It is one of a handful of races that could determine which party controls the Senate next year, but analysts — including the veritable Nate Silver of Five Thirty Eight — agree that the polling in the Alaska Senate race is an absolute mess.
It’s nobody’s fault. Just a historical fact. But it complicates the process of capturing the dynamics of the battle between Republican Dan Sullivan and incumbent Sen. Mark Begich (D-AK).
“Alaska is a hard state to poll accurately,” Silver wrote in his Sunday review of the polling there. He also concluded that in seven statewide races since 2000, Alaska polling has overestimated the Democratic candidate by 7.2 points compared to election day returns.
TPM’s PollTracker average currently gives Begich a 0.6-point advantage.
Silver and the New York Times’s Nate Cohn, who also reviewed the suspect polling in The Last Frontier last week, noted that many of the surveys so far have been conducted by partisan entities or studied a small sample size. A YouGov poll conducted online had 400 respondents. Two of the other most recent polls came from Rasmussen Reports and Senate Majority PAC, a Democratic group.
To underscore their point on the dearth of polling, PollTracker shows seven surveys in Alaska since June. Kentucky has had 15, North Carolina has had 17 and Iowa has had 14 for their respective Senate contests.
The problem, as best that experts can figure it, is the low-response rate that plagues pollsters everywhere — but has an outsized impact in a state with such a small (730,000) and transient population.
Whatever the reasons, the bottom line is that the electoral picture is decidedly fuzzy in one of the most important elections of the year.
“Either way, it is quite easy to imagine that sometime around 1 a.m. on Nov. 5, with most of the national races called, we might realize that domination of the Senate is coming down to Alaska,” Cohn wrote, “and that we have no idea what’s going to happen.”