Modern fire investigation is built on a simple idea: conclusions about how a fire started must be reached through a systematic, testable process rather than intuition, tradition or "rules of thumb." In the United States, that process is described in NFPA 921, Guide for Fire and Explosion Investigations. Whether you work for a fire department, a law enforcement agency or a private firm, understanding NFPA 921 and the scientific method it describes is essential to producing opinions that hold up to peer review and courtroom scrutiny.
What Is NFPA 921?
NFPA 921 is a guide published by the National Fire Protection Association for the investigation and analysis of fire and explosion incidents. It is intended for both public-sector investigators and private-sector professionals who conduct investigations for insurance companies or litigation. It covers topics such as fire science, fire patterns, building systems, electricity and fire, sources of information, planning the investigation, documentation, evidence, origin determination, cause determination, failure analysis, and specialized incidents such as vehicle fires, wildfires and explosions.
The guide is revised on a regular cycle. The 2024 edition, for example, added guidance on an investigator's potential confirmation bias, opinions and expressions of certainty; addressed how fire suppression activities can affect ventilation and flow paths away from the area of origin; relocated fire effects on electrical systems to Chapter 6; added four general classifications of fire cause; expanded guidance on investigator safety, health and wellness; expanded text on hybrid and electric vehicle fires; and revised the wildfire chapter. Investigators should work from a current edition and be ready to explain which edition they relied on.
NFPA 921 works hand in hand with NFPA 1033, the professional qualifications standard, which requires investigators to employ all elements of the scientific method as the operating analytical process. Our certification guide explains NFPA 1033 in more detail.
The Scientific Method, Step by Step
NFPA 921 describes the scientific method as a sequence of steps. In practice the steps overlap and the investigator often loops back as new data arrive, but the logic remains the same.
1. Recognize the need
A fire has occurred and there is a need to determine its origin and cause, whether for prevention, criminal prosecution, civil litigation or product safety.
2. Define the problem
The investigator decides how the problem will be solved: securing the scene, planning the examination, identifying resources and experts, and understanding any legal constraints on access.
3. Collect data
Data include fire patterns, burn damage, witness statements, first responder observations, photographs, building and system information, weather, alarm and camera records, and evidence recovered from the scene. Data collection must be thorough before conclusions are drawn. Useful skills here include fire scene photography, interviewing and evidence collection.
4. Analyze the data
Analysis is based on the investigator's knowledge, training, experience and expertise in fire science. It includes interpreting fire patterns, applying fire dynamics, considering the heat release and fuel load in the compartment, and evaluating arc mapping or other engineering data.
5. Develop hypotheses (inductive reasoning)
Based on the analysis, the investigator develops one or more hypotheses to explain the origin and then the cause of the fire. Each hypothesis should be grounded in the data rather than assumed.
6. Test hypotheses (deductive reasoning)
Each hypothesis is tested by comparing it against all known facts and relevant scientific knowledge. Testing can be cognitive, such as asking whether a proposed ignition source could produce enough energy to ignite the first fuel, or experimental, such as laboratory testing or reconstruction. A hypothesis that cannot withstand testing is discarded or revised.
7. Select the final hypothesis
Only a hypothesis that has been tested and survives is selected. If no hypothesis can be adequately supported, the correct answer may be that the cause is undetermined.
Origin First, Then Cause
A reliable cause determination depends on a correct origin determination. NFPA 921 describes origin as the general location where the fire began, determined using witness information, fire patterns, arc mapping and fire dynamics analysis, often in combination. Investigators should be aware that ventilation, flashover and full-room involvement can create patterns far from the origin, and that suppression activities can alter flow paths. Clearing debris systematically, documenting layers and reconstructing the room before reaching conclusions are all part of disciplined origin work.
Once the origin is established, the investigator identifies the ignition source, the first fuel ignited, the oxidant and the circumstances that brought them together. That combination is the fire's cause.
Classifying the Cause
NFPA 921 uses four cause classifications:
- Accidental: no deliberate human act was involved.
- Natural: caused without direct human intervention, such as lightning.
- Incendiary: deliberately set under circumstances in which the person knew the fire should not be ignited.
- Undetermined: the cause cannot be proven to an acceptable level of certainty.
"Undetermined" is not a failure. It is an honest statement that the data do not support a single tested hypothesis, and it is often the most defensible conclusion when evidence has been destroyed.
Avoiding Bias and "Negative Corpus"
NFPA 921 cautions investigators against expectation bias, which occurs when an investigator reaches a premature conclusion without examining all relevant data, and confirmation bias, which is the tendency to look only for data that support a favored hypothesis. Practical safeguards include:
- Considering alternative hypotheses deliberately and documenting why each was accepted or rejected.
- Separating what you observed from what you were told.
- Requesting technical or peer review of your analysis before issuing a final report.
- Being cautious with information that is not needed for the scientific analysis, such as unverified suspicions about a person.
The guide also cautions against "negative corpus," the practice of declaring a fire incendiary simply because all accidental causes identified have been eliminated, when there is no evidence to support the incendiary hypothesis itself. Every determination, including an incendiary one, must be supported by evidence that can be tested.
Documenting the Process
Your report should allow another qualified investigator to follow your reasoning: what data you collected, what hypotheses you considered, how they were tested and why the final hypothesis was selected. Clear documentation also makes it far easier to explain your work in a deposition or trial. Our article on expert testimony covers how courts evaluate the reliability of methodology.
Putting It Into Practice
Applying the scientific method consistently takes practice. Conference sessions, burn cell exercises and peer discussion are among the best ways to sharpen those skills. Check the Training page for the next Utah Chapter IAAI conference, browse more investigator resources, and consider joining the chapter to learn alongside experienced investigators.
This article is general educational information. Always refer to the current edition of NFPA 921, NFPA 1033, your agency's policies and applicable law.