Larry, Curly and MOE

The polls are flying fast now. Bloggers are rushing to analyze what they mean. But as you analyze, beware the margin of error (MOE). The margin of error is the “oops” factor of research and reflects the depth, or lack of it, in the quality of methodology. Some MOE’s are acceptable and some are not. In educational research the MOE is usually .05 percent or less. That is if the researcher wants to be taken seriously. In medical research the MOE is must be much lower, perhaps .001%. So lets look at the current polling. Today the Pew and Gallop Polls reported completely different results for the national race. How can this be explained? Well, the MOE for both polls was plus or minus 4%. That is a potential eight point swing. What good is that? There is no information there. Can you imagine using the same MOE for medical research? Break out the body bags!

Of course, smaller MOE's would require more money, a bigger sample and more time. In other words a smaller margin of error would require higher standards of research. So give me a poll with a MOE of perhaps plus or minus 2%. Then I’ll pay attention. Until then it is not news, it’s just “for entertainment purposes only.”