MaryAnn Sorensen Allacci on Spatial Analysis

Hello – I’m MaryAnn Sorensen Allacci, Director of Projects for Environmental Health, Knowledge, & Action, Inc. We provide a wide range of evaluation and advocacy services that focus on improving people’s well-being by improving their environments.  As a result, we will often work on community environmental health issues that require our consideration of many levels of organization from different perspectives.  One method we have used is geographic information systems (GIS) and spatial data analysis to understand what is going on in a particular area and how the phenomena are distributed across the landscape.

Hot Tip: When looking at and using spatial or geographic locational data, particularly in evaluation, it is important to be careful about mixing different levels of analysis and expecting that what you find on one level (e.g., a larger scale of data collected) can and will apply to another (i.e., smaller) level or scale of measurement. This is often referred to as “ecological fallacy.”

For example, if you want to look at census tract data, a common source of spatial data, for education level of residents or, say, numbers of people receiving food stamps, the information you obtain at the census tract level may not tell you about specific neighborhoods that lie within the census tract. Census tract data are aggregates of information from smaller units, i.e., census blocks and block groups, and could have different outcomes at the individual block or even street segment level.  Data reported at the census tract level, the larger level of analysis, could mask the phenomena at the smaller neighborhood level by leading the reader to mistakenly assume that the census tract figure is a valid representation of what is going on in the smaller units.  The number of preschoolers who could benefit from an early intervention educational program might be concentrated largely in one less-advantaged neighborhood while the overall census tract count is considerably lower than other tracts.

Cool Trick: As a rule of thumb, for purposes of validity you want to seek data as close to the level of analysis in which you are interested. Sometimes that data can be difficult to find, but you may be surprised at the kinds of data ready availability from different sources in the public realm.  Nested or multilevel studies deal with the complications of different levels of data being represented twice, or co-correlation, using specific study designs and statistical management.

Do you have questions, concerns, kudos, or content to extend this aea365 contribution? Please add them in the comments section for this post on the aea365 webpage so that we may enrich our community of practice. Would you like to submit an aea365 Tip? Please send a note of interest to aea365@eval.org. aea365 is sponsored by the American Evaluation Association and provides a Tip-a-Day by and for evaluators.