Georgiana Onicescu and Yuqian Shen
Department of Statistics
Western Michigan University
Abstract:
The main purpose of this study was to explore the spatial distribution of radon testing patterns at the county level
in the lower peninsula of Michigan and assess its temporal variations, in order to bring awareness about radon and
help determine where the radon interventions should be targeted.
We used a spatial regression model, taking into account the spatial dependence of the data at the county levels.
Several models were considered and compared.
Models including random effects greatly outperformed the currently used radon zoning variable to assess the radon spatial variations. Our best model included county level unstructured heterogeneity random effects and time period effects. Compared to period 1 (1994-2000), there was a .0721 (95% CI=(.6215, .8808)), .1707 (95% CI=(.1098, .2315)) and .2137 (95% CI=(.1529,.2744)) increase in log radon measurements for period 2 (2001-2005), period 3 (2006-2010) and period 4 (2011-2016) respectively. Estimated spatial maps of the spatial random effects showed a great variability, with the majority of counties with higher radon testing levels being found in the southeast part of the state.
---
All statistics graduate students are expected to attend.