Dr. Haim Cohen examines bones in the forensic anthropology lab at the Israel National Center of Forensic Medicine on May 27 in Tel Aviv, Israel. (Corinna Kern/The Washington Post)
Opinion writer

Note: This is part 6 of a 6-part online symposium on the use of forensics in the criminal justice system. You can read the introduction and first four parts here, here, here, here and here. And you can read the full biographies of the panel members here.

Question 6

Suggest three reforms — low-, mid- and high-hanging fruit — that would improve the quality of expert testimony in criminal cases. The first should be a common-sense, slam-dunk reform — something that every jurisdiction should be doing that would improve the system, but at little cost and with few repercussions for those likely to resist change. The second should be a reform that isn’t quite as far-fetched but would require some convincing of policymakers and the public. For the third, assume you are dictator and could make any change you wish, regardless of cost, political feasibility or degree to which it would unsettle the current system.

Sandra Guerra Thompson, University of Houston Law Center, Houston Forensic Science Center

Easiest reforms:

Lab transparency: Forensic laboratories process evidence in order to produce information. As we move forward in an attempt to prevent scandals and wrongful convictions, policymakers should demand transparency by forensic labs. They should post critical documents on their websites, including accreditation certificates, standard operating procedures for each discipline, validation studies, corrective action reports, budget information, and statistics on testing requests, turnaround times, and backlogs. The public has a right to this information, and should not have to file public information requests to get it.

Mid-level reforms:

Independent labs: In my previous answers, I have argued in favor of lab independence from law enforcement. However, the most important reasons for independence have nothing to do with eliminating bias. In my experience with the Houston Forensic Science Center, I have learned that lab independence makes the organization more agile in terms of setting priorities, procurement and hiring. Independence enables the lab to be transparent without the need to get approval from law enforcement authorities. It also eliminates the appearance of pro-police bias.

Independence puts the lab on par with the other critical stakeholders in the criminal justice system so that it can be more responsive to law enforcement, victims and the defense bar. The model used for HFSC was to make the lab a “local government corporation,” and having a corporate board of directors for a government agency has proved to spur innovation and efficiency. A public corporate structure also positions board members to advocate externally on behalf of the lab in obtaining adequate funding, and in handling public relations.

Crime scene units are called into action when serious crimes occur, such as murders and officer-involved shootings. They collect evidence for testing in crime labs and document crime scenes. It is imperative that they follow scientific protocols in handling evidence, and I have come to appreciate the importance of having civilians (not law enforcement) working in these positions. In Houston, we no longer employ police officers to collect evidence in major cases. This work is now done by an accredited crime scene unit that is a division of the laboratory. It is especially important for restoring the public’s trust in the integrity of investigations of officer-involved shootings that the individuals gathering the evidence have scientific training and are not law enforcement personnel.

If I were dictator:

Limitations on testimony. Courts should stop admitting fingerprint and firearm testimony as “scientific” evidence. The Supreme Court declined to make a distinction between “scientific” and subjective expertise, requiring that all types of expertise meet the same reliability test, but the court acknowledged that some of the test’s factors were less relevant to some types of expertise than others. Until we develop the statistical foundations for these disciplines and other comparative (pattern-matching) disciplines, we should make it clear to jurors that the conclusions drawn by these experts, while helpful and worthy of some consideration, are not scientific results backed by objective measurements and standards.

Jules Epstein, Temple University Beasley School of Law, National Commission on Forensic Science

Easiest reform:

Funded forensic teams for statewide indigent defense.

Mid-level reform:

We need the National Institute of Standards and Technology — or some other fully funded forensic discipline institute separate from prosecuting agencies — to set standards for forensic testimony.

If I were dictator:

No forensic evidence in courtrooms (and maybe in prosecution decision-making) without it being fully vetted by an independent entity; and no judicial or attorney involvement in such cases without mandatory education.

Barbara A. Spellman, University of Virginia School of Law

Easiest reforms:

— Implement case management systems that keep forensic examiners away from information that may bias their analyses.

— Forensic reports should be more complete, including information on how the type of evidence can best be questioned.

Mid-level reforms:

— Get the forensics lab out from under the control of the police/prosecutor’s office. And move it physically away.

— Give defense lawyers access to the reports. In fact, tell them when reports have been generated. Suspects should know what evidence there is against them well before trial, e.g., in bail and plea bargaining stages.

If I were dictator:

To every case for which there is forensic evidence, assign three professional forensic experts. They would examine the evidence together but, by lottery, one would be the consultant to the plaintiff, one to the defendant, and one would be the witness at trial. The consultants would give information about content and strategy for questioning the expert witness. The parties could of course hire their own experts to give them such advice as well, but only the court-assigned expert would testify.

Simon A. Cole, Department of Criminology, Law and Society, University of California, Irvine; National Registry of Exonerations

Easiest reforms:

Certification of all examiners and accreditation of all laboratories. The National Commission on Forensic Science already passed these requirements for [Department of Justice] labs. The disciplinary organizations (e.g., IAI, AAFS, etc.) could take public positions that you should not be practicing X discipline with certification from X. Such statements can be used by defense attorneys and/or judges to eradicate uncertified experts and unaccredited providers from the courtroom. At least then we would have a common set of experts/providers, which are all in some sense responsible to and controlled by some body.

Medium-level reforms:

Adoption of an anti-bias protocol akin to what the FBI fingerprint unit calls “linear ACE-V” across all providers and all disciplines (for which it is feasible). Essentially, this mandates: (1) interpretation of unknown samples before exposure to known samples; and (2) documentation of what features were seen in those unknown samples at the time of that “blind” or “masked” analysis.

If I were dictator:

Get forensic science service providers out of law enforcement agencies! People don’t talk about this much anymore, probably because they perceive it as too hard, but you can argue it is at the root of all of the above problems. If forensic science were invented today, we probably would not choose to locate it in law enforcement agencies.

Judy Melinek, forensic pathologist, author of “Working Stiff

Easiest reforms:

Require that all prosecutors get training and experience in criminal defense work, and that all criminal defense attorneys work for a time as prosecutors. This is the basis of the Judge Advocate General Corps (JAG) system in the military. I have noticed, as an expert working in military tribunals, that there is more collegiality and respect between the prosecutors and defense. They don’t have the “win at all costs” mentality I see in civilian trials. They work collaboratively to collect and share the evidence they need to present at trial, and then make their opposing arguments. There also appears to be a lot less prosecutorial misconduct due to inadequate disclosure of evidence. I believe this is largely due to the mandatory lawyer cross-training in the JAG system. Walk a mile in your opponent’s shoes, and an adversarial legal system becomes far less antagonistic and — perhaps — measurably more just.

Mid-level reforms:

Require a minimum per capita expenditure on all death investigation systems nationwide. Right now we have a national crisis in forensic pathology staffing. A hospital pathologist can finish residency and earn, on average, $50,000 to $100,000 a year more for the rest of their career than one who does additional subspecialty training in forensic pathology and goes into that field. It’s hard to attract the country’s best and brightest doctors into my profession if it means they can’t pay back their medical school loans, or that they end up working for an underfunded and understaffed government office where they are doomed to burn out after a few years.

If I were dictator:

Require that all forensic laboratories and death investigation agencies be independent of law enforcement agencies, including the nationwide elimination of the coroner system. Forensic labs will be run by scientists and will be accountable to the public. The public needs to trust that scientists are not being incentivized to aid in convictions, and to trust that they are not “prosecution witnesses” operating as a branch of law enforcement.

Roger Koppl, Forensic & National Security Sciences Institute, Syracuse University

Easiest reforms:

— Don’t let crime labs work on commission! In at least 24 states (as of 2013) public crime lab budgets come at least in part from “court-assessed fees,” which are assessed upon conviction. Thus, if your lab’s work leads to a conviction, the lab gets money. If your lab’s work leads to exoneration, you get no money. This monetary incentive creates an unfortunate bias toward conviction, which can create false convictions. This bad incentive can be eliminated in any state that has it by passing a simple law forbidding courts from collecting such fees and forbidding the state from returning any such fees to the crime lab involved in the case.

— Masking! Forensic scientists should not know whom the police suspect, or even what crime has been committed. If I know that I’m working on the theft of a desk lamp, I won’t be motivated to find a match when there is no match. But if I know that I am working on a heinous double murder, pressure to solve the crime might induce me to see a match when there is no match at all. Certain information should be “masked” from the forensic examiner. And the examiner should thoroughly examine and report on the crime scene evidence before comparing it to the suspect’s sample. Proposals for “sequential unmasking” go back many years. And some labs have adopted sequential unmasking protocols. This is something we can do relatively easily and relatively quickly. Implementing sequential unmasking would reduce unconscious bias.

Mid-level reforms:

— Redundancy! We should sometimes send evidence to three separate labs instead of just one lab. If they send back different answers, we had better find out why. This is a form of checks and balances. The pilot in a jetliner has a co-pilot. Your car has both air bags and safety belts. Roads often have guard rails. And so on. We have redundant safety systems everywhere, but not in forensic science.

— Blind proficiency exams. If we don’t have a good measure of lab error rates, we cannot expect labs to improve their performance over time. Nor can we know how much confidence to put in the lab’s evaluations of the evidence it gets. Blind proficiency tests would give us the information we need to improve operations over time and to evaluate the level of performance at a given time.

If I were dictator:

A real and substantive defense right to expertise. I view this as the lynchpin reform. In any case, one side or the other will have an incentive to reveal any known infirmities in the forensic evidence coming to court. Thus, crime labs will be under constant pressure to improve and to be open and honest about their procedures and error rates. In today’s system, in which the labs tend to be aligned only with the prosecution, there is no such pressure to improve.

Chris Fabricant, the Innocence Project

Easiest reform:

Eliminate bite mark evidence.

Mid-level reform:

State forensic science commissions, staffed by career scientists, that produce rigorous basic and applied research, and, based on that research, issue mandatory guidelines for the introduction of and the testimonial limitations of all forensic evidence.

If I were dictator:

The same, only at the national level.

Itiel Dror, University College London, Cognitive Consultants International

In addition to the solutions I have already suggested in my answers to previous questions . . .

Easiest reform:

Experts need to make their decisions based only on the relevant data, and should not be exposed to irrelevant contextual information.

Medium-level reforms:

— All forensic decisions should be verified (not only identifications), and the verifications need to be totally blind (not knowing who did the first analysis, what they concluded, how, etc.).

— Every state needs to establish a forensic oversight board or a forensic commission (such as those in Massachusetts, Texas and New York). These boards should include representatives from all stakeholders, including experts on bias, and make sure best practices are developed and implemented throughout the state.

If I were dictator:

Move the science out from under the police (they have enough to deal with without doing the science). Instead, establish separate scientific institutions (not private for-profit labs) to take over forensic work. That is, governments should establish independent scientific “forensic science institutions” [that] will work with law enforcement, [district attorneys] and defense lawyers.

Michael Risinger, Seton Hall University School of Law, the Last Resort Exoneration Project

There was never any low-hanging fruit. Perhaps the most low-hanging fruit should have been to eliminate testimony by subjective pattern matching disciplines that “the source of the markings on this bullet (or this crime scene fingerprint, or these tooth marks on the victim’s skin, or whatever) came from defendant’s gun (or fingers, or teeth) to the exclusion of all other guns (or fingers, or teeth) in the world.” There could be no more unjustified overstatement than this. Still, it took a couple of decades to kill it off, and what has been substituted is often not that much better.

Where I do hope to see progress (unfortunately likely to evolve at a glacial pace) is in better judicial understanding of these issues, more deployment of effective protocols to control context bias, and better laboratory management practices that will allow us to learn from and correct mistakes rather than incentivize people to cover them up.

Brandon L. Garrett, Duke University School of Law

Proficiency tests! That would be my main reform. It is extremely important that crime labs error-test themselves using blind proficiency tests as a method of quality control. Right now, only the Houston Forensic Science Center is doing that at scale. Every lab should. The Houston lab is showing that it can be done affordably, even on a large scale.

In most cases, labs generate forensics reports, but the case ends with a plea bargain. That makes it especially important that evidence is analyzed reliably and that the conclusions produced in lab reports be clear and accurate.

Show us the limits of an analysis. Show us the error rates. Jurors will not throw out the evidence, and experts will not be “ruined.” Even for techniques that have not been rigorously error-tested, jurors should know how good a particular expert is at his or her job. And they can only know that if the expert has been subjected to routine, blind proficiency tests.

Keith A. Findley, Center for Integrity in Forensic Sciences, University of Wisconsin Law School

Easiest reforms:

If forensic science wants to claim the mantle of “science,” it must follow fundamental scientific principles, like double-blind testing. Systems should be created to at least shield analysts from domain-irrelevant but contextually biasing information. This can be complicated at times, but it can be done without disclosing the vast array of information that analysts routinely receive today. Proper case management and intake systems can ensure that analysts receive only the information they need, and only when they need it.

Mid-level reforms:

Remove all crime laboratories from the administrative control of law enforcement. This was another important recommendation of the 2009 NAS Report. Science, if it is real science, should be neutral and objective. It should not favor or uniquely serve one side or the other. Only by fully removing all laboratories from the administrative control of law enforcement can the laboratories begin to develop a real ethic of scientific neutrality and objectivity, and can the analysts have a realistic hope of avoiding the pressures and biases when one is perceived as part of the law enforcement team. In many jurisdictions, this suggestion is a nonstarter, as police and prosecutors typically jealously protect “their” laboratories as the source for “their” evidence. But there are examples to show it can work, most notably the Houston Crime Laboratory, which was rebuilt as an independent entity after the old Houston police laboratory imploded.

If I were dictator:

Follow the central recommendation of the National Academy of Sciences in its 2009 report and create a truly independent new federal agency (the “National Institute of Forensic Science”) tasked with establishing and enforcing best practices; establish standards for mandatory accreditation of laboratories and certification of forensic scientists; promote scholarly, competitive, peer-reviewed research; develop a strategy to improve forensic science research and educational programs; fund independent research projects and educational projects; oversee education standards and accreditation of forensic science programs in colleges and universities; assess the development and introduction of new technologies; and the like.

John Lentini, arson/fire expert

This is a fun question.

Easiest reforms:

For the low-hanging fruit, allowing depositions in all criminal cases could be accomplished very easily.

Mid-level reforms:

Judges should be required to consult independent experts when there are competing interpretations of forensic evidence.

If I were dictator:

The law should hold experts liable for opinions that lead to wrongful convictions. This would deter experts from offering unsupported opinions. While they’re at it, the court could reverse the Connick v. Thompson decision and hold prosecutors liable for intentional Brady violations [a Brady violation is when a prosecutor illegally withholds exculpatory information from defense attorneys].

Frederic Whitehurst, FBI crime lab whistleblower, Forensic Justice Project

Easiest reform:

No forensic lab analyst will testify without having fully displayed his academic transcripts, all training materials, and being able to point specifically to something in those transcripts and training that supports his opinion. At this time in this country, privacy trumps due process. Anyone can fail courses in college, sleep through training classes, and still testify without triers of fact having any idea if that analyst has any aptitude at all in his field. Any other field in which one seeks employment will require a full disclosure of academic records, experience and training. That is not true in U.S. courts.

Mid-level reform:

Every crime lab should make its protocols, validation studies and quality assurance documentation easily accessible on the Internet. If a lab refuses, courts should refuse to accept expert witnesses from that lab.

If I were dictator:

Analysts found to have committed malfeasance should be held accountable even into retirement. That retirement check given each month to a serial liar who once worked at a crime lab should be given [instead] to making that liar’s victims whole again. Accountability will help counter the constant pressures put on lab analysts by police and prosecutors.

Read more from Radley Balko:

Part 1 of this series: We need to fix forensics. But how?

Part 2 of this series: How much should juries rely on expert testimony?

Part 3 of this series: How do we reconcile law and science?

Part 4 of this series: Is an adversarial justice system compatible with good science?

Part 5 of this series: How much interaction should we allow between forensics analysts and law enforcement?