In his pre-election analyses of the early vote, Republican consultant Derek Ryan (whose numbers I have used in the past) suggested that there was still a significant number of regular Republican voters who had not voted yet, which could make Election Day redder than early voting was. I thought I’d take a look at the data to see how accurate that was. Short answer: Pretty accurate.
Candidate Early% E-Day% Total% Ratio
=========================================
Abbott 53.75% 57.05% 54.80% 1.06
Beto 45.14% 40.98% 43.81% 0.91
Others 1.11% 1.97% 1.39% 1.77
Note that “Early” here includes mail ballots, as the Secretary of State website combines mail ballots with early in person ballots to give that number. “Ratio” is just the Election Day percentage divided by the Early percentage, which you can interpret to mean that Abbott did about six percent better on Election Day while Beto did about nine percent worse. The Others include the Libertarian and Green candidates plus two write-ins. I am greatly amused by the fact that their voters are the real traditionalists for voting on Tuesday.
If you’ve followed the numbers from Harris County, you know that Democrats overall did at least as well on Election Day as they had done in early voting. I assumed there was a range of outcomes here, so I sorted the data by Abbott’s Ratio, to see where he did best and worst – relatively speaking – on Election Day. Here are a few counties of interest for each. First, where he improved on Election Day:
County Abbott Beto Others
=============================================
Travis Early 24.07% 74.83% 1.10%
Travis E-Day 30.52% 66.96% 2.52%
Bastrop Early 53.93% 44.58% 1.50%
Bastrop E-Day 64.15% 33.53% 2.32%
Williamson Early 47.73% 50.94% 1.33%
Williamson E-Day 54.19% 43.20% 2.62%
Hays Early 42.52% 56.01% 1.46%
Hays E-Day 46.87% 50.30% 2.84%
Bowie Early 73.12% 25.96% 0.92%
Bowie E-Day 80.32% 18.17% 1.52%
Dallas Early 34.85% 64.18% 0.97%
Dallas E-Day 38.08% 60.02% 1.90%
There are numerous small counties in there that I haven’t listed, I’m just highlighting the ones of interest. Travis County was in fact the top Ratio value for Greg Abbott – he did 29% better on Election Day than he did in early voting. This is where I point out that “doing better (or worse) on Election Day” is not the same as doing well (or poorly). That said, Abbott did well enough on Election Day in Williamson County to nudge past Beto’s vote total for that county. Now here are a few where Abbott dropped off on Election Day:
County Abbott Beto Others
=============================================
Fort Bend Early 47.58% 51.07% 1.35%
Fort Bend E-Day 44.72% 52.94% 2.33%
Lubbock Early 70.30% 28.64% 1.06%
Lubbock E-Day 67.54% 30.49% 1.97%
Harris Early 45.06% 53.79% 1.15%
Harris E-Day 43.31% 54.45% 2.24%
Gregg Early 73.76% 25.52% 0.72%
Gregg E-Day 71.09% 27.35% 1.56%
Jefferson Early 56.56% 42.33% 1.10%
Jefferson E-Day 54.61% 43.38% 2.01%
It’s interesting to me to see Central Texas counties filling up that first table, while the Houston area is more present in the second one. I could have included Waller, Wharton, and Chambers in the latter as well. Whether that’s a fluke or a tendency, I couldn’t tell you. But it’s finding weird things like this that makes doing this kind of exercise so much fun.
Does any of this matter on a more macro level? Again, I don’t know. Honestly, I don’t think it matters that much, in the sense that the votes all count the same and from the perspective of a campaign’s win number it doesn’t make a difference. It’s certainly nice to have a bunch of votes banked before Election Day – if nothing else, it mitigates some risk from bad weather and technical difficulties at voting locations. But ultimately, either your voters show up in the numbers you need or they don’t. I think this data is interesting, and it may suggest some strategies for how better to deploy campaign resources. Beyond that, it’s what you make of it.