I’ve been slowly making my way through the Harris County precinct data, with a goal of analyzing all of the interesting races that took place here this year. I’ve got a couple of reports prepped and ready to go, but before I start with those, I want to cover a general point: When the Secretary of State gives the respective Republican and Democratic averages for each Congressional and State Rep/State Senate district, take them with a certain grain of salt. The simple fact of the matter is that the statewide candidates’ average will understate the actual Democratic performance in many, if not all, districts.
How do I know this? By comparing the statewide candidates to countywide candidates in various places. Take Harris County, for example. By my calculation, the average Democratic statewide candidate (not counting Chris Bell; the Governor’s race is a separate post) got 45.12% of the two-party vote. Of 19 contested races, only one Democratic countywide candidate failed to exceed that mark. The average Democratic countywide candidate got 47.77% of the vote. Only Bill Moody, who came close to carrying Harris County, topped that. (Bell did as well for the straight up comparison against Perry.)
You see similar things in Fort Bend (average Dem statewide candidate, 42.73%; average Dem countywide candidate, 45.65%), Galveston (only Bill Moody carried the county, with 48.19% total, but Dems won 9 of 11 contested county races), and Dallas (average Dem statewide candidate 50.02%, with four of ten Republicans carrying the county, while countywide Dems won each of 42 contested races they were in). Even in a Democratic stronghold like Travis County, the pattern holds: The low score among the countywide candidates was Mina Brees with 59.12%; only Moody and JR Molina topped that among the statewides.
After the 2004 election, I argued that using the Bush/Kerry numbers to evaluate districts’ partisan indices was misleading, since Bush was easily the top votegetter in most counties, often outpacing the rest of the Republican ticket by three to five points. I used an average of the other three statewide races as my main factor. This year, I’m going to consider the full statewide averages as a data point, but I think a range of data is going to be needed rather than a single point. I’ll strive to use countywide candidates where I can, and I’ll likely use the Moody and Molina numbers otherwise as the closest approximation of the default Democratic vote. It’s my expectation as I go through the numbers that I’ll see a lot of examples of candidates for Congress and State Rep beating the statewides. We’ll see how that prediction turns out. Stay tuned for more.
I agree. The lack of advertising and other resources Democratic statewide candidates can muster probably misrepresents the real potential for local candidates.
At the end of the series, do you plan to make any comments about the implications of the difference between state-wide and local performance? I hope so. Have been silently dwelling on this question since Election Day. Thanks.
Great idea on the precinct data analysis. One suggestion is that you should post your reports on the the races in Adobe Portable Document Format (PDF). I have access to the full version of the program. I am more than happy to volunteer to help.