Decoding the polls – Yes, there’s anti-Trump bias
Bottom Line: As part of election analysis, and as a way to adjust for faulty sampling within polls, I start with sampling. As I’m inclined to say... If the premiseof anything is false, anything built on top of that false premise will be too. Day in and day out we’re presented with new polling purporting to show nearly all Americans disapproving of President Trump and a news media seemingly ready to hand the election over to Joe Biden. It’s almost like they’re rooting for him... Let’s break down some of what’s going on with these polls.
The first place to start with polling is to look at the partisan breakout of those being sampled. In the case of the national approval polls and head to head matchups with Biden we should evaluate the country as a whole for comparison's sake. When we do, here’s what we find...
Gallup has conducted monthly polling of part is an identification since 2004. In their most recent sampling from June, Democrats held a six-point advantage. It’s the largest advantage of the year for Democrats which supports polls having recently moved in their favor generally, however that’s not what’s reflected in the national polls. Democrats are being given a much bigger advantage. First, not every pollster provides disclosure for who they polled. Some simply assign a “margin of error” and leave it at that. For those that do they often make you work to find it. I kid you not that for two national pollsters I was more than 100 pages into their polling before discovering the identification of who was being polled. Here were the results.
- Politico/Morning Consult favored Democrats by 8%
- The Economist/YouGov was a 9% advantage for Democrats
- Reuters/Ipsos gave Democrats a 12% edge
Three accredited national pollsters. Three samples that significantly over surveyed Democrats. The average sample favored Democrats by 4% more than is appropriate or justifiable. The average approval of President Trump in these polls? 41%. Whether it’s intentional or just poor methodology left unchecked by news media – the onslaught of negative Trump polling is provably overdone. The difference between a 41% and 45% approval rating might not sound like much on the surface but in a historical context in the Electoral College, that 4-point swing literally doubles the odds of President Trump being reelected. To be clear, the news isn’t great for Trump – even when the methodology is corrected in these polls but it’s meaningfully better than what’s being presented.