Search This Blog

22.10.18

Flawed judicial decision oversimplistic, imprudent

Fourteen years later, the wisdom of north Louisiana voters became more apparent when 11th Judicial District Judge Stephen Beasley issued a flawed ruling on a hot issue.

Beasley, who ran as a Democrat for the Second District of the Louisiana Supreme Court in 2004 and for reelection to his current post in 2008, but as a no-party candidate in 2014, sided with the plaintiff in a case concerning the state’s non-unanimous jury standard. At present, concerning major crime only 10 of 12 must vote for guilt except, by the state’s criminal code, for a crime that can carry a capital sentence. A constitutional amendment to be decided next month could change that to requiring unanimity.

In this case, the plaintiff had netted a conviction with the 11 white jurors in the majority, but the lone black juror dissenting. His lawyer Richard Bourke, better known for his strident opposition to capital punishment, also has for an extended period argued against the constitutionality of the non-unanimity requirement.


With Beasley, he found a judge willing to bite. The judge overturned the conviction, despite previous state and federal Supreme Court decisions affirming the constitutionality of the state’s standard, alleging these other courts didn’t have the data available to decide otherwise.

Beasley argued this mainly on the basis of a database compiled by the Baton Rouge Advocate (I write a weekly column for The Advocate). With data points from 35 parishes, including nine of the ten largest in terms of their dockets, these showed that in around 1,000 convictions where the vote count of the jury was known, black defendants were 30 percent more likely than white ones to be convicted by split juries.

The non-unanimity standard itself had come about as a hangover from the 1898 Constitution, where debate over it indicated a racist motive in its making easier convictions of blacks. Slightly modified, the 1973 constitutional convention delegates retained it, despite misgivings about its origins.

Prosecutors put up little resistance to the attempt to connect the two phenomena, the long reach of 19th century racism into 21st century statistics. All the prosecution could muster was a critique of the data, which had some validity, as very likely their incomplete coverage – partially from nonresponse of some jurisdictions including the 11th District and also from records of interest simply not kept – would not have excluded systematic biases. But, significantly, it appeared not to have challenged the research methodology that drove the hypothesis.

Thus, Beasley declared, because of those origins, now with the data showing disparate results, this “could only be explained by some outside force operating on the jury process. The only common denominator in these matters was the use of a non-unanimous jury verdict system. The current scheme in Louisiana has a disparate impact on minority jurors and defendants and therefore violates the Equal Protection Clause of the Fourteenth Amendment and is therefore unconstitutional.” This echoed Bourke’s feelings that “this law was introduced for racist reasons, and it still has a racist effect.”

And both views reflect an elementary error in reasoning often made by students new to understanding research methods: the mistaking of association with causation. An old example launched into the minds of millions of statistics students over the decades suffices to explain.

You observe that at fires, there are firefighters. You also note that the more damage occurs attendant to a fire, the more firefighters are present; and the less damage, the fewer firefighters are present. So, you conclude, firefighters cause fire damage, because wherever they go on duty, the damage from a fire varies with the number of them present.

That’s obviously wrong, with the reason being that association – fire damage varying with the number of firefighters – isn’t the same as causation. There’s an intervening variable – the fire itself, which, if its size is controlled for, statistically would obliterate any covariant relationship between number and damage.

Regarding this case, a number of uncontrolled intervening variables could explain the result that would demonstrate the higher conviction rate of black defendants under the decision rule is not imputable to race. A number of alternative hypotheses present themselves than an intellectually-lazy assumption that white racism explains the observed patterns in the data; that lack of sophistication mirrors arguments such as, for example, higher black unemployment in America must be a cause of a racist economic system.

Regarding the 30 percent higher likelihood of convicting a black with a non-unanimous jury, a number of things far more plausible than racist attitudes – expressed by black jurors more likely holding out for innocence meaning non-unanimity lets racist whites railroad black defendants – can explain that. For example, it’s equally as plausible that black holdouts do so for racial reasons – maybe they hold a grudge against a society they see run by whites to their detriment and so want to give every break possible to black defendants they may consider victimized by that society.

Or, the non-unanimous rule may encourage some black jurors to lodge symbolic protest votes. The way jury deliberations go where one or two black jurors know the whites will vote to convict, they may be convinced of a black defendant’s guilt but also want to make a statement against a society they see as unfair. So, secure in the knowledge the guy will be put away even if they vote to acquit, they do so – meaning if under a unanimity requirement they would have voted for a guilty verdict, so nothing would change insofar as convictions go.

Much less controversially, and far more likely explanations, stem from general demographic differences between blacks and whites in the aggregate, such as blacks typically have lower educational attainments and lower incomes. So, for example, perhaps a basis of comparison should include the perceived status of the defendant, race aside. Black jurors, who more likely come from a lower socioeconomic status background, may feel for empathy for defendants from that background – again, who are disproportionately blacks – and vote to acquit as a result.

I’m sure I could come up with more such explanations, at least as plausible if not much more so that the one forwarded by Beasley, but you get the point. Regardless of what Beasley thinks, we simply don’t have the data – or perhaps haven’t tested thoroughly enough the alternative hypotheses with what we do have – to say one way or the other. And that includes imputing racism to use of the non-unanimity standard.

To draw such a conclusion in this instance displays oversimplification and imprudence. If this exemplifies the quality of decision-making Beasley would have brought to the state’s highest court, voters then wisely instead chose now-retired Associate Justice Jeff Victory.

No comments: