For those who follow Morning Consult’s election polling, you may have noticed something different about our results today. We have shifted to likely voters. Through the election, Morning Consult will release likely voters results along with the broader registered voter population.
Most pollsters do not release details describing how they created their likely voter model, which can make it difficult to determine why polls differ. At Morning Consult, we want to be transparent about our methodology to encourage a discussion about voter turnout and hopefully improve election polling.
How We Determined Our Likely Voter Model
The main challenge of likely voter modeling is determining which Americans are likely to vote on Election Day and which ones are not.
Between 86 percent and 90 percent of registered voters, and between 55 percent and 58 percent of all adults, cast a ballot in the four presidential elections since 2000, according the Current Population Survey (see table below). Morning Consult’s standard national polls each contain 2,000 registered voters, so approximately 10-15 percent of our overall sample will not cast a ballot in November.[table “206” not found /]
Morning Consult has asked about 10 questions about vote history, vote intention, interest in the current election, and knowledge about election topics on each national survey in recent months in anticipation of releasing a likely voter model after Labor Day. We used these questions, along with results from nearly 20,000 interviews throughout August 2016, to subset the broader sample of registered voters to the narrower sample of likely voters – about 85 to 90 percent of individuals.
Our likely voter model includes any registered voters who say they are at least a 7 on a 1-10 point scale, where 1 means an individual will definitely not vote in the November 2016 presidential election and 10 means an individual will definitely vote in November. We selected this model for a series of reasons:
- Proportion of registered voters in sample: In our previous 10 national polls, between 86% and 89% of registered voters fit this screen. This is directly in line with the percent of registered voters who cast a ballot in the previous four November presidential elections.
- Clarity: This single variable prediction model is much easier to interpret than a series of indices (e.g., Perry-Gallup) or probabilistic models that we considered. We found that results from a single variable or an index approach did not vary considerably in our August 2016 polling.
- Directness: The 1-10 scale explicitly asks respondents how likely they are to vote in November. Some indices consider questions less directly tied to future vote intention, such as whether you know where your election precinct is, whether you have voted in previous elections, and how much you are thinking about the upcoming election. These indices are valid alternatives, but we aimed for directness and clarity.
Our previous month’s polling suggests that the differences in reporting between registered voters and likely voters will be relatively small. This is largely because the vast majority of registered voters cast a ballot in November, so the likely voter modeling only reduces the overall sample by 10-15 percent. Unlike previous election cycles, the results suggest that likely voter samples are not uniformly more pro-Republican than the broader registered voter sample. The table below, for example, shows the vote margin in the 2-way horse race for all polls in August broken down by registered voters (Column 2), our likely voter model (Column 3), and a likely voter model based on the Perry-Gallup scale (Column 4).
[table “207” not found /]
What Do You Do With Undecided Voters?
Our current likely voter model includes voters who have not made up their mind. Sometimes pollsters ask undecided voters to select a candidate such as Clinton or Trump. We are not currently doing that for our likely voter model for a few reasons. First, we believe it is important to understand the percent of the electorate that is still undecided because they are still seeking information, are open to a third-party candidate, or feel ambivalent toward to the two historically unpopular major party candidates.
Second, we anticipate that the number of undecided voters will decline as the November election approaches and campaigns ramp up both paid advertising efforts and get-out-the-vote (GOTV) operations.
Finally, the share of undecided voters is considerably lower among likely voters than among the broader populations such as all registered voters or all adults. Morning Consult’s current national polling includes a high number of undecided voters (16 to 19 percent), whereas the percent of undecided voters among likely voters is lower (12 to 14 percent). We plan on releasing the results from the question that forces undecided voters to select a candidate in our normal weekly polling so you can compare them side-by-side.
The table below displays the Clinton-Trump matchup in three separate versions of leaned and unleaned combinations. The leftmost columns show an unleaned question where we do not ask undecided voters to select a candidate. The fourth and fifth columns show results from a leaned question where we ask undecided voters to select Clinton, Trump, or say once again that they are undecided. The final columns, columns 6 and 7, show a fully leaned version where we ask respondents who are still undecided to select either Clinton or Trump.
Importantly, the results across the three versions are not very different. While undecided Americans are slightly more supportive of Clinton than Trump, the margin is close enough that it does not appear to influence the overall horse race result.[table “208” not found /]
Did You Use Two-Way or Four-Way Vote Choice Question?
We are planning on releasing likely voter results from both the two-way (Clinton, Trump, Don’t Know) and the four-way (Clinton, Trump, Johnson, Stein, Don’t Know).
The table below displays results from likely voter modeling for the two-way question and the four-way question. Donald Trump tends to perform slightly better in the four-way match ups than in the two-way race. Additionally, the vote choice question that includes all four candidates (Clinton, Trump, Johnson, Stein) likely overstates Johnson and Stein’s overall level of support, but it has a lower percentage of undecided voters by 3-5 points. The vote choice question including only the two main party candidates has a higher proportion of undecided voters because individuals who support the third party candidates are more likely to say they are undecided when presented with only a few options.[table “209” not found /]
We recognize that likely voter models are at best a pollster’s estimate of what the electorate will look like at a snapshot in time. There is a fair amount of stability in Americans’ voting patterns in presidential elections and in the percentage of adults who cast ballots every four years; however, if the electoral environment changes, we may provide adjustments to our model.
Our general election question largely has a high level of undecided voters because respondents see a don’t know option online (phone pollsters do not explicitly say “don’t know” and some other online pollsters do not show a “don’t know” option to respondents) and because some supporters of third-party candidates will select don’t know instead of picking Clinton or Trump