If you haven't seen it yet, the ever reliable Rasmussen has Ernst leading by a paltry 1 point in their (most likely) final Iowa poll, down from a 3 point lead in their last one. Why is that so important. Well, I looked at the final polls that Rasmussen has conducted for Senate, President, and Governor for the 2006, 2008, 2010, and 2012 cycles, and found that they have shown a tie race or the republican leading by 1-3 points in a total of 18 different races. How many of those races did the republican actually end up winning? The answer is one. One time. That was when they had Missouri tied in 2008, a race that McCain only barely won. In other words, whenever their final poll has a lead for the republican of 3 points or less, the republican has won only 5.5% of the time. Over 4 complete cycles of polling. This doesn't mean Ernst only has a 5.5% chance of winning, who knows they could be much better this cycle, but a win for Ernst would be rarefied air for Rasmussen polling. So lets look at some of their other predictions.
Over the course of the last 4 cycles, they have also found the republican with a 4-6 point lead in their final poll a total of 14 times, and 4 times the democrat actually won that race. So even in 4-6 point races where you should almost never be wrong, they had an almost 30% error. In fact, they've missed at least one of these races in the last 3 cycles when they had Norm Coleman winning in Minnesota in 2008 by 4 points, Ken Buck winning in Colorado in 2010 by 4 points, Bill Brady winning in Illinois in 2010 by 6 points, and Rick Berg winning in North Dakota in 2012 by 5 points. So far this cycle, in this category of 4-6 point leads for the republican, they have Gardner leading by 6, McConnell leading by 5, Sullivan leading by 5, and Cassidy leading by 4. If I had to bet, one of those will probably be wrong, and I would put my money on Alaska for some reason, it just seems like a weird state. Also note that Ras has Georgia tied, although that might not be their last poll there. Anyway, if the democrats win all the seats they where they currently lead in the Rasmussen polls, namely new Hampshire, North Carolina, and Kansas (note Rasmussen as NEVER shown a democrat ahead that has ended up losing), and win the 0-3 point races, Georgia and Iowa, they will retain 50 seats. Throw in Alaska for 51. But who knows, maybe Ras will prove us all wrong and be the best pollster this cycle, and also I have a bridge to sell you...