Though the New Hampshire Senate race has been flirting with becoming competitive for months, no pollster has projected the race in the toss-up realm since January … that is, until now. The University of New Hampshire (Aug. 7-17; 827 New Hampshire adults) finds Sen. Jeanne Shaheen (D) leading former Massachusetts Sen. Scott Brown (R) by only a 46-44 percent clip. On the other hand, if ex-Sen. Bob Smith were to become the Republican nominee, the incumbent would enjoy a 50-36 percent margin.
In terms of public perception, Sen. Shaheen scores a 48:36 percent positive to negative personal favorability ratio, while Brown is upside-down at 36:38 percent. Smith is down to only 15:20 percent positive to negative, with just 35 percent saying they know enough about him to form an opinion. This gives us an idea as to how fast a population can change. The polling history shows that at Smith’s high point during his dozen years in the Senate (February, 2002 in this instance), he recorded a 51:31 percent favorability index with as many as 82 percent of the voting universe knowing enough about him to form a specific opinion.
The UNH ballot test is the good news for ex-Sen. Brown. The bright side for Sen. Shaheen is that they are the polling source. UNH has one of the worst track records in the public polling industry. Their erratic results are currently in evidence when they forecast Rep. Carol Shea-Porter (D-NH-1) to be leading former Rep. Frank Guinta (R-NH-1) 48-32 percent last October, trailing him 39-45 percent in January 2014, ahead once more in April 44-35 percent, and then falling behind the former congressman yet again in July, 43-46 percent. Such huge swings during a time where campaigning is not fully underway suggests polling methodology flaws.
Part of the reason for their past inaccuracies most likely relates to two major shortcomings in their polling processes. First, they typically stretch the sampling period over many days. Their current statewide Granite Poll, the one producing the aforementioned Senate numbers, featured a 10-day polling period. Almost all professional pollsters now complete surveys in two days, if not one. Secondly, UNH does not screen particularly well for registered voters when beginning with a pool of adult respondents. Though they show that 609 of the 827 individuals sampled in this poll are “likely voters”, the methodology does not indicate that the responses came only from those people either eligible or likely to vote in the November election. Rather, the report appears to indicate that the published responses were from the entire adult sample.
When grouping this poll with others conducted throughout 2014, we find a much larger average spread between the two candidates. Of 19 polls conducted by 13 different pollsters, Sen. Shaheen averages 48 percent support versus Brown’s 39 percent. In looking at the last 10 consecutive polls taken before the UNH survey in question, the same 48-39 percent spread occurred, a testament to the race’s polling consistency. This lends further credence to the argument that the latest UNH Granite State poll is an outlier.
Still, the chance of a Brown upset cannot be discounted. New Hampshire voters have swung hard and fast since 2006, inclusive, actually defeating more federal office holders than re-electing during that stretch of time. And, those swings have been ballot-wide, with virtually every officeholder of one party or the other either being swept in or out of office depending upon the year. It is certainly possible for such a trend to again occur here in 2014.