Search

Your search keyword '"Köll, Susanne"' showing total 3 results
Start Over Author "Köll, Susanne"
3 results on '"Köll, Susanne"'

1. Bias Reduction as a Remedy to the Consequences of Infinite Estimates in Poisson and Tobit Regression

Author

Köll, Susanne, Kosmidis, Ioannis, Kleiber, Christian, and Zeileis, Achim

Subjects
Statistics - Methodology, Statistics - Computation
Abstract

Data separation is a well-studied phenomenon that can cause problems in the estimation and inference from binary response models. Complete or quasi-complete separation occurs when there is a combination of regressors in the model whose value can perfectly predict one or both outcomes. In such cases, and such cases only, the maximum likelihood estimates and the corresponding standard errors are infinite. It is less widely known that the same can happen in further microeconometric models. One of the few works in the area is Santos Silva and Tenreyro (2010) who note that the finiteness of the maximum likelihood estimates in Poisson regression depends on the data configuration and propose a strategy to detect and overcome the consequences of data separation. However, their approach can lead to notable bias on the parameter estimates when the regressors are correlated. We illustrate how bias-reducing adjustments to the maximum likelihood score equations can overcome the consequences of separation in Poisson and Tobit regression models., Comment: 8 pages, 8 figures

Published
2021

2. Issues in microeconometric modeling : two-part models, cluster correlations, and data separation

Author

Köll, Susanne and Köll, Susanne

Abstract

Issues that may occur in microeconometric count data modeling are examined within the framework of this thesis. Starting with count data modeling of the innovation process, two-part hurdle models are used to account for potentially differing determinants driving the “first innovation” and “continuing innovation” along with clustered covariances in a quasi-maximum likelihood approach. Continuing with an object-oriented implementation of clustered and panel covariances in the R package sandwich, and based on the new software, providing for empirical illustrations and a Monte Carlo simulation study to assess the performance of various types of clustered covariances for count data and other models beyond linear and generalized linear models. And as a final point, looking at circumstances where data separation occurs in microeconometric models (specifically in Poisson and Tobit regression models), because in such cases the quasi-maximum likelihood approach is not an adequate strategy. An illustration and a Monte Carlo simulation experiment is provided to assess the performance of bias-reducing adjusted score equations as a remedy for separation in Poisson and Tobit models., Susanne Köll, Kumulative Dissertation aus drei Artikeln, Dissertation University Innsbruck 2021

Published
2021

Catalog

Books, media, physical & digital resources
See catalog results