The PDI models utilize variables on the voter file, and commercial data sources, along with completed voter surveys from over 100,000 CA voters, over the course of the last four years, with extra emphasis on more recent surveys, particularly for the new Pro-Choice Model. We built these models to enhance the process for targeting voters that appear harder to differentiate and distinguish key attributes influencing political behaviors. Using the models in conjunction with PDI voter file data and propensity universes provide the most expansive targeting weaponry available.
This model is based on self-described “liberal or conservative” leanings, then modeled across the voter file utilizing partisanship, past partisanship, age, gender, census block group data on income and education levels where the voter lives, and other factors. The model can be used as a means of targeting in addition to registered party. Additionally, this is a good factor to include in polling crosstabs for identifying levels of support or persuadability based on ideological leanings of voters.
To use this model in the PDI, select a score from 1-100, where 1 is the most liberal, and 100 is the most conservative. A score below 40 represents someone classified as a liberal, and a score over 76 represents someone classified as conservative. Within the range of 41-76, where voters are categorized as moderate, scores below 51 will lean liberal and scores above 51 will lean conservative.
- Score 1-3 is correlated to Very Liberal responses
- Score 4-40 is correlated to Liberal
- Score 41-76 is correlated to Moderate (Mid-point is 51)
- Score 77-99 is correlated to Conservative
- Score 100 is correlated to Very Conservative
This model identifies partisan vote choices – who voters are casting ballots for on a partisan basis – rather than trying to use a model to identify political party, which is clearly on the voter file. The basis for this model is a set of questions about which party voters cast ballots for, self-described partisan affiliation (regardless of actual registration), and vote on a generic congressional ballot.
This model can be used to help differentiate the most partisan Democrats and Republicans from those who may swing, and it can assist in placing independent voters into their partisan buckets as we find most independent voters in California strongly lean toward one party or another.
To use this model in the PDI, select a score from 1-100, where ranges closer to 1 will be more supportive of Democratic candidates, ranges in the middle will be swing voters, and ranges going up to 100 will be more supportive of Republican candidates. In practical terms, a score from 0-44 is a voter choosing Democratic candidates at least most of the time, and a score from 80-100 is a voter choosing Republican candidates at least most of the time. For Independent voters in the midrange from 54-77, the exact break from leaning Democrat to leaning Republican occurs at the score of 67.
- Score of 1-44 is correlated to always or mostly Democrats
- Score of 45-67 is correlated to mostly through usually Democrats
- Score of 68-79 is correlated to usually through mostly Republicans
- Score of 80-100 is correlated to mostly or always Republicans
Children in Household
To use this model in PDI, select a score of either 1 or 100, where 1 indicates that models do not predict that the voter has children living with them, and 100 indicates the voter does have children living with them. This model is a binary categorical value, and a person will have a score of either 1 or 100. Either selection can vary between voters in the same household.
Union Member in household
To use this model in PDI, select a score of either 1 or 100, where 1 indicates that models do not predict that the voter has a Union member in their home, and 100 indicates the voter does. This model is a binary categorical value, and a person will have a score of either 1 or 100. Either selection can vary between voters in the same household.
The Union Household model is NOT utilizing any data on actual union membership; instead, it takes self-described union households and models them based on criteria found in the census data and voter files. This model is not automatically in all accounts – please contact your representative to gain access.
Support for Abortion/Choice
To use this model in the PDI, select a score of either 1 or 100, where 1 indicates the voter is not in support of these measures (“Pro-Life”), and 100 indicates that model predicts that the voter is generally in support of abortion being legal and accessible (“Pro-Choice”). This model is a binary categorical value, and a person will have a score of either 1=Pro-Life or 100=Pro-Choice. Either selection can vary between voters in the same household.
As with all models, if a voter has a “0” value, or null value, it is because they could not be scored for some reason. This could be because the voter record is new, and came on after the scores were completed.