by Edward Humes
Ablott took that advice and ended up graduating from the sheriff’s academy with the next class of trainees. After a requisite tour of guard duty at the county jail, where all sheriff newbies begin, he worked his way up from patrol officer to detective rank, joined the homicide unit for ten years, then moved into the arson and explosives detail. There his friend’s dad, Bob Bishop, became one of his mentors, part of an elite group of fire investigators who had been working the job since the 1940s and ’50s. By the time of the Parks fire, Ablott had done thousands of origin and cause investigations, and had been qualified to testify as a fire expert in more than a hundred court hearings and criminal trials.
On April 17, 1989, eight days after the fire, Ablott and his partner made their first trip to the burnt-out shell of the Parks apartment. The Bell Police Department, sufficiently convinced by Kathy Dodge that they might have a murder-by-arson case on their hands, had requested help from the sheriff’s arson unit.
They were met outside the apartment by Bell Police Sergeant William Talbott and Captain William Franklin of the Los Angeles County Fire Department, the initial fire investigator on the case, who had photographed the scene on the day of the fire. Franklin explained to Ablott how he and a novice investigator, Captain Lester “Fuzzy” Fuzell Sr., had used digging tools to excavate some electrical cords covered by a thick coating of ash that hose water had turned into a slurry, now dried. He said he was particularly concerned about the end of an extension cord that had several appliances plugged into it—the “octopus,” as he called it—that was located near the largest V-pattern in the living room, beneath the windows. Franklin said he believed that was where the fire could have started—an accidental electrical fire. He had photographed the octopus of cords, then left them near where they were found. The photos show a couple of small cuts in the cord near the octopus, none of them wrapped in drapery, in an area where shovels had been used during overhaul.
Talbott then briefed the sheriff’s fire experts on Kathy Dodge’s allegations and on the fire a year ago in Lynwood that also involved an electrical cord. He had to remind Ablott that he had investigated that first Parks fire as well, and that he had ruled it accidental.
Decades later, Cohen would argue that, in this briefing, bias contaminated the fire scene investigation at its very start, leaving its findings questionable at best, fatally flawed at worst.
She did not mean bias in the common sense of the word, as a willful and malicious state of mind, synonymous with conscious racial, gender, or religious prejudice. Cohen would suggest an entirely different cognitive bias, a normal, instinctive, subconscious function of the brain. It is not an insult to be accused of cognitive bias. Everyone experiences it in some form every day, as part of the brain’s pattern-recognition system that allows us to anticipate and prepare for events before they occur. One form, expectation bias, evolved as a survival mechanism—the brain filling in blanks in perception and information with what we expect from past experiences and other clues. But during subjective and abstract tasks, such as reading the meaning of fire patterns, the sort of information Franklin and Talbott provided can lead the human brain to perceive the patterns that a person expects to find, or that confirm an existing opinion—the closely related confirmation bias. This brain function explains why most people are so accepting of information—from news reports to political speeches—that supports what they already believe, while feeling deeply skeptical of even the most well-researched and verifiable information that challenges our existing assumptions and beliefs. Resisting such tendencies, keeping an open mind rather than succumbing to “tunnel vision,” is a difficult challenge for anyone, but especially for expert investigators, who tend to be extremely confident in their opinions and abilities.
This common form of cognitive bias is the reason research scientists and medical researchers construct “blind” studies, so that neither their test subjects nor the researchers themselves can be influenced by knowing, for example, which subjects receive a medication and which are imbibing a placebo. Although the virtues of blind scientific studies have been understood for centuries, the pitfalls of not working “in the blind” in criminal investigations only became clear in recent years, with the FBI’s disastrous examination of fingerprints related to the terrorist bombings in 2004 of four Madrid commuter trains that killed 192 people and injured more than 2,000. Three FBI fingerprint examiners separately concluded that a bag of detonators used by the bombers held fingerprints matching a Muslim American attorney in Oregon who had previously defended Al Qaeda terrorism suspects. Brandon Mayfield spent seventeen days in jail before Spanish authorities identified an Algerian terrorist as the true source of the prints. The Justice Department later concluded that “confirmation bias” and a “loss of objectivity” led examiners to see in the fingerprint comparison “similarities . . . that were not in fact present.”
Divining the meaning of burn patterns at a fire scene requires even more subjective judgment than fingerprint comparison. In the Parks fire, Franklin, who initially saw an accidental fire, knew nothing about Kathy Dodge or Jo Ann Parks at the time he examined the physical evidence at the scene. Ablott, however, recalls walking into the house and immediately seeing signs of intent—of seeing that “something was wrong.” A prosecution expert working to uphold the conviction decades after the crime asserted that Ablott also was likely already thinking during that first walk-through that the first nonfatal house fire in Lynwood had been arson, too, but that Jo Ann had fooled him into ruling it an accident.
The forensics community is divided on what can or should be done about the issue of cognitive bias. Those who favor a scientific approach believe investigators trying to determine the origin and cause of a fire should examine the scene without knowing anything about the personalities, motives, or suspicions in the case. That way the physical evidence can be interpreted free of bias. Why, they ask, does a fire expert need to know someone has been accused of being a poor parent in order to interpret a burn pattern? Why take the risk of bias creeping in?
Those who, like Ablott, come to forensics through detective work rather than lab training argue that such a blind approach is impractical—arson investigators must work the scene but also work the people, just like any detective. Omitting the human element, according to this view, would rob investigators of insights they need in order to know where to look and what to look for at a fire scene. Ablott needed to know about the previous fire, according to this view of the issue, because it alerted him to look more closely for signs of arson. Arson detectives say two house fires involving the same victims in the space of a year is a reliable indicator that they were not accidental. The scientists say the forensic analysis of each scene has to stand on its own; making assumptions about the physical evidence in one fire simply because there was an earlier fire is a demonstration not of acumen, but of bias. It is akin to the “Gambler’s Fallacy,” which is the false belief that in a game of chance, such as dicing, the odds of a winning roll increase after one or more losing rolls—the gambler assumes he “is due.” In fact, prior dice rolls have no influence at all on future rolls—the mathematical probabilities on any given roll never vary. Likewise, the existence of an earlier fire may seem suspicious, but that doesn’t change the nature of the physical evidence at the second fire, nor should it alter the scientific analysis of either fire scene.
Ablott would later say he resisted concluding that the Parks fire had been intentional for as long as he could, but that he finally had to accept the terrible reality of what the evidence revealed to him. Cohen asserts Ablott’s perspective of “resisting the inevitable” before declaring Parks guilty is the very definition of cognitive bias; Ablott says his cautious deliberation before reaching a final conclusion demonstrates his unbiased approach.
“I just did not believe in my mind, coming from the family in which I was raised, that anybody could do that to their children,” the seasoned former homicide detective w
ould say years later. “It was just hard for me to fathom, but I finally had to conclude that was what happened.”
Ablott, who retired from the sheriff’s department in 1998 but still works as a private fire scene investigator, says he has always tried to avoid errors and bias by examining every fire scene in the same systematic way. He walks and examines the outside first, snapping pictures in sequence, then moves inside to inspect room by room. He’s left-handed, and so he tried always to circle clockwise around each room, leading with his right side, so it felt a little uncomfortable for a lefty. He felt this kept him more alert to details and patterns that can indicate a larger area of origin for the fire, or allow him to narrow the field to a more specific point of origin.
During his walk-through of the Parks living room, he immediately seized on the power cords that Franklin had pointed out—the extension cord and other wires that had connected the TV, VCR, and fan to a single outlet in the corner. He found bits of charred drapery material wrapped around portions of the cord, most of which had been badly damaged in the blaze. Tracing the path of the cord along the living room wall, he found the remains of a blue plastic crate or basket and underneath it, a portion of the cord that had been protected from the flames. This was a section with small cuts in the cord’s insulation. No drapery wrapped this portion. Ablott decided the various “modifications” to the cord couldn’t have happened during the firefighting and overhaul with metal shovels and other tools. Instead, Ablott concluded that the fire probably had started with those wires—that insulation had been stripped away deliberately to cause a direct short circuit, and that the electrical system was purposely overloaded to heat up the damaged wire and start a fire. The draperies could then have become the fire’s first fuel, allowing the flames to take root and spread, he concluded. This initial finding was remarkably similar to what he opined in the Lynwood fire a year earlier, except he had deemed that blaze accidental and the damage to the cords caused by fire and heat.
For more than a year, Ablott maintained his opinion that a deliberate short circuit caused the fatal fire, even though it should have been obvious from the start that a short circuit with those cut cords was highly unlikely if not impossible. The cuts to the wires were so small that only one conductor was exposed at any one spot on the cords. But both metal wires would have to be bared and touching for there to be a short circuit, heat, and sparks.
Furthermore, there is no evidence that the circuit breakers in the house tripped before the fire—which they are designed to do, automatically cutting power when there is either a short circuit or an overload. This should have suggested to Ablott that either there was no short circuit or, if he were correct about a short, that the breakers malfunctioned. That, in turn, would introduce another possible cause for the fire—deficiencies in the apartment’s conditions. But despite the landlord’s history—cited for violations more than two dozen times in four different cities, including Bell—and despite the absence of legally required smoke alarms in the apartment, no investigation of the landlord’s maintenance practices was undertaken in the Parks investigation.
The only examination of the electrical system was done more than two months after the fire by an electrician hired by the landlord’s insurance company, who found nothing wrong. His findings in favor of those who were paying him were accepted without question by the authorities. An arson expert hired by the landlord’s insurer, a former Los Angeles Fire Department investigator named Steve Takach, was allowed to attend meetings with police investigators and to contribute his photos and opinions, which were so extreme that he concluded Parks had doused the girls’ bedroom with a flammable liquid such as gasoline in order to accelerate the blaze. (That flammable-liquid idea has been rejected by every other expert and investigator involved with the case; Takach apparently relied on the sort of false indicators of arson that John Lentini exposed in the Lime Street fire experiment.)
As he formed his own theory of electrical sabotage, Ablott discounted alternative accidental causes for the living room fire, including the Parkses’ old swap-meet television and other electrical appliances, recalling how he had found the TV outside the fire’s area of origin during his first visit to the house on April 17. Only that sabotaged wire was in the right place to have sparked the fatal fire, he wrote in his arson investigation report.
But in the summer of 1990, Ablott’s theory fell apart. That was when detectives pulled the wires out of evidence storage, got permission to x-ray them at the coroner’s lab with equipment intended for autopsies, then had a forensic electrical engineer look at the X-rays and the wiring in order to confirm Ablott’s opinion of deliberate short circuit. Instead, the engineer, William Armstrong, a former Jet Propulsion Laboratory rocket scientist given to wearing lab coats with the moniker MR. WIZARD on his name tag, said there was no way those wires started a fire. There had been no short circuit or overheating, just partial melting of the insulation from the heat of the house fire—burned or melted from the outside, rather than from the inside, as would occur during an electrical overload. Armstrong also confirmed that if someone deliberately wrapped draperies around the wires to start the fire, the saboteur had, oddly, neglected to wrap the parts of the wire that were cut. Which begs the question: If someone was constructing an incendiary device, wouldn’t the logical step be to wrap the exposed wires with the flammable drapery material?
Investigators and prosecutors responded by rebranding the cords, calling them a failed incendiary device. Armstrong’s report provided ammunition for that. Armstrong, whose expertise was electrical engineering, not edged weapons or differentiating cuts made with knives from those made by scissors or the edges of shovels, nevertheless rendered an opinion that the cuts had been done with a sharp knife. He also opined that a four-year-old child playing with a knife would certainly have cut himself in making such slices in the electrical cords. And none of the children had cuts on their hands. These opinions, though seemingly outside Armstrong’s professional expertise, became key to the prosecution’s argument that the wiring was sabotaged by Jo Ann Parks, despite its inability to actually start a fire. This nonexpert testimony from an electrical expert became gospel in the case, never seriously challenged when, arguably, it never should have been allowed in the first place.
With his original theory on the cause of the living room fire disproved, other experts who later reviewed the case on Parks’s behalf have said Ablott should have reconsidered his area-of-origin determination at that point, given that the wires were at its center. Or he could have reconsidered other possible causes, such as the old television he had discounted at first. Ablott chose a different course. He used the process of negative corpus—after first assuming the absence of possible accidental causes—in order to allege that Parks had to have started the fire with matches or a lighter, because there was no other explanation.
A quarter century later, Cohen would argue that Ablott misused negative corpus, improperly discounted the TV set as a possible fire starter, then falsely testified that all possible accidental causes had been eliminated.
During Ablott’s first walk-through of the house, he also found some unusual fire patterns in the girls’ bedroom that showed flames burning low in the room next to RoAnn’s bed, then moving out of the room into the kitchen. Meanwhile, patterns high on the wall showed fire moving into the room from the kitchen. There are numerous possible reasons for such conflicting fire patterns, including a separate fire set deliberately in the bedroom on the floor, or a separate low fire started by flammable ceiling tiles catching fire and falling to the floor near the bed. The window failing and providing new ventilation, triggering flashover, could also play a role in creating such patterns as the air currents churn and swirl.
In a handwritten report dated June 20, 1989, Ablott concluded that “investigators were unable to establish this area as an area of origin since the damage could have been caused by a secondary fire as a result of radiated heat and fl
ame impingement from the living room fire.”
A typewritten version of the same report, also dated June 20, 1989, included some key edits. Despite the identical dates on the reports, Ablott said the typed version was actually finished in July, a few weeks after the handwritten version. Early that month, Ablott visited the Parks fire scene again with Bell police officers, fire investigators, and the insurance investigator Takach, who was adamant that an arson fire had started in the girls’ bedroom and that the landlords therefore had no liability for the fatal blaze. Ablott’s typed version of the report agreed with Takach: “Investigators established this as a separate area of origin and was caused by the application of an open flame.”
All of these opinions on where and how the fire (or fires) started in the apartment rested on a single keystone finding of Ablott’s: He testified that flashover did not occur in any room of the Parks house, which, if true, would make reading the fire patterns much simpler. Yet every other expert who has looked at the case since, including those hired by the prosecution, says flashover can and does create false indications of areas of origin and should always be taken into consideration in a fire investigation. Those experts include John Lentini, who proved this effect of flashover with his experiments in the Lime Street fire and his observation of false indicators of arson in the Oakland firestorm.
And in the Parks fire, all of those same experts now say flashover occurred in the living room and both back bedrooms where the children’s bodies were found.
Even Dinko Bozanich, the prosecutor who tried Parks and who insisted there was no flashover that night, believed a contrary finding could devastate his case. “If a flashover occurred,” he told the judge in arguing to limit one witness’s testimony about the topic, “that changes everything with regard to the arson investigation presented in this case so far.”
