute. He is affiliated with the Amsterdam School of Economics at the University of Amsterdam.
The late Peter Hop worked for many years as a programmer at SEO Economic Research. William Greene is the author of the bestselling textbook Econometric Analysis, which has appeared in 8 editions to date.
In their article, Van Praag and co-authors develop a new method for estimating Multivariate Ordered Probit models. These models are used to analyse multiple, related ordered outcomes at the same time, such as response categories on questionnaires or rating scales.
The new approach is based on generalized method of moments estimation. It makes it possible to estimate systems of simultaneous equations with correlated errors that are very difficult or sometimes practically impossible to handle with the usual Maximum Likelihood methods. Where conventional Maximum Likelihood estimation can require several hours of computing time for such models, the new method typically produces results within a few minutes.
The method is also useful for Item Response analysis, which forms the basis of psychological test theory and test design. In addition, it can be applied in the sociological analysis of structural models, where complex relationships between latent and observed variables are studied.
Publication details
Estimation of Linear Models from Coarsened Observations: A Method of Moments Approach, published 10 March 2025 in Psychometrika.
Authors: Bernard van Praag, J. Peter Hop, and William H. Greene.