When I wrote up the MPR/Humphrey Institute poll released this morning I was unable to get the sample-size or the margin of error calculation for the portion of the poll that pitted the three DFL primary candidates against one another. Now I have it, and the sample is quite small.
Of the 701 Minnesota adults who were interviewed, 246 said they expected to vote in the primary, according to Larry Jacobs of the Humphrey, who conducted the poll. Using what Jacobs said were conservative assumptions about margin of sampling error, that would produce an MOE of plus/minus 8.7 percentage points.
In the poll, Mark Dayton led by 10 percentage points over Margaret Anderson Kelliher. Taking the most conservative approach of doubling the MOE (because of the plus/minus effect), that would mean Dayton’s lead was within the error margin. Also, consider the large portion (28 percent) of that sub-sample who, said they planned to vote in the primary but who, when offered the choice of Dayton, Anderson and Matt Entenza, said they favored none of the above (I’m guessing many of this group are Republicans) and the number who expressed a preference seems smaller still.