|
|
 Originally Posted by Poopadoop
In this case, the result they got after appropriate weightings was 39% to 34% in favor of Clinton. A result based only on the numbers obtained in the poll with no weightings would have been Clinton 51% to 26% Trump. Obviously they did the weightings.
This is why they need to release multiple renditions of the same data. This instance suggests the poll is deeply unreliable more than it suggests a 25 point slaughter in favor of Clinton. Wacky response results like the above are not rare. Weighting is an art, and its purpose is to best represent the population. The way this poll was weighted has a very big bias towards representing the sample.
|
Reply With Quote