Does “oversampling” definitely mean this is done intentionally to doctor the results?
Or could the higher number of Ds interviewed be a result of a random sampling?
Someone please enlighten me.
“Or could the higher number of Ds interviewed be a result of a random sampling?”
Great question. CNN justified their overpolling of democrats by saying that Republicans were refusing to participate in their polls.
I find it interesting that Trump only does badly in polls linked to news organizations. I wonder why.
It is supposed to reflect the fact that there are more registered Dems than Reps. Which BTW is true. In this case the 9% more is a bit high nationally. Nevertheless a 9% oversampling of Dems that results in a 12% lead for Clinton is not a good result for Trump. Had the result been a 6% Clinton lead there would be a reason for Trump’s campaign not to be disheartened, but 12% is not encouraging.
If you say Trump when a pollster calls, they hang up on you and record your response for Hillary.
So people lie and tell them what they want to hear.
What do you really expect from a liars’ poll?
That’s true here with the ridiculous D+ oversampling and everything else.
Its fiction.
Over sampling can occur, and you can see that in the internals, but what you can not see is the turnout model used by the polling agencies. You can over sample Dems, but if your turnout model has Dems voting at a lower rate than Reps, it will not be that important.
I feel these polls are using a turnout model that favors Dem turnout by a great deal.