Read our COVID-19 research and news.

Forensic Science on Trial: The First of Two ScienceInsider Interviews

In February, a landmark report by the National Research Council (NRC) in February criticized nearly every aspect of the nation's forensics science system, including unreliable techniques for analyzing hair and DNA samples—a problem the U.S. Senate has been addressing in recent hearings.

But the NRC report said almost nothing about arson. So ScienceInsider conducted an email interview with John Lentini (left), a nationally known fire investigator who conducted an outside review of the controversial case of Todd Willingham, a convicted arsonist who was executed in 2004. Tomorrow, we will run an interview with Jay Siegel, a scientist who served on the NRC panel.

Q: What went wrong in Texas, and why? How widespread are such flaws in supposedly scientific forensic investigations?

J.L.: The only thing unique about Texas with respect to miscarriages of justice stemming from faulty arson convictions is that Texas is a profligate user of the death penalty, making it impossible to take back its harshest punishment. Actually, there are wrongful convictions for arson all over the United States. Because there is no DNA involved, and because the investigators only document the scene to the extent necessary to "prove" arson, it is exceedingly difficult to obtain a reversal of a wrongful conviction.

Most fire investigators were trained by mentors who were trained by mentors, who passed on belief systems based on anecdotal experience rather than on chemistry and physics.

Mr. Willingham's lawyers walked through the burned residence in Corsicana and were able to see the "pour patterns" on the floor. These were irregular areas of burning caused by the irregular nature of fire. In addition to the pour patterns, the fire investigators also saw evidence of "high temperatures at floor level," such as melted aluminum.

These investigators held the belief that fire burns up, and should not achieve temperatures equal to the melting point of aluminum at floor level. They're simply wrong. Fully involved compartments frequently achieve temperatures close to 2000°F at floor level, more than enough to melt aluminum. For some reason, fire investigators have adopted the notion that accelerated fires burn at higher temperatures than unaccelerated fires, where there is, in fact, no evidence to support this belief. It is, however, like most fire investigation myths, an appealing notion.

The same is true of the phenomenon known as crazed glass. No less an authority than the National Bureau of Standards [now called the National Institute of Standards and Technology] once published a Fire Investigation Handbook that came right out and said that craze class was an indicator of "rapid heating," though they never defined what they meant by "rapid." Actually, crazed glass only occurs when hot glass is rapidly cooled, but that did not stop many investigators from using this discredited indicator to send innocent citizens to prison.

There are numerous other myths involving artifacts found after fires, including spalled concrete, shiny alligator blisters, annealed bedsprings, and the "normal" angle of a V-shaped pattern. All of these myths were once taught as gospel by the National Fire Academy, which trained most of the fire investigators working in the public sector today. All of these myths have been debunked in the peer-reviewed literature since 1992, but a surprisingly robust rearguard of fire investigators has resisted accepting the loss of the "tools" that they use to determine fire causes and obtain convictions.

These convictions occur because juries believe the fire investigators who pretend to be scientists, but who do not even possess a minimal scientific education. The reason that these individuals are allowed to practice as fire investigators is that we, as a society, have asked them to do so, and have refused to offer sufficient salaries to attract college graduates.

The average fire investigator is a former fireman, who has studied chemistry and physics in high school. Don't get me wrong. These guys are heroes, but the skill set required to extinguish a fire is a completely different skill set from that required to investigate it. As firefighters, these individuals always succeeded in putting out the fire. They developed a mindset that did not allow for failure, and consequently, when the evidence required that a fire be called "undetermined," these individuals were unwilling to make that call.

Q: Is there something particularly challenging about arson investigations?

J.L.: One of the skills that is taught to fire investigators is a skill that just about anyone can learn—the ability to testify convincingly. This is easy to do when the witness believes what he is saying and believes that he has got the right suspect, even if the fire was an accident. Juries have no way of knowing that some of these determinations could have been more credibly made with a Ouija board.

Q: How would you fix the system?

J.L.: Since 2000, there has been a growing acceptance by the fire investigation community of the need for scientifically based determinations, and there have been a remarkable number of experts excluded from testifying in civil cases. Unfortunately, in criminal cases, exclusions of government experts are so rare as to make headlines. Judges do not have the skills to be gatekeepers when the question is one of science. The real gatekeepers in arson cases are the prosecutors.

One way to avoid wrongful convictions would be to hold the state responsible for proving beyond a reasonable doubt that a crime was actually committed before being allowed to introduce evidence of guilt. Most people wrongfully convicted are not pillars of the community, and it is not hard to get a jury to dislike them. If the first 2 weeks of an arson trial are about character assassination, then by the time the state presents its weak science, the jury doesn't care. Mr. Willingham was unfortunate in that his lawyers did not believe he was innocent, and did not put on an effective defense.

Ernest Willis was convicted of arson [in Texas] and sent to death row on almost identical evidence, but he had the "good fortune" of having such an ineffective lawyer that even the Texas Court of Criminal Appeals was forced to admit that he did not get a fair trial.

The National Academy of Sciences (NAS) report focused narrowly on "pattern evidence," which is usually used to associate the defendant with a crime scene. Very little mention was made of forensic science disciplines that answer the question, "What happened?" Fire investigation was mentioned in a single line. One can hope that when the Congress looks at the issues raised by the NAS report, they do not feel constrained to stay within its narrow scope when attempting to strengthen forensic science.