PLEASE NOTE: These are the results as of the close of polls on Sunday, October 26. The countywide numbers may be substantially different at this point.
Amid a generally-dismal political environment nationally, the early voting numbers may be bringing some comfort for the Dallas County Democrats.
As of close of the polls Sunday, just under 86,000 votes had been cast in early voting. Of these, it appears that Democrats have a slight edge over Republicans in turnout among voters having a primary voting history for one party or the other:
If it can be assumed that the voters with a Democrat primary voting history have voted predominantly for Democratic candidates and that the voters with a Republican primary voting history have voted predominantly for Republican candidates, this is likely welcome news for Dallas County Democrats. They’re up by 864 votes (1.3%) in that matchup.
In addition to the 65,374 voters with a primary voting history, another 20,497 voters who have no primary history have cast a ballot . Any of these voters could be aligned with either party, or with none. They are question marks.
Using three different methods of analysis based on prior voting patterns, we’ve generated three estimates of the partisan makeup of these unidentified voters.
The upper bar above is the estimate based on an assumption that the unknown voters in each precinct can be allocated according to the proportions of the vote to Romney and Obama in the same precinct in 2012, recalibrated by observed differences between Presidential and Mid-year voting patterns in Dallas County. (This set of assumptions risks overstating Republican strength somewhat, owing to the fact that it uses a countywide recalibration factor at a precinct level.)
The middle bar above is the estimate based on an assumption that the unknown voters in each precinct have been voting proportionally to the known partisans. Thus, if the ratio between the known Republicans voting and known Democrats voting so far in a precinct is 50:50, the unknown voters in that precinct are allocated in the same proportion. (Given that there are significantly more known Democrat partisans than known Republican partisans in Dallas County, this assumption is likely to at least slightly overestimate Democrat strength.)
The lower bar above is the estimate based on an assumption that the unknown voters in a precinct can be allocated according to the proportions of the vote to Romney and Obama in that precinct in 2012, uncalibrated. (Given the strong historical differences in voting patterns between Presidential and Mid-year elections, this assumption is likely to overestimate Democrat strength significantly.)
According to the first method, the Republicans have a 1,025-vote edge among the unknowns. The second shows the two parties virtually tied, separated by only 69 votes. According to the third method, the Democrats have a 2,049-vote edge among the unknown voters.
CAVEAT: there are a thousand sources of error in any of these estimates. These estimates should be used only to gauge generally how things may generally look halfway through early voting. Most of the votes have yet to be cast.
The only useful take-away from these early estimates? Halfway through early voting, it appears that Dallas County is ANYBODY’S BALL GAME. Neither team appears to be very far ahead. It may come down to one critical question: Who wants it more?