GIS Research Methods. Steven J. Steinberg

Do any geographic features of the area hold special meaning for the people who live there?

      

Are there boundaries associated with your data? How are these defined?

      

Are the data stationary through the timeframe of your analysis, or do they move? How will you account for these changes?

       Questions for analysis

      The following questions will aid in conducting your analysis. They focus on issues of boundaries, time, and place. All of these issues are discussed in greater detail in later chapters, but they are important to have in mind early on. A GIS is useful for conducting a comparative study of groups of people, for example, comparing two neighborhoods of a city. It’s up to the researcher to determine the boundaries of a study. Boundary choice is essential, because changing the boundaries of an analysis can drastically alter the results obtained. Therefore the boundaries of your study are best dictated by the research question you are investigating and should be set in advance, as opposed to during the analysis phase.

      Trying to determine whether geographic variables will be useful to your study can be tricky, but keep the following in mind: almost all units of analysis have an associated geographic location. Social scientists are often interested in measuring basic socio demographic variables, such as age, income, gender, and ethnicity. These core variables can then be associated with geographic locations, such as neighborhoods, cities, and states. The possibilities are limitless. Anything that involves geographic boundaries can be studied using GIS. Sometimes boundaries are artificially created by the researcher, depending on the researcher’s goals. At other times, the boundaries already exist and are based on manufactured geographic features (such as streets) or natural geographic features (such as rivers, lakes, and mountains).

      The issue of boundaries is a very important one for GIS. One thing to keep in mind is that you are the boss when it comes to your data collection and the type of information that you want to include in your study. Therefore, the notion of what are appropriate boundaries in your study may require you to look back at your research questions or hypotheses. It is also important to consider whether you are dealing with socially constructed boundaries, such as political boundaries (e.g., city limits or state lines) or physical boundaries (e.g., mountain ranges, oceans, or rivers). Sometimes the two coincide, but often they do not. Sometimes we use boundaries that are defined more fluidly and conceptually (e.g., a traditional hunting ground, the heart of the community, or a gang’s territory).

      Questions for analysis include the following:

      

What is the primary unit of analysis in your study?

      

What are the physical boundaries for your unit of analysis?

      

What are the social (conceptual) boundaries for your unit of analysis?

      

Are your conceptual boundaries social, philosophical, or economic? Conceptual boundaries can be any parameters that reasonably organize people into groups relevant to the issues you are examining. Although conceptual boundaries are often physically defined, geographic boundaries (e.g., cities or neighborhoods), such boundaries may be self-defined, such as people who identify themselves as part of a particular group because of some perceived shared characteristic (e.g., a particular ethnic or religious group or a group of people with a similar hobby or interest, such as gardening or bird watching). You can identify the boundaries of a group you are studying based on these conceptually defined parameters.

      

What is the hypothesis or driving research question? What is the problem or issue your research addresses?

      

What is the main geographic feature or variable you will examine in your study?

      

How is this feature or issue related to your independent or dependent variable? (Also see the next question to help you answer this question.)

      

Can you identify any geographic pattern by unit of analysis (e.g., group, neighborhood, country) relative to the topic or issue under study? For instance, if you are studying poverty, are certain neighborhoods in the study more poverty stricken than others?

      

If your study involves a comparison of groups, are there differences or stratifications based on geographic location?

      

If your study does not indicate any geographically based patterns, are there geographic variables worthy of exploration that you failed to consider?

      

What themes emerge from your data? Do they emerge by unit of analysis?

      

Are there themes that emerge specific to certain geographic locations?

      

Do specific social, economic, and political data appear to cluster geographically?

      

If this information is not easily available in a spatial database format, could such a database be easily created?

       Steps in the research process

      If you decide that you do want to employ a GIS in your project, you need to accomplish two nuts-and-bolts tasks right away: (1) identify the geographic region and (2) develop a data dictionary (figure 2.12).

      First, you should identify the geographic region (or spatial extent) and important features of your study, including the categories, geographic features, and physical environmental features that you want to consider as part of the study. Essentially, this means drawing a study site boundary, along with having a good understanding of your topic and determining which geographic features you want to measure.

      Figure 2.12 An example of a section of a data dictionary for the US Census (the Census Bureau website uses the more public-friendly term glossary). Concepts included in the dataset, along with the associated terminology, definitions, and units of measurement, are described in detail. This information is critical to researchers when assessing a dataset to determine its relevance to their research needs and to interpret results obtained from its analysis. US Census.

      Developing a data dictionary means developing a set of definitions and criteria for what constitutes your particular categories. In other words, you need to develop descriptions that clearly define each of the attributes that you would be recording in association with your variables. (We explore research design and considerations for incorporating GIS further