Tuesday, November 3, 2009
I have a problem with CNN's John King and his "magic wall." In tonight's New Jersey gubernatorial election, it's impossible to draw any conclusions on voter preference (conservative vs. liberal) based on county-level mapping, which seems to be the only level of mapping that Mr. King has at his disposal. County maps show yield no true spatial information and the comparisons with the national presidential election of 2008 are spatially irrelavent because the mapping is too "coarse." These are relatively large counties and voting blocks are more easily discerned at the ZIP code level where most demographic patterns are established. Look at any of the psychographic profiling schemes established by the leading demographic data suppliers and you will find that you only begin to understand spatial phenomenon at the ZIP level and preferably, the Census Track. The political punditry will draw improbable conclusions if they stick with the County-level maps to understand voting blocks. It's important to look at the psychographic trends by ZIP code if you want to then draw extensions to voting blocks or changes to political preferences.