Introduction

Understanding why something is occurring in a particular location (high levels of a pollutant) or in a population (a measles outbreak) is important to determine which variables may be responsible for that occurrence. Last week, we used cluster analysis to identify positive or negative spatial autocorrelation. This week, we will examine the variables that contribute to why those clusters are there and understand how these variables may play a role in contributing to those clusters. To do so, we will use methods that allow researchers to ask questions about the variables that are found in that location or population and that statistically correlate with the occurrence of an event. One way that we can do this is through the application of correlation and regression analysis. Correlation analysis enables a researcher to examine the relationship between variables while describing the strength of that relationship. On the other hand, regression analysis provides us with a mathematical and statistical way to model that relationship. Before getting too far into the details of both analyses, here are some examples of these methods as they have been used in spatial studies:

Below is a listing of five research studies that use regression/correlation analysis as part of their study. You can scan through one or more of these to see how multiple regression has been incorporated into the research. One take-away from these studies is to recognize that multiple regression is rarely used in isolation. In fact, multiple regression is used as a compliment to many other methods.