Sean Trende over at Real Clear Politics has an analysis of the state of the Ohio Governor’s race that puts my poll analyses to shame. Check out his analysis about why Kasich is favored:
This result is actually consistent with other polls taken in recent days that show a much closer race. Strickland’s numbers in the September polling are remarkably stable: 42, 42, 43, 46, 45, 41, 43, 44, 37, 40, and 43. Kasich’s numbers, on the other hand are all over the place: 50, 43, 45, 47, 49, 47, 50, 51, 54, 52 and 48.
To put it in statistical terms, the standard deviation of Kasich’s numbers is quite a bit higher than that of Strickland’s (especially if you exclude Quinnipiac’s 54-37 result as an outlier).
…we actually see quite a bit of stability in the apples-to-apples poll comparison — in other words, looking at what pollsters find compared to their earlier iterations. This is important, because different pollsters use different likely voter screens and push undecided voters in different ways. If individual pollsters aren’t finding movement, the changes in the RCP Average are likely attributable to changes in the mixture of pollsters in the Average, rather than actual movement.
Here, we see that Rasmussen moved to 50/42 Kasich from 50/43 Kasich. FoxNews/Rasmussen moved to 45/43 Kasich from 47/41 Kasich, with an earlier entry of 48/43 Kasich. The University of Cincinnati poll showing Kasich up four is actually an improvement from their earlier poll showing Strickland up 5 back in May. The only pollster showing real movement was Reuters/Ipsos, which moved from 48/39 Kasich in August to 47/46 Kasich in September.
But Sean’s final point is something you’ve been seeing here on 3BP for quite some time.
…most importantly, the numbers you need to pay the most attention to here are Strickland’s. Incumbents under 50 percent at this point in the game usually do not win; incumbents under 45 percent almost never win. Until Strickland consistently posts numbers in the 47/48 point range, he will be the underdog.
But my readers knew that already.