More words on exit polls
Urban legend has it that Inuits have many more words for "snow" than we do. Apparently that is not true, but apparently we could all use a few more words for "exit polls" that more accurately describe their various stages. The Mystery Pollster, who's no longer a mystery, has a new post that more thoroughly describes the process on election day and sheds a little light on the nature of the changing numbers over the course of the day. They are not just bigger numbers, taking in more of the day; later in the day, they are a different animal all together.

This still doesn't answer the important question of discrepancy, and he makes clear that so long as NEP refuses to release their weighting models and numbers, we won't get any closer to that answer, even with the raw numbers coming out in a couple of months. But it does helpfully problematize conversations that try to refer to "exit polls" as if they were a single thing just because we only have a single phrase for all of them. (link via MyDD)