|Title / Titel||Econometric Analysis of Count Data|
|Abstract (PDF, 14 KB)|
|Summary / Zusammenfassung||The goal of this research project is to analyze discrete data that can take only non-negative integer values, namely dependent count variables. This kind of data is frequently encountered in empirical practice, examples are the number of children in a household, the number of doctor visits, the number of mechanical defects in a production line, or the number of patents applied for at the European Patent Office.
The standard linear model fails to provide a framework that is nearly as rich and complex as is necessary to understand count data. For example, explanatory variables will not only affect the conditional mean of the dependent count variable, but also the conditional variances and each probability itself. This becomes apparent once one recognizes the discrete structure of the data and analyzes them within the framework of full probability models.
Though the field of count data models has been developed for many years, recent developments from semi- and nonparametric econometrics as well as the given progress in IT capabilities seem to have big potentials to refine the existing methods. In particular, we place our emphasis on the interpretation of parameters and models.
|Publications / Publikationen||Winkelmann, R. (2003): Econometric Analysis of Count Data, 4th ed., Springer-Verlag, Berlin.|
|Project leadership and contacts /
Projektleitung und Kontakte
|Funding source(s) /
|Universität Zürich (position pursuing an academic career), Others
|Duration of Project / Projektdauer||Apr 2003 to Oct 2008|