Tuesday, October 19, 2004
What's a Margin of Error in Poll results?
I'm not a math person. I can work out fairly simple things. But statistics and polling methodology are a bit of a mystery to me. Perhaps that's why I'm trying to learn more about these things.
I was surprised to find out that the margin of error in a poll applies to BOTH candidates when people are asked who they will vote for. If a poll says 47% of voters are favoring Bush and 44% are favoring Kerry with a 4% margin of error, this means (assuming the poll is valid in all other ways) the actual count for Bush could be anywhere from 43% to 51% and for Kerry from 40% to 48%. Notice that BIG overlap area.
Find out more with this Google Search: define:margin of error:
Definitions of margin of error on the Web: A measurement of the accuracy of the results of a survey. Example: A margin of error of plus or minus 3.5% means that there is a 95% chance that the responses of the target population as a whole would fall somewhere between 3.5% more or 3.5% less than the
responses of the sample (a 7% spread). www.icpsr.umich.edu/help/glossary.html
Update: Slightly related to this is a small book I've been slowly working my way through. It's not difficult to read but I seem to need a little time to digest it. More Damned Lies and Statistics: How Numbers Confuse Public Issues by Joel Best. It is helping me clear up some of the fog surrounding statistics and especially helping me to evaluate the validity of statistics and charts whenever they are used in the news.