Appendix C: Geographic Targeting Methodology#

This appendix lays out the methodology for measuring the three characteristics important in geographic targeting as explained in Step Three: Targeting the Voters.

Turnout is, quite simply, the percent of eligible voters who actually vote. Because your only clue to a precinct's behavior in the upcoming election its behavior in past elections, you generally predict turnout with the following formula:

turnout = (Votes Cast 1 + Votes Cast 2) / (Eligible Voters × 2)

where
   Votes Cast 1 = Number of votes cast in first  past election
   Votes Cast 2 = Number of votes cast in second past election
Eligible Voters = Total number of eligible voters in the precinct

Thus, this formula is simply the average turnout over two past elections. People doing very careful targeting will give more weight in their formulas to more recent elections.

In measuring expected turnout, it is important that you choose elections of a similar character to the upcoming election.

Performance is the percent of voters in a precinct who consistently vote for candidates or parties of the same orientation as your candidate. In places with strong party affiliation, this may mean looking at the average of all the votes received by all of the candidates of the same party over the course of two or more elections. The following formula would be used to define a precinct's "performance":

performance = (DemLeg95 + DemPres96 + DemGov97 + DemMayor97) / 4

where
  DemLeg95 = percentage of votes received by "Democratic Party" legislative  candidate in the precinct in 1995
 DemPres96 = percentage of votes received by "Democratic Party" presidential candidate in the precinct in 1996
  DemGov97 = percentage of votes received by "Democratic Party" governor     candidate in the precinct in 1997
DemMayor97 = percentage of votes received by "Democratic Party" mayor        candidate in the precinct in 1997

This formula is a simple average of the performance in the precinct of four "Democratic Party" candidates in four elections. Performance formulas are usually more complex, and, again, people doing more careful targeting will give more weight to more recent elections.

Obviously, defining "performance" in areas where party affiliations are less strong will be a little trickier. One possible formula might be:

performance = (demLeg95 + demPres96 + demGov97 + demMayor97) / 4

where
  demLeg95 = percentage of votes received by a similar "democratic" oriented legislative  candidate in the precinct in 1995
 demPres96 = percentage of votes received by a similar "democratic" oriented presidential candidate in the precinct in 1996
  demGov97 = percentage of votes received by a similar "democratic" oriented governor     candidate in the precinct in 1997
demMayor97 = percentage of votes received by a similar "democratic" oriented mayor        candidate in the precinct in 1997

This formula is the average of the votes for "democratic orientation" in four elections. Obviously, choosing which parties and which candidates well represent "democratic orientation" will be a little tricky.

Persuadability is the percentage of voters in a precinct that do not vote in a consistent way. Either they "split" their vote (vote for candidates of different orientations in the same election) or "shift" their vote (vote for candidates of different orientations over the course of two or more elections). It is generally considered that "vote splitters" and "vote shifters" are the voters most likely to be persuaded by a campaign's efforts.

A campaign might use the following formula to measure persuadability. This formula averages the precinct's "vote shifters" (between the legislative races in 1995 and the presidential races in 1996) with the "vote splitters" (between the gubernatorial race and mayors race in 1997).

persuadability = (|DemLeg95 - DemPres96| + |DemGov97 - DemMayor97|) / 2

where
  DemLeg95 = number of votes received by "Democratic Party" legislative  candidate in the precinct in 1995
 DemPres96 = number of votes received by "Democratic Party" presidential candidate in the precinct in 1996
  DemGov97 = number of votes received by "Democratic Party" governor     candidate in the precinct in 1997
DemMayor97 = number of votes received by "Democratic Party" mayor        candidate in the precinct in 1997

Note: Use absolute values for the subtraction results.

Again, where party affiliations are less strong, defining "persuadability" will be a little trickier. One possible formula might be:

persuadability = (|demLeg95 - demPres96| + |demGov97 - demMayor97|) / 2

where
  demLeg95 = number of votes received by a similar "democratic" oriented legislative  candidate in the precinct in 1995
 demPres96 = number of votes received by a similar "democratic" oriented presidential candidate in the precinct in 1996
  demGov97 = number of votes received by a similar "democratic" oriented governor     candidate in the precinct in 1997
demMayor97 = number of votes received by a similar "democratic" oriented mayor        candidate in the precinct in 1997

This formula averages the precinct's "vote shifters" (between the legislative races in 1995 and the presidential races in 1996) with the "vote splitters" (between the gubernatorial race and mayors race in 1997) for "democratic orientation" in four elections. Obviously, choosing which parties and which candidates well represent "democratic orientation" will be a little tricky.