| Interviewing about 1000 americans seems just about right if you want a 3% error, 95% confidence poll of the general view of the US population. Rather, I think the error comes from the selection. I mean, how many normal people have a) the time, b) the patience and c) the trust required to pay attention to anyone calling them up on the phone? If someone calls me on the phone (in the US) and starts asking me questions, I'd be counting the seconds to the point where the sales pitch comes in. Now, if no salespitch shows up, it will be too late to change all the ridiculous answers I've given. If the poll takers are any good, they have a decent lying/faking scale on the test so they know to exclude me. Telemarketing has simply destroyed the telephone as a medium for meaningful statistical polls.
If the poll was made correctly and the sample was chosen properly, you would be surprised at how accurate a poll is in showing the general sentiments of the population. Discrediting polls with "it's just statistics" is plain ignorance.
__________________
The direct use of force is such a poor solution to any problem, it is generally employed only by small children and large nations David Friedman |