Ok, I confess. Part of the Democratic oversample is my fault. I’ve been called at least 5 times by Quinnicpiac in the past two weeks, and I refuse to answer when I see their name on my caller ID.
One of the big pollsters was interviewed yesterday (unfortunately I don’t remember which one)and he said that only 9% of all people called complete the poll. So isn’t it statistically dubious to draw conclusions about the preferences of the other 91% by the 9% who either don’t have caller ID or something better to do than stay home and answer polls?
Zogby was once the King of Pollsters until he started adding his “special sauce.” Look where he is now ... a bad joke. I think he was an MB plant all along, just like his brother. There are plenty of snake-oil salesman who would love to tell you what to believe...
I hung up twice and they hung up thrice
So basically the sample is of people who take time for the survey and who are more likely to associate themselves with Democrats.
I was called once, years ago. I did not finish the poll becasue the questions were so obviously biased toward the desired result.
Poll ‘results’ today are used to skew public opinion, not reflect public thought.
I read one of the interviews
and now I understand the possible source of sample error. They are not deliberately taking the results and adjusting the data down to a sample ratio.
They are predetermining the number of interviews using cell phones and land lines. The cell phone calls will over sample democrats. They claim they are looking for a certain percentage of cell phone calls in order to reach the population that does not have cell phones.
What they are missing is that they are indirectly skewing the polls toward democrats by predetermining the cell phone percentage.
I’m pretty sure this upfront allocation of cell phone calls to land line calls violates the randomness necessary to have a valid poll.