Loading ...Can you spot the fuzzy math in this Columbus Dispatch story?
flockwoodAccording to the Dispatch, pollsters surveyed 50 women to see who they think won the Oct. 15 presidential debate and it was “a major win for Democrat Barack Obama. Seventy-one percent of these undecided voters thought Obama did better in addressing the issues important to them, while only 9 percent felt that way about Republican John McCain. The group slightly favored Obama coming into the debate, but afterward he won support by about a 2-to-1 ratio.”
Q. What is the mathematical red flag in the above paragraph?
A. If 50 women were surveyed, each woman would account for precisely two percent of the total people surveyed. If only one of the 50 women thought Obama won, that would be 2 percent. If 10 of the 50 thought he won, that would be 20 percent. It’s unclear how McCain could finish with 7 percent support…or 9 percent support…or 11 percent…or any number figure that ends in an odd number.
(Click here to read it all…)
October 16th, 2008 at 3:23 pm
You’re absolutely right, Frank, and virtually every percentage given in the article itself is an odd number.
October 16th, 2008 at 3:51 pm
Hmmm. The article doesn’t clearly say that these 50 Columbus women comprised the entire set of pollees. I would expect for them to sample from around the nation, and have a larger number of participants? (The article does refer to one coordinator as a “national pollster”.
Did you ask the Dispatch for an explanation? Let us know.
October 17th, 2008 at 11:11 am
Jose, my reading of the article is that the sampling was limited to one 50 woman sample, all together in Columbus. That may not have been true, but it appears to be what the article says.