Posted on 10/10/2016 10:12:51 AM PDT by TheRef
![]() ![]() |
The YouGov poll, to its credit, provided internals for their poll. CNN has not, although apparently there is more to come.
According to the YouGov poll, Clinton edged out Trump in the debate last night by a 47 to 42 margin. The CNN poll, absurdly, found that Clinton won the debate by a twenty-three point margin, 57 to 34. CNN notes in its article that their results match voter preference from before the debate, where 58% of the debate watchers they are surveying support Clinton.
Sadly, some analysts that people take seriously are lending credibility to these polls, most notably Nate Silver. These analysts shouldn't attach any significance to these polls because they make two fundamental mistakes.
The YouGov poll surveyed registered voters and did not apply a rigorous likely voter screen. They included all voters who did not say that they "definitely" will not vote. Polls of registered voters generally favor Democrats, as even Nate Silver acknowledges. Apparently CNN did not even limit its poll to voters at all, but all debate watchers. Did they poll children? Possibly.
October Surprise | After 1 day | After 2 days | After 3 days | After 4 days | After 5 days | After 6 days | After 7 days | |
National Shift | --- | 1.45 to Clinton | 3.0 to Clinton |
The second big mistake that YouGov made was not adjusting its voter sample to reflect the population at large. The poll assumes that Democrats would outnumber Republicans in this election by 10%. Of course, if that's the turnout, Democrats will win in a landslide.
Because CNN did not bother to limit its poll to voters, the network would have no need for a turnout model.
In summary, the CNN poll is not a political poll but something more like a television analysis. It's just not useful in a political context and shouldn't be cited.
YouGov did an actual political poll, but did not apply a real likely voter screen and heavily oversampled Democrats. Garbage in, garbage out. IGNORE THESE POLLS.
Follow my new account on Twitter for daily updates on the election!
This race doesn’t fit an Obama v. McCain model or an Obama v. Romney model. Motivation among voter blocs is absolutely upside-down from those races. This sampling methodology is wrong, and I believe it’s deliberate.
the pundit class (left and right) are just repeating 538 . com verbatim. they are not thinking or reporting.
I didn’t see but a couple links to polls this time ...????? I’ll vote in more polls if there’s a list of links somewhere like there was last time
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.