Friday, March 21, 2008

Dueling numbers

Jessica McBride has now published the Butler campaign's revised list of cases in which he ruled in favor or against a criminal defendant. The list is supposed to support the campaign's claim that Justice Butler ruled in favor of a criminal defendant 75% of the time. (Actually, the campaign's characterization of the results seems to come to about 70%.)

As I have blogged before, there is limited value to numbers like this. They are best considered in comparison to other justices. Even for those with a fairly conservative view of these matters, sometimes the defendant should win. But the Butler campaign and a group opposing Justice Butler's reelection are offering us some numbers that are, to put it mildly, inconsistent. The Coalition for America's Families, for example, claims that Justice Butler rules in favor of criminal defendants sixty percent of the time. Sixty percent is a long way from 25%.

I don't how CFAC did it's analysis, but the Butler campaign, to its credit, has released it own analysis.

The problem is that a brief review of the analysis raises serious questions. Jessica McBride points to several cases that she feels were mischaracterized. I haven't run down all of them, but she seems clearly right with respect to at least some. Here are a few more based only on cases which I had some familiarity with. These are cases in which the Butler campaign apparently claims that Justice Butler ruled against the "state/inmate" The question here is not whether the outcome was correct, but if that claim is accurate.

Coleman v. McCaughtry - Coleman was an inmate who filed a petition for habeas corpus in 2004 arguing that he had been denied effective assistance of counsel when his lawyer failed to appeal his 1986 conviction. The issue was whether his motion came too late, more specifically, whether it was barred by a legal doctrine called laches that says, essentially, that you can lose your rights by sitting on them for too long. A majority of the court held that some of the elements of laches had been proven by the state, ie., that Coleman had unreasonably delayed in bringing the claim and the state had no knowledge that it would be brought. But Justice Butler disagreed and would have remanded the case for factfinding on those issues.

Coleman nevertheless got some relief because all of the justices felt that there needed to be factfinding on whether the state was prejudiced by Coleman's delay. So, we have a majority issuing a ruling that went against Coleman on two issues (his claim had been unreasonably delayed and the state had not known of it)but gave him a chance to contest the state's claim of prejudice on remand. Justice Butler would have given him a chance to contest the state's finding on all of three issues.

Of course, Coleman has no guarantee of success on remand. He may very well lose. But, clearly, in this case, Justice Butler ruled in favor of a convicted petitioner.

State v. Jerrell C.J. - This is one of the five decisions discussed by Judge Diane Sykes in her 2006 Hallows lecture at Marquette University Law School. It involved the admissibility of a juvenile's confession. All of the justice felt that it was involuntary and had to be excluded. A majority, joined by Justice Butler, went further and adopted a rule that all custodial interrogation of juveniles must be recorded. Justice Butler also would have held that, whenever a juvenile asks to speak to his or her parents, all questioning must stop. None of the justices would have adopted a per se rule that juveniles must be given an opportunity to consult with a parent or interested adult prior to interrogation.

I can't see how this could possibly be characterized as a case in which Justice Butler "ruled in favor of state/against inmate."

McBride also argues that it was wrong for the Butler forces to add cases involving civil commitments. I disagree, in part because, as Illusory Tenant points out, the Gableman campaign or its supporters have raised such cases on this issue and because I think they are pertinent to the larger issue of judicial philosophy which has to do with how you balance public safety with the rights claims of the accused.

But, again, I don't understand the characterization of some of these cases.

State v. Brown is the now infamous case in which the Butler camp and others have pointed out that a claim made by the Gableman campaign that Brown, a sexual predator, has been released is literally false. Brown is still in custody.

But what the Gableman's critics don't point out is that the Supreme Court ordered his release. A majority, joined by Justice Butler, held that the state had not proven, by clear and convincing evidence, that it was more likely than not that Brown would reoffend. It's mandate was as follows:

For the reasons set forth, the decision of the court of appeals is reversed, and we remand the cause to the circuit court with instructions to notify the Department of Health and Family Services that it should submit to the circuit court, pursuant to Wis. Stat. § 980.08(5), a plan for Brown's supervised release.

I can understand why someone would think that the case resulted in Brown's release. Obviously something happened after remand that is preventing it.

The Butler campaign characterizes this decision as a ruling "in favor of state/against inmate." It says that the case involved a "technical issue- remanded to give the State a chance to make its case." Actually, what the court held is that the state hadn't made its case and it remanded for the circuit court to consider a plan for supervised release.

All of this, it seems to me, makes it likely that the "75%" number is wrong. Mike Plaisted says that he did analyses of his own. Once he claimed that his own informal review of 31 cases showed Butler sides with the defendant approximately 42% of the time. That is obviously a much smaller sample than used here, but the rate would be consistent with the idea that there are substantial errors in the Butler campaign's analysis. Yesterday, he threw out a selection of statistics that don't address the question at hand. He says he counts 45 cases which overturned or upheld a conviction (not quite the question here but potentially instructive) and, based on the snippets he offers, it seems that there are up to 20 in which he believes that Justice Butler voted to uphold the conviction and 13 in which he voted to overturn. How I am supposed to conclude, as Mike says, that the campaign's numbers are "fairly accurate" is unclear to me. Maybe they are, but you can't tell that from what he has chosen to share with us.

Again, what would be more instructive is a cross court comparison of these numbers, particularly in cases that are not unanimous. I haven't seen those numbers but my impression is that Butler would vote with defendants more than Roggensack, Prosser and Wilcox (we don't have enough Ziegler cases yet)and less than Abrahamson and Bradley.


William Tyroler said...

I've got a number of problems with this entire effort. For starters, typing a decision as "in favor of a criminal defendant" is often enough an empty exercise. How, just to give one example, should State v. Hamdan, 2003 WI 113 be pigeon-holed in this arbitrarily binary scheme? The court reversed Hamdan's conviction, so it was concretely "in favor of a criminal defendant." However, it did so by recognizing that he was simply exercising a right given to him by our state constitution (and denied him by an overzealous local prosecutor). So, why isn't a ruling in favor of all ciitzens? That said, the court's view of the right was (in my estimation, for whatever that's worth) so stinting as to raise serious doubt as to its practical value. So, it was a ruling in favor of the immediate defendant, granting a potentially important right to all citizens, but construed in such a way that very few will benefit. It's arbitrary to categorize it as "pro-criminal defendant.

Now take a look at Coleman, the case Professor Esenberg's analysis would place on the pro-defendant side of Justice Butler's ledger. As the professor indicates, Justice Butler (along with 2 other justices) would have recognized one more potential basis for relief than the majority. But the relief would have been purely procedural: the thrust of his partial dissent is that "it is clearly inappropriate for either the court of appeals or this court to decide this petition without further fact-finding" (para. 44). Urging a fact-finding hearing, then, is "pro-defendant"? Perhaps, if you're forced into a McBride's binary world.

I'd like to think that a more sophisticated analysis might be forthcoming, especially from a major law school. In the absence of same, it's no wonder we're forced into the current sterile debate.

What if Jerrell C.J. actually leads to a greater rate of conviction (not to say a more factually reliable result)? (The decision, of course, applies to juveniles.) What if the effect of dragging the Milwaukee Police Department from the 19th to the 21st century is to eliminate juror suspicions of police misconduct; eliminate spurious defense claims of same? I'm not saying that that's necessarily so, I'm saying that that's a plausible hypothesis. For that matter, that seems to be what the legislature thought, which in the wake of that decision promulgated sec. 972.115, which at least partially extends the case to adult prosecutions. Is the decision pro-defendant or pro-reliability?

To put it baldly, it is distressing that the issues have been discussed in such a simplistic, polemical way. It's not least a shame because the candidates probably do have very different jurisprudential philosophies; we're not going to get at these differences though, and in that sense McBride is doing the public a genuine disservice.

Finally, I want to express appreciation to Professor Esenberg for devoting his own resources to a public forum. Everyone benefits from his expenditure of time and effort. I am particularly appreciative of his encouragment of robust debate. Thanks much.

Rick Esenberg said...

I noted that there are limits to the value of such a number and, of course, there are cases that are hard to characterize. On the other hand, I don't think that looking at general tendencies is unrelated to getting at differences in judicial philosophy and it is not unusual for academic studies to do so. The idea is that looking at a sufficiently large number of cases will tell us something that might not be as apparent from a case by case analysis.

You can argue that Coleman was correct or that Jerell C.J., although presumably intended to arrest what was thought to be a crisis of false confessions, will actually make it harder to get confessions tossed. But the point is not the merits of any individual case but an attempt to get a date point about a justice's overall approach.

As I also said, it would be most meaningul when compared to the same data for others.

Mike Plaisted said...

Having spent several hours going through every one of these cases, it is indeed hard to catagorize some cases as pro-defendant or anti-law enforcement. The real issue is whether the decision is true to due process and other constitutional rights and the effective administration of justice, but I know that's not as sexy in a campaign.

As a practicing criminal defense attorney, I can tell you that Jarrell and its adult-equivalent statute (requiring the recording of most statements) has had a dramatic impact on Miranda/Goodchild hearings that try to get the statements suppressed. I now get a CD of my client's complete session with the detective, not to mention that of other witnesses or co-defendants. This cuts down quite a bit on the evidence process -- you tell the judge to listen to the CD and they argue from there. Sure, there are still ancillary issues -- what did the cops or my client say before they got in the room; was my client intoxicated, etc. -- but what happened during the interview itself is all there. I think everyone in the system -- including police and prosecutors -- would agree that this is a good thing.

And what about Brockdorf? McBride pretends not to know how Butler's dissent (with Crooks) could possibly be pro-law enforcement, but it is certainly legitimate to call it that. Butler and Crooks both say they want to protect the integrity of police internal affairs investigations, where cops have to talk or be fired.

And, for right-wing issues, how about Fisher? Butler was actually in a 4-3 majority for conviction, but, since it was a CCW case, the three most conservative justices who usually find for the state when they can, would have let the defendant off the hook.

I don't know what the percentage is or how Butler's campaign staff got there, but one thing is for sure -- the Gableman surrogates' 60% pro-defense figure is completely insupportable by any definition. McBride signs off on this figure today for no good reason, just because she wants to. Even after two tries, hers is a crummy hack job.

William Tyroler said...

McBride is working off a sample size, N, of 62 (i.e., she purports to analyze a total of 62 cases where Justice Butler cast a vote). N is, in the first instance, simply too small to draw meaningful conclusions -- at least as expressed the way she wants. Professor Esenberg suggests, if I read him right, that this might be " a sufficiently large number of cases [that] will tell us something." But let's put this number in concrete terms.

The CFA claims that Justice Butler "favors" criminal defendants 60% of the time. That means 36 "pro" votes. The Butler campaign pegs the % at 25, or a total of 15. The controversy over this discrepancy obscures the obvious fact that these numbers accrued over time: the 1st case on McBride's list, Hale, was issued 1/05 or 3+ years ago. This means, then, that even if you take CFA's tally at face value (which I gather even Professor Esenberg thinks a silly proposition), we're talking less than 12 "pro-defendant" votes per year. Is that too many such votes? Should we be shooting for, say, no more than 8? Or 6? 0? What would be acceptable quantum to the CFA and McBride? Is this a meaningful basis to decide a supreme court seat?

But that's not the only implication of this small N. As is true of any small sample, small disruptions wreak statistical havoc. If McBride has mischaracterized 6 "pro" votes, she has altered the % total by about 10. The following figures are purely illustrative. Say she comes up with 24 "pro" votes -- that equates to a 38% success rate for defendants. But if the number is more accurately 18, then the rate declines markedly, to 29%. This isn't necessarily a problem, except that as Professor grudgingly agrees, "there are limits to the value of such a number and, of course, there are cases that are hard to characterize." Indeed so. And without any rigor in the characterization process, no confidence should be placed in the highly contingent conclusion (that Justice Butler profligately favors criminal defendants).

Professor Esenberg, in short, is quite willing to leap into what to me is a void; he would draw conclusions about "general tendencies" and I wouldn't. He seemingly embraces the idea of an election decided by a rush to quantify what to me is, at least presently, unquantifiable.

There's another sticking point, which keeps coming up but is never refuted. The overwhelming number of defendants' pleas for relief are rejected out of hand by the supreme court as a whole (and including Justice Butler individually). These are, of course, petitions for review, and the occasional stray writ. Why don't these summary rejections count for anything? It is surely the refusal to throw this data into the mix that accounts (or maybe better said, allows) for the small N. Throw those in and you've got some data that better support quantification -- and, to be sure, general conclusions uncongenial to McBride's "pro criminal" thesis.

Anonymous said...

Atty Plaisted,

Your point about the positive effects of Jerrel is well-taken. However, wasn't the job of the court to determine whether the constitution required such recordings? It seems like instead of answering the above question the court went beyond what the constitution required and made a policy determination. Good, bad, whatever, it's a policy determination that belongs to the legislature. Obviously, Butler, et al. cloak the decision in the constitution, Miranda, etc, but the truth is that the decision to require recordings for all juvenile statements cannot find support in any legal authority. Encourage the legislature to pass such requirements from the bench or in the opinion, but don't make policy from the bench.

Rick Esenberg said...

I don't recall saying that the election "should be decided" by these numbers. I think I said they are of limited value and are one data point. But they are not worthless and, in any event, have become an issue because both sides have made conflicting claims. It seems clear that the Butler campaign has cooked their numbers. (My guess is that the CFAF may have as well but I don't know where they came from.)

As far as whether this shows general tendencies, I don't think, as someone involved in criminal appeals, that you could deny that Justice Butler is a more sympathetic jurist than Roggensack or Wilcox- something that the campaign is not likely to acknowledge. If the numbers point to that, they have not mislead.

I understand that you believe - and I actually agree - that the public undervalues the rights of criminal defendants. But I think that, if you are going to elect judges, you have to make the case for a judicial philosophy as opposed to pretending that it is other than what it is or acting as if the role of a judge in these cases is mechanical.

Now, it may well be that taking these numbers in isolation does suggest that there is a bigger difference than there really is. But there's a remedy for that as well.

Display Name said...

Yes, I'd like to hear more regarding the value of these numbers. As I pointed out in a previous thread, wouldn't these numbers reflect trends of errors in the lower courts as much as any trends of activism in the highest court? Or reflect some changing trend in law enforcement tactics in the last decade that ultimately get tested and corrected in the courts? Or reflect confusion about poorly written laws created by the legislature whose weak spots get tested by the courts? Or a combination of all three?

We're talking multi-variable analysis of human activity with only a few dozen data points. Get serious!

Some are cheering the partisans who want to distill it into a factoid they can paste into a 30-second commercial. Who is dumbing down what, and who is paying them to do it, and who has been paid to keep quiet?

Do phrases like "tough on crime" and "almost jeopardized" mislead the hordes of the public who do not understand the role of an appellate court?

Mike Plaisted said...


I understand your concern with the Jarrell directive that juvenile interrogations be recorded or be subject to exclusion. Those concerns are all over the opinions of the three justices who partially dissented on that point.

I think it is important to note that that part of the majority's decision is made in the context of a "confession" by an 11 yr-old that all 7 justices agreed had to be suppressed because of coersion -- it has to be a pretty bad situation to bring around a unanimous court on that kind of point. [Sidenote: do you think, given his campaign promises, Gableman would have recognized it? Could he have?] The majority does not take this step lightly. But it was a directive that was successful in other states, easily implimented and prospective rather than retroactive. Certainly it is something the legislature could have done by statute, but that doesn't mean that the Court doesn't have a concurrent responsibility to drag a law enforcement practice into the 21st century in an area of great contention and confusion. That the directive has been greatly successful in achieving the very goals the Court intended is not the entire answer, but it's not nothing.

The bottom line is that the Court doesn't have to sit around and watch a historically difficult situation continue in an important constitutionally-protected area like coerced confessions when there is a simple solution that does not, with clear direction, impose any burden on law enforcement. This is a rare exercise of the Court's supervisory authority, designed to solve a difficult problem. And so it did.


This whole discussion takes place in the context of wild claims by Gableman surrogates and his campaign itself that Butler "ties the hands of law enforcement and sides with criminals". We can play with the numbers, point out how many 4-3 decisions there were in favor of the defendant with Butler in the majority (answer: 4 out of 70), look at the majority, concurring and dissenting opinons he has written, review the kinds of cases the court has faced and, at least, see whether that claim is at all supportable.

Any fair review will find that it isn't. As you seem to agree in your post (he's not as "conservative" as Wilcox or Roggensack; not as "liberal" as Abrahamson or Bradley), he is right in the middle on these cases. He has joined 7 majorities for conviction where Abrahamson and/or Bradley have dissented; he has dissented 5 times where the majority ruled for the state.

However you get there, you have to admit, at least, that the "sides with criminals" line is an unfair charactorization of his record. You seem to want to let those claiming otherwise off the hook because you haven't "seen their numbers", but at some point do you not have to conclude that they have not disclosed their numbers because they don't support their claims?

Anonymous said...

Does Butlers cases ever show a preference for victims, either civil or criminal? It appears to me that he wants to prevent any judicial abuse of the perpetrator/offender but doesn't care about the victims. Apparantly the courts can abuse them all they want without concern of what he may do.

Additionally, Butler has been a Justice for nearly 4 years. Is the figure of 62 cases the total number of cases that he has worked on since becoming a justice? Is that typical workload for our Justices?

As for the dueling numbers I must say that my neighbor that is an accountant is right inthat he claims that lawyers should never deal with numbers because they have enough problems with words.

It's obvious that both camps are using different criteria, however, Gableman has won in being closer to the truth.

William Tyroler said...

they are of limited value and are one data point. But they are not worthless

Alright, then: a single, non-worthless data point of limited value is McBride's rallying cry to derail the Butler campaign. I don't know about anyone else, but to me it makes for a non-edifying election that will be of limited value in clarifying anything.

in any event, [they] have become an issue because both sides have made conflicting claims. It seems clear that the Butler campaign has cooked their numbers.

A phony controversy. No one "cooked" any numbers, not unless you equate a certain amount of sloppiness with cooking data. On the other hand, you might equally want to say that Gableman supporters cook their data, in part with a now-you-see-it-now-you-don't approach to what's criminal and what's civil. (If, as McBride says, you only count "CR" suffix appeals, then you can't throw in juvenile cases [Jerrell C.J.]), SVP commitments [Brown], or most bizarrely 974.06 appeals [Armstrong].)

Besides, this is ultimately an exercise in nit-picking, turning on a hastily put together tabulation. (Had I been advising the Butler campaign I would've said, don't bother doing it at all.) But to Professor Esenberg, this isn't simply a confected controversy but something that reveals an even deeper flaw: "It seems clear," he says, "that the Butler campaign has cooked their numbers." Data is "cooked" when scientists design experiments in advance to obtain a desired result. Quite obviously, it's a serious breach of ethics. Does Professor Esenberg mean to imply, then, that the Butler campaign intentionally falsified the count? I don't know any other way to read his hyperbole. At the same time, it's a silly allegation: the list of cases is public; the process is transparent. That the count may have been wrong simply says that the campaign didn't anticipate the need to hire bean-counters.

Justice Butler is a more sympathetic jurist [to criminal defendants] than Roggensack or Wilcox


If the numbers point to that, they have not mislead.

I don't get it. McBride isn't saying that the numbers merely suggest as much, she's saying two things, both of which seemingly find Professor Esenberg in agreement. 1) The Butler campaign lied. ("Cooked the data.") Professor Esenberg seems content to have an election to our highest judicial seat decided on the accuracy of case-result tabulations. I guess I'm not. 2) Justice Butler is not just "more sympathetic," but is way too "pro criminal." She yokes the ludicrous 60% figure to that end. I don't want to reprise an already lengthy discussion but that gets us back to square one: It's one thing to say that Justice Butler is more likely to support the procedural rights of criminal defendants than Judge Gableman; quite another to say, "Justice Butler is pro criminal 60% of the time!" Put it like this: Professor Esenberg says the numbers generally support a certain (relatively benign and probably incontestable) assumption; McBride suggests that the numbers quantify and thus definitively show just how "pro criminal" he is.

you have to make the case for a judicial philosophy as opposed to pretending that it is other than what it is or acting as if the role of a judge in these cases is mechanical

I never meant to suggest that Justice Butler disguise his judicial philosophy. I actually think the electorate is mature enough to handle someone who intelligently discusses the rights afforded all citizens. (And, for whatever it's worth, I think the current political climate is especially conducive to such an approach.) I'm opposed, rather, to the polemicism of the other side. Professor Esenberg differs on that point, I gather.

As long as we've carried the discussion this far, let me put the question: If one accepts the idea of quantifying a justice's "pro-criminal" philosophy, then what is the pass-fail metric? At what point do you say, too high a percent? 10? 15? 0? Should Justice Gableman commit to a quota on granting relief to criminal appellants?

And, to continue to harp on something, not least because it lies at the heart of the methodological problem (which probably explains why it goes unaddressed): what is the basis for excluding denials of review from the overall count? Because if you do, as Professor Esenberg well knows, you lower the "pro criminal" percentage to something in the area of 95%.

illusory tenant said...

Does Professor Esenberg mean to imply, then, that the Butler campaign intentionally falsified the count? I don't know any other way to read his hyperbole.

That has been, and remains, my question as well. I also think Prof. Esenberg has chosen a very unwise battle to fight here, but that's entirely his prerogative.

Mike Plaisted said...

Annony 9:54 -

The 62 or 70 cases we are talking about are just those having to do with criminal issues -- about 25% of the cases decided by the high court.

As for victim "preference" in criminal cases, I don't know anyone in the criminal justice system who does not have sympathy for victims of crimes, especially violent ones. At an appellate level, though, victims "rights", such as they are, do not play much of a part. The questions is usually whether the accused got a fair trial and whether the state violated any constiutional rights (right to remain silent; right to proper search and seizure of person and property). No judge can or should say, well, the defendant's rights were violated, but this victim got really hurt, so we'll just forget about it and find him guilty anyway. Notice that when a defendant loses a jury trial in a criminal case, he has the right to appeal. If the state loses a case tried to the jury, that's it -- the state can't appeal an acquital. It is the nature of the constitution that it protects the rights of defendants.

I don't know how you and your accountant friend are looking at this, but the Gableman forces aren't anywhere close to the truth with their 60% pro-defendant figure or anything else they have made up out of whole cloth. They haven't even tried to step forward and explain or defend it. I think that tells you all you need to know about them.

Mike Plaisted said...

Another thing about the numbers, gentlemen...

The numbers can also be skewed by the off-the-hook cases that bubble up to the Supremes several times a year. What do you expect them to do with a judge is sloppy at both the plea hearing and the post-conviction motion challenging the knowingness of the plea (James E. Brown)? How about when a juror says he doesn't agree with the verdict when polled (Raye)? These are cases where the unanimous court undoubtably got it right, but it's counted as a "pro-defendant" negative in manipulative campaign rhetoric.