On Judicial Decisions, Part 1

An earlier article sparked active discussion regarding whether or not this model, or some variant thereof, may be applicable to judicial decisions.  In particular, Mark Zuniga asked if the LEO test could be used to identify where particular judges stand, thereby offering a predictive model of judicial behavior.  This question is particularly important now, with DC Circuit Court Judge John Roberts still subject to the scrutiny of the Senate Judiciary Committee and a full vote of the Senate regarding his appointment to the Supreme Court of the United States. 

Conventional wisdom in political science suggests that examining the judiciary is an arduous and frustrating experience.  Rules of restraint prevent judges from expounding upon hypothetical cases, and while a record of their opinions is generally available, those opinions are applicable only to the resolution of specific disputes before the bench.  Nevertheless,  decisions in subsequent disputes may refer to the resolution of previous cases, if those cases share a similar application of law, either in whole or in part.   

With the caveat firmly in place, I set out to apply the LEO test to judicial opinions, if at the very least simply to observe the result.

One of the first cases I examined, and my thanks go out to Mark Zuniga for providing the link, was Roberts' opinion in Ramaprakash v. FAA (2003)  For this test, I employed exactly the same method as for the Sullivan sample earlier.  Here are the results (Click to enlarge):

Raw scores:

Percentages:

 

Suffice it to say, these first scores were disturbing for a variety of reasons.  First of all, if they are valid scores, then the Ramaprakash decision indicates that Roberts holds fast to an extreme establishmentarian position far beyond that of other test subjects such as Rick Santorum or even Michael Peroutka.  However, Roberts own testimony at his confirmation hearing for the DC Circuit Court suggests a level of restraint inconsistent with these scores (see especially p. 80).  This inconsistency suggests that some portion of the LEO test as it was applied to the Ramaprakash case and perhaps to judicial cases in general, is either  incomplete, inaccurate or invalid.

What, then, explains this peculiar result?  Mark suggested a possible explanation--that the legal nature of judicial decisions of necessity incorporates establishmentarian language, for individual disputes are brought before the bench to be resolved in terms of the application of existing law.  This hypothesis is not only reasonable, it can be tested, and my thanks go out to Jeremy Jose of Melbourne University's Political Interest Society for suggesting the means to do so.

If the legal nature of judicial decisions distorts the results of a direct LEO test in favor of Order, then one can reasonably expect most court opinions to have similar results.  Following Jeremy's suggestion, I  tested this hypothesis by "running the test on some notoriously liberal opinions": Brown v. Board of Education of Topeka KS (1954), Miranda v. Arizona (1966), and the controversial Roe v. Wade (1973).    The hypothesis predicts a high percentage of "Order" references for each of these cases.
Below are the results:

Brown et al. v. Board of Education of Topeka, KS
The first Brown decision is best known to Americans as the decision that, through the application of precedent, effectively ended the practice of racial segregation in public facilities in the United States, especially public school systems.  Brown is generally considered the cornerstone of the civil rights movement of the 1950s and 1960s in the United States.  As such it is a favorite case for egalitarians to remember when calling attention to persistent inequalities in American society, in both the public and private sector. 

Now the hypothesis states that the unique legal character of judicial decisions necessarily distorts LEO scores towards "Order".  Using the procedure described above, a LEO test of the Brown decision yielded the following percentages:

 

For each of the test tables, Equality, not Order, produced the highest percentages.  In the first table, almost 60% of ideological indicators captupred by the LEO test were egalitarian in nature. Overall, egalitarian references constituted 50% of all captured ideological indicators for this case. Consequently this one case alone refutes the hypothesis.   Something else is clearly at work besides the legal language of judicial decisions.

Miranda v Arizona (1966)
While the Brown decision specifically addresses issues related to equality and could be described as a "liberal" decision, the Miranda decision focuses on the restraint of government vis a vis the rights of the accused, and is generally understood as a libertarian case rather than a liberal case (one might expect the same of Roe v. Wade, but more on that later).

Unlike the Brown decision, in which the Supreme Court unanimously reversed lower court decisions, the Miranda case inlcudes two distinct dissenting opinions.  Submitting the majority and dissenting opinions separately offers insight into the character of these opinions relative to the LEO model, as they produce measurably different results.  See below (Click to enlarge):

Miranda Majority (percentages):

 

Clark's Dissent in Miranda:

   

Harlan, Stewart and White's Dissent in Miranda:

      

    

Given that the Miranda decision is generally a favorite among civil libertarians, it should come as no surprise that the test of the majority opinion yielded a strong libertarian signature.  Justice Clark's dissent suggests a generally conservative dissent, although it is noteworthy that the first test table indicates that Clark was sure to include a nod to liberty as well. 

Justice Harlan's dissent, joined by Stewart and White, yields several distinct signatures for each test, and from all over the model.  This scattered result may point to a kind of centrist critique of the Miranda case--assuming, of course that these signatures are valid.  The first score appears decidedly liberal, while the second score is decidedly libertarian.  The third score presents liberty and order in equal measure above equality, while the fourth measure is clearly conservative.  Overall, the aggregate score of the Harlan/Stewart/White dissent in Miranda v. Arizona appears to be that of a moderate left libertarian, where equality and order, while they don't take center stage, are nevertheless much closer behind liberty as a primary focus than the majority opinion.

Caveat
On the surface, these scores suggest that the LEO test might in fact be used to analyze judicial decisions.  If these scores are taken as they are, a moderate would therefore find the Roberts nomination especially chilling.   However, the measurement of the next case in this short series suggests that a critical analytical component for judicial decisions is missing; until it is found and applied to these cases, I am not confident to say anything conclusive about them from the perspective of the LEO model.

The decision that throws the entire system into question is, of course, Roe v. Wade (1973).  This decision, along with the much more recent Planned Parenthood of SE Penn. v. Casey (1992), will be the subject of the next article.

Comments

[eowyn]

Great post, Jonathon. I especially liked the way you broke out the Miranda dissents.

[jonathon]

Thanks, Eowyn. I just talked to Justin over the phone about the problems with using landmark posts, but everyone is familiar with at least some part of these decisions. That and Jeremy suggested these particular decisions.

Either way, 'ordinary' decisions may still be likely to support Mark's hypothesis. The trouble of course lies in figuring out exactly what makes landmark cases different.

I am also learning from this exercise just how different judicial decisions are from any other kind of speech. I do not doubt for a minute the suggestion above about there being something missing in the analysis of judicial decisions. I have an inkling about what it might be, but frankly it would take a long time to collect the data.

[Jeremy]

Hey, glad you took me up on my challenge and great analysis there. As noted, comparing the dissents was a wise move.

I'm honestly surprised these preliminary results seem to show that the LEO test work on judicial decisions. I study law and after four years of reading Australian High Court judgements, I have tended to find that the more a judge is taking a radical legal position the more they appeal to order and precedent. These suggest my hypothesis might be wrong. However, I do note that key US legal cases turn upon questions of constitutional importance where appeals to rhetoric and fundamental principles are much more relevant. Our constitution is relatively weaker, so even fairly radical cases like Mabo (which established Aboriginal Native Title rights) have to try and look like they fit in the "skeleton of precedent" and the common law.

Your test is shaping up to look like a more and more accurate predictor, it'll be interesting to see how it holds up as it's applied to more situations :)

[Mark Zuniga]

Compelling findings. Time will tell if Ramaprakash is a telling opinion.

As an aside, it sure feels like the internet is going to alter the way SCOTUS nominees are going to be vetted.

I feel like I do nothing but come up with more work for you. Of course, you can always ignore me. Not an untraveled path. Still, msny people, especially conservatives, criticize the Brown opinion because it is not more in line with the Harlan dissent in Plessy v. Ferguson.
http://caselaw.lp.findlaw.com/cgi-bin/getcase.pl?court=us&vol=163&invol=537
Brown focuses on the rights to go where one wants to go, and critics say that the problem with Brown is that it focuses on feelings of the excluded children, not their rights. Conservatives have said that if Harlan's dissent in Plessy were law, afirmative action would not be. I wonder if the LEO test could differentiate between the two, or if any argument about racial equality necessarily means a higher "e" in the LEO test.

[Mark Zuniga]

By the way, this blog is the only indication that I would prefer having Rick Santorum on the SCOTUS.

[Aaron]

I don't have much background in law, but it seems to me that most standard legislation is written in clear, relatively specific language, and that when considering a case falling entirely within such laws, a court would mostly be concerned with the relationship between the facts and the regulations. Many Constitutional amendments, by contrast, particularly those dealing with civil liberties and civil rights, are vague and ideological in nature, and as such leave themselves open to more opinionated interpretations. That might account, at least in part, for why landmark decisions like Brown and Miranda rely much less on legal, order-based arguments than do more mundane decisions like Ramaprakash.

[jonathon]

Mark,
It's always good to be able to go in a particular direction, to get more results, to put this test "to the test" as it were. So far it's revealed some very peculiar things about court cases. I just did Plessy, and the numbers were certainly not what I expected. But then I read the opinion, and it suddenly made sense.
The majority opinion in Plessy turns on a distinction between social and political equality and therefore yields a high "e" in a direct LEO test.
Harlan's dissent in Plessy appears to make more references to liberty than anything else. In 3 out of the 4 tables, L was highest. Looks like the 'conservative critique' may actually be a libertarian critique. This line from Harlan's dissent bears out this assessment:

"The fundamental objection, therefore, to the statute, is that it interferes with the personal freedom of citizens. 'Personal liberty,' it has been well said, 'consists in the power of locomotion, of changing situation, or removing one's person to whatsoever places one's own inclination may direct, without imprisonment or restraint, unless by due course of law.'... If a white man and a black man choose to occupy the same public conveyance on a public highway, it is their right to do so; and no government, proceeding alone on grounds of race, can prevent it without infringing the personal liberty of each." (from Harlan's dissent)

[jonathon]

Aaron,
First of all, thanks for coming by. Second, I'm inclined to agree with you; this is one of the difficulties when dealing with landmark cases. I imagine that most mundane cases will likely come up as Mark predicted. The missing component deals with a matter of context.

[Mark Zuniga]

Hopping down the rabbit trail . . .

As a lawyer, I have to say that many laws are not written in concise language. The lege, after all, is dealing with generalities, and not with the specific details of the actual case in front of a court.
For example, many states had a rule that said that if the parents do not have a will and they die, the children inherit everything. This is a fine rule generally, but what about the Menendez brothers? I think most of us would want an activist judge (or at least one with little "o") in that case to keep the money from the children.

Many cases involve competing values, of which order is important, but only one. With SCOTUS, the cases are merely more abstract.

[jonathon]

And this is where it might be useful to look first at the opening and closing arguments of both plaintiff and defendant, capture some data to establish a baseline, then figure out its orientation in LEO space. After that, one may compare the language of the judgement's position relative to that baseline as if the midpoint of that line were an origin. This modification could also measure whether one or another basic principle is emphasized, obscured, or rendered irrelevant to the particular case.
From what I can tell at the moment, that is the missing piece. Just as judicial decisions are made on an individual basis, so too must an analysis of those decisions. An aggregate of decisions thus analyzed should provide us a reasonable picture of the judicial actor's preference.
It hurts to think about it. That means pouring over original court records, doesn't it?

[Mark Zuniga]

It's harder than that. You also have to look at the appellate briefs filed by any party.

[Jeremy]

I don't have much background in law, but it seems to me that most standard legislation is written in clear, relatively specific language, and that when considering a case falling entirely within such laws, a court would mostly be concerned with the relationship between the facts and the regulations.

Hehehe, you'd hope that wouldn't you. Having a background in law, I'm afraid to report that most legislation does not, and indeed cannot, fit that description. The issues which you think are/should be simple have occupied many hundreds of thousands of pages of writing over a couple of centuries.

That means pouring over original court records, doesn't it?

Older case reports (at least in Australia and the UK) often had a good and fairly brief extract of counsel for both side's arguments, otherwise, yeah, that could be some nasty research.

[jonathon]

Yeah, that's what I was afraid of. Mark's that I'd have to look at all the appellate briefs for a given case are particularly daunting.
One thing I have learned about legislation, having poured over bills and resolutions, is that the legislative process itself does not lend itself to either clarity or precision. This may be why a direct LEO test works as well as it does on sponsorship records (see the case study on John Kerry for details). In addition, the short ones are often vague, the longer ones contain seemingly exacting yet often obscure language, others contradict themselves or another bill, usually through nearly successful attempts either to saddlebag or transform the bill before it ever gets out of the legislature. That complexity lends itself quite well to the methods I'm using.

With the judiciary it's even more complex, largely as a result of the nature of judicial power, which is to resolve disputes in law and equity. One has to wonder, then what would happen to these landmark SCOTUS decisions if they were submitted to this other hypothetical method of establishing a relative field for each case (described above) using the record and appellate briefs, and then deriving a new adjusted score.

[The Citizen Bob]

Howdy Ho J,

Please... You've got my attention cornered. I wanna see you break down Roe v Wade w/ LEO. I am the furthest thing from a lawyer and I'm getting interested in law. :)

[jonathon]

Sorry about the long delay. So much has happened in the last couple weeks that I haven't had time to do the write-up. There was the beginning of classes and then DC trip over this weekend. Not quite sure how I'm going to manage all that I have on my plate at the moment.