Political Data has consistently created likely voter universes for statewide and local contests and post-election analysis has pointed to their effectiveness at helping campaigns and pollsters identify those voters that will turn out on election day.
A “Likely Voter” is exactly that – someone who has shown the propensity to cast a ballot in an upcoming election. The most-effective driver in this turnout modeling is past election performance – and PDI is the only data vendor in the country that has statewide, local, and special election data obtained directly from the counties and cities that administer each election.
For any given campaign, several different models can be used. For example, an expensive mailer could be targeted to a very tight universe, or a subset of a tight universe, in order to keep costs down. But that same campaign could build a phone program to a wider universe, or a walk program to a universe that is wider still. As the strategy of each voter contact is decided, campaigns use PDI likely voter universes to determine how many will be reached, and through what means.
The upcoming San Diego Mayoral special election will be a unique contest. The primary is set for mid-November without any other issues on the ballot. The last time the city saw an election like this was in 2005. That seems fairly recent, but it is a significantly different electorate with 286,000 of the city’s 677,000 voters registered to vote since that election.
The voter universes provided in counts below build on primary election turnout, utilize the recent San Diego special elections that overlapped a fraction of the city, and incorporate for the first time those who have donated to political campaigns. The act of giving a political contribution to a local, state or federal campaign is so highly correlated with voter turnout that we are now using this to supplement the models.
In addition to adding political donor status, these counts have some additional new features, including:
- A new GOTV universe captures those voters who are likely, but have some infrequency. This allows campaigns a quick way to efficiently target GOTV resources to those voters who may need a little nudging to the polls, but avoids both (a) those who are extremely unlikely and (b) who are so high-propensity that they don’t need a push.
- A new WRAP universe has been created to meet the needs of the Independent Expenditure community. A principle of some IEs is to mail voters that are not being targeted by the internal campaigns. This universe wraps around the most commonly reached voters and helps independent groups target those who are not likely being contacted.
- Counts with DemPlus and RepPlus. In San Diego the number of voters without political party is increasing, and could be the largest group of voters in the coming years. Yet, many of these voters had a party affiliation at one time, or pull a partisan ballot. Some even have contributed to partisan campaigns. This data is combined with the actual party registrants to make DemPlus and RepPlus. In some areas the number of actual “nonpartisan” voters will drop in half once those with partisan actions can be placed into DemPlus or RepPlus.
The following chart shows the size of each likely voter universe, the relative strength of DemPlus, RepPlus and Nonpartisan voters in each set. 
As can be seen in this chart, all but two of the likely voter universes are larger that 2012 Primary turnout. This is not a prediction of total votes that will be cast in November, but the segment of voters that is most likely to participate. The tighest universes may see 85% turnout while the loosest universe may only see 65%. That doesn’t make one better or worse for campaign purposes – the universe to be used must be judged with the overall strategy. And sometimes it makes more sense to contact the hyper-performance voters while other times a campaign must contact voters who are 50-50.
The following count sheet provides breakdowns for each universe by partisanship, ethnicity, income, ages, gender, household type, voter history and more. The counts description provides a breakdown of what was used to construct each universe.
Likely Voter Universe Counts: cr_Co37SanDiego13SDP
Likely Voter Universe Description: [2013 SD MAYOR]