Thursday, October 6, 2016

Why We Love Polls

Because with little or no effort you can get any answer you want. A lot of them right now are heavily skewed to discourage targeted voters from voting.

Reuters was reporting Trump up 2 this morning which would be a significant jump from being down 6 yesterday. It makes a good headline, but look under the hood. To get to Trump by 2, the sample included 46% democrats, 32.5% Republicans, and 13.9% Independents. In the last election, Coloradoans cast more Republican votes than Dem by a small margin. Expert gerrymandering in 2010 got the Donks the House and Governorship.

Other polls suggest that about 11% of each party is planning to vote the other party, being unable to stomach the selected candidate. If the I's break 100% for Trump (unlikely) the poll would show a 46/46.4 in Trumps favor. If the I's break 50-50 (more likely), the results look like 53/39 in favor of Hill. Right now no one sees that happening either. Adjusting for the disgruntled would move about 5.8 points from D to R and 4.3 points from R to D making the margin 51.7/40.3.

In order to get to a 2% Trump win, about 8 points of the D's would have to switch to the R side, or voter turnout would have to look more like it usually does around here. Still, if it took that big an oversampling of Dems to get Trump 2 points up, think what an approximately even split must yield.

No comments: