Feb. 5, 2016 — A late-breaking Harper Polling New Hampshire survey (Feb. 1-2; 425 likely New Hampshire primary voters) finds ex-Florida Gov. Jeb Bush surprisingly claiming second place within the Republican presidential field, but he’s still far behind leader Donald Trump. There are, however, three reasons to question the results.
According to the new data, half of which was gathered after the Iowa Caucus results became known, Trump commands first position with 31 percent preference. Bush is second registering 14 percent, followed closely by Ohio Gov. John Kasich at 12 percent. Sen. Marco Rubio (R-FL) posts 10 percent, with Iowa winner, Sen. Ted Cruz, not faring particularly well in the Granite State, dropping to nine percent support.
All of the remaining candidates –- and still including Sen. Rand Paul (R-KY) and former Sen. Rick Santorum (R-PA) who both suspended their presidential campaigns Wednesday morning -– find themselves landing only in the mid-to-low single digits.
There appears to be methodological flaws in the survey, which was conducted through an Interactive Voice Response mechanism. First, the favorability indexes are curious in that the only candidate with a positive ratio is Donald Trump. All of the other Republican contenders, remembering that the respondents are GOP primary voters, are seriously upside down.
Aside from Trump’s 51:47 percent rating, it is Gov. Kasich and Bush who have the best ratios at 45:50 percent favorable to unfavorable. Such results seem unfathomable, but the IVR technology seems to elicit higher negative ratings of candidates than does live calling. Therefore, because the favorability results are highly questionable one should also be skeptical of the overall reliability factor.
The second potential flaw is the presence of self-identified Democrats within the sampling universe. New Hampshire uses a modified primary system that closes for individuals belonging to a political party -– meaning the people can only vote in the primary of the party for which they are registered –- and open for the non-affiliated Independents, or those expressing no preference for party membership. Therefore, to include the responses from an eight percent segment of Democratic Party voters who cannot vote in the Republican primary seems to provide further evidence of an unreliable conclusion.
Third, the geographic sample appears non-reflective of the overall New Hampshire voting population. Using media market distinctions from other states, i.e., Portland-Auburn (ME), and Burlington-Plattsburgh (VT and NY) as the region labels for eastern and western New Hampshire does leave one questioning exactly how the sample was derived, even though the rural area draw seems numerically correct.
The sample then deviates in the most populous sections of the state, however, the south central and southeast regions. In actuality, the Hillsborough/Merrimack County (Manchester, Nashua, Concord) corridor houses approximately 41.6 percent of the New Hampshire population. Rockingham County, containing the state’s Seacoast communities, accounts for 32.1 percent of the state’s people.
But, the Harper sample exaggerates the emphasis, giving Hillsborough/ Merrimack 53 percent of the respondents, and the Seacoast region only 21 percent. A skew this large can certainly produce a different result than voter history would suggest.
Do all of these potential flaws help Bush to a degree that he can leap-frog several candidates into a second place standing when never doing so before? Quite possibly.
We will undoubtedly see more polls before the Granite State votes are cast on Feb. 9. Confirming data will be necessary in order to adequately gauge whether former Gov. Bush is experiencing such dramatic upward mobility.