The Predictor
I learned today about a method of analyzing House races that may be able to predict winners (no warranties, expressed or implied) in close races for the Texas House of Representatives. The idea is to determine whether Democrats have a chance to win certain races, based on primary turnout of the two parties and Obama’s percentage in the presidential primary.
Take Michael “Tuffy” Hamilton.
Republican primary turnout in the district: 4,714 = 6.2% of registered voters
Democratic primary turnout in the district: 25,465 = 33.4% of registered voters
Looks pretty good for the Democrats, right?
But: only 22.9% of Democrats voted for Obama in the presidential primary
Conclusion: mostly white conservative voters. Hamilton wins.
(Hamilton is an icon in his community. We didn’t need a formula to tell us he is going to win.)
Zedler-Turner
R primary turnout (rounded) 9,600
D primary turnout (rounded) 16,900
Obama %: 61.6
Big trouble for Zedler
Barrett-Shelton
R primary turnout (rounded) 13,000
D primary turnout (rounded) 14,000
Obama % 52.
Too close to call
Goolsby-Kent
R primary turnout (rounded) 4,800
D primary turnout (rounded) 7,300
Obama % 57
Could be trouble
It doesn’t work for all districts. In Dan Branch’s district, R’s cast 5,300 primary votes, D’s 8,400. Obama got 53.5%. But some of those votes were from moderate/disillusioned R’s. Branch isn’t going to lose.
Anderson-Miklos
R primary turnout (rounded) 7,500
D primary turnout (rounded) 8,800
Obama %: 55
Could be trouble
Tagged: bill zedler, dan branch, michael tuffy hamilton, Texas house, turner.
Dr. Hotze's RX Manufacturing says:
Paul
Great Teaser but don’t be selfish with your love now..
What about the following: Jim Murphy; M Homer; Delisi; Bohac; Haggerty; J Davis; R. Cook; just to name a few?
Reply »
Anonymous says:
This is offensively stupid.
The same flawed prejudiced logic could be used to suggest which non-Mormon Republicans in Utah will lose to Democrats based on the margin between McCain and Romney. That religious stereotyping would lead to conclusions that are both as inaccurate and as tasteless as the racial stereotypes inherent in your model.
If you must engage in racial stereotyping, then at least admit several counties where Obama fared most poorly versus Clinton (Hidalgo County, for example), are demonstrably not populated by “mostly white conservative voters” as your bizarre model assumes.
Please kindly acknowledge the stupidity of this 6″11$#!+ and swiftly repudiate the dubious racial stereotyping inherent in it.
Reply »
Anonymous says:
Please… Your telling me that the primary turnout is representative of the number of voters for each party in the district???
1st- you can vote in any primary you wish, no party registration needed.
2nd- what reason did Republicans, or what I really mean is (moderate republicans), have to vote in the Rep. primary. Their race was over, why vote??
So here is the example of why Paula Burka favors the left.
Shelton versus Barret. So in Barret’s primary, Obama get 52%, thats about 7,000 of the 14,000 votes for barret according to this analysis. Shelton runs in a FOUR way race and gets 7,700 of the votes, straigh up. So now you telling me that the 7,000 other Rep. voters are going to magicaly vote for Obama like they didn’t in the primary (but could have) and somehow also vote all Dem and give Barret the advantage??
So by your analysis… To close to call??? What is THAT???
Am saying Barrett or Shelton, could, should, might win… NO. Am I saying the race is decided or too close to call… NO.
What I’m saying is that Burka should actually do the anyalysis as he says he is instead of looking at his data, becoming upset and making up conclusions.
What I do want to say is that I think this is a nonsensical way to try and perdict how a precient will vote.
Reply »
houtopia says:
I can give you the HD 133 (Murphy) numbers.
R primary turnout — 6,900
D primary turnout — 12,900
Obama %: 56
Not sure whether I buy the model, but…
Reply »
paulburka says:
To various critics:
This is not my model. It is a model that is being used to analyze the races by serious people with a stake in the battle for control of the House — a tool to determine Republican vulnerability. Take it or leave it. It’s nonsense to say that this is “racial stereotyping.” It’s just numbers.
Anonymous at 10:02 p.m. criticizes the model by using Hidalgo County as an example: The low percentage of Obama voters indicates that the county is populated by white conservative voters. If Anonymous wishes to be deliberately stupid, that is his privilege. We are allowed to know what we know: that Hidalgo County is populated by Hispanics who preferred Hillary Clinton to Obama. As I said, the purpose of the model is to determine Republican vulnerability. Since there are no Republican incumbents in Hidalgo County, the model isn’t going to be applied there.
Reply »
Emptyk says:
In defense of what may appear to be political astrology, the standard predictive tools for analyzing Texas legislative elections are flawed. The Optimal Republican Voter Strength analysis lacks the necessary data to be of much value because the Democrats have not been competitive with well-funded statewide candidates in the most recent elections.
I think of this analysis of primary voting as demonstrating the enthusiasm gap between the core voters of both parties. If the Sarah Palin ‘enthusiasm bubble’ of early September had not burst, this analysis would be less relevant. With Palin’s shrinking support among independents and weak R’s this analysis regains some relevance.
Reply »
Anonymous says:
Acknowledging that Burka is presenting this with no warranties or endorsement, the level of stupidity here with this model seems to know no boundaries. A couple of random historical examples:
2004 HD 1
Demo Primary Vote: 18,455
GOP Primary Vote: 4,006
Shouldn’t be a race, right?
Final Result -
D – 28,485
R – 25,463
2006 HD 118
Demo Primary Vote: 5,706
GOP Primary Vote: 1,928
Final Result:
D – 10,982
R – 10,082
Before anyone says it, I realize that these past examples have no “enthusiasm gap” to point to; they just show that apples don’t compare to oranges to begin with. They are unique.
But an enthusiasm gap? Everyone and their mother are going to go vote. We’re electing a new President with no incumbent running for only the 4th time in 40 years. Turnout is going to be huge on both sides.
This also fails to take into account that the Texas primary was a political lifetime ago and people know a lot more about Barack Obama in October than they did in March.
The real issue is the nitwits and numbskulls who are either advising people to donate to candidates based on this cockeyed theory or who might be running campaigns based on this.
Burka ought to release the names of those people who advocate and are using this screwball theory so they can be removed from the political process – and they can meet the shortage of workers at McDonald’s, Burger King and FedEx/Kinko’s – they have no business being paid for what they are doing.
The people using htis are not serious – they are people who take themselves seriously – but are seriously stupid when it comes to politics.
Reply »
Burka's Research Assistant says:
I found this post pretty entertaining, but think you might want to make it more obvious that it is unlikely to be useful.
The most interesting thing about your race is that there are people who are paid to do politics who take this approach seriously. You should seriously name names, as above anon indicates.
Reply »
James White says:
I think that HD102 numbers are wrong. I have:
15,331 votes for Prez in HD-102 in the primary. Obama got 8,759,
which is 57%.
Reply »
Anonymous says:
Hey Paul,
Using this method, what does it show in the Garcia-Hunter race?
Reply »
Anonymous says:
I think this method definitely concludes that Tampa will win the World Series.
Otherwise, that’s all you’ll get out of it.
Reply »
Anonymous says:
Burnt Orange Report has an even more ludicrous model that implies that 100% of Democratic Primary voters will vote for the Democratic nominee.
That would sure surprise all the Republicans who were in line with me as we all were going in to vote early for Hillary!
Oh those Burnt Orange kids, maybe they should quit blogging and go get real jobs so they can begin learning – before it’s too late.
Reply »
paulburka says:
I’m not going to release names, period.
Reply »
Phillip Martin says:
To anon at 2:57pm –
Which model says that? There’s a model from the Lone Star Project that uses a base vote of ‘04 Kerry numbers. I think that’s very reasonable. But I don’t see any model on our site that presumes to make a prediction based on 100% Democratic primary turnout. Please feel free to point me to what you’re referring to.
Also — its already been heavily documented that this “tons of Republicans voting against Obama” was marginal at best. Read the Quorum Report archives (if you have access) for the numerical analysis.
Reply »
GZA says:
My favorite thing about this comments chain is when Burka chimes in – dead serious – that He Will Not Release the Names of his Precious Sources. That’s real elite journalism, baby.
Reply »
Anonymous says:
Probaly because he would be embarressed if he told us where he got this from.
Reply »
Burka's Research Assistant says:
I think it is legitimately newsworthy who actually thinks this is a good method. If they are advising clients, they shouldn’t be ashamed of their method. I can understand you told a source you wouldn’t mention his/her name, but why not go check with whoever is using this method and see if they’ll let you release their names.
Reply »
GZA says:
Don’t you see that Burka’s source is….. Burka.
Reply »
Anonymous says:
To GZA and the rest:
Just because this is a blog does not mean that journalistic principles are somehow inapplicable. (1) I always set ground rules for interviews before asking questions, so the source always has the option of whether or not to be on the record. (2) I am not going to reveal confidential sources, period. (3) I will state that the methodology is not mine.
Reply »
GZA says:
Sorry, I was just teasing.
Reply »