Off the Kuff Rotating Header Image

Early vote totals, Days Four and Five

Here’s your EV totals for Thursday and Friday, plus the 2009 spreadsheet, the 2007 daily EV report, and the Erik Vidor spreadsheet. And here’s the cumulative summary for the first five days:

Type 2005 2007 2009 2011 ===================================== In Person 15,996 10,767 19,366 14,349 Absentee 1,891 3,555 3,801 4,340 Total 17,887 14,322 22,967 15,689 09 Pct 77.9% 62.4% 100% 68.3%

The relative shares of the 2009 vote is up for the 2005 and 2007 elections, and down a bit for the 2011 election. This is because Friday was relatively busier in the other years than this year.

So with five days’ worth of data, it’s time to start thinking about projections. Here’s how the five day EV totals compared to the final amounts in each of the last three elections:

Year 5 Day Ev Total EV 5 Day Pct ====================================== 2005 17,887 78,585 22.8% 2007 14,322 50,264 28.5% 2009 22,967 80,516 28.5% 2011 15,689 55,001 28.5%

As noted before, the final day EV totals in 2005 were much higher than on other days, thanks in large part to the anti-gay marriage referendum. In a more normal year, it looks like 28.5% of the early vote is cast in the first five days. That gives us a projection of 55,001 early votes for this year.

We then compare the EV totals to the final totals, bearing in mind that this is for all of Harris County. This is how it looked before:

Year Early Total EV Pct ====================================== 2005 81,007 332,154 24.4% 2007 52,476 193,945 27.1% 2009 82,978 257,312 32.1% 2011 57,000 233,606 24.4% 2011 57,000 210,332 27.1% 2011 57,000 177,570 32.1%

The Early totals are higher here because absentee ballots continue to arrive between Friday and Tuesday. In each of the previous three elections, about 2000 or 2500 late absentee ballots came in. I added 2000 to the projected total for this year, then extended it out to a final Harris County tally based on the EV percentage of each previous election. It can be a pretty wide spread, depending on what you believe is the early vote share. Note that these numbers are generally lower than for even-numbered years, which I suspect is related to the older demographic that votes in municipal elections.

Finally, we discount these totals to reflect Houston votes only. In previous elections, the Houston share of the Harris County vote has been between 63 and 69 percent. To simplify things, let’s assume that the low end is 60% and the high end is 70%. Applying those numbers to the high and low vote totals gives a range of 106,542 to 163,524 Harris County votes for the Houston elections. Throw in another 2,500 votes or so from Fort Bend and Montgomery for the final totals, so in round numbers 109,000 to 166,000 total votes. Again, a pretty wide spread, but my guess is the lower end is where the actual mark will be. Let’s adjust my over/under number to 130,000 at this point, which is to say a slight improvement over 2007. Nothing to write home about, but at least it wouldn’t go into the record books as a new low.

Anyway. Here’s the Chron story, which makes the fundamental mis-assumption of comparing the totals to 2009 only. I’ll say again, the model for this election is 2007, and by all indications we’re right on target for that. Have you voted yet?

Related Posts:


  1. Ursula says:

    Are your 2011 values right? Doesn’t 14 k+4k=18k?

  2. […] may notice is that the share of the 2009 vote total for this year is up considerably from what I reported for the five day totals. There’s a simple reason for this: I screwed up the arithmetic on […]