One of the less publicized aspects of the work pollsters do is an evaluation of their work “after the fact”; in other words, how accurate was their polling?

Several weeks ago, JMC Analytics and Polling polled a party runoff in eastern Oklahoma. This is an area that was once the Democratic bastion of the state, regardless of whether you were talking about Presidential or local races. However, the unpopularity of the Obama administration in the Deep South/border states, combined with a Congressional vacancy, has led to an open seat that the Republicans have a chance to pick up in November.

Both the Democratic and Republican parties held runoffs in this race, and in an August 2 poll that we conducted, plumbing contractor Markwayne Mullin had a 57-26% lead over state representative/cleaning contractor George Faught (Mullin led 42-23% in the primary). Mullin ended up winning the runoff 57-43%, and now faces Democrat Rob Wallace in the November election.

What we noticed in the analysis was that, relative to the poll we conducted, Rep. Faught performed significantly better in the central (which is his home) and southeastern parts of the district. Other than that, the Mullin numbers (and our weighting by region) were within the margin of error.