A federal judge on Wednesday threw out Democrats’ effort to reinstate the straight-ticket voting option in Texas.
Siding with the state, U.S. District Judge Marina Garcia Marmolejo found that Democrats lacked standing to challenge Texas Republicans’ decision to kill straight-ticket voting ahead of the November general election. The judge dismissed the federal lawsuit after ruling that Democrats’ claims of the electoral fallout that could come from eliminating straight-ticket voting were too speculative.
The Texas Democratic Party — joined by the chair of the Webb County Democratic Party and the Democratic campaign arms of the U.S. Senate and House — filed the lawsuit in March on the heels of Super Tuesday voting that left some Texans waiting for hours to cast their ballots.
They claimed the elimination of straight-ticket voting is unconstitutional and intentionally discriminatory because the longer lines and waiting times it is expected to cause would be disproportionately felt at polling places that serve Hispanic and Black voters.
[…]
In her order, Garcia Marmolejo ruled that that Democrats’ predictions about the negative effects the lack of straight-ticket voting would have on voters and the election process were “uncertain to occur.” She also found fault with their assumptions that the Texas secretary of state and local officials would not work to “ameliorate the situation.”
Garcia Marmolejo also pointed to the likelihood that in-person voting would be transformed by the new coronavirus, which has led to long lines in other states where elections have already occurred during the pandemic, regardless of whether straight-ticket voting was eliminated.
“Considering the pandemic has already caused long lines at polling-places, many Texans will endure longer lines at polling places indefinitely, irrespective of any order issued by this Court,” she wrote. “And other Texans will experience shorter lines given that voters have been encouraged to steer clear from in-person voting where possible.”
See here for the background. I thought this case was weak, and I am not surprised by the ruling. I do find it ironic that the judge is citing vote by mail as a mitigation of the concerns raised by the plaintiffs. From your lips to John Roberts’ ears, Your Honor. Anyway, there’s still a lot of legal action going on out there. We’ll hope to get ’em next time.
In her order, Garcia Marmolejo ruled that that Democrats’ predictions about the negative effects the lack of straight-ticket voting would have on voters and the election process were “uncertain to occur.”
Garcia Marmolejo is either innumerate or lying. Queuing theory clearly predicts that there will be adverse effects on the election process.
Let me put on my social scientist hat and say that I find the reasoning shocking facile, if not fallacious, at least on a first glance:
DELAY IN BALLOTING IS ENTIRELY PREDICTABLE EVEN WITHOUT STATISTICAL EVIDENCE
Filling out a long ballet (especially in large urban counties with many offices, such as judicial races) obviously takes longer than straight-ticket voting, so there is necessarily a time lag and this effect is entirely predictable without even the need for further evidence (predicting the strength/size of the effect, however, would require more). That’s true even if the voter just uses the new method to register his or her preference for each candidate of the voter’s preferred party without thinking or knowing anything about the individual candidates other than party affiliation. To the extent a certain percentage of no-longer-straight-ticketers engages in more sophisticated voting, it will likely take even longer because they will have to apply their knowledge and evaluation of specific candidates upon making the choice. If they use non-party decision heuristics, such as candidate sex or inferred ethnicity based on name, it will also take longer.
A historical basis for comparison is readily available: past elections with straight-ticket voting, which provide a benchmark for the incidence of straight-ticket voting (number or percentage using what was then an option) that will no longer be taking place under the new regime in the forthcoming election. So the size of the effect can be gauged in a location-specific manner based on historical patterns.
But this historical statistical evidence is not even necessary to confidently predict that individuals will take longer making multiple selections as opposed to making a single one, even if they do so only mechanically. There is nothing speculative about this factual prediction. It can be made based on common sense alone without resort to other evidence (which is not to say that the other factual allegation of predictive nature as as easily proven true or reasonable).
But the other conditions that the judge finds “speculative” have no bearing on this simple causal fact (that marking a choice for each race takes longer than registering a preference for all choices en bulk at one time), although some will ALSO contribute to the prolongation of the voting process, and the length of the lines, if any.
ISOLATING CAUSAL VARIABLES
When you try to assess (or predict) the effect of a change in public policy on human behavior you try to do so in a ceteris paribus framework (all other things being equal), so your attention is focused on the single variable of interest (the change in policy or law) and its impact on the dependent variable (which, in a retrospective analysis, can be done, though never perfectly, through multivariate statistical analysis models). So, if other factors or (changes in) conditions also have an impact on the variable of interest, you have to eliminate those sources of variation from the causal analysis, or analyse their impact separately. Under some circumstances, there could be an interaction of variable, or conditional relationships.
MULTIVARIATE CAUSAL MODEL
While it is true that the conditions of pandemic will ALSO increase the lines (because of cleaning protocols, physical separation, etc that slow down the processing of the voters at the polling place), all those measures ADD to the delay. At the conceptual level alone (without the need for empirical evidence) it is clear that they do not cancel the delay that will predictably be caused by each voter taking longer to register numerous choices separately in the voting booth, rather than voting straight ticket and be done pronto. That is not to say that this cannot be counteracted with additional voting booths, but that is a question separate and distinct from the factual proposition that straight-ticket voting is faster (or that elimination of straight-ticket will slow down the ballot-casting process, which is merely another way of stating the causal proposition).
LENGTH MATTERS
An ostensible fallacy here is to treat “long lines” as a dichotomous variable (either present or absent). That conceptualization of the problem is highly unreasonable because it makes a huge difference if the line consists of 2 dozen people, of several hundred of in-person voters meandering through several city blocks, possibly in heat, humidity, or rain. So, to say that voters will have to endure long lines and delays (and associated risks) because we have a pandemic going really skirts the issue, and impermissibly undercuts a serious analysis of the empirical component of the Plaintiffs’ legal theory.
THE DISPARATE IMPACT THEORY IS EMPIRICAL AND REQUIRES EVIDENCE
As for the partisan effect (adverse to the Democrats, of which they complain in the lawsuit), that is a more difficult empirical proposition to assess and prove, but not an impossible one. It would require social scientific evidence about voting behavior and the associated social scientific theories (that examine the relationship of voter characteristics and propensity to vote, and vote choice).
But at the motion-to-dismiss stage the court takes the allegations in the pleadings as true. So why does this judge reject an allegation of a causal relationship out of hand as too speculative without evidence, thereby precluding any evidence from being adduced that might show that the alleged skewed effects are actually quite predictable, although there might be statistical uncertainty about the SIZE of the effect. This is merits question that can and should be addressed through evidence at trial (or an evidentiary hearing). For standing purposes it should be enough for the Democrats to allege that there is a partisan effect based on extant knowledge about voting patters and correlations with voters attributes such as race and educational attainment, etc.
Note that there is always uncertainty in forecasting models, and there is always statistical uncertainty when data from a sample is used (typically represented as the margin of error/confidence level). Almost all studies of voting behavior at the individual level (as opposed to patterns in aggregate results) is based on samples. There is really no way around it because the voting is secret. So there will never an ability to prove with precision what the judge seems to be asking for. Not to mention that the votes cast could not prove the adverse effects in terms of voters that stayed away. Even if ballots are available for examination after the election is over to determine the extent of straight-ticket voting (or not) and phenomena such as rolloff, that cannot be linked to individual voter characteristics because of the secrecy of the ballot.
IS ALL FORECASTING SPECULATIVE?
Further, all predictive models of future values on metrics/measures of interest (such as here in-person turnout and votes for one party rather than the other) either rely on certain factual assumptions and/or on historical data, or a combination of both. Surely, the position that all social science is too speculate to yield any useful predictions cannot be taken seriously. Making informed predictions about likely effects of alternative courses of action is inherent in all government action (public policy choices) and in much judicial decision-making likewise.
The contrary proposition reeks of SCIENCE DEFIANCE that has become all too prevalent of late.
So much for my first impression from reading the order. I may have more later.
I’m all for the Voters of this State actually having to read the ballot and make some informed decisions.
Neither party deserves our unweighed support. Having seen voters pull for a convicted murderer, a LaRoucher, and countless nobodys who happened to have a famous name, I’m happy to see at least one vehicle for thoughtless voting eliminated.
Charley Hoarse what in the world makes you think that having to vote individually make people go and do research on all the candidates or any more research than they now?
There were hundreds of candidates in the Democratic primary in Harris County this year.
Millions of people voted for Donald J. Trump, think they did some research on what he was really like? Or you think that maybe those that voted for him are mostly racists, bigots, and religious fanatics.