The purpose of the first essay is to determine the impact of the North American Free Trade Agreement on location decisions of U.S. confectionery manufacturers using county level data from 1993 to 2005. A random effects panel data Tobit model is used to test the null hypothesis. The second essay analyzes factors leading to the growth in employment and location decisions of new firms and expansion of existing firms in southeast Oklahoma. A generalized least squares model is used to measure factors affecting employment growth while a heteroskedastic Tobit model is used to measure factors affecting new firm locations and expansions of existing facilities from 1997 to 2005. The final essay focuses on whether or not students' first impressions of instructors and courses are lasting. An ordered probit model is used to test for factors that drive initial impressions and a multinomial logit model is used to determine the factors that lead to observed changes final evaluation scores.

The North American Free Trade Agreement is found to have a negative impact on location decisions of confectionery manufacturers in the U.S. from 1993 to 20005. However, other factors such as income, population, and other food manufacturing establishments are more important. The second essay finds that higher sales tax levels are prohibitive in explaining observed employment growth from 1997 to 2005. Counties in the state of Texas are also more likely to experience employment growth but less likely to attract new firms than counties in Oklahoma. The ten counties in Oklahoma that recently formed a regional consortium are significantly different from all other counties in Texas and Oklahoma at attracting new employment. Areas with higher per capita fire service expenditures and more manufacturing establishments are likely to attract new firms while factors affecting expansion of existing facilities include the population density. The final essay finds that students do change their minds on courses and instructors over the course of the semester. Gender bias is not present in determining initial impressions of an instructor/course although grade expectations can play a role in determining evaluations scores at any point in the semester.