Since 1950, Gallup has used likely voter models to identify Americans who are most likely to vote in a coming election. These models involve asking poll respondents a series of questions about their interest in the coming election, their past voting behavior, and their current intention to vote in the election.
Likely voter models are necessary in pre-election polling because a substantial proportion of eligible voters do not end up voting in U.S. elections. Thus, reporting voter preferences on the basis of all national adults or all registered voters does not generally provide the most accurate estimate of the vote in a given election.
Gallup scores each respondent on a scale ranging from 0 to 7, using Gallup's standard set of seven likely voter questions that indicates his or her likelihood to vote (questions are listed below). The respondents with the highest scores (usually a 7 and often a 6) are considered likely voters. Gallup develops models of the likely voter electorate based on several assumptions about turnout rates among the adult population.
In midterm elections, turnout typically ranges from 35% to 40% of the adult population. For the 2010 congressional midterm elections, Gallup will provide two estimates: one based on 40% turnout and the other based on a higher turnout scenario.
By the time of Gallup's final pre-election polling, the likely voter estimate will be narrowed to the most probable turnout rate, based on all available indicators.
The likely voter model has proven to be accurate in Gallup's estimates of presidential and congressional midterm elections.
Computation of Gallup's Likely Voter Models for Midterm Elections
In each pre-election survey, Gallup identifies a subsample of likely voters by taking into account each respondent's score on the 0 to 7 scale. The top scorers on this scale -- equal to the projected turnout -- are identified as likely voters.
I. Computing the Likely Voter Score for Each Respondent
The seven questions that make up the likely voter scale are as follows (see the full wordings at the end of this page). Respondents get one point on the likely voter scale for each question to which they give the response listed in parentheses (with a maximum of 7 points possible).
- Thought given to election (quite a lot, some)
- Know where people in neighborhood go to vote (yes)
- Voted in election precinct before (yes)
- How often vote (always, nearly always)
- Plan to vote in 2010 election (yes)
- How certain to vote (absolutely certain)
- Voted in last midterm election (yes)
The following adjustments are made to these raw scores:
- Respondents who are not registered to vote are assigned a score of 0.
- Respondents who do not say they plan to vote (see item 5 above) are assigned a score of 0.
- Younger respondents' scores are adjusted to account for their ineligibility to vote in some or all past elections. In other words, even though the model identifies voters on the basis of past voting history, younger voters are not penalized for not being of voting age in past election years.
-- If aged 18-19, their scores are converted as follows: 1=2, 2=4, 3=5, 4+=7
-- If aged 20-21, their scores are converted as follows: 1=1, 2=3, 3=4, 4=6, 5+=7
II. Computing Likely Voters as the Proper Proportion of the Sample to Equal Estimated Turnout
Gallup begins by assessing various turnout scenarios, which can be adjusted close to Election Day, depending on Americans' reported interest in the election and voting intentions. Thus, Gallup models the turnout level under various scenarios by counting only the responses of that percentage of all eligible voters applicable to each scenario.
It is important to point out that the likely voter sample calculations are based on weighted sample sizes. Gallup usually reports unweighted sample sizes. The two can vary, sometimes substantially, depending on how the group of people scoring 6 are weighted in order to reach the projected turnout level. In some instances, this group of voters is given a statistical weight that lowers its overall representativeness in the likely voter pool. The unweighted sample size includes all respondents whose voting intentions are taken into account in a likely voter group, regardless of their proportionate value in the likely voter pool.
For a sample size of 3,000 national adults, the likely voter sample will have a weighted sample size of between about 1,200 and 1,800 voters. The unweighted sample size for likely voters will be about 1,800 voters.
1. How much thought have you given to the upcoming election for president -- quite a lot, some, or only a little?
1 Quite a lot
3 Only a little
4 None (vol.)
5 Don't know
2. Do you happen to know where people who live in your neighborhood go to vote?
1 Yes, any response given
3 Don't know
3. Have you ever voted in your precinct or election district?
1 Yes, any response given
3 Don't know
4. How often would you say you vote -- always, nearly always, part of the time, or seldom?
2 Nearly always
3 Part of the time
5 Never (vol.)
6 Don't know
5. Do you, yourself, plan to vote in the election this November, or not?
3 Don't know
6. How certain are you that you will vote -- [READ 1-3]?
1 Absolutely certain
2 Fairly certain, or
3 Not certain
4 DON'T KNOW
7. Thinking back to the elections held for Congress in November 2006, did things come up that kept you from voting, or did you happen to vote?
1 Yes, voted
2 No, did not vote
3 DON'T KNOW
Note: (vol.) = Volunteered response