Quote:
Originally Posted by northcrunk
The problem with polls has always been the people who actually answer them. You don't get a full demographic picture of peoples opinions. You end up with the opinions of whoever is willing to sit there on the phone with pollsters answer questions for 30 to 60 min so it is usually older people with land lines who respond. We would have a better cross section of people but less reliable data doing online polling. I don't know what the best way is.
|
That's a bit of a mischaracterization of the problems with polls. Good polls (which Quinnipiac is), attempt to correct for any problems that they have in sampling cellphones vs. landlines, by sampling both and then correcting for the number of households that have each.
Nonetheless, even in a good poll, the important thing to look for isn't the exact number, it's the trends and where change is coming from (in this case, a slightly downward trend), and the subtabs (in this case, losing support amongst whites without a college degree), both of which suggest that Trump's hard floor of support might not be as hard as previously thought.