Real-life criminal investigations are not conducted like an episode of “CSI.” Toxicology results on a homicide victim are not produced within minutes and fingerprint analysts aren’t breaking down doors to arrest a suspect after finding his print at the crime scene.
Forensic examiners are generally not peace officers with guns and the power to arrest people; the roles are separate.
The most common exception to this is in fire investigation, which in the public sector is usually done by firefighters who get peace-officer powers after several weeks of training. Often the person looking at fire patterns and collecting samples to test for accelerant also is interviewing witnesses and suspects, making arrests and presenting cases in court.
Is there something wrong with this model?
A small portion of the fire investigation profession thinks there is, including the supervisor of the newly formed Stanislaus Regional Fire Investigation Unit, Lt. Dave Hutchinson.
“We want both disciplines working to their core strength,” he said.
“Just like ... if the DoubleTree Hotel caught on fire we wouldn’t send four patrol officers to go grab a hose and a breathing apparatus and run up to the top floor,” he said. “It’s fine to have fire investigators from the fire side to investigate origin and cause ... but when it comes to the court process, the legal process, search warrants, interviews and interrogation, dealing with juveniles, dealing with attorneys, case law updates and all of that (which) we are so used to in law enforcement, unfortunately the fire investigators were left to their own devices and they just weren’t getting that information.”
Hutchinson said the goal also is to mitigate cognitive bias, in which statements gathered during the criminal investigation affect the way the forensic investigation is done.
He gives the example of a fire investigator arriving at a house fire and interviewing a witness who says he overheard the homeowner’s boyfriend say he was going to burn her out of her house the night before.
“No matter how hard you try to get that out of your mind, it has imparted bias on you so subconsciously you are going to be looking for things that fit that expectation that this was a crime,” Hutchinson said. “You might ignore things that tend to point that this was an accidental fire.”
So the Stanislaus County District Attorney’s Office, working with top officials in local law enforcement and fire agencies, created a new model for fire investigation: a partnership between law enforcement and fire personnel.
In October, the county Board of Supervisors approved the Stanislaus Regional Fire Investigation Unit, changing the way fire investigation had been done in the county for decades and straying from a model still used nearly everywhere else.
The fire investigators are tasked with determining the origin and cause of a fire, while a criminal investigator follows up with interviewing witnesses and potential suspects, and preparing a case when arson is suspected.
Hutchinson, a lieutenant with the District Attorney’s Office, supervises the unit composed of a criminal investigator with the District Attorney’s Office and four fire investigators – two from the Modesto Fire Department, one from Ceres Fire Department and one from Stanislaus Consolidated Fire Department. Turlock is the only city not covered by the Stanislaus Regional FIU.
The unit is funded by staffing and other resources contributed by partner agencies and a financial contribution from the less-than-countywide fire tax, which is associated with property tax and used for fire services such as investigation, prevention, communication and training.
It took five years to get the unit off the ground because of funding issues during the recession and a shift in organization with the disbanding of the Modesto Regional Fire Authority.
During that time, former Modesto landlord George Souliotes was freed after spending 16 years behind bars for arson and murder in the the deaths of his tenant and her two young children in a house fire in 1997.
In 2013, a federal judge overturned the conviction, ruling that Souliotes had been convicted based on outdated fire science.
The judge ordered the district attorney to release him or begin a new trial within a few months without the testimony of the fire investigators or of an eyewitness who put Souliotes and his RV at the scene.
Ultimately Souliotes accepted a plea deal from prosecutors for three counts of involuntary manslaughter for failing to maintain smoke detectors in the rental home and was released with time served.
Hutchinson wouldn’t comment on the case, citing a pending civil suit brought by Souliotes, but District Attorney Birgit Fladager maintained in a Modesto Bee op-ed at the time of Souliotes’ release in 2013 that he is guilty and that all the evidence used to convict him 13 years before should have been allowed in the new trial.
Paul Bieber, founder of the San Jose-based Arson Research Project who worked as an investigator for Souliotes’ defense team, believes that if the new fire investigation model were in place during that deadly fire, the outcome of the investigation would have been much different and there likely wouldn’t have been an arrest.
Coverage of the Souliotes case most often has focused on the outdated fire science that was used to determine the fire was arson, but cognitive bias also played a large role, Bieber said.
When the fire investigator, since retired, arrived, the home was still burning. He spoke to a witness who reported seeing an RV that left just before the fire broke out and took note of a “For sale” sign in the yard.
The investigator called his brother-in-law, a real estate agent, to learn the owner of the home was Souliotes. When Souliotes showed up at the scene, the fire investigator asked the Modesto Police Department detective on the case whether he matched the description of the man the witness described and if he could drive by Souliotes’ home to see if he had an RV.
“This fire investigator, just like most (cases) leading to wrongful convictions, knew who started the fire before he ever stepped foot in the house,” Bieber said. “The easiest way to hijack any forensic examination is to tell the examiner the answer before they ever ask the question.”
He gives the example of a fingerprint expert testifying to finding the suspect’s fingerprint on the handle of a gun used in a crime.
“Imagine if the next question asked was ‘What was the defendant’s mood when he touched the gun? Did he touch it on purpose or by accident?’ ” he said. “Why is that different in fire investigation?”
He said the fire investigator can determine a fire started on a couch by a lit cigarette but shouldn’t then decide whether it arrived there intentionally or accidentally.
“We don’t need to fix the investigator, we need to fix the process, and that is where Dave Hutchinson is really unique,” Bieber said.
The National Fire Protection Association, a private group that is the leading authority on fire investigation, published a guide in 1992 that debunked techniques used to determine if a fire was accelerant-propelled.
But it took years for fire investigators to embrace the findings – some of the techniques were used in the Souliotes case – and the association’s finding about bias will be met with the same resistance.
The association first mentioned the need to mitigate bias in its 1992 guide.
It was a single sentence then, but the most recent direction in the 2014 publication expands on the topic. It explains the danger of reaching premature conclusions and using the scientific method to prove a hypothesis, effectively failing to consider alternative hypotheses or discounting contradictory data.
The association says it is important to use the scientific method properly but doesn’t offer suggestions on how to avoid presumptions or premature conclusions.
Hutchinson believes the best way to accomplish that is with this new model, even if it is counterculture.
He stressed, though, that the former model did not lead to innocent people going to jail; if anything, potentially guilty people were getting away because fire investigators didn’t have the help and resources to follow up on as many cases.
Investigator Ed Bartley acknowledged he and other fire investigators needed help so they could focus on their expertise, origin and cause, and the law enforcement could work to their strengths by conducting interviews, writing warrants and conducting surveillance.
Bartley said on busy days they would sometimes have back-to-back fires and no time for investigations. They were hesitant to ask for help from patrol officers or even follow up with detectives because they weren’t on the case from the beginning and weren’t invested in it.
“With the model the way it is now, it frees us up more for vetting evidence and getting hypothesis,” he said.
In its first eight months, the unit responded to 442 incidents resulting in 160 full investigations. Twenty-three cases were sent to the district attorney, who filed charges in 19 of them. So far four have resulted in prison time for the defendants, four in jail time, four were committed to mental hospitals and seven cases are still pending.
One case involved a woman who started her downtown Modesto apartment on fire because of a fight with another tenant, and in another a man caused a series of structure and vegetation fires in west Modesto. Both pleaded guilty, and each were sentenced to three years in prison.
The boards that govern the involved agencies – the county, Modesto, Ceres and Stanislaus Consolidated Fire Protection District – all agreed in July that the model is working, signing a memorandum of understanding for the continuation of the Fire Investigation Unit.