This would absolutely not happen in a valid sample survey. These people are deliberately skewing the results.
It would be GREAT if we could find someone who gets called, responds 'wrong way' (the response they want), then gives the same age group. If that respondent don't get thrown out, I'd say that's rock-solid evidence of the dishonesty of the pollsters.
doesn't get thrown out. Duh.
"This would absolutely not happen in a valid sample survey. These people are deliberately skewing the results."
That's not entirely true. I used to work for a company that die phone surveys, and they have quotas of many different categories. For example - the client wants a 250 respondents who think "right way," and 250 respondents who think "wrong way." Then these 250 have to be broken down into a certain number male and female, and certain age groups. It depends on how your calling goes, and who is willing to talk to you. There were many times when we would have an evening of getting lots of women to complete surveys, and our supervisor would have to stop everyone and make us ask for a male. (Not fun for a female!!!) Most times our demographic questions were at the end of the survey, but if they are at the beginning, they are screening questions. SO I think those questions are valid - why do a whole survey only to find out it's not the group you needed.
On the other hand, I think the real skewing comes from the way the questions are asked, no matter who they ask. That "is the country headed in the right direction or wrong direction" is open to so many interpretations. The person who answers it may feel a certain way for a reason, and when the network interprets it, they will make it out to mean something entirely different.