Author(s): Fred Durso. Published on March 4, 2014.

IT’S NOW REGARDED AS the premier resource for fire investigations, but when NFPA 921, Guide for Fire and Explosion Investigations, first arrived on the scene in 1992, some investigators were rooting for its early demise.

The problem was the guide’s shift toward science as the basis for analyzing fire incidents. Prior to NFPA 921, the fundamental principles for determining the causes of fires did not involve science per se, but rather experience-based hypotheses that were not tested to determine their validity. This process, known as “negative corpus,” relied on a process of elimination rather than supporting evidence or scientifically supported conclusions, and some investigators feared that a science-based approach would establish criteria of proof that would be too difficult to meet. Following a crucial U.S. Supreme Court ruling in 1993, though, U.S. courts began referring to NFPA 921 as the “standard of care” for evaluating expert testimony regarding fire investigations.

A member of the NFPA 921 Technical Committee on Fire Investigations since the release of the first edition, Randy Watson has investigated approximately 3,000 fire incidents using the guide. He’s now the committee chair for the guide, and works as a senior fire investigator and regional manager for S-E-A Ltd., in Atlanta, Georgia. Watson spoke with NFPA Journal about the guide’s evolution, updates to the newly released 2014 edition, and the role of NFPA 921 in contesting arson convictions.

Before NFPA 921 was released in 1992, how were fire incidents investigated?
For fire investigations, 1992 is the dividing line. Prior to that, there was no definitive document by which a fire investigation could be judged. Most of the time in court, investigators would talk about their experience, but there was nothing to challenge them. Also, when investigators talked about their conclusions, they might say, “I’ve eliminated everything else, so this is the cause”—the essence of negative corpus. There was very little challenge, and very rarely was an expert testimony excluded. You never heard an expert talk about using a scientific methodology. There was a lot of resentment about referring to fire investigation as a scientific endeavor.

Members of the fire investigation community felt that a scientific methodology would hold them to something they couldn’t adhere to. Testing a hypothesis was unheard of. When the first edition of NFPA 921 was released and it included the scientific method, it was a hard pill for fire investigators to swallow. Some felt that if the document was ignored, it would simply fade away.

But it didn’t. What happened?
The Daubert v. Merrell Dow Pharmaceuticals decision was handed down in 1993, where the U.S. Supreme Court made judges the gatekeepers of expert testimony. The ruling said that judges have to determine if expert testimony is reliable, and that using a methodology that has been peer reviewed, published, and accepted was part of that reliability test. So in cases that included fire investigations, courts began gravitating towards NFPA 921, a consensus document from a group of leaders in the fire investigation community.

How widely used is the guide today?
It’s universally utilized. It’s printed in French, Spanish, Korean, and Chinese, to name a few. In almost every case involved in litigation, an expert will say, “Yes, I did my investigation in accordance with NFPA 921.” More and more you’re seeing experts’ reports directly referencing NFPA 921. When someone is giving a deposition or testifying, you’re either going to be challenging the expert because he didn’t follow NFPA 921 or you’re supporting your own opinion because it’s consistent with NFPA 921. One court case referred to it as the “bible of arson forensic science,” and another called it the “gold standard.” The International Association of Arson Investigators released a position paper last year supporting NFPA 921. The National Association of Fire Investigators releases a resolution after each edition comes out, which talks about it being the standard of care in the industry.

That being said, some investigators still support the negative corpus model over the science-based approach. Why?
Some fire investigators, just like in other professions, don’t like someone challenging their opinion or forcing them to change their ways. They have always done it that way, so they don’t see a need to change.

Has there been talk about making NFPA 921 a standard?
One of the things the committee voted on last year was that I, as chair, would explore with NFPA the idea of moving NFPA 921 to a standard. But here’s the problem: a guide contains nonmandatory language — “You may do this or that” — but a standard says, “You shall do this or that.” The standard contains declarative statements without much explanation, unless there’s an Annex that expands upon the statement. One of the purposes of a guide like NFPA 921 is to provide the kind of explanatory material that you won’t find in the body of a standard. The reason NFPA 921 can’t be a standard is because it’s an investigative document, and no two investigations are the same. One investigation might be a house, the next might be a car, the next might be a boat. You can’t have definitive statements saying you shall do “X” on every investigation because every investigation is different.

How does NFPA 921 work with other NFPA standards?
One example is NFPA 1033, Professional Qualifications for Fire Investigator. The standard identifies the minimum job performance requirements for fire investigators, requiring them to be trained in the scientific method and to utilize it when conducting an investigation. As the companion document, NFPA 921 describes the elements of the scientific method and provides additional guidance. So NFPA 1033 focuses on the skills of the investigator and NFPA 921 focuses on the process of the investigation. That’s how they work hand in hand.

How are you applying the scientific method to investigations?
Let’s say you begin collecting data on the origin of a fire. As you analyze data, you begin to develop scenarios and hypotheses. Then you start testing the hypotheses, validating what you’ve come up with. You test it and you try to disprove it. Through the testing process, you end up with one hypothesis that’s consistent with all the data. There are times that as you develop a hypothesis, you end up with two or three and you can’t eliminate any of them based on the data. You’ll end up with an origin that’s undetermined. You can apply that same logic to cause and ignition sources. 

The 2011 edition of NFPA 921 refers to negative corpus as “not consistent with the scientific method, inappropriate, and should not be used because it generates an un-testable hypothesis.” Was there backlash to this language?
That section [in Chapter 8] was titled “Inappropriate Use of the Process of Elimination.” There was a lot of uproar because people didn’t understand the section. They felt we were saying the process of elimination was a bad thing. That’s not the case. What we were trying to communicate was that if you misuse the process of elimination, it could be bad. In the new edition, we added some introductory language to address that the process of elimination is an integral part of the scientific method. But if you’re making a determination for which you have no evidence to support, that’s not consistent with the scientific method. Whatever determination you’re going to make in regards to the cause of fire, you have to have evidence to support it.

How will other updates to the 2014 edition of NFPA 921 — particularly the addition of color photos — aid your investigations?
Fire investigation is a very scientific endeavor, and it’s also a very visual one. Having the ability to include color images in the document greatly enhances our ability to show examples of the kinds of things to look for at the scene. A good example of the benefit of color is Chapter 6, on fire patterns — you’re able to see the contrast of damage and of charring. The color images demonstrate how the patterns can be more easily recognized and more visible.
There’s also a new chapter on fire protection systems.
A fire protection system can affect fire — it can slow fire or put it out, or if it’s not working properly the fire can spread as if there were no system. The 2014 edition of NFPA 1033 includes these systems as one of the 16 knowledge-based items for investigators. Discussing this topic in NFPA 921 supports requirements in NFPA 1033. Now there’s more information for fire investigators on the systems themselves, how they operate, their components, and what to document.

There have been recent incidents of arson investigations being challenged or overturned. What’s causing the trend?
An organization called the Innocence Project [which works to exonerate the wrongfully convicted] has taken an interest in arson cases. They are going back and evaluating the science or lack of science in investigations and comparing that with what’s available today. You’re not necessarily condemning the investigators, because they may have been using what was considered state-of-the art at the time. The goal of the criminal justice system, however, is to get it right, so in some cases there may be reasons to go back and apply current state-of-the-art science and see if mistakes were made.

Is NFPA 921 being used as part of the reevaluation process in any of these cases?
Yes. NFPA 921 is the document that has promoted the scientific method and testing of hypotheses. If the Innocence Project is looking into a case, they have to have something reliable that opinions can be based on.

Interview conducted and edited by Fred Durso, Jr., NFPA Journal staff writer. Follow him on Twitter @FredDursoJr.