Key Research Poll General Election Predictions

Don’t go to bed early on election night if you’re interested in the outcome of the Love/Matheson race.

BYU’s Center for the Study of Elections and Democracy is analyzing a poll conducted by Key Research, a survey and market research company in Provo, about the upcoming general election. We consulted on the questionnaire and sample design and Key Research collected the data.

For this survey 500 voters were sampled from the state’s file of active registered voters; 100 in each of the four congressional districts with an extra 100 oversampled in the 4th Congressional District. Using the voter list for sampling allows information in the file to be used to make the sampling more effective and efficient. Much research in political science confirms that voter turnout is strongly related to voting in past elections, age, and party registration status and all of this information was used to draw a sample of likely Utah voters for a telephone survey.  For more information on this method, see here.

Here are the vote questions along with the results.  For the statewide results, the results have been weighted so that each congressional district is equally represented.  For each question, a very small number of people said that they wouldn’t vote in that election.  Those responses have been omitted.

President
In the November 6, 2012 General Election for U.S. President will you vote for: (READ CHOICES)
71%,  Mitt Romney and Paul Ryan, the Republicans
20%, Barack Obama and Joseph Biden, Jr., the Democrats
9%, Don’t Know / Someone else (DO NOT READ)
n=500

U.S. Senate
In the November 6, 2012 General Election for U.S. Senator will you vote for: (READ CHOICES)
22% Scott Howell, Democrat
61% Orrin Hatch, Republican
17% Don’t Know / Someone else (DO NOT READ)
n=500

Governor
In the November 6, 2012 General Election for Governor of Utah will you vote for: (READ CHOICES)
65%, Gary Herbert and Greg Bell, the Republicans
19%, Peter Cooke and Vincent Rampton, the Democrats
15%, Don’t Know / Someone else (DO NOT READ)
n=500

U.S. House
In the November 6, 2012 General Election for U.S. House of Representative will you vote for: (READ CHOICES)

District 1
20%, Donna M. McAleer, Democrat
62%, Rob Bishop, Republican
18%, Don’t Know / Someone else (DO NOT READ)
n=100

District 2
20%, Jay Seegmiller, Democrat
46%, Chris Stewart, Republican
34%, Don’t Know / Someone else (DO NOT READ)
n=100

District 3
68%, Jason Chaffetz, Republican
15%, Soren Simonsen, Democrat
17%, Don’t Know / Someone else (DO NOT READ)
n=100

District 4
43%, Mia Love, Republican
43%, Jim Matheson, Democrat
14%, Don’t Know / Someone else (DO NOT READ)
n=200

First things first.  The 4th Congressional District race is close!  Keep in mind that the relatively small sample size of 200 for that district has a margin of sampling error of + or – 7 percentage points.  The public polls released during the summer and during the last couple of months show a race that has been going back and forth between Matheson and Love.  With a “don’t know” response that exceeds the margin of sampling error, this one is still up for grabs.

The other races offer no surprises.  Utah is a Republican state and the Republican candidates are clearly ahead.  The statewide races, with the full sample of 500, have a margin of sampling error of + or – 4.4 percentage points and leave no doubt that Mitt Romney, Orrin Hatch, and Gary Herbert are leading in the polls.  But Republicans are also clearly leading in the other three congressional districts, even where a sample size of only 100 has a margin of sampling error of + or – 10 percentage points.

Across all of the races, one important question is what to make of the relatively high “don’t know” response. The easiest way to get a sense of what might happen on Election Day is to allocate them in the same proportion as voters who have made up their minds already. Basically, this means throwing out the undecided voters and recalculating the percentages only for those who have stated a choice.  However, doing so ignores the fact that in general elections Utahns regularly give 3-4% of their votes to candidates from minor parties.

For example, throwing out the “don’t know/someone else” on the presidential race produces a 79% Romney, 21% Obama prediction, but that seems out of reach when you look back at past Utah presidential election results, even for Mitt Romney.  A safer bet is probably to assume that at least 3% of voters will select another candidate leaving you with 76% Romney, 21% Obama split on Election Day.  Assuming the same for the Senate race, you get 71% Hatch, 26% Howell.  For Governor it’s 75% Herbert, 22% Cooke.  For District 4, you would simply give 47.5% to each candidate.  Don’t go to bed early on election night if you’re interested in the outcome of the Love/Matheson race.

Stay tuned.  We’ll post more results from this survey in the next few days.

Survey Methodology
The sample was drawn from the publicly available file of Utah registered voters.  A model of general election turnout was estimated using age, party registration status, length of registration, and past election turnout. This model was used to estimate a probability of voting in the 2012 general election.  A Probability Proportionate to Size (PPS) sample was draw using this turnout estimate such that voters with a higher probability of voting have a higher probability of being selected in the sample.  This produces a sample of likely voters.  The sample was then matched to a database of telephone numbers and sampled voters were administered a questionnaire over the telephone by Key Research. The survey field dates were October 9, 2012 – October 13, 2012.  The statewide sample of 500 produces a margin of sampling error of 4.4%.  The margin of error is larger for questions that some respondents chose not to answer or for the vote questions only asked of voters within each congressional district.  Of course, sampling error is only one possible source of error in survey research.  Results can also be affected by measurement error (e.g. question wording and question order), coverage error (e.g. counting as “likely voters” survey respondents who will not vote), and non-response error (e.g. the people who responded to they survey are systematically different from people who refused or were not reachable).

**Update: Click here to download a topline report that includes the full survey questionnaire, frequencies for each question, and a detailed methodological report (including details about the sampling as well as response rates and cooperation rates).

Possibly related posts:

About Quin Monson

Quin Monson is Associate Professor of Political Science and a Senior Scholar with the Center for the Study of Elections and Democracy at Brigham Young University.
This entry was posted in Everything and tagged , , , , , , , , , , , , . Bookmark the permalink.