October 17, 2019 Feature

Interview with Judge Jed Rakoff

The well known federal judge discusses eyewitness identification, science education for judges, and more.

Ashish Joshi

Download a printable PDF of this article.

Judge Jed Rakoff has been a federal judge in the Southern District of New York since 1996. Before that, he was both a federal prosecutor and a white-collar criminal defense lawyer. That past experience has informed his time on the bench, where he has presided over a number of cases involving novel issues about criminal justice, our financial system, or, as it sometimes happens, both.

Judge Rakoff also teaches law at Columbia and trains other federal judges on the latest developments in forensic science in the courtroom. And he manages to still find the time to regularly contribute to the New York Review of Books, writing on such issues as access to the courts, the death penalty, mass incarcerations, and neuroscience.

The following are excerpts from an interview of Judge Rakoff by Ashish Joshi, which took place at the judge’s chambers in downtown Manhattan.

Ashish Joshi: Judge, you have written and presented a lot on the topic of wrongful convictions. How big is the problem? As a federal judge, what do you see?

Judge Rakoff: The only accurate answer is no one really knows. We have some very strong indications. The Innocence Project, for example, by now has exonerated over 350 people. Those were people who were convicted of very serious crimes, usually murder or rape. Perhaps 10 percent of them pleaded guilty even though they were actually innocent. On appeal, their convictions were affirmed, often by courts saying that the proof was not just beyond a reasonable doubt but was overwhelming. Yet, in fact, they were innocent. There is also something called the National Registry of Exonerations, which is a broader base that includes state and federal exonerations for the last 20 years or so. I think they’re now up to something like 1,800. So that’s a lot of people.


AJ: Did you find the problem to be systematic, systemic, or more prevalent in particular jurisdictions?

JR: While there were certain jurisdictions where it was more likely to happen than others, it’s primarily a function of systemic problems, some of which are not easily cured. For example, eyewitness identification. There are about 80,000 cases a year in the United States—state and federal—in which eyewitness identification testimony plays a major role. We now know, in a way that we didn’t know even 20 or 30 years ago, that often those identifications are inaccurate, even as many of them are accurate. Distinguishing between the two is not very easy because the inaccuracy often derives from subtle differences in the makeup of human beings, as opposed to improper police techniques, bad lighting, or things like that, that a layperson can recognize.

So the studies really have advanced our knowledge about how inaccurate eyewitness identifications can be, but they haven’t given us any great tools for figuring out, at the time of the testimony, whether a specific identification is or is not accurate. It’s worse, because these days very few cases go to trial, so you don’t even have whatever testing would be done through cross-examination, maybe Daubert-type hearings in advance of trial, and things like that. In the overwhelming majority of cases, nothing like that even occurs. The real place where those cases are resolved is in the prosecutor’s office through plea bargains. But it’s extremely difficult for both the prosecutor and the defense counsel to know at that stage just how good or bad the eyewitness identification really was.


AJ: Eyewitness identification makes for dramatic courtroom testimony. What exactly is wrong with it?

JR: First, we know that it carries a real risk of being inaccurate that’s almost never a question of lying. These are typically people who have no motive to lie, which makes their testimony all that more powerful. But, again, for example, of the 350-plus people who have been exonerated by the Innocence Project, 70 percent of those cases involved inaccurate eyewitness identification. If you look at the broader index of the National Registry, it’s closer to about 30 percent, but that’s still a very large number of people. So what are some of the things that we’ve learned? One is that it’s very difficult for people of one race to be good at perceiving the fine features of a face of someone of another race.


AJ: What’s known as the “other race effect”?

JR: Yes. Very well documented in the literature, though no one is quite sure whether its source is genetic or environmental; there are many different ways it is accounted for. What is agreed on is that, all other factors like lighting and so forth being equal, a person who sees a person of the same race commit a crime will have a much better memory of the facial features of that person than if the person is of a different race.


AJ: Interesting. Does it matter if the person who is doing the identification has been exposed to diversity or lives in a cosmopolitan area with many races?

JR: It does not appear to, but some of the studies that are being done now, still not completed, focus on very early childhood and whether people who grow up from infancy in a family with parents of different races or in a community that is racially balanced may not display this problem, which would suggest that it’s something learned at an early age or caused by other factors. We don’t yet know the answer to that.

Another example is the so-called merger factor. A victim or bystander witness first is shown a lineup or a photo array. These days it’s usually a photo array, because given the computer, they’re much easier to construct than finding seven more or less similar people to put in a live lineup. The victim or witness is asked, “Do you see the person who you saw commit the crime?” And if they’re properly instructed, they are also told, “The person may not be here at all. Don’t assume anything.” Some people will say, “I think it’s number three”; other people will say, “I’m quite sure it’s number three.” That difference—between “I think” and “I’m sure”—is largely a function of personality rather than anything else.


AJ: Not confidence?

JR: Some people, by their nature, are very diffident. Other people, by their nature, are very confident. My joke is that the people who are always, by their nature, very confident, we call federal judges.

But, either way, it will not be what it appears to an outsider to be—“This one’s not really sure. That one is very sure. So the second identification must be much better than the first identification.” It’s not that at all. Rather, it’s the function of personality.

That’s just the start. Let’s assume that the witness initially has sort of a fuzzy memory of the facial features of the person they saw commit the crime. If we had a way of looking into the witness’s brain at that moment in time just before they’re shown the photo array, we’d see that—a fuzzy memory. They only saw the culprit for a minute or two, he was holding a gun, and they were focused on the gun, or whatever—they have some memory, but it is not a really sharp memory.

Then the typically good witness looks over the array very carefully, all seven photos, studies them carefully. Finally, the witness says, “I think it’s number three.” OK. Now fast-forward to they are testifying to a grand jury or to many months later when they are testifying at trial, and they say, “I am very sure,” even if they previously said they thought so but they were not very sure.

A good prosecutor will ask, “Why is that?” And the witness will say, “Well, among other things, the guy had that very distinctive scar right over here on his—near his right temple.” What is actually being remembered is the scar seen on photograph number three from the initial photo array. There may not be any actual memory of a scar from the incident itself—we said that the original memory was too vague, too fuzzy—but what’s being remembered instead is seeing the scar in photo number three. Number three is the guy who was picked in the photo array. At the time of ultimately testifying, those two memories—the initial fuzzy memory and the later more vivid memory—have merged; not consciously, of course, but it happens nonetheless.


AJ: Can’t a good defense attorney pick that apart by asking, “When you were first interviewed by the police, there was no mention of the scar?”

JR: Sure. But it’s often things that are more subtle than a scar. I’m just using that as an example. The point is, if the defense lawyer asks, “Are you sure that you noticed the scar when you first saw the guy?” most witnesses will honestly answer, “Yes,” because—through no conscious act of their own—their memories have merged together into one memory and they think they’re remembering having noticed the scar initially.


AJ: They are unconsciously filling in the gaps?

JR: That’s right, too. That’s, in fact, often called the “fill in the gap” problem. So those are just a few examples. There are many others. And what’s most disturbing is we don’t really have a good way to correct for them. When an eyewitness identification might be the function of bad police practices—like when the police officer gives a hint (“Take a good look at number three”) or something similar—that’s something we actually can identify and correct. So, for example, the National Academy of Sciences report on eyewitness identification, in which I had the privilege of participating, recommends that the person who conducts the photo array or lineup be someone who knows nothing about the case, so that they can’t, even subconsciously, give a hint.

That’s an example of a correctable problem, but problems like filling in the gaps or the racial effect on memory, there is no straightforward fix, at least not that we know of. The most we can do is, as you said, foster cross-examination to try to bring out some of these problems, or put an expert on the stand to testify about some of these problems, or—as in New Jersey—have a court instruction that notifies the jury of some of these problems.

So far, none of those approaches seems to work very well. It’s all too soon to say with any definitiveness, but there have been some studies done, for example, of the New Jersey instructions, and usually the jury gets a different message from the one intended. According to the studies, the message they usually get is that “the judge doesn’t believe the eyewitness, so we better disregard them altogether.” In those cases, unless there’s evidence other than the eyewitness, the likely result is acquittal because the jury is getting the wrong message.

The judge is supposed to be telling them, “Don’t be too quick to assume the eyewitness is right, just because they have no motive to lie,” but the jury is getting a different subliminal message. In a case of experts, the early studies—and, again, these are very preliminary studies, so I don’t want to overstate them—indicate that where there are experts on both sides of this kind of stuff, they generally are seen as balancing each other off and the jury disregards them altogether.


AJ: How are the courts wrestling with this? If we were to look at the 1967 breakaway decision of the Supreme Court and compare that to the post-Daubert 21st-century world now, have the courts changed in their approach to eyewitness identification as a piece of evidence?

JR: Not nearly as much as they should, and it varies tremendously from jurisdiction to jurisdiction. New Jersey has been the leader in adjusting to the new science or at least in attempting to adjust to the new science. In the federal courts, it’s very slowly permeating down through the judges, but it hasn’t yet led to any major decisions in the assessment of eyewitness identification. The federal courts are less important here than the state courts. Mostly eyewitness cases are street crimes and things like that. They tend to wind up in state court.

There is one thing that happens in virtually every federal case. The government will ask the eyewitness, “Do you see, here in this courtroom, the person known to you as John Jones?” and of course they will have pointed him out—they will point right to the defendant. It’s not like a lineup; they’re being asked to identify only one person. In the normal case, there’s only one person sitting there right next to the defense counsel. It doesn’t take a rocket scientist to figure it out—it’s got to be that guy—so that’s really more theater rather than reality. It’s a little bit of a charade, but that’s a very small part of the problem.


AJ: What can the legal system do about this?

JR: I’m concerned that there isn’t as much as can be done in other areas. For other areas, where we’ve had a lot of revelations through science in the last years, as in the case for most forensic sciences, the solution is to devise a better science. For example, the one really good science in the forensic science area is DNA. Because DNA exposed so many errors, there has been a push, greater in some jurisdictions than in others, to tighten up the standards for other kinds of forensic science. In my view, some forensic science, like bite marks, should be eliminated altogether. I think that it’s not really science at all. But others, like fingerprints, there’s enough there and they can be made more scientifically reliable—they could be tested even more. In recent years, to its credit, the FBI has been doing more testing and more tightening when it comes to fingerprints.

The problem with eyewitness identification is that it doesn’t lend itself to that kind of improving, because you’re not dealing with scientific tests. You’re dealing with just an everyday person who happened to see whatever he or she saw. So it’s much more difficult to cure.

In my view, it is getting addressed ironically, in many cases, by the increasing existence of videotaping. So many crimes that I see that, in earlier days, would have been eyewitness identification cases do not even require an eyewitness identification. The camera was there at the back; the camera was there at the bodega; the camera was even there running when the guy ran down the street. So we can look not just for the 5 seconds or 15 seconds or 5 minutes that the eyewitness saw, but for the several hours that the tape was running. We can detect what features the person has, what they were wearing, and a hundred other things. Again, that doesn’t mean that mistakes can’t be made, but it’s much more reliable.

Of course, sometimes someone is wearing a mask, and there are still a lot of crimes in the home that occur that are not videotaped. But I think, ironically, that videotaping is increasingly replacing eyewitness identification as the key way of solving some of those crimes when there is no other kind of circumstantial evidence. The other thing that I think hasn’t been done nearly as much as it should be is to educate the police and prosecutors to some of the difficulties of eyewitness identification. In the Innocence Project, many of their cases involved situations when there initially were two suspects and, based on eyewitness identification, the police went after suspect number one and pretty much forgot about suspect number two. Then, perhaps 12 years later, the DNA test came along and, lo and behold, it turns out that suspect number two did the deed rather than suspect number one.

In many cases, particularly murder cases for which there is no statute of limitations, the prosecutors can go after the guy who actually did the crime, but in the meantime, poor suspect number one has spent 12 years in prison. That’s still the injustice. So, if the police and prosecutors were more aware of the limits of eyewitness identification, I think they would be better at pursuing what might appear to them to be secondary leads but which might really pan out in the end if they did pursue them right at the time of the crime.


AJ: Let’s talk about New Jersey. What’s the gist of the Henderson case [State v. Henderson (2011 N.J. Lexis 927)] as it pertains to this particular topic?

JR: The New Jersey court, grappling with the very issue that you and I have been discussing, determined that eyewitness identifications were more problematic than previously realized and were leading to inaccuracies and even wrongful convictions. They were hopeful that by having judges alert juries to some of the problems with eyewitness identification, they could prevent it from happening—prevent the wrongful convictions. So they devised patterned instructions that have been given in a number of cases by now.

First—I don’t want to be misunderstood—I thought that it was very laudable on the part of the New Jersey court. They were way ahead of virtually any other jurisdiction in both recognizing the problem and recognizing that we must do something about it.

There are at least two problems that evade that approach, however. First, most of the cases are resolved in the plea-bargaining arena, so educating the jury doesn’t even address that. You have to educate the prosecutor. You have to educate the police. In the plea-bargain setting, there is no jury involved. And, as you may know, the statistics show that over 95 percent of all criminal cases in the United States are resolved by plea bargain. Second, the preliminary studies show that even in those cases that do go to trial, the instructions tend to convey a message that is different from what the court intended to convey.

No one claims that these are final, complete, and definitive studies, but the early studies show that the New Jersey Supreme Court intended that those instructions would communicate to a juror, “We’re not saying that eyewitness identification is wrong or right. We’re just saying, don’t take it as an automatic. Consider some of the things that make eyewitness identification sometimes problematic, and see how they relate to this particular case.”

If that’s what the juries did, that would be great. That’s what they’re supposed to do. But what those early studies suggest is that the jurors are instead getting a very different subliminal message—“Why is the judge, who didn’t spend even 10 minutes on any other special issue in this case and just gave us general instructions on the elements of the crime, now giving us 10 minutes of specific instruction on eyewitness identification? It’s must be that he’s telling us that he doesn’t really trust the eyewitness.”

So the juries are drawing the wrong inference. Perhaps not one that is so surprising, given how jurors look to judges for guidance, but figuring out how to deal with that is still very problematic.


AJ: What about being liberal with the defense to allow them to present expert testimony on the pitfalls of eyewitness identification? Should the courts be more liberal in allowing the defense to present that evidence?

JR: Well, there are different aspects of what you’re talking about. One would be at the trial itself, where the defense could call an expert who could point out some of the limitations, and it would not have the same automatic force as a judge giving instructions, but it might, ironically, accomplish what the instructions were supposed to accomplish in alerting the jury to some of the pitfalls of eyewitness identification. There are problems with that: It’s expensive, many jurisdictions will not pay for such experts for indigent defendants, and there are a limited number of such experts.

Second, again, the preliminary studies suggest that what the defense does is matched by the government. They call an honest expert, who won’t deny there are some problems with eyewitness identification but says those problems really don’t apply here. Then the jury will typically just treat that as a wash and not do the hard work of trying to distinguish whether any particular factor might have been especially prominent in that case.

The third thing is that none of this is really well within the jury’s ken. These are all things that science has developed. Let’s go back to the merger effect or the filling in the blanks—those are not dynamics that people know as a matter of common sense or their everyday experience. Yes, jurors can be alerted to it, but how do they know how to apply it to the specific case before them? Not so easy.

And the final thing, of course, I keep harkening back, much as I don’t like this fact: It is a fact that 5 percent or fewer of cases go to trial. All the rest are done through plea bargaining. Now, I think a more helpful thing would be for the courts to have the equivalent of Daubert hearings.

AJ: Gatekeeping?

JR: Yes, before the guilty plea is finalized. In my own court, I try to identify cases involving not just these kinds of scientific issues but also other issues, and I sort of alert the lawyers early on that this may be a case where we’ll need to have a Daubert-type hearing before I can accept a guilty plea.


AJ: How does that go over with the defense attorneys?

JR: Well, you know what happens is the government then always comes back and offers a guilty plea to things that don’t involve the eyewitness identification, so they finesse it away.


AJ: But why is eyewitness identification even relevant at that point?

JR: Because the defendant often is saying that because he can’t face the huge penalties that are imposed when going to trial. This goes back to the question of why do innocent people plead guilty? We know that happens. We know that happens with some frequency, though no one can say what the percentage is. But even if it’s only 5 percent—even if it’s only 2 or 3 percent—we’re talking about tens of thousands of people who’ve pleaded guilty to crimes they did not commit. Why does that happen? I think, by far, the biggest reason for that is because defendants face a much greater penalty, much greater punishment, if they go trial and are convicted than if they don’t.

Take your everyday federal case involving a drug conspiracy. These typically would involve maybe 10, 12, 14 defendants, all in one indictment because the government is going after the entire conspiracy. Under mandatory minimum legislation, if the defendants go to trial and are convicted of that conspiracy, they would face, depending on the amount of drugs, 5, 10, even 20 mandatory minimum years. Plus, under the guidelines in drug cases, 70 percent of the calculation is the weight of the drugs, and now you’re looking at the weight of the entire conspiracy.

AJ: Also, if weapons were involved . . .

JR: Well, weapons would add still another major mandatory minimum; but just on the weight of the drugs, you’re not looking at what any given defendant distributed. You’re looking at what the entire conspiracy comprised. That’s always way up there—40 or 50 years—and if you go to trial, that’s what you’ll be tried on. So, if you’re a defense lawyer, you go to the prosecutor and say, “Mike, I was at a pretty low level. Why are you hitting on me?” And the prosecutor says, “You know, I have some sympathy for you. I’ll tell you what, I will let you plead guilty to a superseding information, a superseding charge, where we’ll narrow the conspiracy as to him to the part that he was really personally involved in.” And, suddenly, instead of the 20-year mandatory minimum, it’ll be a 5-year mandatory minimum; and instead of the 40-years guideline, it’ll be a 6-year or 7-year guideline.

So you go back to your client—and I was a criminal defense lawyer for 15 years, so I have some feel for this—every client will tell you initially they’re innocent. And notwithstanding the attorney-client privilege, which is designed to get out the truth, eventually it does, but it takes a while. You have to build trust with your clients or whatever it all comes down to. So your client tells you. Now you come back—two weeks, three weeks later—and say, “You’ve got a 5-year play; but if you’re innocent, Mr. Client, I of course can’t let you plead guilty. We’ll go to trial.” Now, there is such a thing as an Alford plea, but in the federal government, it’s not recognized at all as a matter of practice. In some states it is, but usually you don’t have that option.


AJ: You’re referring to maintaining factual innocence while pleading guilty?

JR: Yes. But that’s to say it varies tremendously from jurisdiction to jurisdiction, so typically you don’t have that option. You have your client, who may already have a number of priors and who may have a cynical attitude toward the whole process. Even if, in fact, this time he’s innocent, he nevertheless may well say to you, “OK, what do I have to say to satisfy the court that I’m guilty?” And what he’s really saying is, innocent or guilty, this is the deal that I want. And then you will craft a plea allocution that will be as minimal as possible, which is good for other reasons as well from a defense standpoint because you want to give yourself maximum leeway at sentencing to say what mitigating factors there may be. And most judges will not inquire very much.

Partly because of all this stuff that I’ve become aware of, I do inquire more. But it’s very rare. I have had it happen. I have had two cases now in which defendants, who were prepared to plead guilty when I really started questioning, said that they were innocent.


AJ: What happened?

JR: What happened in those cases? The government went back and offered them a better plea bargain and they took it.

AJ: There are a lot of jurisdictions in this world. Many of those jurisdictions look at our plea bargains as “devil’s pacts”—an unholy bargain—and they’re astounded about the amount and the magnitude of power that state and federal prosecutors in America have. You’ve been a defense lawyer, a federal prosecutor, and a federal judge. What’s your view of the plea bargain? Should we look at it with a fresh perspective?

JR: I don’t think we should ban it. There are positive aspects, and the system really could not operate if we banned it altogether. India has a huge backlog in its legal system, and that’s not a good thing; it’s gotten out of hand. Because of the high risks of going to trial, almost no one goes to trial. The real sentence is negotiated with the prosecutor in secret in the prosecutor’s office.

AJ: So where are the checks and balances?

JR: They’re not operating as well as they should be. The solution is not very difficult to see. We should eliminate mandatory minimums. We should eliminate excessively high penalties that we impose through the guidelines. The huge problem that is connected to, and dwarfs, everything that we’ve been talking about is mass incarceration. Even though the U.S. crime rate has steadily declined since 1996, we continue to incarcerate over two million people a year and that has only slightly declined. Why? Because we have these laws where everyone goes away for a long time and where you can’t take the risk of going to trial because if you do and you lose, you’ll go away for an even much longer time.

So it is that draconian, punitive mandatory sentencing legislation that has given the prosecutors the ability to turn a reasonable thing—a plea bargain—into a weapon that could be used abusively, even to convict innocent people.

Now, I want to hesitate—I want to make clear that I don’t think the prosecutor thinks he or she is convicting an innocent person. But, typically, they’ve heard only one side. The agent has come to them and said, “We’ve got two eyewitnesses.” The prosecutor hasn’t done the analysis and may not even know about the problems with eyewitness testimony that we’ve just discussed, much of which I know only because I was on that committee. So he says to the defense lawyer, “Your guy’s dead in the water.” Then it turns out, 20 years later, the DNA test shows that the guy was innocent, but he spent 20 years in prison in the meantime.

The source of the problem is not plea bargaining per se, but the excessive penalties that have allowed it to become too much of a weapon as opposed to simply a useful tool. The statistics are that for most of the 20th century, somewhere between 75 percent and 80 percent of all criminal cases were resolved through plea bargains. That’s not a modest amount. It’s a substantial amount. But that still means that 20–25 percent of those cases went to trial, and the prosecutor never knew, at the beginning at least, which cases were going to plea and which were going to go to trial, so he or she had to assume each case would go to trial.

That in and of itself kept the system more honest. Plus, we had enough trials to put the system to the test, so to speak. But now those things no longer are the case. We can trace with near-perfect correlation the increase in the penalties beginning in the 1970s with mandatory minimums, career offender statutes, and things like that, and the decrease in the number of cases going to trial, so that now, instead of 20–25 percent going to trial in the state systems, it’s about 5 percent; and in the federal system, it’s at about 3 percent. That’s a huge decline.

AJ: Do you think that when there were more trials, it also caused the prosecution to pause before overreaching in their charging decisions?

JR: That’s a fair inference, because they never knew in any given case whether it might be the one in which they’d get challenged. They had to be really sure that they had their ducks in a row, their proof strong. An interesting statistic that I don’t have would be whether there were more cases dismissed by the prosecutor before those laws came into effect or afterward, on the theory that they used to have to look at it more carefully and therefore often would find then that “yeah, we really don’t have the proof and, therefore, we dismiss.”

One of the reasons I would not eliminate plea bargaining as a whole, aside from the fact that as a whole the system would not be able to deal with the volume of cases, is that, for better or worse, we have a lot of crime in the United States even with the decreases, and it would be very hard to manage that. We would need a lot more judges. We would need a lot more prosecutors. We would need a lot more defense counsel.

But there’s an additional benefit to plea bargaining, which is to get cooperation. That’s the way you get the bigger fish, so to speak. A good example of this is Brazil. I know about it only because I had a case in my own court involving Brazil. So Brazil has just gone through what they called “Operation Carwash,” which was a huge corruption scandal that already has led to the conviction of two former presidents of Brazil and over 200 people altogether, and has exposed incredibly widespread corruption in the Brazilian government and in Brazilian business. It was the subject of a parallel class action in my court that settled about a year ago for $3 billion, called In re Petrobras. One of the things I learned in our discussions in that case was that Brazil had a military dictatorship for a time and then it became a democracy again.

In 1988, Brazil passed one or two laws that are very important. The first law, which was part of their constitution, provides for merit selection of judges and prosecutors. There’s still, at the very highest level, some political aspects to it, but it’s much more of a merit system than it used to be. That freed up the judges and prosecutors to be more independent.

The second development, which didn’t occur until 2013, was they had their first laws authorizing plea bargains that granted leniency in return for cooperation, and that proved to be critical to Operation Carwash. That began, as the name suggests, as a small little money laundering case involving a carwash in a small city in Brazil where the carwash was the front and, behind it, there was some money laundering. The now independent prosecutors and now independent judges, by making use of the cooperation agreements, were able to push it up all the way to the very highest levels of Brazil.

The point is, plea bargains are not without their benefits, too. It’s a question of degree.

AJ: I was looking at some of the picture frames in your waiting area, and I was pleased to see a couple of prints by Honoré Daumier in France, which led me to think, have you ever studied the civil law systems and how they have treated eyewitness identification as a piece of evidence?

JR: I have not; and you raise a very good point. I really should. The civil law system, first of all, doesn’t involve juries except sometimes. In France, for example, murder cases aren’t tried by a jury. There are three judges who sit along with the laypeople. So it’s a very different system. But all the other crimes are just tried by the judges. Secondly, to an incredible extent, the judges play a quasi-prosecutorial role.

AJ: Right. A much more powerful role than we see here.

JR: Yes. Yes. From an American perspective, that looks like a conflict of interest. But it is ingrained in the civil law system. And the third thing is that when it comes to trial, a great deal of the trial is done on paper statements, on affidavits and other statements—typically; not always, but typically. The witnesses have to be available to be questioned if the judge wants to question them, but they submitted a statement and that’s usually all that happens. So it’s a very different system in a hundred ways. But how do they deal with eyewitness identification? I don’t know that answer.

AJ: What about judges? What do you think of science education for judges?

JR: I teach a course at both Columbia and NYU Law Schools on science in the courts. One of the themes is that you can see from their opinions how little judges know about science. I always identify myself as the number one example, because I was an English major and, as did most judges, I took only one science course in college—the required freshman course—which in my case was known among the students as “Physics for Poets.” So I was totally ignorant of the great variety of scientific issues that come into the courts in any of a hundred ways.

AJ: It’s ironic that after Daubert the judges are the gatekeepers.

JR: Daubert is one of the really great decisions. There was an awful lot of junk science coming into the courts before that. But, of course, Daubert is only as good as each judge who takes the time to really get up to speed. We may not see judges arrive on the bench with lots of scientific knowledge, other than perhaps on the federal circuit dealing with patent law. But in the American system, judges often are former trial attorneys. One thing almost all trial attorneys have in common is being quick studies—one has to be, as a trial attorney. So judges, if they take the time, really can get themselves up to speed.

While I’m not certain how to deal with this, if I had a magic wand, I would say to my fellow judges, “I know that when you have a case that involves a scientific issue and a technical legal issue, your natural instinct may be to see if you can resolve the case on the technical legal issue, but you’re not really advancing the law as well as could be done if you would take the time to address the scientific issue. That technical legal issue may never come up again; and even if it does, it doesn’t really get to the merits. The scientific issue is much closer to the merits of the case. If you can advance how judges think about science on any given particular issue, you will be doing well, and it’s a great service.

It’s much easier for a federal judge to say this than for a state judge to say this, because our dockets are so much lower. I’ve had Daubert hearings that have gone on as long as two weeks. A state judge, even if he or she wanted to, probably couldn’t do that, because their case load is so huge that they must keep the system rolling. So it’s easy for me to preach this. And—I don’t want to be misunderstood—I don’t think that judges are consciously saying, “I’m going to ignore my obligations as gatekeeper” or anything like that. It’s just that, because it’s a new language to them—a new area to them—it’s not as easy. If they can, they often say, “This was a deficient pleading; I’ll knock out the case on that.” Of course, then there’s an amended pleading, and before you know it, the case goes on for another six months or more. Often, I think, the law could be greater advanced by addressing the scientific issues directly.

We see that happen more in the civil cases than in the criminal cases. That’s because generally the parties can afford meaningful scientific experts. Another whole aspect of this is that we don’t see anywhere near incisive Daubert opinions written in the criminal area as we see in the civil area, even though the stakes are higher, so to speak, because it’s someone’s liberty. The first reason is that, as a group, defense lawyers—as with most lawyers—just don’t know that much about science. Second, not every criminal defendant can afford a really good scientific expert. For indigent defendants, some states will pay for experts, but other states will not. Lacking quality expert support, defense lawyers will, by necessity, focus on other issues instead.

Occasionally, judges may question the strength or validity of the government’s experts even when the defense hasn’t directly done so. The prosecutors tend to get defensive when their forensic science is criticized. “Well,” they’ll say, “we’ve always done it this way. I can’t believe you’re attacking it. I didn’t hear any defense counsel making that argument. Who are you, the National Academy of Sciences?”

That’s a shame, when it happens. Daubert was—and is—a great decision. It provides a great tool for judges to really separate good science from bad science. That would be immensely helpful in the criminal justice area, but it hasn’t always happened that much.

AJ: What I see in state settings is not enough Daubert hearings. Many times, the defense doesn’t even pose the challenge. And state judges often don’t see Daubert gatekeeping as a mandatory, proactive function; rather, as one to be brought up only if a party objects. That’s unfortunate.

JR: I agree; I totally agree. And you have to remember, state judges mostly are elected—most are former prosecutors—and while I’m not suggesting there is a conscious bias, there may be a tilt or inclination there, especially if the defense counsel is not raising a really serious scientific challenge that impresses them. Civil cases don’t present the same stakes—it’s not someone’s liberty at stake; and it’s not going to affect the election one way or the other, if they decide for one civil litigant versus another civil litigant.

Ashish Joshi

The author is with Joshi: Attorneys + Counselors, Ann Arbor, Michigan, and is a senior editor of Litigation.

Copyright © 2019, American Bar Association. All rights reserved. This information or any portion thereof may not be copied or disseminated in any form or by any means or downloaded or stored in an electronic database or retrieval system without the express written consent of the American Bar Association. The views expressed in this article are those of the author(s) and do not necessarily reflect the positions or policies of the American Bar Association, the Section of Litigation, this committee, or the employer(s) of the author(s).