Political targeting has become more refined over the years, particularly in California where the PDI database has provided consultants and candidates with access that can only be dreamed of in other states. As a result, campaigns can more efficiently target the highest turnout universes for mailers, while segmenting GOTV programs to lower turnout supporters, and avoiding spending any resources on the dead weight.
The common targeting methods utilize specific voter behavior – when that voter participates, if they are a new registrant, if they vote by mail, and even if they registered online. It can also include person information like ethnicity, or census-based information like the income or education level of their neighborhood.
But as campaigns get more refined we find other characteristics, like Renters versus Homeowners, people who live alone versus those who live with others… these also matter in determining turnout. And these Household Type measures have different impacts in high and low turnout elections.
As the following table shows, something like “lives with a political donor” or “in a Domestic Partnership” shows much higher turnout for that type of household, while someone living alone or in an apartment shows lower. Interestingly, those households with kids were higher turnout in the 2012 General, but lower than average turnout in the Primary. Is this because parents are just too busy for the Primary, or is it because the General Election had a hotly contested education issue, while the Primary did not?
The following table gives the raw values.