The Perils of Predicting Political Affiliation by ZIP Code, Age and Income
Midterm elections are coming up. If you are advertising on behalf of a particular political candidate or special interest group, it would be natural to want to target a subset of all possible households in your area that fit your target profile. This could be, for example: “Republican,” “Democrat,” “Independent,” or “Undecided.”
But, how do you best go about determining which households contain which types of individuals?
With all of the demographic data available from mail houses today, it is tempting for some people to try to guess the political affiliation of your target audience based upon demographic data alone. For example, you may believe that political party designation categorized by geographical area (e.g., “red” or “blue” county), household income range, or voter age together could be used to guess a given household’s political affiliation.
However, our research blows a hole in that assumption. The truth is, we have found that you can find Democrats, Republics, Independents, and Undecided voters with given any combination of geography, age and income attributes.
Now, it is true that some geographical regions, for example, do skew toward one or another party affiliation. And yet, doing marketing based solely on the political tendency of the members of a given ZIP code or other large area still risks sending 20-30% or more of your direct mail to the wrong people. Or, placing a billboard on the wrong street corner. Or advertising in the wrong local paper. In all of these cases, it is a pure waste of marketing dollars.
The best way to advertise or get the word out to members of a particular political persuasion is to simply purchase a mailing list that is already coded by the said designation. Of course, no such “self report” data is ever 100% accurate or up to date. But, going this route is certain to give you a much higher response rate for every dollar you spend.
Bonus tip: go a step further and cross-tabulate your targets (already segmented by political affiliation) by additional filters such as segmentation cluster, income, geography or other factors.