Yvonne Palm and Tue To
"Estimating Asthma Prevalence Rates in the United States"
Advisors: Shonda Kuiper (KuiperS@grinnell.edu), Tom Moore (MooreT@grinnell.edu), and EPA client: Thomas Brody
The purpose of this research is to develop models to predict asthma prevalence rates in the United States in order to identify areas in need of intervention by the Environmental Protection Agency. This project incorporates geospatial demographics with National Health Interview Survey (NHIS) responses to produce models that estimate asthma prevalence rates at the census tract and county level. Three of the resulting models based on small area analysis and logistic regression are compared. These modelsí predictions are then compared to selected sample county data from the Center for Disease Controlís Behavioral Risk Factor Surveillance System (BRFSS) in an attempt to supply better localized estimates of the asthma epidemic. Finally, the differences between the model predictions and the BRFSS data are compared to environmental pollutants of concern across the country to identify areas where intervention may be successful.
Lin Ji and Demetrio Rojas
"A Geographical Approach to Analyzing Student Admission Patterns"
Advisors: Shonda Kuiper (KuiperS@grinnell.edu), Tom Moore (MooreT@grinnell.edu), and Institutional Research client: Scott Baumler
Our objective is to study geographic, demographic and socio-economic factors affecting student admission at Grinnell College. Our focus is on county demographics and high school characteristics to predict levels of student interest throughout the nation. We use multivariate regression techniques to show that areas with similar socio-economic and demographic characteristics are likely to have similar interests in Grinnell College. After separating the US into four regions (West, Northeast, South and Midwest), we use stepwise regression models to identify these key socio-economic and demographic factors affecting the number of students interested in Grinnell College. Applying this model to all of the high schools and counties (including those high schools and counties that have shown little or no activity), we predict student admission patterns at Grinnell College. The variables that are taken into account are based on high school records from the College Board, individual student records from the Office of Admissions at Grinnell College, and socio-economic and demographic data from census tracts, elections and national surveys.