User:Perk101/my sandbox

Assigment #1 had us create a personal Sandbox page.

Here is my report for Assignment #1:

Recognizing MAUP in Research

By David Perkins

As the old saying goes, “statistics lie” or “you can make statistics say or mean whatever you want”, that is how I interpret the Modifiable Areal Unit Problem. One can change, or modify, the areal units in a study to fit their analysis or results. Thus, artificial spatial patterns arise from researchers using data at varying scales. For example, average household earnings are going to vary at the state level, county level, district level, census tracts, and down to neighborhoods and communities, and thus results of a study using average household earnings are going to vary at the different levels. With that being said, one way MAUP relates to the “Big Sort” article is that it points out how political elections should not be discussed at the state level anymore, but rather at the county level. During Presidential Elections, the media likes to use maps at the state level to show the results of an election, but the real action or turning points are now occurring at the county level and lower. A certain state might still be classified as Democratic or Republican as a whole, but there are now predominate counties in said States that has an opposite party affiliation, which will now have to become the focus of campaigning by candidates. Also, the “Big Sort” article used county date for its research, which proved its findings of how people are sorting themselves into counties of similar political and religious beliefs, as well as similar economic and education status. But, it also showed how the sorting might actually be occurring on an even smaller scale: at the neighborhood or community level. So, while demographic data at the county level proved the “Big Sort” was occurring, it might be even more pronounced (or not) at the neighborhood level. The article also makes reference to Los Angeles County, CA sending the most people to Colorado between 1981 and 2004, but I would be most interested to see if it was actually mostly people from certain parts or neighborhoods in the county migrating to CO, or was it evenly spaced throughout the county as a whole.

Another way MAUP relates to the “Big Sort” is from the study using census data at the county level. The Openshaw article discusses this problem, which is one of the main issues of MAUP, that “census areas do not have any intrinsic geographical meaning”. What this means is that geographical boundaries used in a census (tracts, counties, districts, states, etc.) are all defined by political and government administrations. They are arbitrary in nature, never mind that some of them change from year to year. The Openshaw article goes on to state that “if results and analysis depend on these definitions of areal units, and the areal units are arbitrary and modifiable, then the value of any work based upon them must be in some doubt”.

Finally, when discussing MAUP and the “Big Sort” another issue that crops up is the Ecological Fallacy Problem. Now, it’s possible the Big Sort might not fall victim to the ecological fallacy problem since areas are becoming more homogenous in nature due to like-minded individuals are migrating near each other. That is individuals with similar backgrounds such as religion, political affiliation, earnings, and education. Counties are becoming more homogenous in terms of political affiliation due to the increasing number of counties with landslide victories during elections. But, it still falls under the Ecological Fallacy Problem because “spatial stationarity is assumed across an entire area” (Lecture 1). This relates to the issue of MAUP in that the “Big Sort” study was using counties as the boundaries for individually collected data, but the problem comes from counties not being the same size across the United States. Counties come in all sizes and this could possibly skew the results. Ultimately, this sums up the issue of MAUP in that there are no standards for aggregating areal data (collecting individual data at the county level), and no rules to govern the size and shapes of chosen for the areal units in a study. The “Big Sort” study using county data is an example of the zonal problem associated with MAUP. Counties will have different shapes and sizes; they are not uniform across the U.S.

References:
1. Bishop B. & Cushing R. (2008). The Big Sort. In Teixeira, R. Ed.: Red, Blue, & Purple America.; pp 50-75

2. Openshaw, S. (1984). Modifiable areal unit problem. Norwich: Geo Books; pp. 1-16.

3. Lecture 1 (MAUP); GES 679: Professional Seminar on Geospatial Technologies: Analytical Problems with Geographical Information Systems (GIS)