In an earlier blog post, I cited a Suffolk University poll on the New Hampshire Democratic primary race that showed a 10 point difference between Hillary Clinton, the front runner, and Bernie Sanders. In that poll, Sanders does particularly well with voters on the western side of the state, nearest to Vermont, where he leads Clinton by 21 percentage points. Unfortunately for Sanders, he does not do as well in the more heavily populated southern section of the state. On the other hand, he does better where he is better known, and he has six months to become more familiar to New Hampshire voters. At this point, only 12 percent of New Hampshire voters said they never heard of Bernie Sanders, while one third have never heard of Martin O’Malley.
Where Sanders may run into some trouble (apart from facing challenges that may prevent him from being on the ballot at all) is that while he is the top choice for 31 percent, he is second choice for only 17 percent. Vice-president Joe Biden, as yet undeclared, is second choice for 21 percent. Among those whose first choice was Clinton, 34 percent choose Biden as their second choice, 29 percent go to Sanders, and 30 percent move to undecided. Among those whose first choice is Sanders, 54 percent choose Clinton as their second choice. In other words, Sanders voters are more likely to select Clinton as an alternative than Clinton voters are to select Sanders.
In the large Republican presidential field leading to the 2012 election, there seemed to be strong competition between Mitt Romney (the front runner) and Not Romney, the search for an alternative to the front runner. That alternative changed throughout the primary season (Bachman, Perry, Cain, Gingrich, and Santorum), but there does not seem to be a Not Clinton movement among Democrats, at least not in New Hampshire.
A poll released by Suffolk University in June, 2015, showed Vermont Senator Bernie Sanders trailing Secretary Hillary Clinton among likely New Hampshire Democratic primary voters by only 10 percentage points. In the June poll, about 41 percent favored Clinton, while 31 percent favored Sanders, with 15 percent remaining undecided; no other declared or potential candidate received more than 8 percent support.
Among those who do not support Clinton, the top reason given in the Suffolk Poll was that voters do not trust her. A non-probability poll conducted by Castleton Polling Institute through social media asked respondents if they support Sanders’ bid for the Democratic nomination and what word they would use to describe Sanders. Among those who support Sanders, the most common descriptor was “honest.” (Figure 1, below, is a word cloud that illustrates Sanders’ supporters’ responses.) While Clinton supporters would likely use adjectives like “competent,” “hard-working,” “capable,” and “strong leader” to describe Clinton, it is not likely that “honest” would be a common description of the former Secretary of State. While we do not have the descriptive data for Clinton comparable to that which we have for Sanders, the narrative attached to Clinton does not indicate that honesty is an associated quality, and the Suffolk Poll supports that view.
So while many pundits suggest that the real challenge that Sanders offers Clinton is that he is more in touch with liberal Democrats and occupies all the space to the left of Hillary, the bigger threat may be that he is seen as more trustworthy, more genuine, and more consistent in his views throughout his political career. The trust deficit has been the Achilles heel of both Clintons, while, at the same time, trustworthiness seems to be one of Sanders’ defining features. Even among those who do not support Sanders, “honest,” “integrity,” and “sincere” came up as words most closely associated with him.
The word that non-supporters of Sanders most cite to describe the Vermont Senator is “Socialist,” (see Figure 2). This will have implications that I’ll raise in another blog post.
Appendix A: American Association of Polling Organizations (AAPOR) Transparency Initiative Details
This project was sponsored by the Castleton Polling Institute.
Castleton Polling Institute collected the data.
The project was funded internally by the Castleton Polling Institute.
The questionnaire was fielded via an online (web-based) survey. Please see Appendix B for the questionnaire.
The survey was conducted through a convenience sample of adults aged 18 and older.
A sample was not drawn for this study. Requests for participation were solicited via Facebook (one request), LinkedIn (one request), and Twitter (three requests), each using the Institute director’s accounts. The survey was administered as an open, anonymous web survey. The following are examples of the requests for participation:
As this was conducted using convenience methods, a sample was not purchased from any third party.
Respondents self-selected into the study. Only adults (age 18 or older) were eligible. Participants were recruited via Facebook, LinkedIn, and Twitter and provided a generic link to the survey. We used a screening question at the start of the survey to screen out any respondent under 18 years of age.
This study is a non-probability, convenience sample.
The number of completed survey is 280. Due to the utilization of a non-probability method of selection, it would be inappropriate to estimate sampling error, including the margin-of-error and confidence level, or provide any statistical description of precision.
The data are not weighted.
The sample size for reported subsets of the data: Sanders supports = 202, Non-supporters of Sanders = 71.
The survey was administered in English via web survey during the period of April 30, 2015 to May 11, 2015. The actual instrument can be observed at Sanders2016.
Public opinion is central to a democracy, and polling is one of the major means by which we measure public opinion. This blog will discuss issues related to polling and public opinion, particularly in Vermont. In addition, we will explore issues related to higher education and preparing students to be active citizens and intelligent consumers of social science research.
Vermont has a long history of active self-government at the local level, and the practice of town meetings gives a platform for citizens to be heard. In that spirit, the Castleton Polling Institute serves as another means for bringing citizen views to the public debate at the state and local level.
Not only will public opinion research give voice to ordinary citizens, it also serves as a way to highlight our differences. While Vermont may appear to the rest of the country as a homogeneous, rural collection of liberals, we have more diversity than meets the eye. What we lack in racial and cultural diversity we make up for in ideological diversity. Public opinion research can highlight differences in views based on gender, partisan affiliation, age, number of years one lives in Vermont, or a host of other demographic variables. Understanding our differences and diversity is important in understanding our politics.
Please check back to this blog to explore issues around public opinion and polling, especially as they pertain to our collective lives in Vermont.