| Research | Sandrah (Sandy) Eckel |
|
Generally, I am interested in modeling interactions, assessing predictions from models
for binary outcomes and machine learning methods such as classification and regression
trees, boosting, and random forests. I have mainly worked on statistical applications
in aging and the health effects of air pollution.
Thesis projects Identifying modifiers of the health effects of ambient air pollution The first component of my thesis centers on developing a method to identify effect modifiers of the relationship between daily mortality and ambient air pollution levels in 50 of the largest cities in the publicly available National Morbidity and Mortality Air Pollution study data. In our two-step analysis, the first step consists of "fractionating" the standard city-specific estimated air pollution effects into month-year-city specific effects. In the second step, we use weighted linear regression and weighted regression trees of the month-year-city air pollution effects on predictors such as temperature, relative humidity, CO, NO2, O3, SO2, season, year, and other city-specific characteristics to identify potential effect modifiers. You can find the working paper associated with this project here. A surrogate measure for the low physical activity component of frailty in older adults Frailty as a modifier of the health effects of ambient air pollution |
home
contact teaching research cv [.pdf] |