Midterms in Retrospect: Recapping the 2014 HPU Poll Results

HIGH POINT, N.C., Nov. 11, 2014 – Ballots were cast and races called. As dust settles from the midterm elections, how did the HPU Poll fare in its 2014 findings?

“Midterm elections can be tough to poll,” says Dr. Martin Kifer, assistant professor of political science and the director of the HPU Poll. “In this case, we assumed that this would be a smaller electorate, more like the last midterms in 2010 than 2012 presidential election in terms of its composition and voting behavior.”

This semester, HPU fielded surveys of likely voters in three highly competitive Senate races in North Carolina, Colorado and New Hampshire to gain a snapshot of campaigns that determined the U.S. Senate’s leadership.

You may have seen results from this round of HPU Polls in the Washington Post, the Washington Times, the Huffington Post, Yahoo! News, New York Times, Fox News, ABC News, USA Today, MSN, the Denver Post and more. In case you missed it, here’s a rundown of what the polls found:

  • Neck-and-neck Senate race in North Carolina: In the closely watched North Carolina Senate race, HPU twice reported a dead heat between the two candidates, incumbent Senator Kay Hagan and North Carolina Speaker of the House Thom Tillis. Each received 40 percent for a poll fielded Sept. 30 – Oct. 9, and 44 percent for the final survey fielded Oct. 21 – 25. In both cases, Tillis received a few more respondents’ support: 0.9 percent in the first survey and 0.4 percent in the second. The final margin between the two candidates was 1.7 percentage points.
  • Republican Cory Gardner to win over Democratic incumbent Mark Udall in Colorado: The final margin between the candidates was 2.5 percentage points (48.5 percent for Gardner and 46 percent for Udall), and the Oct. HPU Poll margin was 3.6 points (46 percent for Gardner and 42.4 percent for Udall).
  • Incumbent Senator Jeanne Shaheen to win over Republican Scott Brown by a narrow margin: The final margin between the candidates was 3.2 percentage points (51.6 percent for Shaheen and 48.4 percent for Brown), and the Oct. HPU Poll margin was 2.1 points (48.1 percent for Shaheen and 46 percent for Brown).
  • Low approval ratings for elected officials: In North Carolina, the most recent HPU Poll found that 41 percent of North Carolinians approved of President Barack Obama’s job performance, while 42 percent approved of incumbent Senator Hagan. In New Hampshire, 41 percent of likely voters approved of President Obama’s job performance, while 47 percent approved of incumbent Senator Shaheen. In Colorado, 40 percent of likely voters approved of President Obama’s job performance, while 39 percent approved of incumbent Senator Udall.
  • Ambivalence about a possible Hillary Clinton presidential run: Nearly half of likely voters in North Carolina (48 percent) and Colorado (49 percent) said there was either some or a good chance they would vote for Hillary Clinton if she ran for president in 2016, while the same respective percentage in each state said there was no chance they would vote for her. In New Hampshire, 50 percent of likely voters said there was either some or a good chance, while 46 percent said there was no chance.
  • An increase in awareness and concern about the Ebola virus: Over the course of a four-week period, the HPU Poll found a 26 percent increase in North Carolinians who were following news about the recent outbreak of Ebola in West Africa and the cases that have occurred in the U.S. either extremely or very closely (44 percent for the Sept. poll, and 70 percent for the Oct. poll). The poll also found a 22 percent increase in concern that there will be a large outbreak of Ebola inside the United States within the next 12 months (32 percent for the Sept. poll, and 54 percent for the Oct. poll).

 

Further results and methodological details from the most recent HPU surveys and past studies can be found at the Survey Research Center website at http://www.highpoint.edu/src/. The materials online include past press releases as well as memos summarizing the findings (including approval ratings) for each poll since 2010.

For the most recent North Carolina survey, the High Point University Survey Research Center contracted SurveyUSA to interview 862 state of North Carolina registered voters Oct. 21 through Oct. 25, 2014, using Registration Based Sample (aka Voter List Sample) purchased from Aristotle in Washington, D.C. To be included in the sample, a voter needed to have voted in both 2010 and 2012, or needed to have newly registered to vote thereafter. Of the 862 registered voters, 802 were determined to be likely to vote or had already voted in person or by absentee ballot in the Nov. 4, 2014 general election. Approximately 10 percent of the respondents said they had already voted. This research was conducted 100 percent by telephone. Respondents reachable on a home telephone (72 percent of likely voters) were interviewed on their home telephone in the recorded voice of a professional announcer. Respondents not reachable on a home telephone (28 percent of likely voters) were contacted by live operators, who hand-dialed the telephone, secured the respondent’s cooperation, qualified the respondent, conducted the interview, and remained on the phone until the call was completed. Crosstabs of this study are available at http://www.highpoint.edu/src/files/2014/10/PDF-North-Carolina-Political.pdf.

The High Point University Survey Research Center fielded its second North Carolina survey of the election season with live interviewers calling between Sept. 30 and Oct. 9, 2014. The responses came from 584 likely voters with landline (277 interviews) or cellular (277 interviews) telephones. First, the HPU Poll identified registered voters using a Registration Based Sampling (Voter List Sample) system that selected possible respondents from a statewide, North Carolina list of registered voters that had landline and cell phone numbers appended by a contractor. Likely voters were estimated by asking a screening question: “On November 4, North Carolina will hold its general election for U.S. Senate, U.S. House of Representatives and other offices. How certain are you that you will vote in this election? Are you almost certain to vote, you probably will vote, your chances of voting are 50/50, or you will not vote in the November 2014 general election?” The only registered voters who passed the screen were those who responded “almost certain” or “probably” to the screening question AND voted in the 2010 general election in North Carolina, registered in North Carolina after 2010 and voted in the 2012 general election, or registered in North Carolina since 2012. The Survey Research Center contracted with Survey Sampling International to acquire this sample, which was originally compiled by Aristotle (Washington, D.C.). The survey has an estimated margin of sampling error of approximately 4.1 percentage points for this population of respondents. The data are weighted toward estimated turn out figures for age, gender and race based on North Carolina Board of Elections data and exit polls from past campaigns. In addition to sampling error, factors such as question wording and other methodological choices in conducting survey research can introduce additional error into the findings of opinion polls.  The frequencies for this study and a comparison with the New Hampshire and Colorado studies are available here: http://www.highpoint.edu/src/files/2014/10/32memoC.pdf

The High Point University Survey Research Center fielded its first North Carolina survey of the election season with live interviewers calling between Sept. 13 and 18, 2014. The responses came from 410 likely voters with landline or cellular telephones. First, registered voters were identified using a Registration Based Sampling system that selected possible respondents from a statewide, North Carolina list of registered voters that had landline and cell phone numbers appended by a contractor. Likely voters were estimated by asking a screening question: “On November 4, North Carolina will hold their general election for U.S. Senate, U.S. House of Representatives and other offices. How certain are you that you will vote in this election? Are you almost certain to vote, you probably will vote, your chances of voting are 50/50, or you will not vote in the November 2014 general election?” The only registered voters who passed the screen were those who responded “almost certain” or “probably” to the screening question AND voted in the 2010 general election in North Carolina, registered in North Carolina after 2010 and voted in the 2012 general election, or registered in North Carolina since 2012 passed the screen. The Survey Research Center contracted with Survey Sampling International to acquire this sample. The survey has an estimated margin of sampling error of approximately 5 percentage points for this population of respondents. The data are weighted toward estimated turn out figures for age, gender and race based on North Carolina Board of Elections data and exit polls from past campaigns. In addition to sampling error, factors such as question wording and other methodological choices in conducting survey research can introduce additional error into the findings of opinion polls. The frequencies for this study are available in the memo and press releases here: http://www.highpoint.edu/src/hpupoll/.

For the New Hampshire version of this study, SurveyUSA interviewed 850 state of New Hampshire registered voters Oct. 4 through Oct. 8, 2014, using Registration Based Sample (aka Voter List Sample) purchased from Aristotle in Washington, D.C. To be included in the sample, a voter needed to have voted in both 2010 and 2012, or needed to have newly registered to vote thereafter. Of the 850 registered voters, 824 were determined to be likely to vote on or before Nov. 4, 2014. This research was conducted 100 percent by telephone. Respondents reachable on a home telephone (72 percent of likely voters) were interviewed on their home telephone in the recorded voice of a professional announcer. Respondents not reachable on a home telephone (28 percent of likely voters) were contacted by live operators, who hand-dialed the telephone, secured the respondent’s cooperation, qualified the respondent, conducted the interview, and remained on the phone until the call was completed. Crosstabs of this study are available at http://www.highpoint.edu/src/files/2014/10/PDF-New-Hampshire-Political.pdf.

For the Colorado version of this study, SurveyUSA interviewed 876 state of Colorado registered voters Oct. 4 through Oct. 8, 2014, using Registration Based Sample (aka Voter List Sample) purchased from Aristotle in Washington, D.C. To be included in the sample, a voter needed to have voted in both 2010 and 2012, or needed to have newly registered to vote thereafter. Of the 876 registered voters, 800 were determined to be likely to vote on or before Nov. 4, 2014. This research was conducted 100 percent by telephone. Respondents reachable on a home telephone (72 percent of likely voters) were interviewed on their home telephone in the recorded voice of a professional announcer. Respondents not reachable on a home telephone (28 percent of likely voters) were contacted by live operators, who hand-dialed the telephone, secured the respondent’s cooperation, qualified the respondent, conducted the interview, and remained on the phone until the call was completed. Crosstabs of this study are available at http://www.highpoint.edu/src/files/2014/10/PDF-Colorado-Political.pdf.

You can follow the HPU Poll on Facebook at http://www.facebook.com/SurveyResearchCenter and Twitter at http://twitter.com/HPUSurveyCenter.

Dr. Martin Kifer, assistant professor of political science, serves as the director of the HPU Poll, and Brian McDonald serves as the assistant director of the HPU Poll.

Share Button

Related Posts