The display of secondary data, and especially Census data, in map, or GIS format has become commonplace. As director of the Minnesota State Data Center at the time of the 1990 Census of Population, Dr. Randy Cantrell, was a pioneer in the presentation of Census and other government data using political maps (see for example Cantrell, 1993). These depictions of demographic and economic data are typically presented at the county level. Counties are a convenient unit of analysis, having similar definitions in all states and being the smallest unit of analysis available from sources such as the Bureau of Economic Analysis and Department of Agriculture.
County level data have the disadvantage of masking characteristics of communities and neighborhoods through aggregation. In recent work, Cantrell (2005) and others ( Winchester , 2002) have argued that the demographic characteristics of rural areas are better represented through the analysis of data reported at the Minor Civil Division (MCD) level. MCDs are essentially towns and townships. These data are especially useful in identifying micro-regions and rural neighborhoods where specific and sometimes unusual or even idiosyncratic conditions may be responsible for anomalous demographic and economic activity. Acreage development around trade centers provides an example of the sort of detail provided by MCD level analysis.
MCD level analysis is relatively rare in the literature because the geography is quite difficult to work with. The difficulty is found in the number of observations involved. Nebraska , for instance, is home to 93 counties with relatively stable boundaries (unchanged since 1920). Nebraska 's MCDs, on the other hand, number over 1,700: Including 537 incorporated places and 1,200 townships, precincts, unincorporated communities and Census designated places. This large number of observations provides a rich wealth of detail for research. However, the same numbers make tabular analysis of the data so complex as to be meaningless for all but the most sophisticated observers.
GIS representations of MCD data are both valuable and informative for the majority of viewers. However, these are also quite difficult to produce, especially when analyzing change, because political boundaries at the MCD level can be quite unstable as the result of mergers, dissolutions and annexations. Between 1990 and 2000, the boundaries of over 30 Nebraska MCDs changed. In order to describe change geographically between two points in time, political boundaries must be standardized; a process that often requires specific local knowledge.
In this project, GIS depictions of in-migration and associated characteristics down to the local (MCD) level will allow the project team to identify residential, social and economic characteristics of rural micro-regions that would either be invisible or overlooked with other methodologies. For example, population dependency ratios, levels of educational attainment and other economic growth metrics will be attainable to a very local scale of resolution. Those insights will, in turn, guide the process of gathering primary data by identifying target locations and shaping lines of inquiry.
The GIS output will also provide a spatial analysis of the heart of the Buffalo Commons for the year 2000. This product, while suitable for publication in its own right, is of especially great importance because the data from which the study will be constructed come largely from the Decennial Census Sample Survey. That questionnaire will never again be distributed, having been replaced by the American Household Survey, a monthly random sample of 250,000 households (U.S. Bureau of the Census, 2004). While the American Household Survey is a fine statistical tool, its value is likely to prove limited for small populations, such as those found in small rural communities, where the demands of statistical reliability require that a very large percentage of the population be polled. At the very least, point in time data of the type proved by the Decennial Census will no longer be available for such variables as income, employment, poverty and migration for small places. Thus the importance of establishing baseline representations of data from the 2000 Census to this scale of resolution is especially important.
