Rich Ulrich
Fri, 27 Oct 2000 11:43:37 -0700
On 27 Oct 2000 07:26:43 -0700, [EMAIL PROTECTED] (Wolfgang Rolke) wrote: < snip, about polling ... > > A very nice > discussion of the issues involved can be found at > http://slate.msn.com/framegame/entries/00-10-26_92147.asp . An explanation of the > methodology of the Gallup tracking poll is at > http://www.gallup.com/poll/faq/faq000101.asp , where they say that the +-4% is > for a 95% confidence interval. I would like to pose another question, though, > namely what a reader should make of all these polls. There is of course the > additional problem of the horse race question: if Bush's percentage is estimated > at 46%+-4% and Gore gets 44%+-4%, than in fact Bush is 2% ahead of Gore, but this > difference has an error of +-8% and so is clearly not significant. Now if a poll < snip, rest > No, you have it wrong. Think about the case where the two candidates/options add exactly to 100% -- the "4% error" is exactly the same 4% for both, since one goes down exactly as much as the other goes up. This is "correlation" among the proportions. With a few percent as "Missing" or "Nader", the Gallup correlation of support for major candidates is not quite -1.0, but it is close enough. Oh, by the way, when you do have error terms to add, it is almost always the squared values, "variances," that get summed. So two Errors of 1.0 add up to one error of 1.41, since that is the square root of (1-squared+ 1-squared). -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html ================================================================= Instructions for joining and leaving this list and remarks about the problem of INAPPROPRIATE MESSAGES are available at http://jse.stat.ncsu.edu/ =================================================================