New poll shows Begich up by 5 percentage points. Read with caution.

A new poll of the U.S. Senate race by Harstad Strategic Research who’s polling for Senate Majority PAC  shows that U.S. Sen. Mark Begich is leading GOP challenger Dan Sullivan by 5 percentage points. Here are the numbers:

begich poll


Polling guru Nate Silver of FiveThirtyEight strikes a cautionary note:

As other commentators have noted, Alaska is a hard state to poll accurately. What we haven’t seen remarked upon is how those misses have come in one direction, almost always overestimating the performance of Democrats.

Here’s Silver’s chart detailing that historical bias:

Silver dem bias

The poll does effect the FiveThirtyEight model, but by less than if it weren’t paid for by the Senate Majority PAC.  It says that Sullivan has a 62 percent chance of winning the seat, which is down from 69 percent as a result of this poll.


3 thoughts on “New poll shows Begich up by 5 percentage points. Read with caution.

  1. Reed

    I read the same FiveThirtyEight article and I question if the bias that Nate Silver identifies here is one driven by incumbency rather than party affiliation. That is to say: do Alaskan voters realize that the more senior their elected official in Washington are, the more ability they have to “bring home the bacon” and vote accordingly? Excluding the gubernatorial elections (where the value of seniority is negligible) every example used by Silver to prove this bias is with incumbents (who happen to be Republican) and Democratic challengers.

  2. Lynn Willis

    I gleaned from this report that the polling industry is doing well. Does three months of positive margin for Begich mean anything? Perhaps it does if it creates an image of pending victory; however, then you might want to look at who asked for the poll.
    Years ago being contacted by a pollster was more significant. Today, you can’t beat them off with a stick. People are starting to “wise up” about the power they have when contacted by a pollster – they lie which skews the results or, like me, they ignore the pollster so the random sample isn’t so random any longer which also skews the result.

Comments are closed.