BMaker11 said:
Ugh...not this again. "Just 5,000 people so ignore it" (is the gist of what you're saying). #1 the poll was nearly 6500 people. And #2 presidential poll trackers like Rasmussen, Gallup, NPR, and the Washington times use less than 1500 respondents in their polls (some of those are less than 1000). The resulting percentages as >3% margin of error, and that's based on the sampling to be indicative of a 125M people. 1500 respondent sample size. Less than 3% margin of error representing a population of 125M voters. 6488 is an extremely large sample size. |
And - much in difference to the presidential poll trackers - the sample isn't randomized. That says a lot more than the pure sample size.
