Some people have asked me why the margin of error on the latest Minnesota Public Radio News/Humphrey Institute poll is so high. It's plus or minus 5.8 percentage points. I asked similar questions to Larry Jacobs from the Humphrey Institute both before and as we were working on our stories about the poll.
He said the pollsters deliberately took a conservative approach while calculating the margin based on the response rate, the design of the survey and the methods they use to weight the responses. In other words, the margin of error is on the high side compared to what other pollsters might use, which would put the margin at about 3.7 points.
The margin of error gets bigger when the poll considers smaller samples of people, such as likely voters in the DFL primary. So when you look at the results that show Mark Dayton leading Margaret Anderson Kelliher by 10 points, the margin of error is between 6.2 and 8.75 percent. Jacobs notes that Dayton's lead is still outside the high end of that margin. He believes the bottom line is that the pollsters are being extra cautious and Dayton's lead holds up.
Dayton's lead is NOT outside the MOE. The MOE applies to both candidates. As in, Dayton support could be 6 points higher or lower than it is in the poll and MAK's support could also be 6 points higher or lower than stated. Both of these margins are independent of each other. Therefore, there is a possibility that MAK is actually leading. Finally, in a poll with so many undecideds, leading now means little.
Thank you, James. It seems like a serious problem, in such a poll-obsessed political climate, that we don't know how to properly read polls.