Lies, Danged Lies, And Mark Dayton

The New York Times publishes a morale booster for the DFL (Minnesotan for “Democratic Party”), a week after the Minnesota Public Radio/Humphrey Institute poll showed the Minnesota bubernatorial race a dead heat.

The analysis – from “Five Thirty Eight”, a left-leaning statistics blog run amok (without mentioning anything about “Lies” and “Damned Lies”, for some reason) says Mark Dayton has a 78% chance of winning the gubernatorial election; they claim the current stats point toward a 46-41 Dayton win on November 2.

The DFL Chanting Point Bots are duly repeating that “78%” figure as if it is the be-all and end-all of the story.

They’re ignoring the part where the sausage gets made.

Look at the polls they select for their sampling.

The sample gives full weight (“.99″) to an August 2 Survey USA poll that came out about a week before the DFL primary.  When the DFL was outspending Emmer 16:1, and flogging their astroturf campaign against Target to maximum effect.

It gives 2/3 weight (“.66″) to last week’s MPR/Humphrey Institute poll, which showed the race a dead heat.

It gives .4 credit to the un-creditable Strib/”Minnesota” Poll, which showed a ten point Dayton lead, and a tiny little fringe of weight to Rasmussen and Survey USA polls from earlier this summer that showed Emmer inside the margin of error and long before Dayton started his orgy of spending on his slime campaign.

In other words, the stats that the DFL and media (in Minnesota, one must constantly pardon the redundancy) are jumping up and down and caterwauling about could hardly be better cherrypicked to show Dayton in a commanding lead; they are, to say the least, a misleading sample given a questionable weighting.

Take this number seriously at your own risk.

Cross-posted at Shot In The Dark