Ever since my Statistics classes back in my time at University, it has always amused me that people do not understand what "margin of error" actually means.
In particular by the brain dead media.
A 5% margin of error refers to the confidence level of the survey results as a whole. As in there is a 95% chance the survey results are an accurate representation of the entire population.
Even if it meant what you thought, comparing the difference between the two results versus the margin of error % figure is meaningless.
At best if 5% actually meant that each of the 2 results are off by up to 5%, the possible range would be that Johnson is ahead by up to 12 points (54 to 42) or conversely Feingold is ahead by up to 8 (52 to 44).
Ever since my Statistics classes back in my time at University, it has always amused me that people do not understand what "margin of error" actually means.
ReplyDeleteIn particular by the brain dead media.
A 5% margin of error refers to the confidence level of the survey results as a whole. As in there is a 95% chance the survey results are an accurate representation of the entire population.
Even if it meant what you thought, comparing the difference between the two results versus the margin of error % figure is meaningless.
At best if 5% actually meant that each of the 2 results are off by up to 5%, the possible range would be that Johnson is ahead by up to 12 points (54 to 42) or conversely Feingold is ahead by up to 8 (52 to 44).
You can not deduce anything from that.
Same poll has Walker up by 9%, so do you conceed that Barrett is going to lose?
ReplyDeleteTo Anon. #2: I'm not conceding anything.
ReplyDeleteIn addition to not conceding James, you are also not responding to anything of substance.
ReplyDeleteGuess it is easier to just ignore reality.
So lets get this straight Rowen, the poll that supposedly shows Russ in a tie is significant.
ReplyDeleteBut the same poll that shows Barrett getting beat badly is not.
One does have to admire your ability to rationalize and parse reality.