… Before you got involved in this study for the National Academy of Sciences, how did you think about or conceptualize forensic science?

I think I, and many of my colleagues, assumed that the forensic disciplines were based on solid scientific methodology, were valid and reliable. I don’t think that we assumed that there was anything seriously amiss.

We assumed there might be mistakes, but I don’t think that we had been forewarned in any way that there were the serious problems that the committee uncovered. …

When you saw forensic science, did you have any reason to question it?

No.

You just assumed it was —

You assumed that the methodology was valid and reliable and that the work in putting the evidence together and in offering the testimony was proper.

That was your assumption. There could be mistakes, and you understood there could be challenges, but you didn’t assume what we later on uncovered, which was that there was systemic, serious problems with respect to certain of the disciplines.

Systemic, serious problems. Explain.

Among the biggest problems that we uncovered in the report is the lack of scientific research to support the disciplines; that is, the absence of the application of scientific methodology to determine whether or not the discipline was valid and reliable, as was done with DNA.

DNA is the principal example of real science at work. … DNA is really the only discipline among the forensic disciplines that consistently produces results that you can rely on with a fair level of confidence when you’re seeking to determine whether or not a piece of evidence is connected with a particular source.

The kind of work that was done by scientists to put the DNA discipline together had not been done in a number of the other disciplines.

… Five years ago, I would’ve thought … that fiber evidence, prints, all the various things we see on television, those are valid procedures that have a basis in fact.

It depends on which discipline you’re talking about, before you reach any conclusions.

Fingerprint evidence has qualities that are perfectly fine, for example. … There’s a lot about fingerprint evidence when you’ve got decent prints — you’re not talking about latent prints, that has some difficulties — that if you have a good practitioner, it works perfectly fine in most cases.

But the cases we’re [talking] about is where we have partial prints and smudge prints. And when we’re told that they’ve not yet done the scientific research to confirm that their testimony is valid and reliable, that’s problematic. …

You mean that you discovered that they, in fact, openly admitted that they had no testing or basis for what they were saying?

They claimed that they had a methodology they’d been following for years that they thought was consistent and reliable. But when pressed on that point, when you come back to the hard cases, the blurry print, the partial print —

Real life.

Real life, the tough latent print cases. It’s a very different proposition to say: Here are the scientific studies that we’ve used to show you that we have some real level of confidence in the testimony that we’re going to give with respect to these prints — because here’s the science, here’s the research that we’ve done to confirm that we’re accurate in what we say with respect to a partial print.

Because after all, a partial print is partial, so you’re not in a situation where you can comfortably say: I have a clear picture of a print and I have a clear picture of the suspect’s print, and a good practitioner can match them up.

It’s a lot harder when you don’t have that, and so it’s a terrible problem when you don’t know whether or not what you’re saying is accurate. The bigger problem was the courts had been misled for a long time, because we had been told, my colleagues and I … had been told by some experts from the FBI that fingerprint comparison involved essentially a zero-error rate.

The report says that’s nonsense. There’s no such thing as a zero-error rate in science, and there certainly was no such thing as a zero-error rate with respect to partial and smudged prints.

The other thing we realized was that there hadn’t been any serious studies. I’m told that the FBI is now engaging in this research … trying to quantify the likelihood of mistakes, given certain types of prints. And they’re trying to look at the error rates to see whether or not they can come up with data that they can use in the future, both for training and to ensure that the testimony that they give is good. …

One of the problems was the experts came into court and were frequently testifying that there was a match between the evidence and a source. And that’s not accurate. The most they could say is something like: Given the evidence that we have, we think the likelihood of there being a match is –, and then fill in the blanks. And we’re finally beginning to get some change with respect to how people testify.

… How can they testify about the likelihood if they haven’t done the studies?

That is the problem. Your testimony should be even more modest until you’ve done the studies.

Modest?

Careful. You’ve got to remember there are some situations where you’re testifying to a match that is evidence source, and there’s some testimony which has to do with classes of evidence.

You could say, for example, “I think this red paint came from a Toyota.” Well, there are lots and lots of Toyotas; that’s useful information to have at a trial. But that’s very different from saying, “This red paint came from Joe’s Toyota.” That’s what we were trying to get.

They routinely testify, and apparently some still do, that it’s this person’s fingerprints to the exclusion of all others.

Yes.

Or it’s a 100 percent match.

Right.

As the FBI wrote until recently in their own manuals and pamphlets, fingerprint analysis is infallible.

That’s what the courts have been essentially told when I look back at the cases. You had judges who were absolutely naive and ill-informed writing in these opinions.

The first case in which it came up was the FBI expert told us, and we wrote it — not I personally, but judges in another circuit — said zero-error rate. Then another circuit court picked up the same information and said it’s a zero-error rate when you talk about fingerprint comparisons.

That’s repeated over and over again without us ever understanding that’s completely inaccurate, and you don’t have the science yet to support it. And the claim is invalid with or without the science because a zero-error rate makes no sense.

It’s impossible.

In science, it makes no sense.

Are you telling me that the federal appellate courts of the United States of America were accepting the idea of a zero-error rate and the infallibility of fingerprint matching on faith?

We thought that an expert [who] testified in the trial court didn’t discredit the testimony. And so the court of appeals, which had limited review, accepted it as part of [the] record, and said the record shows that there’s a zero-error rate. That’s essentially what was happening.

When you get to the appellate court, the standard of review is limited. The trial courts had been led to believe that these experts were basing their testimony on good science.

We recently interviewed a well-known fingerprint expert, not in the FBI, and we said, “What about all this?” And he said: The problem is you can’t really do a mathematical model for this kind of evidence. It’s too complicated.

Fortunately, there are a couple of people trying now. The National Institute of Justice finally gave some money to [UCLA School of Law] professor [Jennifer] Mnookin and … [Itiel] Dror, who are now trying to do exactly that — that is, to take a body of information and get some data on fingerprints and see whether or not they can come up with error rates that will be useful.

I’m really hoping that they’re successful. They ran into some resistance initially in just trying to get the data. But if they can do the work, we’ll have that information and it will be immensely useful.

The FBI … says that they no longer testify that fingerprint analysis is infallible, or that it’s a percent, or that it’s to the exclusion of all others. Is that good enough?

It’s a step in the right direction. … You’re at this curious stage now where the problems have been exposed, and now it’s going to take time to do the research. So what do you do in the interim? Do you throw it all out and say you’re done? No, that’s not likely to happen. …

Now, at least with respect to certain disciplines, we’re trying to do it better and eventually get to a place where we have, we hope, scientific methodology undergirding a number of the disciplines. But we’re not there yet.

We’re getting there. The studies can be done. It takes time and money, but they can be done, and you have to have a willingness and a commitment. The president’s budget, I’m told, that was just submitted to Congress proposed the $5 million allocation to NIST, the National Institute of Standards and Technology, to do some serious standards research in the forensic discipline.

It’s a small step, but it’s a step forward.

I don’t want to overstate it, but you’re saying that there basically have not been hard standards when it comes to forensic sciences.

You have to look at the discipline, so I don’t want to overstate it either. It varies from discipline to discipline.

DNA.

DNA is very solid analysis.

Meaning testing for chemicals.

Yes, and that’s traditional science. There’s not confusion there, and you have the great research institutions that have been working in those areas a long time. So there’s no surprise that DNA, which was well supported nationally, and drug analysis, which is the work of some of the major universities, is good.

Fingerprints?

Fingerprints is good in a lot of respects. It’s just that we need the research to show that when you have the hard cases, you want to know what the likely error is. What kind of confidence should we have in your assessment? And we want to know that you’re testifying honestly.

We also want to know that the practitioners who are doing the analysis understand the limitations of their work. That’s a big problem. If you don’t have people who’ve been well-trained and certified, then they don’t even understand the limits within their work.

But if you have people who are well-trained and have been certified, and they’re working pursuant to good performance standards, they understand they have a hard case. They understand that when it goes to the prosecutor, they should say: “Here’s the most we can say. We cannot say this is a match. We cannot say we know that it’s 100 percent certain that this fingerprint matches that person.”

So it’s very helpful that the FBI and others are backing away from that.

But the National District Attorneys Association and some of the more prominent fingerprint examiners that we’ve gone to, they’re not really backing away from that. They’re saying, “We’re fine, and fingerprints have worked fine in court for decades, if not a century.”

Working fine in court is in the eye of the beholder. …

What they’re saying is, “Look, there’s an error rate in everything, but this works, and if there are any errors, it’s an anomaly.”

… If some people are saying it works because we’ve gotten convictions, that is to say nothing more than juries and judges have believed that experts knew what they were talking about. And so they bought it and they convicted. That’s not proof that the discipline is undergirded by serious science.

For somebody out there watching who’s not deeply into this, are you saying that when experts have testified that this is a 100 percent match to the exclusion of all others, they didn’t know what they were talking about?

I’m telling you that the expert who testified before the commission, before my committee [National Academy of Sciences National Research Council Committee on Identifying the Needs of the Forensic Science Community], when asked the question, “If you’ve got a blurred print or a partial print, do you have any scientific basis for confirming an expert’s testimony that it’s a match?”, he said, “The research has yet to be done.” That’s what I’m saying. They have not had research to support that claim.

Well, they say it’s based on practice, that it’s based on years of experience. That’s valid, isn’t it?

No, of course not. If your experience or practice [has] been inaccurate and wrong for many years, it doesn’t become better because it’s many years. It’s just many years of doing it incorrectly. …

The stakeholders, they have an interest in their work. No one likes to be told that what you’ve been doing for a long time is questionable, and we want to take a look at it.

That’s part of what the pushback has been. People are resisting. The folks who’ve been doing the fingerprint work, I can understand why they would push back even more, because there’s much of what they do that’s pretty good. …

The thing I’ve never understood, though, is good science serves both prosecution and defense in criminal trials. We’re confirming that convictions are accurate where they ought to be. Or we’re confirming that we have the wrong person on trial. That’s a good thing for a prosecutor as well as the defense. It serves society well. …

You don’t want wrongful convictions for two reasons. The stakes are too high. You don’t want to put the wrong person in jail or, much less, execute the wrong person. And you don’t want to leave the person who should have been tried, out as a threat to society. It cuts both ways.

My hope is that all of the stakeholders will get behind it and say we have to have the research to support what we’re doing. Let’s do it.

And it’s not just limited to fingerprints?

No. … For example, there was a report on bite marks recently, … saying there’s just no science to support a lot of what the so-called experts have been assuming with respect to the validity and reliability of bite marks. …

So bite marks are a problem.

I’m going by what the report has. … We identified bite marks as an area that needs some research.

Because there have been some problems.

Yes.

Hair matching.

The second big moment for me during our committee deliberations was when I was listening to testimony from a very serious, more thoughtful expert about hair analysis.

I listened, and I said: “Let me ask you [it] this way: Let’s assume your teenage child was on trial, and the only significant evidence was testimony from a so-called expert who says there is a match between your child’s hair and the hair that we’ve found. How would you feel about that?”

He said, “I would be horrified, because in the absence of mitochondrial DNA to back [it] up, that’s not good evidence.”

That’s not good enough. It might be good enough to show classification. In other words, this is brown hair; it came from someone who has brown hair, or something like that.

But to say a match? This is an expert in the field, and I’m sitting there listening, and he said, “No, I would be really distressed if my child was subject to that kind of testimony and that was going to determine whether the child was convicted or acquitted.”

They wanted DNA backup.

They wanted the mitochondrial DNA to back it up. …

I think there have been enough acquittals, post-conviction releases in hair cases, and there has been enough science now to show that if you’re not backing it up, you should not ever use hair. And the report says that you should not use microscopic hair analysis alone to claim a match.

Fibers?

There are questions with fibers as the report says.

And we could go down the list.

Yes. We identified fibers, tool marks, the questions raised about handwriting; we need more confirmation there. Debris and fire.

Arson.

Yes, arson investigations.

That’s not to say there is nothing there, especially when you’re talking about class evidence as opposed to matching evidence.

[What] we are saying [is] with so many of these disciplines, you haven’t pointed us to the serious research undergirding your claims. That’s all we’re asking for. …

Is there any estimate how people have been testifying about this, what manner, for how many years?

No. … There’s no way for us to estimate. I don’t like the risk because you’re sending people to jail. …

Because if you send the wrong person to jail, then you stop looking.

You stop looking for the right person. Exactly right. …

We care deeply about how our justice system works. We don’t want bad evidence, and we don’t want ill-trained, uncertified experts appearing in court and testifying to things that they’re really not sure about.

… Should fingerprint evidence be allowed in court?

Yes.

But the testimony should be qualified.

Yes. The committee was not prepared to say you toss everything out until we’re done with all the research. That’s not a viable position.

Some people have told us that the reason the National Academy study happened, or a major reason was the Brandon Mayfield [fingerprint misidentification] case. What was its impact?

The Brandon Mayfield case got a lot of play, but you know what’s interesting — the impetus for the creation of the committee. The congressional legislation came from the forensic science communities.

A consortium of forensic science organizations went to Congress and said, in essence, “We need help.” They understood better than most of us who ended up on the committee how serious their problems were. …

It’s a very fragmented community. There’s no serious oversight. People often look to the FBI as a kind of model that’s pervasive. It’s not a pervasive model. The FBI lab is special. It has tons of money and it lives in rarefied air. [It does] things that most labs can’t begin to do.

They have training that’s very different. They have quality standards that are very different. That’s why you see [that] a lot of laboratories have problems across the country, and that’s what we saw.

There were accreditation programs. Not all labs are accredited. There are some, but a lot of labs have failed. You have problems in North Carolina, Houston, Detroit, serious problems.

So we’re not talking about isolated problems. … Overall, [the] forensic science community goes to Congress. Are they saying they have a systemic problem?

They’re saying they have a large problem: the fragmented nature of their community, the lack of certification, no mandatory accreditation, even though there are accreditation programs, no oversight. …

There are inconsistent practices.

There are inconsistent practices and inconsistent standards. Labs in one set of standards that are not followed by labs in another place.

And one of the biggest problems that the forensic people pointed out to us in the committee is the too-close connections between law enforcement and the … forensic disciplines, and that was news to most of us.

What they explained to us was in most instances, the labs are controlled by either the police chief, the prosecutor, law enforcement, so to speak. And the forensic people said it’s a mistake because there’s too much pressure put on the forensic people to deliver results and get them done quickly, and introduce the possibility of contextual bias for example. We’re trying to study that. …

The court is set up to allow for adversarial proceedings, where the defense has an ample chance to challenge the evidence and the scientific basis for forensic science itself.

So here’s the committee’s view on that: It’s a myth to suggest that the difficulties that we’re facing with the forensic disciplines are going to be corrected by the court system.

The judges don’t have the expertise to do it. … You have defense attorneys who are often strapped, because of lack of resources, to be able to pursue the questions that they ought to be pursuing. … We don’t even know the right questions to ask.

That is what’s been going [on] for years. It’s not a lack of willingness; it’s that most of us are ill-informed. We’ve been led to believe, for example, there’s a zero-error rate for fingerprints, which was not accurate. We bought it.

Now you could say to me how could you all [have] been such dummies? That just makes the point. We assume that the people who are coming to us from the scientific disciplines know their business, and their disciplines are undergirded by the scientific methodology. …

Just because they have a lab coat on.

It doesn’t mean if you give us the report, we’re not going to accept it at face value. A defendant has the right, if he or she chooses, to bring the person in who prepared that report and question him [about] what led to that. …

The American Bar Association just came out — the House of Delegates approved recommendations suggesting that trial judges should get better hold of their trials on expert testimony, on confirming that jurors really understand forensic disciplines and the evidence that’s being presented to them. …

What the ABA is saying is we haven’t been doing it correctly in the adversarial system. It’s an adversarial system, that’s the problem. …

In an adversarial system, once you decide to go to trial, your interest is in prevailing. So you’re not looking to make it easier for the other side. You’re not going to find scientific truth in the adversarial process. That work has got to be done by the scientists. …

… The National District Attorneys Association’s director said: “Nobody’s ever asked me about the hundreds of thousands of cases where it does work, where good forensic scientists testify. We just come up with these bizarre examples and say that represents the system, and that’s not true. We get it right most of the time.”

I don’t know how you measure “we get it right” unless you’ve done the research that can satisfy those of us who are looking at you. …

Another way to look at it: We’ve gotten away with a system for many years, and [in] a lot of those instances the results may have been correct. Very different than saying we know what we’re doing in the hard cases, and all of our disciplines are based on science. …

Do we go back and look at all these cases?

Not necessarily, because there are a lot of cases that are based on more than just forensic evidence.

That has been [a] worry and caused some prosecutors [to be] hesitant to step forward and say, “Yes, we should do the research.” Because I think what they’re thinking is it opens Pandora’s box, and that means every case gets reopened. I don’t think that’s the natural consequence. …

But we don’t know today how many cases there are out there where people have been convicted or allowed to walk based purely on forensic evidence.

I don’t think anybody could tell us that number, no. And I think they’re probably right in saying that in a lot of instances, the result was correct, but for me, that’s not [an] answer. …

You said, and the report said, that there’s a problem because there should be mandatory certification. What is there now?

There isn’t mandatory certification now. The training is very limited, and that’s one of the things the forensic people called on Congress for help. …

You don’t have any of the great research institutions, the universities, who’ve been seriously interested in forensic science. So they haven’t geared themselves up to think about the issues that we’re talking about. That’s most unfortunate. …

You have some schools who’ve introduced programs, but it’s a small number, relatively speaking. You have certification programs, most of them not mandatory, but there’s not follow-up; there’s no enforcement.

For example, I can be disbarred if I do certain things, if I violate certain ethical rules of the profession. You don’t have that kind of a system in the forensic sciences.

We find that you can get certified as a forensic expert of various kinds online.

So I’m told. …

You’re familiar with the research of Dr. Dror?

Yes, I am.

Amazing that fingerprint experts who apparently said they were 100 percent certain found different results on the same prints.

Same print. If you change the circumstances.

Explain.

For example, if … they said that it was a 100 percent match, and then … you give the same prints to those same fingerprint experts sometime later without their knowing it, and you put in a setting where the prosecutor says, “We’ve got someone here who has, we think, been involved in child rape.” I don’t know what his examples were, I’m trying to give —

He gives a different context.

Right, whatever the context is. You create the context. And now that examiner unknowingly is looking at the same print that he or she previously said [is a] 100 percent match, and in a significant percentage of those cases, they changed.

So it’s not infallible.

It’s not infallible, and contextual bias is terribly important.

I heard people testify from the disciplines saying, “Oh, that’s nonsense.” It’s not nonsense. Contextual bias is important in my work. In all of our work, we need to understand how we might be affected by factors that should not directly bear on what we’ve been asked to do. We shouldn’t run from that. We shouldn’t be afraid of someone saying, “Have you thought about this?” …

… Are there national standards for who’s an expert?

… You can join associations, and as I understand, you pay your dues and you swear to follow whatever their rules are. But again, there’s no oversight.

No policing?

There’s no policing.

Hard to get thrown out?

At least as I understand it, yes.

We couldn’t find anyone who tracks the track record, for instance, of an expert.

… That’s one of the things that I inquired about. … And I’m not sure that I ever understood the answer.

There’s no mandatory accreditation for labs. That’s voluntary as I understand it.

That is voluntary, although I think some people say a lot of the labs are accredited. …

The important question with the lab is: What are their own internal performance standards and who are they hiring? And then what does the accrediting agency do when they come in? …

… Justice [Antonin] Scalia, possibly the most conservative [Supreme Court] justice, … [quoted the NAS study in a 2009 opinion] [PDF]: “The forensic science system, encompassing both research and practice, has serious problems that can only be addressed by a national commitment to overhaul the current structure that supports the forensic science community in this country.” Is this getting bipartisan attention?

That’s what we were hoping for. That was our recommendation, … we called it NIFS, the National Institute of Forensic Science, because we felt you had to have national oversight. The disciplines are too fragmented.

We felt that the only way you could really make some progress is if you had a national organization that was responsible. Or set some standards for certification and accreditation and training and performance standards and the like. Essentially a watchdog, with funds to back it up, to encourage the states to make the kinds of changes that we’re talking about. …

… You’re talking about the law enforcement agencies saying, “We may have been doing it wrong, and we have to change, and by the way, nobody’s going to pay for that change.”

There’s another way for law enforcement … to think about it. You don’t have to say we’ve been doing it wrong. What you should say is, there are ways for us to do it better. … We did the best we could, and science evolves. … We’re never at the final point in science; it continues to evolve. …

The prosecutors say: “We weren’t on that commission. We didn’t have anybody sitting there.”

They had [the] opportunity to offer whatever they want. There were people that were from the prosecution community. There was a member of our committee who was a former U.S. attorney. The prosecutors were not excluded.

The coroners say they were not on the commission.

There was a coroner who testified.

What they’re saying is, this was the defense bar combined with some experts who want a change.

That’s just not accurate. The committee was a cross representation of all of these groups: lawyers, prosecution, defense, scientists, people from the forensic discipline policy. And anyone who had information to offer was allowed to testify and could submit to us anything that was useful.

We told every discipline: Wherever you think there is research, refer it to us. We looked at everything, and that’s why we were quite comfortable in saying when we did, there was a paucity of research, because we looked.

How realistic have you been about this? You want to abolish coroners, who are local elected officials, who have political weight themselves, and in some states dominate the investigation of the cause and manner of death.

I don’t think we’re politically naive. I think we understood that in recommending that there be a National Institute of Forensic Science, that it was going to be a huge political hurdle to jump, especially when we went into an economy that hit a downturn.

But I don’t think it was naive for the committee to say to society: This is a really serious problem that we as a society have to think about. … I totally reject the suggestion that this was pro-defense. This was pro-American system of justice. That is, do it right. That’s what we’re supposed to do.

… One of the surprises to us … was to come in and find out that an expert is who[ever] a judge says is an expert.

There’s supposed to be standards within communities of experts that a judge should be able to look to [in order] to be able to make the determination as to whether this person is qualified.

You may not always agree, but you have to understand, no one was telling society that a number of the forensic disciplines were fragile insofar as the —

Reliability, right?

They were fragile. They didn’t have the research to back up, but they were the ones who were explaining who they were.

So maybe some judges got misled, just like we got misled when we got told fingerprint comparisons have a zero-error rate. That’s an expert coming from the FBI saying that. No one doubted that he was an expert, but it was not accurate testimony.

You multiply that a lot of times, and you had judges accepting, not unsurprisingly, folks coming in from these disciplines and saying, “I’ve done this for a long period of time.”

One piece of the legal doctrine that can control this was: if the practice has been recognized … for a long period of time, then a judge is bound to accept it. Well that turns it on its head. …

Florida is, for example, on that [Frye] standard [of] “if it’s been around for a long period of time.” And [the nation] watched the Casey Anthony case, and they let in testimony from an expert on a science that had never been accepted before in court: the “odor of death.”

… At least the public awareness is going up. … There are even some cases where judges have been a little more insistent about the testimony that’s being offered in saying, “No, I need more than that to satisfy me before I let it in.”

But I still believe you can’t count on the system of justice to correct the science. The science has got to do that. … They’ve got to make it right. …

You wanted to say something about the forensic science community.

For all the problems we’ve been discussing, one of the good things that I did see was some truly smart, committed people in the disciplines who cared deeply about the issues and understand the need for change. …

I don’t want anyone to think that I mean to damn all the disciplines and all the people working in them across the board. I don’t. I’ve met some really able people in forensic sciences who are trying very hard to advance. They want certification. They want better research. They want better labs. So there are some good people working. They’re doing the best they can, and they’d like to see the change too.

… Is it true that the defense rarely has the same resources as the prosecution when it comes to forensic evidence?

That’s been my limited experience, and that’s been my understanding from people who I know on that side of the business. That is, that they’re often outgunned.

A lot of times they’re single practitioners, they don’t have the resources, and the person they’re representing may not have the resources for them to call the experts. …

It is an adversarial system, yes, but we need the science first, and then let them fight about what the good science means, as opposed to struggling with disciplines that don’t have good science undergirding them. We’re not equipped to do that.

That was the Supreme Court’s motion in Daubert [v. Merrell Dow Pharmaceuticals, Inc.]. I think it’s now turned out to be a little bit naive, that is, that the adversarial system will do it, will have smart lawyers on both sides, and they’ll duke it out, and they’ll figure out the right answer.

If the problem is that there isn’t the resource there with respect to a number of these disciplines, we got to go a step back and get that right first before we start duking it out in court over the nuances of good science. …

If there is a certification group out there that is certifying 99 percent of the people who apply, is that an adequate standard for the forensic sciences?

… The 99 percent wouldn’t bother me if you said to me this certification group has a three-year training program, they have written tests, they have follow-up, they have an oral exam, and then they’re so good at what they do that 99 percent of people pass.

That’s fine, but that’s not my understanding of what’s going on. We have certification in so many other professions. Why not here? …

You mean barbers get certified by the state, so what’s the resistance to having forensic examiners certified by the state or by a federal agency?

… I think there are a lot of people who would not resist it, but there hasn’t been anyone who’s said, “I have the money, and I have the authority, and here’s what we’re going to do.”

I think if we had that, the fight would be over what would be the nature of the training and the certification, and not a resistance to certification. We just haven’t had the champion yet. …

Is this the tip of a bigger iceberg?

I honestly don’t know the answer to that. … To be very honest with you, it doesn’t matter to me. It matters to me in the sense that there may be problems that we haven’t addressed, but how large that group is doesn’t matter to me.

As someone who works in the system of justice, I want to limit the mistakes as much as we possibly can. So you don’t have a convincing argument to me when you say, “Well, it’s only 10 percent of the cases as opposed to 50 percent,” because 10 percent is too high for me. If you said 5 percent, I would say, “That’s too high.” Whatever.

Let’s say it’s less than 1 percent.

It’s too high, because we’re sending people away unnecessarily, or we’re not finding the right people when we could. And the stakes are too high when you’re talking about prosecution, incarceration. The stakes are too high. …

There’s really no way to tell today what the certifications mean. There’s no standard for that.

… There are certifiers, but it’s not what you and I are talking about. That is, real licensing programs, real certification programs, where if someone came into court, and counsel or judge [asked if you] are certified by “the,” and “the” is a group that trains, gives serious tests, and follows up and will revoke your license and affect your job and your ability to testify in the event that you do something wrong or fail. No, that doesn’t exist now. …

By the way, why are you talking to us? Federal judges don’t usually jump in front of the camera.

You asked me politely if I, as a former co-chair of the National Academy of Sciences, would share some of the committee’s thinking on matters that I think are incredibly important. And then I was willing to talk to some of my colleagues on the committee, and they felt it would be perfectly appropriate to do that.

The additional thing for me as a professional matter, … forensic science is the handmaid of the justice system. We have to know that the experts who are testifying know what they’re talking about, and that the disciplines that they are representing are supported by validity and reliability. We have to know that. …

I’m willing to say that the National Academy of Sciences, with a very diverse, smart, thorough group, took this issue on at the request of Congress. We studied it very carefully, and we found some very serious problems. And we think that the American people ought to understand and worry about the problems that we found and do something to correct them.

Because the wrong person may be going to jail, and the wrong person [may] be out on the street.

Exactly. It cuts both ways. It is not pro-defense, it’s not pro-prosecution. It’s pro-justice system to get it right. That’s what we should be trying to do, is to get it right. And we should never run from that. …