math lesson
By Diogenes (articles ) | January 07, 2008 10:22 AM
Today, class, we'll be doing a simple problem involving statistics.
An American Research Group poll, released on the eve of the New Hampshire primary, shows Senator Obama leading Senator Clinton among the state's likely Democratic voters, by a margin of 11%.
A Franklin Pierce University poll, released the same day, shows Obama leading Clinton by just 3%.
The American Research Group pollsters claim that their results are accurate within 4%; the Franklin Pierce survey gives a 4.9% margin for error.
Question #1: Can these two poll results be reconciled, given the margins of error provided by the pollsters?
Question #2: If so-- if an 11% margin might really be only a 3% margin-- what is the predictive value of these polls?
|
An appeal from our founder, Dr. Jeffrey Mirus: Dear reader: If you found the information on this page helpful in your pursuit of a better Catholic life, please support our work with a donation. Your donation will help us reach five million Truth-seeking readers worldwide this year. Thank you! |
Easter Campaign: Progress toward our Spring 2013 goal ($34,555 to go):
$80,000.00
$45,444.82
43%
57%
|
All comments are moderated. To lighten our editing burden, only current donors are allowed to Sound Off. If you are a donor, log in to see the comment form; otherwise please support our work, and Sound Off!
-
Posted by: -
Jan. 09, 2008 4:53 PM ET USA
Clearly, they were all wrong Bad side: Hillary Won Good Side: Media and pundits had egg on their faces Overall its a wash
-
Posted by: -
Jan. 09, 2008 4:54 AM ET USA
Who cares. Hillary won.
-
Posted by: -
Jan. 08, 2008 8:34 AM ET USA
Good point, but not quite stastically accurate. An 11% margin might really be only a 7% margin, and a 3% margin might be up to a 7.9% margin, but an 11% margin can't be a 3% margin. Putting the two polls together actually provides the much more precise margin of only 7-7.9% for Obama. It's only useful, of course, if the pollsters accurately represent the margins of error and the polls were well-conducted otherwise, which is by no means a guarantee of accuracy and predictive value.
