Erik Nordheim

Professor of Statistics and Sociology
1110 Med Sci
(608) 263-5812
nordheim@stat.wisc.edu

Research Interest Statement
As leader of CDE’s Statistics subcore, Nordheim provides provided statistical consulting to researchers in CDE. He has recently worked with Prof. Giovanna Merli on a study of respondent driven sampling (RDS), to support her ongoing research on female sex workers in Shanghai China. The findings to date suggest that violations of the rather stringent assumptions underlying RDS can lead to inaccurate estimation with this method. Thus, substantial improvement in methods for sampling hidden populations is needed. Efforts in this direction are underway. A second project is work with Lane Burgette, a graduate student supported by CDE, developing new methods for Bayesian multinomial probit switching models. This work, which was motivated by a problem stemming from research being done by Prof. James Raymo on individuals’ future interests regarding work status (full-time, part-time, not working). The solution to this problem has required extending some of the existing selection model work in new directions with quite intricate modeling using Markov Chain Monte Carlo methods. This work expands the statistical toolbox for complex switching models.

Nordheim has organized several workshops on statistical issues relevant to demography research (see Progress Report), such as propensity scoring and Bayesian Model Averaging. These are aimed at researchers in CDE along and also in allied fields.

CDE Research Theme Working Groups
Data and Methodology