Quote:
Originally Posted by Kelt
A couple of other things I always like to watch out for: When a certain publication says something like “135 of our readers responded”… Well, we know that those readers are probably already pre-biased by their selection of publication. Also, I always look for the margin of error which in politics is frequently +/-3%, If we use this poll as the example taking away 3% from the higher figure and adding 3% to the lower figure would completely reversed the outcome.
This poll did not specify any margin of error. The sample group was selected from “registered voters“. Obviously not all of them and no other selection criteria was listed. Leading me to the conclusion:
|
Both the poll I linked to and the one Martina did both specified a margin of error, so I'm not sure what you are referring to.
One polls is simply that - one poll. It is the trend over time or what a group of respected polls are saying that is important - or you can just think of them as malarkey if you wish, lol. I also look at the 538 website to see which polls they rank as most reliable.