Whatever Selzer's final poll shows is what my prediction will be. In 2010, their last poll for Senate was spot on and Governor was pretty close. In 2012, they had Obama up 5 while others had it much closer.
Even Selzer missed in 2010, even tough they were closer. Everybody focuses on Colorado, Nevada as D support missed, Iowa is worse. Every election since 2008 pollsters have missed badly to one side.
Final Results -- -- 53.0 43.3 Branstad +9.7
RCP Average 5/25 - 10/29 -- 52.3 37.3 Branstad +15.0
Des Moines Register 10/26 - 10/29 805 LV 50 38 Branstad +12
Rasmussen Reports 9/23 - 9/23 500 LV 55 37 Branstad +18
PPP (D) 5/25 - 5/27 1277 LV 52 37 Branstad +15
To be fair, pollsters are more accurate in distinguishing a +1 lead from a +5 lead than they are at distinguishing a +9 lead from a +13 lead. I don't know why, but that's pretty well-established.
I agree, however, in Iowa, they miss either way. Missed in 2008, 2010, 2012 by sizable margins. Dem support in polling averages is underestimated in Iowa, their is a clear pattern.
Yep, I agree. At this point, I still believe Braley will win by a narrow margin. But, past performance is no guarantee of future results.