The purpose of the Perceived Voter Model is to predict campaign strategies and their subsequent effects on voters by paying attention to the perceptions that political campaigns have of their electorates. In Chapters 2 and 3, I explained why public records of personal information, such as those that originate in the voter registration system, contribute to campaign perceptions. I offered specific hypotheses that explain how the provision of data about voters' racial identities or party affiliations, and the lack of data predictive of persuadability, will affect campaign strategies. In Chapter 4, I reviewed the data sources that will allow me to examine these hypotheses.

In this chapter, I begin to test hypotheses by focusing on the provision of public data about voters' partisanship. States differ in their public collection of personal data about partisanship. Some states collect data from voters that, when transmitted to campaigns, provide campaigns with effective tools for predicting which voters are going to support Democratic or Republican candidates in upcoming elections. Other states do not collect these data. In all states, voters cast secret ballots, but in some states, a campaign can use public records to predict a person's choices in the voting booth with a high level of accuracy. In this chapter, I examine how the availability of public records relevant to predictions of partisanship affect how campaigns perceive the electorate, how they strategize given those perceptions, and how their strategic choices affect voters.

PUBLIC RECORDS OF PARTISAN SUPPORT

There are two forms of party information that are available to campaigns on voter files in different states. The first form is party registration, whereby voters have the option of registering as a member of a political party and this designation becomes public information. The second form is party primary data. For each primary in which a voter chooses to participate, the party's ballot that the voter selected is noted on the public record. In 2010, in seventeen states, both forms of partisan information were available.