Publicly many were putting on a brave face—we need to count all of the votes. Privately, however, most conceded what we believed—Measure G, the parcel tax that would add $3 million for teacher compensation, had lost.
It’s a cliche that we have to count every vote. I certainly accept that as a legal and moral imperative for our democracy. But as a prognosticator, for the most part, we know who is going to win long before the final vote is counted. In part, that is because most elections are not close.
Take Measure Q for example. It was at 80 percent after election day and needed just a majority. That one was safe to call.
We also knew that the Angel Barajas and Duane Chamberlain race, which was separated by 90 votes on election day, was too close to call. Small swings in votes would have—and in fact did—change the outcome.
But Measure G was a gray area. The first results had it at 63.9 percent of the vote—good but short of the two-thirds, and by the end of Election Day it was at 65.1 percent.
I like to run the numbers because it shows us what happens when trends continue and how much they have to change to affect the outcome. Over the years, it has been extremely reliable.
But that may be changing. One person I spoke to on Saturday noted that in a lot of the congressional race, the late votes, often people registering and voting the same day, leaned heavily to the left and they swung many of those close races into the blue column.
Indeed, it appears that many of the precincts on and near the campus are the ones that pushed this over the top. Whereas the more “conservative” and more “peripheral” precincts were still in a majority, but failed to reach the two-thirds majority.
Our initial take on March 4 was we found it would take about a 68.8 percent yes vote percentage in those additional 10,000 votes we estimated were remaining, in order to push the ballot measure to victory.
The verdict: “Our view is that is not exactly impossible—other tax measures have cleared that number. But we view it as unlikely. That would be two full percentage points higher than the election day vote percentage. While not impossible, it seems like a lot of ground to make up.”
The running assumption is that in general the voting make up of those who turn in their absentee ballots on election day mirrors the electorate. And since this was a district-wide tally, there was no reason to believe that the votes would vary markedly from what was counted.
That might need to be revised in elections that are likely to attract a high number of same day registrations.
One astonishing factor is that more people voted late than voted early. The number of ballots counted after election day has now exceeded the number that were turned in beforehand plus the number that voted in the polls on election day. We have not seen that before.
A big factor here might have been the nature of the Democratic Primary, which had whittled down from a huge number to a handful of active candidates by the time we reached California.
Still, what happened was pretty remarkable.
We projected it would take at least 68.8 percent. Given that only 63.9 percent supported it in the early ballots, and 66.9 percent on election day, that seemed a high hurdle—but again that was assuming similar voting populations, and I think if we can ever dig deep enough here, we will find that they in fact weren’t similar voting populations.
So what happened was that instead of the 68.8 percent needed, it received 69.4 percent of those late votes.
In addition, we had projected about 10,000 new votes, and that number is currently 13,563, and likely to end up at 14,500 and perhaps as high as 15,000.
As it turns out, that number is not that important; what was important was the 69.4 percent yes vote.
- Pre-election day votes: 63.9
- Election Day votes: 66.9
- Up to March 13 votes: 69.2
- Last 3300 votes March 27: 74.5
The 7282 votes tallied over the last three weeks come to 71.5 percent.
These numbers appear to be too strong to be mere random occurrence. It appears that the voting population got steadily more progressive as the election cycle wore on. The voters who voted by mail early were the most conservative and those who voted late were the most liberal.
We learn something new all of the time. Next time, when I see something within two percent of the target, I will be less likely to make the call.
—David M. Greenwald reporting