2016 Postmortem
In reply to the discussion: Fun with California ballot numbers [View all]onenote
(42,581 posts)Last edited Mon Jun 13, 2016, 02:33 PM - Edit history (1)
It doesn't seem to impact your estimates by very much, and maybe I'm missing something, but it looks like you're assuming that 68 plus percent of the unprocessed ballots will be Democratic presidential primary ballots (based on your calculation that of 5.6 million presidential primary ballots counted 68.7 were Democratic presidential primary ballots). But just as only 5.6 million of the approx 6.5 million ballots processed were presidential primary ballots, so too must an assumption be made that a certain percentage of the unprocessed ballots are not presidential primary ballots. It gets a bit tricky because there are three categories of unprocessed ballots: uncounted mail in, provisional and "other". Assuming that the same proportion of each category are presidential primary or non-presidential primary ballots is probably inaccurate (just as it probably is inaccurate to assume that the same proportion of provisional Democratic primary ballots were cast for Sanders as the proportion of mail in ballots).
I think you've actually accounted for the fact that the unprocessed ballots include non presidential primary ballots indirectly through your "rejected" ballots adjustment. But non-presidential primary ballots (which includes ballots cast by NPP voters who did not elect to participate in any of the presidential primaries and presidential primary ballots in which the voter left the choice for a presidential candidate blank and voted only in the down ticket races) are not "rejected" ballots and given the claims that some have been making about 1 million "discarded" ballots, I think its pretty important that these matters be described clearly and accurately.
Thus, what we know is that out of the total number of processed ballots, 58.95 percent of the 6.4 million unprocessed ballots were cast for a Democratic presidential primary candidate. If that same percentage is applied to the 2.6 million unprocessed ballots, it would mean that there are around 1.53 million unprocessed votes left to be counted that will be Democratic presidential primary votes. This assumes, as noted above, that all three categories of unprocessed ballots contain the same percentage of presidential and non-presidential ballots, which may not be the case. It also assumes that none of the provisional or other votes will be rejected. If we apply a pretty conservative rate of 10 percent to the provisionals (719,000) and other (72,000) and assume 100 percent of the mail-ins will count, we can subtract 79,100 ballots from the available pool it leaves around 1.3 million, or just half, of the unprocessed ballots available for Sanders to make up a 475,000 deficit.
That means Sanders would need roughly 68% of the remaining ballots to catch Clinton.
I think these calculations, while not precise and with a number of admittedly uncertain assumptions, present a pretty accurate picture of the hill Sanders would have to climb. But if there is some glaring error here, please let me know.
Edit to fix error where I failed to distinguish between all presidential primary ballots and Democratic presidential primary ballots and subsequent errors resulting from that error.