Title: James-Stein Estimators in Factor Analysis

Author: Elissa Burghgraeve

Affiliation: Ghent University

Abstract:
In the social and behavioral sciences, latent variables are ubiquitous. The
conventional factor analysis model relating these latent variables with their
indicators, assumes that the relation between the latent variable and its
indicators is linear. An assumption that might not be fulfilled. We propose a
flexible way to estimate linear and nonlinear factor models relying on
semiparametrics. We consider James-Stein estimators which were originally
proposed for measurement error models. We apply them in our setting of
estimating parameters in factor models and extend them to accommodate for
nonlinear relations.

References:
James, W. and Stein, C. (1961). Estimation with quadratic loss. Proceedings of
the Fourth Berkeley Symposium on Mathematical Statistics and Probability, 1,
361-379.

Whittemore, A.S. (1989). Errors-in-Variables Regression Using Stein Estimates.
The American Statistician, 43(4), 226-228.