Wasn't the theory that a large number of Republicans had begun to identify as independents without changing their voting habits widely accepted long before election day?
Well, it was accepted by those who took an objective look at the numbers. However, most of the right wing never took such an objective look. In survey after survey, the general shift in party self-identification showed a drop in R, a gain in I, and not much change for D (leading to the D+6 or so advantage). It didn't take a rocket scientist to conclude from this shift that a bunch of former R's were now identifying as I's. From that conclusion, it's an obvious corollary that the new batch of I's would likely be very pro-Romney. This would then skew the overall preference among *all* I's toward Romney -- which is exactly what the data showed.
The fundamental math error by the right wing is that they rejected the D+6 advantage in ID, but accepted the pro-Romney support among I's. However, these two trends were inextricably linked; both were created by the R->I shift. By rejecting the first trend and embracing the second, the right wing was, in effect, counting the pro-Romney R->I's *twice* -- once as Indepedendents, once as part of the "unskewed" Republican turnout. With those numbers, it's no wonder that they believed they could win. But that's what happens when you choose to accept only the facts that are beneficial to your side and reject those that aren't.
