>Doug:
>> Phone polls undersample young people (18-35 - esp those who don't
>> live with older family members), nonwhite people, and men. The first
>
>
>This problem has been known for some time. Other factors compromising
>phone sampling include:
>- multiple phone lines which increase the probability of selection of
>those who have them (i.e. those who can afford them are more likely to
>be oversampled than those who cannot);
>-telemarketer fatigue - the intrusiveness of telemarketing produces high
>levels to refusals to participate in any kind of survey; again the
>probability of refusal is non-randomly distributed, which results in
>oversampling and undersampling of different groups
Yes, it's been known for some time, but it appears to be getting worse. Also, lots of people, esp young ones, are dropping land lines for cell phones (which can't be randomly dialed by pollsters), also complicating things. The problem is especially problematic if younger people vote in greater numbers than usual this time.
Doug