The Magnitude of Sanders’ Win
It comports with the polling.
In the comment thread of James Joyner’s post on Sanders’ Nevada caucus win, the issue of the magnitude of the win was raised as an explanation of what was driving the media coverage of the outcome. To wit: most reporting shows Sanders at 46% while he was polling at 32.5% in the RCP average and 30.5% at FiveThirtyEight (as James noted in his post).
At first glance, that 46% number looks like a major out-performance of the polls. However, a second glance is needed because the 46% is not of the vote, it is of the county delegate count. Moreover, that delegate count is the result of final preferences being taken into account, not just first preferences:*
If we look at the first preference vote total (33.2%) we see that the polling was pretty accurate (especially for the notoriously difficult to poll Nevada caucus process). The RCP average was off by 0.7% and 538 was off by 2.7% (note that 40% of the vote is yet to be counted as per the NYT report above, so these are not the final numbers).
So, James’ headline, Nevada Caucuses Go as Expected is, to quote Mona Lisa Vito from My Cousin Vinny, “dead-on balls accurate” (to use, no doubt, the political science term).
Really, what we are seeing in the media regarding Nevada is a version of the same thing I noted in Iowa: a simplistic view of the process. Now, unlike in Iowa, there is no doubt that Sanders is the big winner here (although, in terms of delegates, not the sole winner). But, there is a lack of sophistication in how the process is reported and understood broadly (whether it be delegate counts or the fact that the results are based on final alignments and not just first preferences). While, yes, Bernie is the big winner, his actual first preference position is about a third of support of Nevadans who participated. Even when get to the final preference count, the support is still less than 40%.
I would note that the system continues to be semi-proportional due to the 15% threshold.** I would note, too, that the ranked aspect is incomplete because the process does not actually reallocate lower-ranked candidates enough to get to a 50%+1 winner. (Indeed, the ballots only required ranking the first three and the reallocation was less ranked-choice and more just realignment in the room).
To be clear: I am not trying to downplay the fact that Sanders was successful in Nevada. I agree with James’ other post that he is in the driver’s seat for the nomination at the moment. But I cannot help but continue to be more than a tad frustration at the coverage (recognizing that it is the normal way this stuff is covered, although every four years the drama seems to tick up) as well as with the insane nature of the process itself.
On that last point: it is difficult to look at a process that is so heavily predicated on sequential events in Iowa, New Hampshire, and Nevada and say that it is anywhere near to a rational means to select a nominee.
*For detailed info on the Nevada process, see: Learn how to Caucus in Nevada. See, also, Precinct 1612: How Democrats caucused in one classroom at Durango High School.
**For more on that, see Matthew Shugart’s post: The strategic voters’ nightmare that is US Democrats’ “proportional” system