With the Primary Election finally coming to a close, each of the counties has provided PDI with new voter files including vote history for the June election allowing us to create the new 2014 General Election Likely Voter Universes.
These likely voter universes are a mainstay for campaigns in California. Their use ranges from doing mail files, to field work, polling, and even post-election research to determine effectiveness of campaign efforts on different segments of the voting population.
This year’s likely voter universes include six main sets of graduated voter groupings, along with some additional shades in between. We are also bringing back the new GOTV and WRAP universes that were successfully used in the recent San Diego Mayor Special Election.
As always, the PDI universes lean heavily on the extensive vote history available only on PDI, and then include some smaller data elements such as absentee voting history, household party types, political donor behavior, and other factors to make the best likely voter universes in the industry.
Unlike a “voter model” these universes are based heavily on specific voter history flagged to each record. No voters slip into a universe because of a mysterious score from some unknown algorithm; they are in a universe because of their actual documented behavior.
Looking closely at one of the most commonly used universes, our 14G3, like the Primary Universe 14P3, gives a good balance of key demographics such as Party, Ethnicity, Age and Geography while also providing a great medium-density mail or field voter contact universe.
This universe is listed with its election date variables, as follows:
(Voted 6/14 or Voted in either 11/10 or 6/12 or Voted in 2 of 11/06, 6/10) & (Voted 11/12 or Voted in any election in 2013 or 2014 or Re-registered after 6/14)
This gives voters several ways to qualify as a “likely” voter in 14G3 based on participation in statewide elections, in concert with either local / special elections, or, more commonly, other past statewide elections. It also gives a chance for voters who are recent re-registrants to get into the likely voter universe with a slightly wider screen.
And these universes are dynamic. At the statewide level 14G3 captures 53% of all voters, but then in a very low turnout community, such as Calexico, it is only 39% of all voters, while in a high turnout community like San Rafael, it is 74% of registrants. So, you don’t have to say “well, I’m in a low-turnout community, so I need to adjust my universe.”
And you don’t have to stick to just one universe for a campaign. You can change up the voter universes for different types of voter contact. For example, a mailing universe should probably be fairly narrowed to the most likely voters, but emailing and robocalling are cheap, so why not hit a wider universe? The greatest cost of walking is on the front-end: getting volunteers, printing kits, assembling materials. Once precinct walkers are in the field, you want them hitting as many good voters on a street as possible, so widen the universe. And for the candidate’s own neighborhood, why not have her hit every decent voter? Maybe someone with a bad voting history will still turn out to support the teacher down the street in her first campaign.
In addition, using the more advanced flags in the data can allow you to target based on message – like picking the right neighborhoods for an issue, mailing your Republican endorsement to a house with only Republican voters, or picking voters modeled to have kids for an education piece.
And, of course, these universes above could be split into Early Absentees as a first pass beginning around a month out when mail-in ballots are received, and a second pass to late Absentees and poll voters closer to Election Day.
The final universes, 14GGOTV and 14GWRAP are unique to PDI and have only been used in one prior election cycle.
The 14GGOTV is a universe which captures the types of voters that should really be the focus of a GOTV effort: All somewhat likely voters, minus the very high likelihood voters. There are three versions:
- 14GOTV1 is a wide universe, minus voters who are in 14G1
- 14GOTV2 is a wide universe, minus voters who are in 14G2
- 14GOTV3 is a wide universe, minus voters who are in 14G3
If your campaign can identify voters that are extremely likely supporters based on partisanship, ethnicity or other criteria, this universe could be used to get Yes commitments from these voters, then turn them out on Election Day.
Some campaigns could even blindly turnout a segment of these voters. For example, a Filipino Democrat could make the strategic decision to just push hard on a 14GGOTV2 Filipino Dems universe, without even identifying them as a “yes” first based on polling or past election experience. Or a Republican in a race against a Democrat could decide to use 14GGOTV3 to all over-50 Republican voters who don’t live with a Democrat.
Are we suggesting the 14G1 voters don’t need to be contacted? Of course not. But a GOTV effort is built around driving a turnout program. But someone could question the cost-effectiveness of a campaign spending limited resources “turning out” voters who have never missed an election.
The 14GWRAP is not something used by most campaigns, but can be a secret weapon for Independent Expenditure campaigns. As can be seen in the example campaign plan above, the most common universe used by many campaigns is our 14G3. Some field campaign, phone calls and emails go to wider universes, but the bulk of the general mail is hitting the same kind of voters.
By targeting a very wide universe of active but infrequent voters who aren’t in 14G3 and don’t live with anyone in 14G3, the campaign can wrap their campaign effort around the most likely campaign effort of the candidate they are supporting. Instead of duplicating efforts, they are bringing something entirely new.