March 14, 2008
Canvass report - Voting downballot

I'm taking a look through the precinct-by-precinct canvass data for the Democratic primary in Harris County to see what I can learn. The first thing I wanted to look at was voting in the downballot elections. Did the supporters of one Presidential candidate or the other vote more heavily for other candidates as well, or did they tend to be one-race voters? And if they voted all the way down the ticket, for whom did they vote? Let's take a look.

I'll note up front that I deleted the sixty or so precincts that had fewer than ten votes in them. I also threw out the tiny handful of votes for Presidential candidates other than Obama and Clinton. Finally, while there were some undervotes in the Presidential race as well, I simply used the sum of Obama and Clinton's vote totals as my "Presidential turnout" base figure. For each downballot contest, I then took the sum of the votes in that race and divided it by the Presidential total to get a percentage of that. In this study, I'm looking for the precincts that had the highest such percentages for each race.

I considered the four contested statewide primaries - Senate, Railroad Commissioner, and Supreme Court - as well as the three contested countywide primaries for non-judicial offices - County Judge, Sheriff, and Tax Assessor. We'll start with the Senate race. For this one case, I took the additional step of filtering out the precincts in Rick Noriega's State Rep District 145, on the grounds that I figured they'd vote heavily in the Senate race regardless. Which they did - every single precinct in HD145 had at least an 80% participation rate in the Senate contest. That left a total of 67 other precincts for which the total vote in the Senate race was at least 85% that of the Presidential. Here's how the vote broke down:

Candidate Votes Pct
Obama 8284 39.8
Clinton 12545 60.2
Noriega 12098 67.0
Others 5946 33.0

As you can see, the precincts that voted the most in the Senate race favored Hillary Clinton, and gave a solid majority to Rick Noriega.

Next up is Railroad Commissioner, where 53 precincts had at least 75% of the Presidential votes:

Candidate Votes Pct
Obama 15462 64.0
Clinton 8681 36.0
Noriega 10102 49.5
Others 10287 50.5

Take that for what it's worth. I think very few people knew who they were voting for in the RRC race, and some of those folks weren't too informed about the Senate race, either. I didn't bother breaking down the votes for each RRC candidate, as I didn't think there was anything useful to learn there.

Moving on to the Supreme Court. There were 86 precincts with 75% of the Presidential vote for Place 7, and 104 such precincts for Place 8. Here they are, listing the Place 7 totals first:

Candidate Votes Pct
Obama 11949 42.3
Clinton 16329 57.7
Noriega 15621 64.3
Others 8686 35.7
Houston 9943 45.8
Cruz 11764 54.2
Yanez 14690 67.5
Criss 7060 32.5

Candidate Votes Pct
Obama 18010 45.4
Clinton 21703 54.6
Noriega 21329 62.8
Others 12610 37.2
Yanez 20624 67.7
Criss 9832 32.3
Houston 14057 46.7
Cruz 16070 53.3

I think the Sam Houston/Baltasar Cruz results interest me the most here. These appear to be mostly Hispanic and Anglo boxes. How much did Sam Houston's name help him? I'd say a fair bit. Perhaps Dale Wainwright ought to be worried.

On to the countywides. Generally speaking, the number of people who voted in the countywide races was less than that of the statewides. For the County Judge race, there were 93 precincts in which the total vote for that race was at least 65% of the Presidential race.

Candidate Votes Pct
Obama 18563 52.0
Clinton 17243 48.0
Noriega 18490 61.6
Others 11548 38.4
Mincberg 18643 77.8
Hassan 5315 22.2

A slight preference for Obama here, along with a hearteningly solid majority for Noriega. Mincberg ran a bit better in these precincts than he did overall, which is no small thing given the size of his win.

Next up is Sheriff, where the vote total was the highest in any countywide race. There were 88 precincts that voted at 75% or higher of the Presidential level.

Candidate Votes Pct
Obama 8739 29.1
Clinton 21260 70.9
Noriega 19278 74.6
Others 6555 25.4
Garcia 19545 82.1
Others 4251 17.9

Now that's what I call a strong preference! Adrian Garcia probably had the highest name ID among the countywide candidates, and that definitely helped him in his three-way primary. Many of these precincts are in State Rep districts that overlap or are proximate to Garcia's City Council District H - HDs 140, 143, 145, and 148 in particular.

The last countywide race to look at is Tax Assessor. I wanted to set the cutoff for percentage of the Presidential vote at 70%, but that produced fewer than 40 precincts, so I went to 65%, and wound up with a total of 150 precincts. Better too many than too few, I guess.

Candidate Votes Pct
Obama 46198 62.5
Clinton 27751 37.5
Noriega 33170 54.2
Others 28036 45.8
Trautman 32986 66.6
Webb 16561 33.4

These precincts were all over the county. I was a little worried going into the primary that Diane Trautman's name wouldn't be as well known as it could be, and that this could cause her problems, but I needn't have fretted. She did just fine.

Can we draw any conclusions from this? Mostly, I don't think you can say that downballot participation was the province of one Presidential candidate's voters or the other, and I think that's reassuring. I think it suggests that whoever tops the ticket in November, we'll do fine on turnout either way. What do you think?

Finally, though it's not directly related to what I'm looking at here, I want to flag this Matt Stiles post about turnout in the Presidential primary. It's good stuff.

Posted by Charles Kuffner on March 14, 2008 to Election 2008

Where did you get the pct data? I'd like to take a look too. Thanks!

Posted by: Presentense on March 14, 2008 9:52 AM