After the murder of George Floyd, some states looked to independent agencies to examine deaths in police custody. But dozens of cases handled by the Texas Rangers show the approach has flaws.
When the Texas Rangers learned that a woman had died in a jail south of Dallas, they put Adam Russell on the case.
He found that there had been a struggle between the woman, Kelli Leanne Page, 46, who was being held on drug charges, and two guards, who entered her cell because she would not stop banging a hairbrush against the door.
One jailer threw her to the floor, punched her in the face while they scuffled and piled atop her as blood streamed from her nose. The other, a trainee weighing 390 pounds, pinned her down until she stopped breathing.
Six hours into his investigation, Ranger Russell indicated in his notes, he was not inclined to blame the guards for her death. And when an autopsy later determined that Ms. Page was the victim of a homicide — having died on Oct. 8, 2017, of a form of asphyxiation — Ranger Russell appeared not to reconsider.
Instead, he obtained a second opinion from a retired chief medical examiner, who read the forensic report and said he believed that heart disease might have led to her death while she was being restrained. Ranger Russell later testified that the initial autopsy was a rush to judgment and that “something inside Kelli” had killed her. The guards were not charged.
For many years, Texas has made its state investigators, the fabled Rangers, available to review deaths that occur in the custody of local authorities, filling a need that is especially pressing in rural areas. After the murder of George Floyd last year by a Minneapolis police officer, at least seven other states have embraced a similar approach to Texas’, reasoning that outside inquiries are more likely to hold wrongdoers accountable.
But state agents do not necessarily lead to better investigations or greater accountability, according to a New York Times examination of the record in Texas. Drawing on dozens of interviews and more than 6,000 pages of investigative files, autopsy reports, police records and court filings, The Times found that state investigations can be afflicted by the same shortcuts and pro-police biases that outside interventions are meant to eliminate.
State investigators in Texas review deaths in custody more often than in any other state, records show, making it a rich laboratory for studying how the investigations play out. The Rangers, in response to a public records request, provided The Times with roughly 300 case files completed since 2015, while other documents were obtained from local police departments, medical examiners and court proceedings.
Some of the Rangers’ investigations offer a textbook example of dogged police work, like the one into the overdose death of James Dean Davis, 42, in the La Salle County jail in 2017. Ranger Randy Garcia meticulously documented the neglect by guards who mocked Mr. Davis as he screamed for help from the floor of his cell, interviewing more than a dozen witnesses and reviewing video footage and audio recordings — and securing indictments against two jailers for tampering with government records.
In other instances, the Rangers fell short of basic standards. They did not speak to all relevant witnesses, delegated investigative tasks to the agencies under review and failed to follow up on signs that officers were negligent or behaving dangerously. In the death of Ms. Page, Mr. Russell sided with the two guards over 10 pathologists who conduct autopsies in Dallas County, home to one of the state’s largest medical examiners’ offices.
The Times shared its findings with a half-dozen veteran homicide detectives and policing experts in six states, all of whom emphasized that death investigations range widely in difficulty and circumstances. When officers shoot and kill someone, for example, many of the facts are not in dispute, particularly the manner of death, and the pressure mostly falls on prosecutors to decide whether to treat the killing as a criminal act.
It gets more complicated when no shots are fired, they said, and there is a struggle in which the person in custody stops breathing. A deep examination of those investigations offers hints about the thoroughness of the outside police work because state investigators must retrace how officers used their hands, feet and body weight at every turn — and determine whether those actions were appropriate.
The Times identified 29 cases the Rangers had investigated since 2015 in which a person stopped breathing after struggling with local authorities. None of those inquiries led prosecutors to charge anyone in law enforcement. In two-thirds of the cases, The Times found shortcuts, missteps or judgment calls that some veteran homicide detectives said might indicate a lack of effort on the Rangers’ part. For example:
After Genaro Rocha II, 47, died in an Amarillo jail in 2019, bound in a harness and left in a cell because guards said he’d kicked them, the Ranger did not record a single interview in his case file.
The Ranger who reviewed the 2018 death of Andrew Carmona, 36, east of San Antonio, started his investigation 11 days afterward, because local officials told him they did not need him there right away. He never visited the scene, a front yard in which an officer had held Mr. Carmona by the head and neck because he was acting “frantically.” And he conducted one interview — with a toxicologist.
The Ranger investigating the death of Michael Cassel, 41, in Tyler County in 2016 provided video footage to lawyers for the sheriff’s deputies involved before taking their statements. The deputies, who had struggled with Mr. Cassel in the woods near a road, heard him complain that he could not breathe before he lost consciousness and died, records show.
The Texas Rangers who handled the various cases declined or did not respond to requests for interviews, and the Rangers’ parent agency, the Department of Public Safety, would not make any official available to answer questions. A spokeswoman did not respond to multiple requests for comment.
It is a reality of policing in Texas and elsewhere that people sometimes die in custody, through no fault of the arresting officers. When officers do cross the line, investigators play an important role in holding them responsible. But the cases are often a lower priority than other duties because of the many demands placed on police agencies and the general reluctance among law enforcement officials to assign blame to their own.
“I guarantee you this is not a sought-after assignment for the Texas Rangers,” said Adam Bercovici, a former homicide lieutenant for the Los Angeles Police Department who works as a consultant and reviewed case files for The Times. “Nobody wants them, because 99 percent of the time it’s just an unfortunate set of circumstances. But it’s a lot of work to dot all the I’s and cross the T’s.”
There was no shortage of facts to examine in the death of Ms. Page. Guards at the jail had a history of misconduct, court records show, including one who put pepper spray in an inmate’s food, landing him in the hospital. The day before Ms. Page’s fatal encounter with two other guards, she angered one of them by splashing him with a cleaning solution through the food slot in her cell. She suffered a black eye when he restrained her.
Surveillance cameras captured video but no audio of her death, and Ranger Russell made no mention in his report of interviewing witnesses.
None of the medical examiners involved in Ms. Page’s autopsy were called to testify at a hearing about the death, and when the chief pathologist later asked for permission to speak to The Times, the official who had presided — the Coryell County justice of the peace — refused to grant it. The pathologist said in a brief statement that the doctors stood by their ruling: homicide by mechanical asphyxiation.
All of it combined to leave the impression that Ms. Page, who had five children and had long battled periods of sadness with drugs and alcohol, was just another addict who had died in jail, said one of her daughters, Tiffany Gruwell. When Ms. Gruwell thinks of her mother, a different image comes to mind, she said: one of a grandmother, clean, sober, beaming, cradling Ms. Gruwell’s daughter Mari.
“She’s not just her police record, but so many people look at it that way,” Ms. Gruwell said. “I feel like there are a lot of unanswered questions about what happened.”
A Patchwork Approach
In the past decade, flash points of police violence across the country have focused attention not only on aggressive tactics and racial disparities in law enforcement but also on the patchwork way in which the encounters are investigated.
Some states, like California, have rarely sent their agents to review in-custody deaths, leaving those inquiries to local police and sheriff’s departments.
Washington State has recently relied on a hybrid model, dispatching teams of local and state investigators. Officers must recuse themselves if their team is investigating a home agency.
Other states like Georgia and North Carolina deploy state agencies by longstanding custom rather than legal requirement when local departments request help. Texas operates much the same way, although the law since 2017 has mandated independent investigations of jail deaths.
In 2014, Wisconsin became the first state to require independent investigations of deaths involving police officers, spurred by family members of Michael Bell Jr., who had been killed by the Kenosha police. The death of Eric Garner on Staten Island that year led to an executive order requiring the New York attorney general to investigate certain in-custody deaths.
The push to mandate independent reviews has grown since Mr. Floyd’s murder in 2020. At least seven states have passed measures requiring local agencies to turn over in-custody death inquiries to state officials or other outside investigators. California’s law applies to shooting deaths of unarmed civilians, while New York’s expanded to require attorney general investigations of any deaths possibly caused by law enforcement. Maine and Maryland have similarly turned to their attorneys general, while Connecticut created a dedicated state office. Colorado and Florida also imposed new measures, and a handful of other states, including Arkansas and Illinois, are considering the same.
“Let’s face it — if you come from the same organization, you’re going to have some level of bias no matter how objective you try to be,” said Ashley Heiberger, a former police officer in Bethlehem, Pa., who advises departments on the use of force. “We’re all human. There’s a very understandable desire to protect our own.”
The 29 deaths reviewed by The Times happened throughout the vast state of Texas, from the plains of the Panhandle to the marshes of the Gulf Coast. Most occurred outside major metropolitan areas, and more than half of those killed were white. The rest were largely Hispanic, and just three were Black — figures that roughly tracked the regions’ populations.
The review suggested an uneven approach by the Rangers, though the record is open to interpretation. Jay Coons, who spent 36 years at the Harris County Sheriff’s Office in Texas, reviewed the case files at The Times’s request.
“Their actions appear to be methodical, in keeping with current investigative techniques, and I found no evidence to suggest shoddy work or a lack of commitment in seeking the truth,” said Mr. Coons, who is now a criminal justice professor at Sam Houston State University.
Mr. Bercovici, the former Los Angeles lieutenant, reached a different conclusion, noting gaps in some of the inquiries, including a failure to press for immediate interviews with the officers involved.
“You’re giving them a chance to adjust their story,” he said. “You need to be in a room with somebody. You need to look them in the eye. You need to see their body language.”
In at least 16 of the cases, the Rangers failed to interview everyone who witnessed the encounter. They sometimes relied on written statements or video footage instead, and in at least five investigations, they interviewed no witnesses at all.
The Times found 12 instances in which the Rangers delayed going to the scene or did not visit it. Delbert McNiel, 48, who had a history of mental illness, died in a struggle with Hamlin police officers after they restrained him and then shocked him 11 times with a stun gun in December 2017. The Ranger assigned to the case went to the hospital, photographed the body and conducted six brief interviews with the officers and paramedics involved. But his report reflected no interviews with other witnesses or a visit to the convenience store outside of which Mr. McNiel died.
Rob Bub, a 33-year veteran of the Los Angeles Police Department who now works as an investigation consultant, said a failure to examine the death location was generally a sign of “lazy investigating.”
“You have to go out to the scene,” he said. “I don’t care if it’s a 30-year-old cold case. You go out to the scene.”
In three cases reviewed by The Times, the Rangers assigned important tasks to employees of the agencies under investigation.
After Robert Geron Miller, 38, died following a brief stay at the Tarrant County jail in July 2019, Ranger C.H. McDonald learned that he had tussled violently with his jailers. Mr. Miller sustained head injuries when one threw him to the ground, and he complained of chest pain after another guard blasted him with pepper spray. He was dragged facedown to a cell where another jailer saw him splayed out on the floor, splashing toilet water on his face. Twelve minutes later, the Ranger noted, Mr. Miller stopped breathing.
Ranger McDonald described conducting four brief interviews in his file and left the task of watching video footage to a detective with the county sheriff’s office — the agency responsible for the jail — who reported back that he did not see any “suspicious or criminal behavior.” The Ranger closed the case about nine months later, on learning that the medical examiner had ruled that Mr. Miller’s death was caused by a sickle cell crisis.
Three cases had red flags that might have led detectives to conduct more interviews or seek additional evidence if they weren’t investigating law enforcement. Instead, the Rangers closed them without much scrutiny.
One involved Coy Walker, 41, who had done prison time for robbery and assault. Mr. Walker had struggled with mental illness and drug use over the years but had seemed on an upswing by spring 2015, his family would tell the Rangers. Then came the night that May when Mr. Walker showed up at his parents’ house near Weatherford, about 30 miles west of Fort Worth, in a wild panic.
Unable to calm her son down, Mr. Walker’s mother called 911, asking for help. The Parker County Sheriff’s Office arrived soon after. An investigative report filed by Ranger McDonald, the same officer who handled the 2019 death of Mr. Miller in Tarrant County, documented what happened next.
The first sheriff’s deputy said she found Mr. Walker in the house, on his back and surrounded by overturned furniture and broken glass, swinging a curtain rod. She jolted him with a stun gun and was struggling to get him into handcuffs when her backup, Corporal Ethan Stark, charged in. Corporal Stark leapt on top of Mr. Walker, put a knee on his neck and punched him in the face, according to the Ranger report. While atop Mr. Walker, the deputy “choked him with both hands,” cursed at him and told him to obey commands — before realizing he was not breathing, according to the Ranger’s interview with Mr. Walker’s father, who witnessed the encounter. Mr. Walker was taken to a nearby hospital but could not be revived.
When Ranger McDonald interviewed the deputies, both said that Corporal Stark’s hand had slipped and accidentally grasped Mr. Walker’s throat. But an autopsy suggested more than casual contact, records and interviews show. The pathologist’s report showed that Mr. Walker’s hyoid, a delicate bone in the upper throat, was broken — an injury well known among homicide investigators as a possible sign of strangulation.
The same autopsy report showed that Mr. Walker also had a large amount of methamphetamine in his system. The cause of death was labeled “sudden death during physical restraint with neck injury and methamphetamine intoxication.” The manner was listed as “undetermined.”
Mr. Bercovici, the former Los Angeles lieutenant, and other experienced death investigators said that even if a person was found to have drugs in his system, a broken hyoid and an eyewitness account of strangulation ought to give an investigator pause. “If I was running that investigation, I would step back,” Mr. Bercovici said. “That wouldn’t pass the smell test for me.”
Records suggest that Ranger McDonald did not pause after learning of the new evidence. On the same day he received the autopsy report, he presented his investigation to a Parker County grand jury, which brought no charges.
The other cases were also closed despite troubling autopsy results.
The Ranger who reviewed the death of Lorenzo Juarez outside Austin learned from pathologists that he had petechial hemorrhaging — tiny pin pricks of blood in the eyes that also suggest strangulation — while being arrested in 2018. Mr. Juarez, 47, was on meth and had been swinging a metal pipe by the side of a road when the Bastrop County Sheriff’s Office was called to the scene. One deputy bore down on Mr. Juarez with a forearm against the base of his neck, and others put weight on him as he was given a dose of Versed, a sedative, by paramedics.
But after the autopsy report, the Ranger, Brent Barina, documented no effort to re-interview the deputies or otherwise seek an explanation for why capillaries in Mr. Juarez’s eyes might have burst. The medical examiner determined that the death was an accident caused by methamphetamine but noted a “component of asphyxia.”
The Ranger who investigated the death of Wesley Manning, 40, turned in a detailed review of that case, which unfolded on Rattlesnake Point Road in Aransas County, near Corpus Christi, in 2015.
In a 59-page report based on 12 interviews and multiple videos, he recounted how sheriff’s deputies and police officers who believed that Mr. Manning was preventing his girlfriend from receiving medical care shocked him twice with a stun gun, doused him with pepper spray, threw him to the ground, dug a baton into his neck, put him in handcuffs, pushed a knee into his back, jammed fingers into the pressure points behind his jaw, punched him and restrained his legs before he stopped breathing.
The Ranger, Antonio DeLuna, also noted that the medical examiner ruled the death a homicide caused by “sudden cardiac death following restraint procedures.” But he did not include an additional detail from the forensic exam: that a piece of the cartilage around his voice box was fractured, another warning sign that death investigators are trained to look for.
Despite the homicide ruling and the otherwise thorough investigation, the outcome was the same as in the other cases. The Ranger presented to a grand jury, which issued no charges.
A Lack of Specialists
Despite its large population and land area, Texas employs about 165 Rangers statewide — the Houston Police Department, by comparison, has 5,300 sworn officers — and most of the Rangers work as generalists, investigating major crimes and public corruption cases and in-custody deaths depending on the day.
In contrast, most major police departments have specialists, with at least one detective unit dedicated to solving murders and other violent crimes. That focus allows detectives to gain knowledge through repetition.
“If you’re not bringing in somebody who has had a certain amount of training, you’re doing a disservice,” said Mr. Bub, the former Los Angeles detective. “There’s an expertise that goes with it.”
Clete Buckaloo, who spent 20 years with the Rangers before retiring as a captain in 2007, described a system in which Rangers are largely assigned to tasks by geography.
“He’s responsible for his area. He’s responsible for responding to those calls,” Mr. Buckaloo said of the typical Ranger, nearly all of whom are men. “He may be working a public integrity investigation, he may be working an officer-involved shooting that he’s been working on for several weeks, and now he gets a custodial death, and now he just gets that much more in his caseload.”
The Rangers declined to discuss their training and investigative tactics with The Times. But records and interviews show that Rangers are typically recruited from the ranks of the Texas Highway Patrol and might not have investigated a homicide before signing up. Once hired, all Rangers receive extensive training, Mr. Buckaloo said, and many learn on the job.
But training guidelines and technical best practices cannot account for every situation an investigator will encounter, and sometimes judgment calls must be made that can affect the outcome. That was the case in the death of Michael Garrett in Comal County in 2019.
As the police were arresting Mr. Garrett, 18, on a drug possession charge that December, an officer suspected he had swallowed some meth. Instead of putting Mr. Garrett into an ambulance, records show, he took him to the county jail, where guards tried to strip-search him. When Mr. Garrett resisted, they forced him into a restraint device and left him outside the jail’s control room. Soon after, a guard noticed that Mr. Garrett had turned purple, and he was taken to a hospital.
Medical tests revealed meth in his system, and he was on life support when the jail contacted the district attorney’s office, which issued a sworn statement indicating that Mr. Garrett would not be charged. The jail was then able to release him from custody and bypass rules for reporting the episode as an in-custody death. The Comal County district attorney, Jennifer Tharp, could not be reached for comment.
Met with deep respect among local officials, the Rangers often advise on criminal proceedings, and it would not have been unusual for one to suggest an autopsy in a case such as Mr. Garrett’s. But when the investigating Ranger, Joseph Evans, learned of the legal maneuver, he made no objection. Since Mr. Garrett was no longer under arrest when he died, Ranger Evans noted, “an autopsy could not be ordered and would only occur at the expense of Garrett’s family.”
He concluded that Mr. Garrett probably died because of his “intentional ingestion of methamphetamines,” and then he closed the case.