Elections, Misc, Politics, Spring 2015 Articles, Topline Results

Students Undecided on 2016 Presidential Election

Last week (February 5th-6th), PackPoll fielded a survey of State undergraduates asking them about the 2016 presidential election. While the election is 21 months away, and the first caucus is 11 months away, polling about the 2016 election began almost the day after the 2012 election was decided. However, history has shown that early polling results often reveal very little predictive information about the upcoming race. Most potential candidates who poll well early, for example, do not win their party’s nomination. With that in mind, PackPoll sought to see how students feel about the primaries and the eventual election now as an early benchmark for student opinions.

Political scientists claim that early presidential polling results—such as those that are increasingly available now—tend to measure candidate name recognition more than they will predict the future. Name recognition, for example, or the lack thereof, is why Barack Obama polled poorly and was excluded from many early polls, even though he went on to win the Democratic nomination and the presidency. Another criticism is that polls “manufacture” preferences by prodding people to give an answer, even if they haven’t thought much about the race. One way they do this is by not letting survey respondents know it is ok to say they haven’t decided yet, and not providing this as an explicit answer option. We wanted to test the notion that early candidate preferences in polls mainly reflect name recognition, and the claim that most polls manufacture preferences that really don’t exist yet. We tested this in an experiment in question wording that we embedded in the poll.

After asking students if they were Democrats, Republicans, or Independents, we asked only Democrats about their potential nominees and only Republicans about theirs. Partisans were then asked, at random, one of three versions of a possible election question. Some were asked how familiar they were with potential candidates. Others were asked which person from a list they would prefer as their party’s nominee, but they were not given the option to say they hadn’t decided yet. A third group was given the same list, but also told that they could say they hadn’t made up their mind yet. Democrats, unsurprisingly, were most familiar with Hillary Clinton (mean of 1.49 on an inverted 1-4 scale). They were also familiar with Joe Biden (1.68), but others were not known. Elizabeth Warren (3.05) was a distant third, edging out Andrew Cuomo and Bernie Sanders. Republicans, on the other hand, are less well-known overall. Of the 10 potential candidates polling at least 3% or better in national polls, none came close to Clinton or Biden’s name recognition. On the other hand, most did better than the lesser know Democrats. Mike Huckabee and Paul Ryan were tied for the most familiar among Republicans.

Table 1. Candidate Familiarity, among Democrats

Potential Candidate Very familiar Somewhat familiar Somewhat unfamiliar Very unfamiliar Total Responses Mean
Hillary Clinton 37 13 6 1 57 1.49
Joe Biden 28 21 6 2 57 1.68
Elizabeth Warren 6 13 10 28 57 3.05
Andrew Cuomo 1 13 14 29 57 3.25
Bernie Sanders 5 7 8 37 57 3.35

Note: Entries are the number of respondents picking each answer option.

Table 2. Candidate Familiarity, among Republicans

Potential Candidate Very Familiar Somewhat familiar Somewhat unfamiliar Very Unfamiliar Total Responses Mean
Mike Huckabee 17 17 15 16 65 2.46
Paul Ryan 11 28 11 15 65 2.46
Rand Paul 11 25 8 21 65 2.60
Chris Christie 10 23 12 20 65 2.65
Jeb Bush 8 24 11 22 65 2.72
Rick Perry 9 20 13 23 65 2.77
Marco Rubio 11 14 12 28 65 2.88
Ted Cruz 6 14 14 31 65 3.08
Ben Carson 8 8 5 44 65 3.31
Scott Walker 2 10 9 44 65 3.46

Note: Entries are the number of respondents picking each answer option.

On the Democratic side, candidate familiarity correlates highly with vote preference for the primary. Clinton is the best-known candidate, and she is the preferred nominee regardless of whether we prodded students to choose a candidate. Without the undecided option, a majority chose Hillary Clinton; however, her support declined when the undecided option was presented. The outliers for vote preference compared to name recognition were Elizabeth Warren and Joe Biden. Elizabeth Warren pulled in more voters than expected compared to her familiarity level, but is considered by many a Democratic sweetheart. On the other hand, Joe Biden had vastly fewer voters than his familiarity level would suggest, possibly due to opinions of the job he has done as Vice President.

Table 3. Democratic Nomination Voter Preference

Potential Candidate Without Undecided With Undecided
Hillary Clinton 52% 40%
Elizabeth Warren 17% 15%
Joe Biden 17% 3%
Andrew Cuomo 9% 3%
Bernie Sanders 5% 1%
Have not decided yet N/A 37%

On the Republican side, voter preference is mostly undecided and widely dispersed. For example, Mike Huckabee and Paul Ryan were preferred less than simple familiarity with them would predict, although the race for the nomination is close and bunched. However, unlike with the Democratic nomination, when provided an undecided option, no candidate stands out as an obvious leader. Rand Paul, while leading over the other potential Republican candidates with 22%, still lags far behind the percentage of respondents who have yet to decide (47%). This variance in voter preference between when an undecided option is available and when one is not mirrors the level of unfamiliarity students have with the whole field of potential Republican candidates. Unlike with the Democratic nomination in which Democrats already know Hillary Clinton and Joe Biden and their stances and may feel comfortable in throwing their support behind a candidate this early in the race, Republicans are facing a wide open field with many candidates that have not had an opportunity to connect with the American people recently. This results in a heavy drop off in undecided voters for all candidates except Rand Paul, who is presumably drawing support from the strong Libertarian base on campus that will already know Rand Paul and his views. This huge proportion of undecided voters draws into question the validity and purpose of polls on the Republican nomination that do not include such an option.

Table 4. Republican Nomination Voter Preference

Potential Candidate Without Undecided With Undecided
Ben Carson 18% 3%
Rand Paul 18% 22%
Paul Ryan 12% 7%
Chris Christie 11% 1%
Jeb Bush 11% 9%
Mike Huckabee 8% 4%
Rick Perry 6% 0%
Ted Cruz 6% 1%
Marco Rubio 6% 4%
Scott Walker 5% 1%
Have not decided yet N/A 47%

In the last part of the survey, respondents were randomly assigned to a potential presidential race between one of three Democratic candidates and one of three Republican candidates. The three candidates from each party were selected from national polls measuring the leaders for each nomination. Further analysis on these match-ups will be conducted throughout the week, but our initial results are posted below. What may be worrisome for Democrats is that Hillary Clinton is behind in all three match-ups, and her possible races have fewer undecided voters. Check back in with Pack Poll later this week for further analysis of the potential presidential races.

Table 5. Hillary Clinton Potential Presidential Races

Potential Candidate Republican Hillary Clinton Undecided
Jeb Bush 35% 23% 42%
Rand Paul 43% 34% 23%
Mike Huckabee 41% 39% 20%

Table 6. Joe Biden Potential Presidential Races

Potential Candidate Republican Joe Biden Undecided
Jeb Bush 47% 22% 31%
Rand Paul 33% 35% 33%
Mike Huckabee 27% 35% 37%

Table 7. Elizabeth Warren Potential Presidential Races

Potential Candidate Republican Elizabeth Warren Undecided
Jeb Bush 33% 23% 44%
Rand Paul 32% 28% 40%
Mike Huckabee 17% 31% 52%


NOTE ON METHODOLOGY: This survey included multiple split-ballots, with sample sizes as low as 47 amongst the potential presidential race polling questions. This decrease in n results in an increase in margin of sampling error. While sampling error is a potential issue present in this survey, it is not the only source of survey error. This survey could face survey error from response (those without strong opinions who received a survey about the 2016 presidential election may have been less likely to respond), measurement error, and other types of survey error.

Click on the link below to download a PDF of the results from this flash poll, and the SPSS file for the data. Feel free to email us if you have any questions.


2016 Early Presidential Flash Poll (Topline Report and Data)

Leave a Reply