Quote:
Pool: 130 million internet users across 5,400 web sites . . .
2,200 respondents . . .
1.5 percentage point margin of error
|
I don't quite understand how this poll was conducted. But this doesn't sound like any good polling methodology I have ever heard of. Weighing by sex and age? That's part of a good methodology, but useless if the data quality coming in is poor.
According to their powerpoint, more people have purchased ten or more books in the past year than have purchased none -- even if you include "uncertain" with none! This does not meet the smell test.