Monday, November 7, 2016

The "Comey Effect" Myth; There Never Was Any Poll Impact By The Comey Report

The Comey report, which exercised so much media and DNC angst and much Republican jubilation because of the suspected massive "October surprise" effect on the polls and the campaigns, turns out to be, yet another, campaign myth.

It may have been an"October surprise" but a very damp squib of one in the end

Such luminaries as Cher tweeted that, since Clinton was exonerated, as she and no doubt countless Democratic supporters see it, Comey "may have cost Clinton the election." 

If Trump does win no doubt these same voices will claim it is a 'stolen election caused by a Republican (such as Comey is) conspiracy and use it as a stick to beat the Trump administration for four years.

However lets look at the actual facts;

Clinton took over the lead from Trump in the aggregate of five tracking polls on October 24th. I do not consider the MSM "one off" polls in this exercise because of the outlandish Democratic sampling advantage nor the preposterous Reuters "poll" and, until they fixed it 10/31, the ABC Washington Post Tracker (the sixth tracker from then) with it's prior and ludicrous "Clinton+12" results.

Clinton's highest aggregate was 0.7 10/26-27, pre-Comey, and her decline clearly started after that. On 10/29, a day after Comey she went up in some polls and down in others. 

The only subsequent day where there was any measurable change in Clinton's aggregate polling was 10/31 where she was down 1.6 points behind Trump which could easily be just  polling noise which seems clear as being the case by her being only 0.7 down the next day.

By 11/4 Clinton and Trump are in a dead tie, including the ABC/WashPost tracker which seems determined to boost her again, coming in at Clinton+5 from being tied on 10/31. As with aggregation this is balanced out by the USC/LATimes boosting Trump by 5.4 points.

So for all the furor, cries of bias and cheating Clinton was ahead on 10/28 by 0.3 she's tied on 11/4 a decline of 0.4 points. 

That Trump has moved into the lead in a number of key states and the "one of polls have tightened to have Clinton with a 1-2 point aggregate lead reflects voters making up their own minds without due influence from Comey. 

Comey, in the end, appears to have simply reinforced anti-and pro Clinton camps views of her rather than changing minds and the non-trackers Clinton decline being due their changing the way they compute their polls  "herding" to reflect reality rather than the bias their paymasters and/or their own ideology wish them to present.

NB; Rasmussen doesn't release polls on weekends