A Quinnipiac University Presidential Election Poll released on October 26, 2011, directed by Douglas Schwartz, Ph.D, indicates that voters in Ohio disapprove of President Obama, yet still prefer him over the three frontrunners for the Republican nomination. The same poll, conducted over October 17 – 23, shows Herman Cain leading among likely Republican primary voters, edging out Mitt Romney by seven points, and ahead of Rick Perry by 30 points. Although this poll only represents results from one state, and is not meant to predict national results, the methods used in conducting the poll comply with ethical polling guidelines, for the most part.
Quinnipiac University Poll is operated by the Quinnipiac University Polling Institute in Connecticut. The polling institute is larger than most other academic polling operations, employing eight full time staff members in addition to hundreds of student researchers. Quinnipiac University Poll focuses specifically on swing-states, and is well regarded for its accuracy in predicting presidential elections. Although some academic polls must be questioned due to lack of resources and experimental delivery methods, this particular poll was conducted by an institution that is highly regarded amongst major pollsters.
Delivery of Questions
Poll questionnaires were delivered to 1668 registered voters and 542 Republican primary voters, who were contacted by live interviewers via phone calls to landlines and cell phones. Phone numbers were dialed at random using computer software designed to give both listed and unlisted phone numbers equal chance of being dialed.
Poll data does not reveal whether weighting was used to account for multiple voters in a household, or other demographic factors. Raw demographic data indicates 26 percent of respondents considered themselves Republican, 33 percent considered themselves Democrat and 33 percent considered themselves Independent. Also, 83 percent of respondents were non-Hispanic whites. This would certainly indicate bias in a national poll, however 2010 Census data reveals 82.7 percent of Ohioans are non-Hispanic whites. Poll data also states 11 percent of respondents were black, as compared to 12 percent blacks listed in the Ohio Census data. Hispanic representation in the poll was 2 percent, compared to 3.1 percent reported in the latest Ohio Census data. If demographic weighting was not used, the sample is still closely representative of the Ohio population. The polling institute’s website states it has Spanish-speaking interviewers on hand, and that phone calls are made during weekdays and weekends, from 10 A.M to 1 P.M. and from 6 P.M. to 9 P.M.
According to the Quinnipiac University Polling Institute website, “likely voters” are self-identified registered voters. The margin of error for the sample of 1668 registered voters was +/-2.4 points. The margin of error for the sample of 542 Republican primary voters was +/-4.2 points, which is slightly high, as most national polls aim to be below 4. However, since this was a state poll, the sample was significantly smaller, making for a larger margin of error. As with all polls, there is some bias present in the sample, since a small percentage of the population does not own a phone, and respondents typically have some level of political interest if they consent to participate in a poll.
Presidential Poll Questions: Wording
Poll data shows the order in which questions were answered, and the “for whom would you vote” questions were asked prior to any other questions. This is considered a good polling practice, since subtleties in the wording of other poll questions may temporarily alter respondents’ opinions. Registered Republicans were asked for whom they would vote among a list of all current Republican candidates, as well as from a list of Perry, Cain or Romney. Registered Republicans were also asked for whom they would vote in the Ohio Senate primary. General voters were asked, in three separate questions, for whom they would vote in a hypothetical 2012 presidential race between Obama and Perry, Cain, or Romney. Additional questions assessed the respondent’s enthusiasm about voting in 2012, and their approval of Barack Obama and Ohio Senator Sherrod Brown.
Biased Questions in Presidential Poll?
There was one question in this poll that reflected a bias, which may have affected the answers to the question. This question was given towards the end of the poll, and was worded as follows:
“As you may know, Herman Cain has never served in public office. Does the fact that Herman Cain has never served in public office make you more likely to vote for him for President, less likely to vote for him for President, or doesn’t it make a difference?”
This question provided the respondents information about the candidate before asking them to assess that candidate’s electability. Since this question was given late in the poll, it is not likely that it affected the responses to earlier questions, since interviews were conducted live, and respondents did not have the opportunity to change previous answers.
The final two questions of this poll included a description of specific economic policy proposed by candidate Herman Cain, and asked respondents whether they’ve heard of the proposal and whether they think it is a good or bad idea. In this case it was necessary to include a description of the plan to prevent respondents from answering a question about which they have no knowledge. It is common for respondents to feel embarrassed when asked about an issue with which they are not familiar, and many will provide an answer in an attempt to make themselves sound knowledgeable.
Understanding the Poll Results
This poll provides an abundance of information, and the polling institute broke down results of many of the questions into demographic groupings such as gender and age. The pollster also compared results to past surveys that included the same questions. Conclusions reached show that, while voters generally disapprove of President Obama and do not think he deserves reelection, he fares better against all Republican candidates whether the results are analyzed by themselves, or broken down by demographic.
Among Republican primary contenders, Perry’s support is dropping as Cain’s is rising, indicating many former Perry supporters are switching candidates in favor of Cain. While Romney’s support is relatively stagnant, he remains strong among Republican women.
Peter Brown, assistant director of the Quinnipiac University Polling Institute, summed up the poll results, stating, “Ohio voters are not happy with the president’s performance and don’t think he deserves a second term. But elections are about choices. At this point none of the Republicans are able to take advantage of these presidential negatives. The next year will determine if the GOP is able to nominate a candidate who can do so.”
It should also be noted that a poll conducted in one state is not representative of the entire country. However, since Ohio is a swing state with a solid mix of Republicans and Democrats, this poll may be a fairly accurate indicator of nationwide sentiment.
Lapidos, Juliet. What’s With All The Quinnipiac University Polls? Accessed October 26, 2011.
U.S. Census Bureau. Ohio. Accessed October 26, 2011
Quinnipiac University. Polling Institute Contacts and Information. Accessed October 26, 2011
Decoding Science. One article at a time.