Quantifying Political Support

June 14, 2014 • John Vandivier

This article discusses the relationship between political support, political favorability, and name ID.

The first thing to note is that each of these ideas is both a concept in theory and also refers to a particular practical counterpart. In other words, there are very specific and standard political industry operationalizations for each of these concepts.

Theory:

\"Support\" is whether or not a person will vote for a candidate, issue, or other political element being supported. It can also refer on a population scale to the number of individuals in a population, or the % of a population, that will vote for a candidate. It is either a yes or a no at the individual level, and between 0 and 100% at the population level.
\"Favorability\" is the degree to which an individual \"favors\" an element. It can also refer on a population scale to the distribution of favorability of a candidate or issue across a population. It falls along a spectrum from infinitely negative to neutral to infinitely positive. In theory, one problem is the subjective value of the term \"favor.\" It is taken to be positive sentiment, and perhaps also as willingness to support, but it may not be indicative of actual support if the cost to support is greater than the willingness to support, even if the willingness to support is positive. \"Favor\" is an arbitrary unit of support, defined by the supporter and normative to the population. The ability to measure this is derived from the assumption that the definition is stable in the short run and long run, even if not the medium
\"Name ID\" is name identification, or a measure of the ability of a person to recognize a person based simply on their name. It is a yes or no. Population wide name ID is a % of the population that knows a person based on their name. One large problem here is mistaken identity which can be due to several people having the same name, a person thinking they know a name when they are thinking of someone else, or other causes.

Practice:

Support is usually measured by one of the following question types:
- Straight-support type question.
  - \"Can [Candidate Name] count on your support in the [Office] election on [Date]?\"
  - \"Can [Candidate Name] count on your vote in the [Office] election on [Date]?\"
- Head-to-head support type question.
  - \"I am now going to read a list of candidates for [Office]. At the end, please select which candidate you would vote for if the election were held today. For [Candidate 1], press 1. For [Candidate 2], press 2...Please make your selection now.\"
Favorability is usually measured by the following question:
- Would you say that your opinion of [Candidate Name] is very favorable, somewhat favorable, neutral, unfavorable, very unfavorable, or undecided?
Name ID is usually measured by the following question:
- Have you ever heard of [Candidate Name] before this call?

The goal is votes. This is best measured by head-to-head support questions. If head-to-head support is low, the issue can be broken into either an ID issue or a favorability issue. In theory, support is determined by practical cost of voting considerations (weather, distance to nearest precinct, and so on), comparative favorability, and name ID. If support is greater than some arbitrary threshold determined by the voter they will vote. On an economic view, the arbitrary threshold is explained as an expected value in return for the cost of voting.

Some key points:

If perceived benefit of voting is greater than the perceived cost, by person Y for person X, then Y will vote for X. Including opportunity cost derived from choosing not to vote for another candidate.
If someone doesn't know candidate X, then their expected favorability is the average favorability for candidate X.
Practically, support can be raised to about the level of average favorability by increasing name ID.