Читать книгу GIS Research Methods - Steven J. Steinberg - Страница 87

When explanation is the goal of your research, you are seeking to examine relationships between variables. A researcher conducting an explanatory study is interested in developing a scientifically based understanding of why things happen. Explanatory studies still involve description, but with a different goal in mind. One goal might be to develop an understanding of connections that explain why something occurs in some locations but not in others. How does GIS fit into this? It enables the researcher to consider spatial relationships as one potential component that helps to explain why an observed relationship occurs.

For example, let’s say that you are interested in understanding why crime rates are higher in some neighborhoods than others. Is it due to socioeconomic issues, such as income and education? Are there spatial relationships, such as proximity to ATMs or convenience stores? Or are there negative correlations to locations of police stations, well-lit streets, or other physical factors? As part of your study, you could create a geographic picture of these variables within the areas that experience varied crime rates. The GIS will show you where the different variables intersect with higher or lower crime rates and quite possibly show that certain variable combinations have a stronger correlation with crime than others. To avoid committing the ecological fallacy (see discussion later in this chapter), you might wish to follow up the analysis by conducting interviews or surveys of residents and viewing trends in crime in the context of the national economy and crime in other areas to develop a complete understanding of why these higher crime rates occur in specific types of neighborhoods.

Of course, one must be cautious when using a GIS to avoid the temptation to draw conclusions based solely on appearance. Although visualization of the data is a powerful method for showing relationships, one must be careful to follow this up with sound analysis—be that spatial analysis in the GIS, spatial and nonspatial statistical analyses, or other means appropriate to the data. To help in this process, we provide step-by-step guidance for two different approaches to research: deductive and inductive.