Sunday, October 28, 2012

Who's ahead?

Are you getting frustrated with the deluge of polls--national and local?  I am.  I've really just about had enough of these ridiculous polls that claim Romney is ahead by .7% or Obama is ahead by .6% with a error margin of +/- 3%.  Hello?  Nobody can know who is ahead when the tallies for both candidates fall withing the margin of error.  The margin of error is the "we can't tell" zone of a poll.  All you can properly say is that both candidates are statistically tied. 

If Smith tallies 47% and Jones tallies 46% and the margin of error is +/- 4.5%, Smith is NOT ahead of Jones; neither is Jones trailing Smith.  The only proper interpretation to be made is that Smith and Jones are statistically tied.  You can't claim an accuracy finer grained and more detailed than that which the polling results permit.  If you do, then you're a fool because there's simply no statistical basis on which to make that claim.  The uncertainty (otherwise known as the margin for error) simply won't let you make such a precise claim.  Imprecision rules.

Right now in the battleground states Obama and Romney are in statistical ties in many races, like Ohio.  No possible poll can tell you who is "ahead" (whatever that may mean).  People who claim otherwise--like bloggers and other internet bloviators--are idiots.  They don't understand elementary statistics.  Tune them out.  Also tune out the all-too-confident chatterers on TV news shows or on YouTube or on whatever venue you find them.  They can't know what they claim to know.  Nobody can.

Unless something causes a sudden break away from one candidate to the other, we're going to have to wait until November 6 to find out "who's ahead."  For now, it's neck and neck.  And that's all you can say about the race.  Period.

No comments: