There is a lot that one can read at this blog about what is wrong with forensic science, but there are also great efforts to reform and make forensic science better and scientific. As I stated a long time ago when I started this blog, I was going to be critical of those disciplines and people that are not scientific, but I was also going to praise those who are. One of the best organizations that is doing the correct thing (putting science back into forensic science) is the National Fire Protection Association (NFPA). In my opinion, NFPA is a model for forensic science reform.
As we discussed here in Fire Whisperers-the Story of Bad Arson Investigation, the world of fire investigation had been notoriously full of unscientific “fire whisperers,” those who had no real scientific training and simply opined as to origin and cause of a fire based upon subjective and purely wildly speculative means. Unfortunately there are many jurisdictions that still employ these non-scientific “fire whisperers.”
After great struggle, NFPA formed a committee to promulgate a document that would force fire and explosive investigations to be more scientific. This resulted in NFPA 921, Guide for Fire and Explosion Investigations. It is universally hailed as the industry standard. (A brief discussion of the history and the background of NFPA 921 can be found at the www.NAFI921.com website.)
The latest version of NFPA 921 has been published just this month. In the 2011, we find the last vestiges of the “fire whisper” abolished: negative corpus. For years, the false methodology and the illogical basis to support a negative corpus conclusion as far as origin and source had been well exposed (See The Pitfalls, Perils and Reasoning Fallacies of Determining the Fire Cause in the Absence of Proof: The Negative Corpus Methodology By Dennis Smith and in Dr. Gerald Hurst’s opinion letter regarding the Terri Hinson case-March 16, 1998). However, it is only with this most recent document NFPA 921:2011 that we see direct language that addresses this. The document reads in short regarding Negative Corpus NFPA 921-18.6.5
The process of determining the ignition source for a fire, by eliminating all ignition sources found, known, or believed to have been present in the area of origin, and then claiming such methodology is proof of an ignition source for which there is no evidence of its existence, is referred to by some investigators as “negative corpus.” Negative corpus has typically been used in classifying fires as incendiary, although the process has been used to characterized fire classified as accidental. This process is not consistent with the Scientific Method, is inappropriate, and should not be used because it generates un-testable hypotheses, and may result in incorrect determinations of the ignition source and first fuel ignited. Any hypothesis formulated for the casual factors (e.g., first fuel, ignition source, and ignition sequence), must be based on facts. Those facts are derived from evidence, observations, calculations, experiments, and the laws of science. Speculative information cannot be included in the analysis.
I say let us all take a minute and praise NFPA and the arson community for doing the right thing and putting science back into forensic science!
Fire Investigator Mason says:
This addition to 921 was written by a criminal defense attorney, so it is appropriate that you would be supportive of this. However it is fact that this new statement added to 921 is the most poorly written, under explained piece of work ever included in a version of 921. First off section 18.6.5 is Titled “Inappropriate use of the process of elimination”, this tells one that there is an appropriate use of the process of elimination. Yet in the text the appropriate use is not given. As a fire investigator I understand the reasoning behind this text. There are fires that occur that this methodology holds no basis. I have seen fires in which damage was so extensive that a real determination can not be made, and in many cases was, using negative corpus. Here is an example….A fire started in a kitchen damaging the kitchen to a magnitude that it appeared as a “black hole”, throughout the kitchen ceiling is 100’s of feet of electrical wiring, both romex and knob and tube. An investigator goes into the fire room chooses a small area of origin, based on floor burn patterns and has an electrical engineer remove electrical equipment from that area for laboratory examination. Both the investigator and engineer leave all of the wiring in the ceiling intact without ever examining it. Some wiring was also consumed by the fire and therefore could not ever be ruled out as an ignition source. However, the artifacts that were removed were examined in a laboratory and ruled out as a potential cause, and the fire was deemed incendiary. This example I gave is exactly what this section of 921 was written for. The investigator in this example formed one hypothesis in his mind, he never considered any alternate hypothesis. Could wiring in the ceiling faulted, overheated causing fall down debris to cause the pattern on the floor? Could a fault have occurred to the wiring which was consumed by the fire? These are other hypothesis that were never even considered in this case. This is exactly what is described by this section of 921. But now lets look at another example in which process of elimination is appropriate. A fire occurs on a desk top, there is one distinctive burn pattern directly above the desktop to the ceiling and on the wall that the desk is positioned along. The pattern is a truncated cone pointing directly to the remains of a small debris pile on the desk, which appears to be the remains of a pile of crumpled paper. This fire was extinguished in its incipient phase. So what possibilities do we have as a source of ignition for this fire? Smoking, electrical, spontaneous heating, lightning, a gas leak ignited by a static electricity discharge, burning candle, or a deliberate human act? The investigation discovers that, the electrical supply as well as gas was disconnected prior to the fire, no one in this office smokes and no evidence of anyone smoking in this office is found, no one had been burning any candles and no candles are found anywhere in the building, there have been no recent storms producing lightning, and even though electrical supply was disconnected and electrical equipment is undamaged, all electrical equipment is removed and examined. No electrical cause is found. The gas system is tested for leaks and none are found. The debris pile is sent to a lab and tested for saturated fats to rule out the possibility of spontaneous heating…none are found. The only possibility left is that someone ignited the pile of crumpled paper with an open flame and took their source of heat with them. All other potential sources of ignition have been ruled out. According to this section of 921 this is not the scientific method? I disagree, the scientific method states collect data—-analyze data collected—develop cause hypothesis—test the hypothesis—select final hypothesis. We collected data at our scene…burn patterns, fuels, possible sources of ignition, witness statements. We analyzed the data we collected….looking at ignition temperatures of the fuel on the desk, potential ignition sources, air as the oxidizing agent, possible ignition sequences. We developed multiple cause hypothesis….electrical, smoking, lightning, gas leaks, spontaneous combustion, human factor. And finally we teated the possible hypothesis….testing discovered that the only probable cause for an ignition source was a human ignition factor. Here is a more simplified version….a person walks into a metal container and there is nothing in this container but a pile of brand new newspaper on the floor. There is no electrical, no gas…nothing. The person walks into the container takes a bic lighter out of his pocket, sets the pile of newspaper on fire, puts the lighter back into his pocket and walks away. A witness watches him enter and exit the container. According to this section of 921 this fire should be undetermined? It states to theorize an ignition source is unscientific because it is un-testable. How is this scenario untestable? In this instance I would state that the ignition source was an open flame, I cannot simply guess and say it was a match, or a lighter. For all I know this individual may have had a small torch in his pocket. But you can test the hypothesis of an open flame…..will a lighter ignite a pile of newspaper, will a match ignite a pile of newspaper, will a torch ignite a pile of news paper….yes all three are possible. So then it is not arson because he took the ignition source with him? I hope you understand my point here. The addition is on the right track, but needs much refinement. All this has created is a tool for defense attorneys to get arsonists off the hook. Thank god NFPA 921 is merely a “GUIDE”, or there would never ever be another arsonist convicted for his crime again. They would have free reign to burn you me and our families out of our home with no repercussion.
Justin J. McShane says:
I want to thank you for your post.
It is a great complement that you took the time to write a post.
We appreciate your passion and your point-of-view.
As you can see, I have published your comment in full, without editing.
However, I do have to address directly one aspect: To just simply brush off my analysis simply because I am a criminal defense attorney is dangerous. To assume bias is also wrong.
For example, if someone else were to do similarly to you (Eric Mason of Ohio) as you have done with me, that person might come up with this impression of you:
–Another who would make assumptions would point to your multiple references to your apparent dislike of criminal defense attorneys in this post and that person who makes assumptions could fairly conclude that you are a biased pro-prosecution and/or pro-insurance investigator.
–Another who would make assumptions would also conclude that you are a person who has very powerful and very conservative political ambitions as evidenced by your declaration on another website that you will run for Congress on a Tea Party platform (anti-civil libertarian agenda).
–Another who would make assumptions would also conclude that you assume that every fire is an arson regardless of a lack of tangible evidence. In other words, no verified proof would be necessary.
–Another who would make assumptions could look at the fact that you are a member of the Woodville Township Fire Department and that this might make it appear to another that you have bias in favor of the police as they are a fellow first responder, and are therefore the perfect candidate for what others have framed the unscientific “old school” approach of fire investigation perhaps accurately called “the fire whisper.”
–Another who would make assumptions would look at the fact that you have achieved certification through NAFI and not be terribly impressed as the current CFEI exam is based on the NFPA 921: Guide for Fire and Explosion Investigations and simply consists of 100 Multiple Choice and True/False style questions that can even be done through the mail with a “proctor.” Another who would make assumptions would discount the certification as it only requires on the multiple choice a score of 75% or better for passing. There are no essays, no practical component, and no actual investigation into the applicant’s current quality of investigation of arson cases required to sit for the exam. Further, another who would make assumptions would point out that the membership of the association and the certification itself has no minimum educational requirements other than a GED or high school degree. Another who would make assumptions would point out that there is no requirement for university level study or any sort of formal study in science at all such as formal studying of thermodynamics, engineering, mechanical engineering, physics, chemistry, analytical chemistry, metallurgy, tribology, fluid dynamics, kinetics, and so on—other than reading and guessing correctly based upon NFPA 921, filling out the application, and paying the $125 application fee.
–Another who would make assumptions would also point to the fact that even NAFI endorses NFPA 921: 2011 as they write on their website: “It is the official position of the National Association of Fire Investigators that NFPA 921 is the standard to which all fire and explosion investigations should be held by the courts and the profession. NAFI believes that NFPA 921 is the professional standard of care for fire and explosion investigation as the only peer-reviewed, science based, consensus text on fire investigation. NAFI has been a zealous supporter of NFPA 921 since the document’s very inception. NAFI has had official representation on the NFPA Technical Committee on Fire Investigators since the Committee’s very first meeting in 1986. Each of the three previous editions of NFPA 921 and the currently proposed 2001 edition have been adopted by NAFI resolutions at annual general membership meetings.” Another who would make assumptions would point to the fact that the current iteration of NFPA 921 and likely future editions would require more formal education than perhaps you have and therefore you might be biased against the document for that very reason.
–Another who would make assumptions would also point to the fact that you work at Sadler & Associates, Inc. and also show how a large volume of their business is from insurance defense where in the case of all fires it is always in the financial interest of all insurance companies to deny the claim and hope that there would be a finding of arson so as to make profit.
But I am not that other person who would make such assumptions. Instead, I would carefully consider the substantive part of your argument and divorce it from all issues of bias, accusation and non-substantive surplusage.
In doing so, I think the thrust of your position could best and most efficiently stated as:
1. NFPA 921:2001, in whole, or in part, is a criminal defense attorney created document.
2. There are legitimate uses of the negative corpus approach to conclude and scientifically present a fire as an arson.
3. NFPA 921 is simply a guideline and not to be considered a scientific standard.
As to your substantive point 1:
To correct one of your assertions, NFPA 921:2011 was not written by a criminal defense attorney. As you should know, it is a consensus document that requires an affirmative vote by those who are on the NFPA 921 technical committee, and then adopted by the executives of NFPA. NFPA 921 is written by the NFPA Technical Committee on Fire Investigations. The Technical Committee consists of up to thirty principal (voting) members and their alternates, appointed by the NFPA Board of Directors from those “vitally interested, qualified, and active” in the area of fire and explosion investigation. Of the thirty voting members there is not one criminal defense attorney. Maybe you are thinking of Hal C. Lyon. He was/is Chief Investigator for Robins, Kaplan, Miller & Ciresi, MN. Robins, Kaplan, Miller & Ciresi is primarily known as an insurance defense firm, although it appears as if they do handle some plaintiff’s work mostly in complex bad faith claims when insurance coverage is unjustly denied. He is not an attorney. If you are thinking of Stuart A. Sklar, he is not a criminal defense attorney, but rather is in plaintiff’s work. Further your own association has endorsed NFPA 921. According to their NAFI website: “It is the official position of the National Association of Fire Investigators that NFPA 921 is the standard to which all fire and explosion investigations should be held by the courts and the profession. NAFI believes that NFPA 921 is the professional standard of care for fire and explosion investigation as the only peer-reviewed, science based, consensus text on fire investigation.”
As to your point 2:
It seems as if we both agree that there is a large potential for abuse in fire investigators as to the negative corpus process. Your first example where NFPA 921 is basically ignored and there is an improper investigation would lead to what I would think that everyone would agree a non-validated conclusion. That is a clear abuse. We agree.
I suppose when one fairly examines your second example, we have to consider what the true goal of our analysis is. Are we going to be scientific or unscientific?
Most importantly to this analysis is that while you acknowledge that the scientific method should be the sound basis for the process of all investigation, you fail to highlight the key ingredient of the scientific method in that the hypothesis must be falsifiable. This falsifiable hypothesis is the key to the scientific method, not simply that the hypothesis is testable. As all true scientists and admirers of empiricism and science know: no series of experiments can ever prove a hypothesis, but a carefully designed experiment can disprove a hypothesis. This is what Albert Einstein meant when he said, “No amount of experimentation can ever prove me right; a single experiment can prove me wrong.”
There is an interesting entry that I think is instructive on Wikipedia that reads as follows:
When we use a negative corpus process, we violate the necessary falsification component of the true scientific method. You have committed the error of confirming the consequent in your example.
Further, with the second example that you argue would be an appropriate use of negative corpus is totally dependent upon an eyewitness account. The eyewitness may have bias, may be honestly incorrect, or lying. The account of the eyewitness cannot be verified. This eyewitness account is at most circumstantial evidence. It is clear that the scientific method does not allow for circumstantial proof.
Attempting to use circumstantial evidence in science is utterly unscientific. Arguments based on probability (circumstantial evidence) are subject to certain biases and interpretation. These biases come from the tendency to prove a hypothesis or scenario rather than to falsify it. Three forms of bias are particularly common. A tendency to select the first alternative that comes to mind consistent with the evidence — a bias known as satisficing — may prevent the analysis of other alternatives. Even when more than one alternative is considered, the limitations of the human mind to form and remember lists may limit the explanations to only those available to the imagination. The bias of availability limits the number of considered possibilities. Finally, if one scenario is considered and additional evidence shows it to be incorrect, the next scenario that is considered may not differ much from the first one. This is because of anchoring. The truth may not resemble either the first or second scenario at all, but the tendency of the human mind to anchor additional hypotheses to the first one will limit the considered possibilities.
Evidence can be either empirical or circumstantial. Empirical evidence is capable of being verified or disproved by observation or experiment while circumstantial evidence is in need of more information but gives a general direction for which observation and experiment can go. Both are admissible in the courtroom and are useful at looking at the world, but it is clear in science one must realize the difference. Circumstantial evidence guides us but is not proof so that it is always open to interpretation.
It is clear that the idea of negative corpus as a legitimate means of investigation was being abused. Something had to be done. The elimination of the negative corpus process as a legitimate means of concluding arson is the best practice. Proof is scientific. Faith or conjecture is not.
As to your third point:
Your own organization disagrees with you as NAFI clearly proclaims: “NAFI believes that NFPA 921 is the professional standard of care for fire and explosion investigation as the only peer-reviewed, science based, consensus text on fire investigation.”
For your sake as an investigator, and in the age of the internet where there is a lot of information available to anyone who simply enters your name and keyword terms and therefore can find this blog post, I hope you did not mean as you wrote as to this point. If so, you are going to face a very long deposition or cross-examination one day because legally if you believe that either the document is simply a “guide” and therefore no standard exists in your community (which is what you published here on the internet) or that you follow no standard in your particular investigation despite NFPA 921 being your organization’s acknowledged standard, then you are opening yourself up to a very easy Daubert/Frye challenge in terms of the admissibility of your particular technique of investigation. If you acknowledge that either NFPA 921 is not a standard (and therefore a standard does not exist) or you personally follow no standard, but rather a guide, then there is no method of scientific-based analysis (Daubert) and/or there is nothing in what you specifically do that is generally accepted in any scientific community (Frye). Jeez, I hope in your passionate writing that this was an error. If so, post again, I will give you fair reply.
Again, thank you for your post.
Jay VandeBerg says:
Mr. McShane,
I stumbled on your website as a result of a Google search for “negative corpus”. I believe your arguments and the arguments of Fire Investigator Mason both have some validity to them.
While preparing a case for a class, I read 18.6.5 Inappropriate Use of the Process of Elimination (NFPA 921, 2011, p. 174). While I understand the spirit of the passage, I pose some questions that I have asked my colleagues (and am awaiting answers) regarding the process of elimination. I am wondering what your opinion is, not as a criminal defense attorney, but as a practicing legal expert.
To determine a cause of a fire, “the first fuel ignited, the ignition source, the oxidizing agent, and the circumstances that resulted in the fire” (NFPA 921, 2011, p. 169) must be identified.
To simplify things, I will use one of F.I. Mason’s examples. A crumpled piece of paper, ignited by a flame source that is taken from the area of origin.
Pre 2011, I would have approached this situation by eliminating possible ignition sources. With no electrical sources, no evidence of lightning, no evidence of focused sunlight (you’d be surprised) and no other possible heat or ignition sources, except for the possibility of an open flame provided by the actions of a person.
Thinking scientific method, the only hypothesis that cannot be disproven is the application of open flame. If there is no lighter or match left in the hand of the offender, as witnessed by both a police officer and a fire investigator, is it expected that the cause will be required to be ruled “undetermined”?
A hypothesis of the application of open flame (regardless of source), the oxidant naturally consisting in the air, and the deliberate act of a person to bring the fuel, an oxidant, and a heat source together to light the fire could be argued by some as a non disputable scientific explanation that cannot be disproven.
Wondering your thoughts, and if you feel this gives defense attorneys an easy out to completely negate any investigation involving an origin and cause determination now with the 2011 edition of NFPA 921?
And to give you some of my background, I have been involved in Fire Investigations since 2005. I am certified by my state as a fire investigator, I am certified by PATC in Fire Pattern Recognition, I am trained by my state as an evidence technician, I have associates degrees in fire science and criminal justice, and just completed my bachelors in fire service management. I teach the basic firefighter program, including relevant topics such as fire behavior and evidence preservation/origin and cause awareness. I work entirely in the public sector.
Fire investigator Mason says:
To answer your questions in the last paragraph……
1) I do realize that 921 is a standard. And that is should be followed as closely as possible. However I do also realize that this standard is constanly changing. As soon as a version is published the next version, which may contain corrections is being written. I believe that in a future addition some changes will be made to the negative corpus addition. The title itself indicates that there is an appropriate use for the process of elimination, hence the title “the inapropriate use of the process of elimination.” If an inappropriate use of exsists, than also an appropraite use of also exists. However the document currently does not address the appropriate use of.
2) The title of the document 921 itself calls it a guide. It has been proven in court that deviations can be made from the document, as long as the investigator can explain why such deviations are made.
3) I do not believe that 921 is a criminal defense document. It is a well reputed document, that I as an investigator follow as closely as possible. What I am simpily stating is that, if you took the negative corpus addition word for word without deviation, no arsonist would ever be charged. How many proven arson fires, is the match or lighter left behind? What I am stating is that without further explanation to this section, it is a criminal defense attornies best friend. Do you understand the point I am trying to get across?