Making Sense of Presidential race polls: the scoop

Posted on: August 13th, 2012 by Ari Wubbold

Tim Hibbitts, Principal and Founder, August 13, 2012

If you are reading this blog, you probably are an active consumer of polling data.
If you have a particular interest in the Presidential race; here are some things to keep in mind when looking at polling data:

1) First, don’t put too much weight in any given national poll. Right now, there are seven different national polls on the Presidential race that have come out within the last week. The range of the polls is anywhere from a 3 point lead for Romney (Rasmussen) to a 9 point lead for Obama (Fox). When you have a number of credible polling firms producing a variety of results, the best thing to do is average them all, which is likely to give you the best measure of where things actually are. Purists would blanch at that suggestion (techniques are different; samples are different; you shouldn’t average polls that use different methodologies; etc.) but the simple truth is that if you go back to the last three Presidential elections, the averaging method has given an excellent snapshot of both where a race stands and what the eventual outcome is likely to be. Real Clear Politics and Pollster.com are just two of the sources that average national polling.

2) Second, be aware that many national firms are still polling all eligible adults and not ‘likely voters.’ After Labor Day you can expect that all firms will be shifting their polls to likely voters. This is likely to help the Romney campaign a bit; as most years drawing the sample in from all ‘eligible voters’ to ‘most likely voters’ produces a bit more conservative and Republican sample.

Lastly, for all of the back and forth between the two campaigns, not much has happened over the last several months. Obama has led by between 1-4 points since March, so be very wary of any individual poll that suggests a dramatic change; regardless of who benefits from it; unless there are events that could explain the change.

Comments are closed.