Minimizing Cognitive Bias in Fire Investigation Reports

When you enlist a forensic expert to analyze the aftermath of a fire, you’d ideally expect them to be impartial in their investigation. However, nobody is completely immune to drawing initial conclusions. Investigative bias has been recognized in the field for as long as fire investigators have existed.

This guide explains how fire investigators keep cognitive bias out of their reports.

Types of Cognitive Bias#

When dealing with fires that may or may not have been intentional, investigators must contend with two distinct types of cognitive bias. These biases, when left unchecked, can impact the decisions the investigator makes and affect the conclusions they draw.

Expectation Bias

When a forensic fire analyst draws a conclusion before they’ve gathered all available evidence, they are exhibiting an expectation bias. It’s also commonly called a “preconceived notion.” For example, an investigator may give a cursory glance at the scene in question and assume that the fire was accidental.

Expectation bias can stop an investigation in its tracks. If investigators don’t collect crucial data because of an assumption, they often can’t go back and collect it later. Failing to properly investigate a fire scene due to expectation bias also violates the scientific method, as they are not adequately testing the hypothesis.

Confirmation Bias

This type of bias generally occurs during the testing process and can also halt an investigation. When a fire investigator happens upon a single theory and believes it to be true, they often fail to collect data that may contradict their theory.

Adequate fire investigation methods require testing for many different ignition styles to confirm or rule them out. However, if a forensic expert strongly believes that the fire started in a particular way, they may only focus on data that supports their theory.

NFPA 921, the gold-standard guide for fire investigations, states that they must test hypotheses with the intent of disproving them. Confirmation bias does the opposite: it puts blinders on the investigator and encourages them to test a single hypothesis with the intent of proving it correct.

Collecting All Available Data#

How can a forensic fire analysis expert push past their innate biases? The answer is simple: collect all the data and evidence, not just that which may prove them right. Forensic investigators must think scientifically and be willing to be proven wrong.

There are several types of evidence that fire investigators often collect, generally under one of the following three categories.

Demonstrative Evidence

Physical (real) evidence: the tangible items recovered from the scene (appliance and equipment remains, conductors and arc artifacts, receptacles, debris and comparison samples). NFPA 921 calls for physical evidence to be documented in place before it is moved, collected and packaged so that it is neither altered nor contaminated, and tracked through an unbroken chain of custody. Demonstrative evidence is different: it is the diagrams, scaled models, photographs, 3D captures and animations prepared to explain testimony, and it is admitted on a showing that it fairly and accurately represents what it depicts rather than on chain of custody.

Maps of the area impacted by the fire and accurate models and photographs of the damage are also considered demonstrative evidence.

Documentary Evidence

As the name implies, this type of evidence is on paper. Fire insurance policies, sales receipts, and the fire investigator’s notes all comprise documentary evidence. An unbiased investigator will note every piece of data, regardless of whether it supports their suspicions.

Testimonial Evidence

This form of evidence is verbal and given by a competent witness who is relevant to the investigation. Witness testimony is generally taken under oath, and forensic fire investigators are often called as expert witnesses in the courtroom.

Statements taken from witnesses to the fire also fall under this category, as eyewitness testimony is crucial to any fire investigation.

Analyzing and Corroborating Available Data#

Collecting all the data available is a crucial preliminary step for fire investigators to avoid drawing biased conclusions. When they have access to a large amount of data, they can test different theories and disprove the ones that don’t have evidence to back them up.

A fully qualified fire investigator will test all plausible theories, regardless of any cognitive bias they brought in with them. They will collect as much data as possible and pay close attention to every fact, as even the smallest detail can overturn an investigation.

For Example…

Let’s say a house fire looks like it started in the laundry room, right by the clothes dryer. A closer look at the dryer finds that the lint screen and nearby vents were severely clogged with lint and debris before the fire started. A biased forensic expert may unofficially end their investigation right there: lint is highly flammable, and it’s plausible that the clog could have started the fire.

Meanwhile, an unbiased investigator will keep collecting data and may follow up with the local utility company. What if the local power company reports that the homeowners in question had their power shut off due to non-payment? The dryer couldn’t have been running at the time of the fire, so investigators must keep testing theories about how the fire really started.

Thinking Like a Scientist#

Forensic fire analysts are scientists and must conduct their investigations according to the scientific method. To avoid cognitive biases during the investigation, forensic experts use the following set of steps:

• Define the problem: There’s been a fire, and investigators need to find its origin.
• Collect data: As discussed above, forensic experts must collect all available evidence.
• Formulate a hypothesis: Based on the evidence gathered, investigators put together a theory of how the fire could have started.
• Test the hypothesis: Fire investigators must test their theories with the intent of proving them wrong.
• Revise the hypothesis: If testing proves the initial hypothesis wrong, investigators edit their theory to re-test.
• Analyze the data (inductive reasoning): interpret the collected data on the basis of knowledge, training, experience and expertise before forming any hypothesis. Select the final hypothesis: the hypothesis that withstands testing against all of the data becomes the conclusion, reported with the level of certainty the data actually support. If no hypothesis survives testing, or two or more remain equally supported, the correct result is “undetermined,” reported as such rather than resolved by choosing the most likely-sounding option.

Unbiased forensic fire experts often go through multiple incorrect hypotheses before arriving at the correct one. Continuous revision of hypotheses and corroboration of data help strengthen their ultimate conclusion.

To minimize cognitive bias as they examine fire scenes and collect data, fire investigators have to think and work like scientists. Narrowing their focus to the facts, not their feelings, helps to mitigate the effects of expectation and confirmation bias in fire investigations.

Dreiym Engineering employs a talented and detail-oriented team of forensic engineers who can help you locate the source of a traumatic fire. Our experts aim to collect as much accurate data as possible by getting to the scene quickly and documenting every inch of it. If you need a team of experienced, unbiased investigators to determine how a fire started, contact Dreiym today!