All polls have the same problem. They take a non random sample of people and sprinkle in some magic and call that reflective of electorate.
How they try to get a representative sample of likely voters from all the people that sign up or that they call leads to differences beyond the statistical variance of the poll. So while the polls statical accuracy is +/-3% 19/20 that assumes a random sample of the population. None of these polls are random samples.
Since we know there is sampling error in our polling the best way to counteract that affect is to take all of the differen pills and average. It isn’t which poll is right. They are all wrong in their own way. Polls being the same could reflect that poll is accurate. It would also suggest it has the same error.
The other interesting thing about the Janet Brown poll is that this is an unreleased poll done for some private group which then someone with access to the data released some of the numbers to the media. Now Janet Brown is a high quality pollster so it’s unlikely the questions or methodology was skewed to drive towards a result based on the poll sponser but one might suspect if the result was different the leaker wouldn’t necessarily leak it. Meaning if this poll wasn’t advantageous for certain groups the data wouldn’t have been leaked which makes it skew data. So I would like to see who commissioned the Janet Brown poll and was it intended to make it public no matter what.
|