Who’s Winning Dallas County So Far?
No one knows for sure, of course, but there is reason to believe that the overall countywide turnout may be running neck-and-neck.
As of close of the polls Friday, the total number of ballots cast stood at just over 170,000 votes.
We can’t, of course, know exactly how many of those votes were cast by Republican-leaning voters and how many were cast by Democrat-leaning voters, but we do have some ways of making some educated guesses.
For starters, we can determine how many of the early voters have a prior voting history indicating affiliation with one party or the other. According to our analysis, 116,413 of the voters who cast ballots in the first five days of early voting fall into this category, as follows:
Known Republican: 48,100 (41.3%)
Known Democrat: 68,313 (58.7%)
If we assume that the partisans vote for the members of their own party, that certainly doesn’t look too encouraging for the Republicans. The good news for the Republicans is that the identified partisans comprise only a portion of the early voters. The remaining voters have no easily-identified partisan alignment. The question becomes, then: can we reasonably predict the partisan leanings of the remaining early voters? The answer: maybe.
Using three different methods, there is reason to believe that the Dallas County Democrats may hold a slight edge over the Republicans, based on the turnout patterns in the first five days of early voting. A summary of each is included below.
Method 1: Precinct-by-Precinct
One way of making a rough estimate of the partisan alignment of the remaining voters is to compare the overall turnout of the precincts in which known Republicans are outvoting known Democrats to the overall turnout of the precincts in which the known Democrats are outvoting known Republicans. Surely there are Republican voters in Democrat precincts and Democrat voters in Republican precincts, but it’s not unreasonable to think that these voters may, on balance, tend to cancel one another out.
According to this analysis, approximately 81,000 voters have cast ballots in the precincts in which known Republicans are outvoting known Democrats, while approximately 90,000 voters have cast ballots in the precincts in which known Democrats dominate. By this admittedly-rough measure, the Dallas County Democrats come out with an edge, 52%-to-48%.
Method 2: Weighted Total
The strength of Method 1 is in its simplicity. The weakness of Method 1 is its crudeness. Surely there’s a better way to estimate the partisan alignment of the unknowns.
We could make a number of different assumptions regarding the political leanings of the unknowns. We could assume that the unknown voters are evenly-divided, or that the partisan leanings of the unknown voters are roughly proportional to the partisan leanings of the known partisans, but neither assumption would align with known voting patterns in Dallas County. In the November 2008 election, the Democrats won Dallas County decisively overall at the top of the ticket (with ~422 thousand votes for Obama to ~311 thousand for McCain,) but this was due to a massive turnout by known Democrat partisans. Among voters without a known partisan alignment, the Republicans actually came out ahead by a couple thousand votes.
If we weight the early voting figures by historical election patterns and renormalize the numbers, there is reason to think, based on these prior patterns, that perhaps a little over 60% (~35,000) of the voters without a partisan affiliation may be Republican-leaning voters, while just under 40% (~20,000) of the unknowns may be Democrat-leaning. Applying these percentages to the unknowns gives rise to a prediction of ~83,000 (~48%) votes for the Republicans and ~89,000 (~52%) for the Democrats. Interestingly, this estimate aligns with the 52-48 number we calculated above using our much-simpler “total turnout” method.
Method 3: Weighted Precinct-by-Precinct
With the benefit of years of prior years of voting history, we can make a reasonable guesses as to the future behavior of a precinct. Unfortunately, we don’t have any voting history for many of the newly-drawn precincts, and can only guess as to how they are likely to vote. We can, however, make predictions as to the precincts which were not substantially changed in redistricting and some educated guesses as to those which were.
Using a precinct-by-precinct analysis, the total number of Republican-leaning independents can be estimated at approximately 35,000 and the number of Democrat-leaning independents can be estimated at approximately 19,000.
Adding these numbers to the known partisan totals, the totals come out to ~49% for the Republicans and ~51% for the Democrats.
After running the numbers three different ways, the various results came out surprisingly close to one another. Whether these estimates ultimately align with reality remains to be seen. Until the votes are counted, there will be no way to know for sure what will happen in Dallas County.
The above analyses are intended as educated guesses only based on historical voting patterns and an early sample comprising perhaps 25% of the overall votes that will be cast this election. Even if these estimates are close to the actual votes as they stood halfway through early voting, approximately three out of four votes have yet to be cast.
With all those caveats in place, the above numbers give us reason to believe that neither party is totally dominating the polls as of yet, and Dallas County is, at this point in time, still “anyone’s ball game.”