I noticed that the in the education component of the Atlas-Intel sample, the college degree or higher - at 46-48% - seems maybe unduly high? True, turnout is higher for that demographic than for lower education levels, but lower education voters trend heavily for Trump.
I noticed that the in the education component of the Atlas-Intel sample, the college degree or higher - at 46-48% - seems maybe unduly high? True, turnout is higher for that demographic than for lower education levels, but lower education voters trend heavily for Trump.
If they are underrepresented in the sample results could be skewed in favor of Harris, no?
One other thing - Trafalgar also has a very respectable rating, did you check for or note any discrepancies between their findings and Atlas-Intel's?
Criticism of polls often comes down to whether or not a particular demographic is oversampled or undersampled relative to actual voter turnout.
And with good reason. If the composition of the voter turnout deviates significantly from the composition of the sample population, the poll is going to be skewed.
Did AtlasIntel oversample college-educated voters? That is of course a possibility, and that could skew the numbers towards harris.
In Pennsylvania and Wisconsin, AtlasIntel's poll sample favored Republicans over Democrats by 3.6pp. In Michigan, the sample was more evenly distributed. AtlasIntel also projects a smaller Independent voter demographic in Pennsylvania than in Wisconsin or Michigan.
Last month Gallup released a party identification poll which showed people tended to identify more as Republican than as Democrat by about 3pp (31%-28%) with Independents being the largest identifying demgraphic. If we go by the Gallup distribution, Pennsylvania undersampled Independents and oversampled both Democrats and Republicans. while Michigan and Wisconsin likely undersampled Independents.
Nationwide, Trump does very well among Independents, according to AtlasIntel. In Michigan, Pennsylvania, and Wisconsin, Kamala Harris does slightly better than Trump among Independents.
We won't know until the final vote tallies where the oversampling and undersampling actually occurs.
As regards Trafalgar Group's polling, their latest Michigan and Pennsylvania polls show Trump leading by about the same margins as AtlasIntel. Trafalgar was the second most accurate pollster for 2020, so their numbers do tend to reinforce AtlasIntel's.
Browser issues - or something - aren't letting me like replies - thanks for responding.
BTW - interesting House race to watch (don't know of any polling) - Ohio CD-7, probably the only one where an independent has a shot.
Novice China oriented Dem v. rabid Zio-supporting Republican incumbent Max Miller v. peacenik veteran former (Dem) Rep. Dennis Kucinich, running as an independent.
Not getting hopes up, but it would be great to see the district deliver a big FU to the Uniparty there.
Hi, Peter - nice work.
I noticed that the in the education component of the Atlas-Intel sample, the college degree or higher - at 46-48% - seems maybe unduly high? True, turnout is higher for that demographic than for lower education levels, but lower education voters trend heavily for Trump.
If they are underrepresented in the sample results could be skewed in favor of Harris, no?
One other thing - Trafalgar also has a very respectable rating, did you check for or note any discrepancies between their findings and Atlas-Intel's?
Criticism of polls often comes down to whether or not a particular demographic is oversampled or undersampled relative to actual voter turnout.
And with good reason. If the composition of the voter turnout deviates significantly from the composition of the sample population, the poll is going to be skewed.
Did AtlasIntel oversample college-educated voters? That is of course a possibility, and that could skew the numbers towards harris.
In Pennsylvania and Wisconsin, AtlasIntel's poll sample favored Republicans over Democrats by 3.6pp. In Michigan, the sample was more evenly distributed. AtlasIntel also projects a smaller Independent voter demographic in Pennsylvania than in Wisconsin or Michigan.
Last month Gallup released a party identification poll which showed people tended to identify more as Republican than as Democrat by about 3pp (31%-28%) with Independents being the largest identifying demgraphic. If we go by the Gallup distribution, Pennsylvania undersampled Independents and oversampled both Democrats and Republicans. while Michigan and Wisconsin likely undersampled Independents.
Nationwide, Trump does very well among Independents, according to AtlasIntel. In Michigan, Pennsylvania, and Wisconsin, Kamala Harris does slightly better than Trump among Independents.
We won't know until the final vote tallies where the oversampling and undersampling actually occurs.
As regards Trafalgar Group's polling, their latest Michigan and Pennsylvania polls show Trump leading by about the same margins as AtlasIntel. Trafalgar was the second most accurate pollster for 2020, so their numbers do tend to reinforce AtlasIntel's.
Browser issues - or something - aren't letting me like replies - thanks for responding.
BTW - interesting House race to watch (don't know of any polling) - Ohio CD-7, probably the only one where an independent has a shot.
Novice China oriented Dem v. rabid Zio-supporting Republican incumbent Max Miller v. peacenik veteran former (Dem) Rep. Dennis Kucinich, running as an independent.
Not getting hopes up, but it would be great to see the district deliver a big FU to the Uniparty there.