One of the records inside a massive and rapidly growing Republican National Committee database contains three predictive numbers about President Barack Obama’s political leanings. There’s a 95 percent likelihood Obama will vote in the 2016 general election, the database predicts, based on computer modeling. It also shows an 83 percent chance Obama will side with his party’s nominee, while suggesting a 10 percent shot he’ll back the Republican candidate. It’s an extreme example—Obama, a two-term Democratic president, has repeatedly said he plans to support his party’s nominee and has been highly critical of Donald Trump—but it illustrates how the RNC has attached a score to each of the 192 million registered voters in America as part of a massive push to regain parity with Democrats in using data to win elections. Democrats have used similar scoring for several election cycles, but this will be the first in which the RNC has used such a system in a presidential election.
The party won’t share all of the ingredients it uses to cook up the scores, but they include voter registration, participation in past elections, political giving, property and neighborhood information, periodic surveys, door-to-door contacts and other field work, and data from consumer research giant Acxiom Corporation.
By crossing the values of the three predictive scores, the party infers voting behavior and uses the information to create more customized voter contact programs that it hopes will help its candidates more efficiently utilize resources. It’s the sort of sophisticated consumer and market analysis corporations have done for decades.