Back to Contents

November / December 2009
Volume 21, Number 6

Lining Up the Science Behind Eyewitness Identifications
By Maria Sprow

Cory Session knows the power held by an eyewitness, especially an eyewitness who faces a jury and says, ‘That’s the man who did it.’

Such an eyewitness caused his brother, Tim Cole — a United States Veteran, a law school student, an upstanding citizen with spotless record — to be convicted in 1986 for a rape he did not commit.

Years later, Session still tears up just thinking about the injustice and how his brother dealt with the charges against him and having his freedoms taken. Dealt, past-tense, because Tim died behind bars in 1999, waiting for an exoneration that came too late. He had spent 13 years in prison.

But Session is determined that his brother’s conviction and exoneration will not be in vain, and he is aided by state legislators and other leaders who are using Tim Cole’s case to implement changes in the ways Texas brings the accused to justice.

A group of prosecutors, lawmakers, defense attorneys and other stakeholders during the last legislative session worked to draft Senate Bill 117, which would have required that law enforcement departments have written guidelines related to eyewitness identification procedures that fall in line with scientific “best practices,” namely, blind sequential line ups and the recording of the identifications. Though the various groups were mostly able to reach a consensus to support the bill, the political standoff on voter identification killed its passage.

Voter identification also killed several other innocence-related bills. One, SB 116, asked that “when practical,” law enforcement agencies tape interrogations. Another, SB 1976, would have allowed habeas corpus writs for cases involving discredited science. Yet another, SB 1864, would have restricted the reasons that judges can deny post-conviction DNA tests.

That doesn’t mean that the issue died. The Legislature passed the more controversial House Bill 498, which created the Timothy Cole Advisory Panel on Wrongful Convictions. The commission was charged with investigating the weaknesses in the criminal justice procedures used to prosecute the cases that eventually became exonerations and with identifying programs to prevent wrongful convictions.

Those against creating the innocence panel opposed for a number of reasons: they believed the panel’s power and authority were too broad and that the problems leading to the known 38 exonerations have already been fixed over the course of many years, since those cases are mostly decades old. Others believed the commission would be redundant and add another layer of bureaucracy to the problem, copying or possibly slowing the work of the Texas Criminal Justice Integrity Unit, established in 2008 by Court of Criminal Appeals Judge Barbara Hervey. Several of the Tim Cole panel members also serve on the Integrity Unit.

At its first meeting, held last month, the Tim Cole panel’s members all expressed disappointment at the Legislature’s inaction, and at the lack of progress made overall.

“We have talked and danced around things way too long,” said Mary Ann Wiley, who represents the governor’s office on both the Integrity Unit and the Tim Cole panel.
“We’ve known about these issues for some time,” said panel member and Wichita County District Attorney Barry Macha, discussing the Criminal Justice Advisory Council, which had been created in 2005 for nearly the exact purpose. “This (SB 117) had wide support, it’s been studied and studied and studied. … The sad fact of the matter is that too often these issues get politicized and instead of following through and getting the legislation passed, they don’t for other political reasons. … That’s the fault of the process.”

Exonerations in Texas
According to the Innocence Project of Texas, DNA evidence has led to the exoneration of 38 individuals since 1994. All but three of those exonerations occurred after 2000, though the majority of the cases were more than 10 years old by the time the incarcerated individual was exonerated. Most of the original cases date back to the 1980s and 1990s, and many law enforcement supporters and officers say that investigative procedures have greatly improved since then.

The lone exception so far has been the case of Ricardo Rachell, who was convicted of child sexual assault in 2003 and sentenced to 40 years in prison. The case stemmed from an attack that occurred in late 2002, when two 8-year-old boys were solicited by a man riding a bike and wearing a scarf on his face. The man promised the boys cash if they helped pick up trash; the boys agreed, but the man abducted the victim and took him to an empty home, where he held him at knife point and assaulted him before finally leaving.

The next day, the boy’s mother saw Rachell riding his bike down the street and asked the boys if Rachell was the attacker. Both boys said he was. They agreed again nearly two months later, when officers conducted separate photographic lineups with each boy.

Police obtained an oral swab, the child’s clothing and a sexual assault kit, but none of the DNA evidence was tested at the time; according to an internal investigation conducted by the prosecutor’s successor, “the DNA Lab was embroiled in controversy during the investigative period of this case and was closed December, 2002.” In any case, Rachell’s defense attorney didn’t seem to recognize the evidence was there and the prosecutor apparently didn’t need further evidence to get a conviction. The boys’ eyewitness testimony was enough, though neither mentioned to police that the attacker had been severely disfigured. Rachell has been severely disfigured since the mid 1990s, when a shotgun blast tore apart the right side of his face and left him struggling to speak.

It’s imaginable that in the mind of a scared and shaken 8-year-old boy, Rachell’s disfigurements made the boy believe the scary man who attacked him must have been the same as the scary man riding his bike down the street.

However the mistake occurred, the evidence was finally tested late last year. According to the prosecutors’ internal investigation and subsequent report, in 2007, Rachell actually mailed the police department a letter stating the name of the true rapist, Andrew Wayne Hawthorne, dubbed “Yellowstone Park Serial Rapist,” who used a similar method of operation as had been used in Rachell’s case. (Rachell began following news reports of similar rapes and cutting newspaper clippings when he originally became a suspect; he sent most of the clippings to his defense attorney.) He also filed a court request asking that the DNA evidence be tested. He was released in December 2008, after serving five years in prison. The DNA evidence identified that the rapist had indeed been Hawthorne.

In its investigation of the causes of wrongful convictions, The Justice Project looked at the contributing factors of each of the 38 Texas cases. Eyewitness error was the leading contributor, cited in 32 of the cases. In addition, false or faulty forensics were cited in 18 cases and snitch testimony was cited in five cases (the 81st Legislature did pass a bill related to that cause). Four cases involved false confessions and six cases included other errors or oversights.

Hindsight shows that the incorrect convictions have ramifications that fall beyond the system’s search for justice. Not only does each exoneree get paid for each year of their life spent in prison, but public safety has often been impacted when the true criminals remain free. In Rachell’s case, Hawthorne was eventually convicted of the rape of another 8-year-old boy that occurred on Nov. 16, 2002, as well as two other child sexual assaults with matching methods of operation. He pleaded guilty to those three cases on April 8, 2004 and received 60 years in prison.

The Tim Cole conviction also had public safety consequences.

“After the authorities arrested Timothy Cole for a series of rapes near the Texas Tech Campus, Jerry Wayne Johnson abducted a couple after a party. He raped the woman in a cotton field. He then later raped a 15-year-old girl at knife point. He did that after the investigation into the rape for which Timothy Cole was convicted was shut down. And so the faulty eyewitness evidence led the authorities off his track, and he was free to commit more crimes,” testified Edwin Colfax, the director of The Justice Project’s Texas office, during the SB 117 public hearing before the Senate Criminal Justice Committee.

Better Eyewitness Identification
Had it passed, Senate Bill 117 would have required every law enforcement agency in the state to adopt a written policy on the administration of photographic and live lineup identification procedures.

The policy was to follow a model policy written by the Bill Blackwood Law Enforcement Management Institute of Texas at Sam Houston State University, which would have taken into consideration the most up-to-date scientific research on eyewitness memory. The written policies would have been required to address how line up photographs and participants are chosen, the verbal instructions lineup administrators give to witnesses prior to the lineup and the documentation and recording of eyewitness statements. They also would have required that departments have policies for “alternative procedures designed to prevent opportunities to influence the witness.”

While Nacogdoches County Sheriff Thomas Kerss, president of the Sheriff’s Association of Texas, said he believes SB 117 and its supporters harbored good intentions, he said he is concerned about knee-jerk reactions and overly legislated solutions and doesn’t believe that having written policies regarding eyewitness identification procedures will make much of a difference in today’s criminal justice system.

The cases from which the problems stem are decades old or aberrations, which can happen with even the best of intentions and policies in place.

Though lineups and eyewitnesses often add drama to police shows on TV, his own department maybe conducts six photographic lineups a year, Kerss said. In those cases, prosecutors nowadays will demand more evidence than eyewitness identification, and law enforcement officers now know that eyewitness identifications, though possibly convincing to a jury, aren’t made of stone.

“I’ve been in law enforcement 28 years. I probably haven’t seen more than a handful of live lineups conducted in all that time,” Kerss said. “We don’t do a lot of lineups period. Most of the time, it’s not that necessary… sometimes (the lineup is) part of the probable cause to get your arrest warrant to locate and arrest the individual to start with. Every case is going to be different, but if you can make an identification on a suspect, you’ve got other criteria to follow that up with.”

In cases where lineups are possible, Kerss said he believes law enforcement agencies will follow and are following the most up-to-date scientific recommendations without the Legislature’s involvement. After all, that’s what prosecutors want to secure a conviction. “As a law enforcement officer, we have as much responsibility to prove someone’s innocence as we do to try and prove their guilt. The standards that we operate by are always held under close scrutiny,” he said. “Does it hurt if there is legislation put in place? Certainly not, but my opinion is that today’s standards in law enforcement have probably already changed enough that those same scenarios don’t play out to be true.”

According to The Justice Project, written policies around the state don’t illustrate the changes and are not currently following the practices promoted in Senate Bill 117. During testimony given to the Senate Criminal Justice Committee, Colfax discussed what the group found by researching the exonerations and law enforcement policies regarding eyewitness identification procedures.

In 2008, The Justice Project surveyed 1,034 law enforcement agencies in Texas, asking the departments if they had written guidelines to follow when conducting photographic or live lineups. Only 750 departments responded, and of those, only 88 departments said they had written policies that The Justice Project identified as having followed “at least some” scientifically-endorsed “best practices.”

According to the report, when assessing whether written policies followed best practices, The Justice Project looked at the four best practice areas laid out in the SB 117 legislation: whether the administrator tells witnesses that the perpetrator may not be in the lineup; whether the policy requires uniformity in presentation of people or photographs; whether the policy requires an administrator who does not know which person in the lineup is the suspect; and whether the policy requires either an audio or video recording of the lineup and the witness’s initial degree of certainty. Those are the four best practice areas laid out in the SB 117 legislation.

Of the submitted policies, 53 required fairness in lineup composition, seven required blind administration procedures, 31 required that an administrator deliver cautionary instructions prior to the lineup and four required “full documentation” of lineup procedures and a confidence statement from the witness.

“The quality of these policies varied widely among departments across the state,” the group’s report states. “This lack of standardized protocol indicates that Texas is failing to reap the benefits of systematic scientific research on eyewitness error or to follow the best practices recommended by the U.S. Department of Justice, the International Association of Chiefs of Police, the American Bar Association and other organizations.”

Agencies recognized for following at least one best practice in their written policies include the sheriff’s departments in Cameron, Ellis, Guadalupe, Harris, Jack, Lamar, Liberty, Tarrant, Tom Green, Wichita, Williamson and Wise counties. The written policy used by the Llano County Sheriff’s Department received positive remarks across the board.

In its report Eyewitness Identification Procedures in Texas, The Innocence Project gave some above-and-beyond suggestions for policies that law enforcement agencies can enact and which The Innocent Project believes will help protect innocent people from being the victim of an eyewitness error.

For instance, the Lewisville Police Department requires that officers not only select fillers who generally fit the witness’s description of the suspect but also cautions officers to “avoid using fillers who so closely resemble the suspect that a person familiar with the suspect might find it difficult to distinguish the suspect from the fillers. According to The Innocence Project, “the policy attempts to minimize incorrect identifications while preserving the good identifications.”

The University Park Police Department written policy states that officers should “create a consistent appearance between the suspect and the fillers with respect to any unique or unusual feature (e.g., scars, tattoos, facial hair) used to describe the perpetrator by artificially adding or concealing that feature.” The Innocent Project believes that policy helps “to prevent claims of suggestive, and therefore inadmissible, lineups.”

Llano County Sheriff’s Department was praised in the report for being one of seven departments that require a blind administrator for lineups, which The Innocence Project believes keeps the investigating officer from accidentally clueing an eyewitness into who the suspect is. While many departments’ policies stated that officers should not assist a witness, The Innocence Project believes the cues are “often inadvertent and unintentional” and that written policies against such types of actions “fail to safeguard” lineup participants.

The Llano County Sheriff’s Department was also praised for being one of four departments with written policies that require full documentation of lineup procedures and ask for a witness confidence statement after the lineup has concluded, two steps that go above and beyond the average policy, according to The Innocence Project. The average policy asked officers to photograph or photo copy the lineup and document the identification procedure in the report, which “does little to provide jurors with the full context of an identification procedure,” states the report.

Shannon Edmonds, who handles government relations for the Texas District & County Attorneys Association and who had a hand in the final Senate Bill 117 language, said he supports the four best practice areas promoted within the legislation.

“The prosecutors who got involved on SB 117 during the last session supported that bill in part because they think it will help improve the quality of eyewitness identifications and that’s important because they want jurors to be able to have faith in the result of an eyewitness identification,” he said. “When we talk about this issue, we go, look, this is going to happen, it just got derailed because of a political issue unrelated to it, but it’s going to happen, so get involved, figure out what is going on and get educated about the issue.”

Edmonds said he expects the bill to pass during the next legislative session, but added that any changes to its language could create problems.

“The original version of the bill had several problems with it. Prosecutors were concerned that it micromanaged the investigative process, which is something we don’t do in statute,” he said. “It also would have resulted in lots of evidence or identifications that were otherwise valid being excluded from evidence based on technicalities, and that was prosecutors’ biggest concern.”

If law enforcement agencies have been cautious or slow in implementing the best practices outlined in Senate Bill 117, as innocence-related agencies may allege, it’s partly because even the smallest changes to how officers conduct investigations can have a wide effect and negative consequences; there are always pros and cons to balance.

Just look at the controversy surrounding sequential and simultaneous lineups, which, while not included in the Senate Bill 117 legislation directly, could fall under the category of “other procedures or best practices supported by credible research or commonly accepted as a means to reduce erroneous identifications and enhance the objectivity and reliability of eyewitness identifications.”

In a letter to the Tim Cole Advisory Panel on Wrongful Convictions, Senator Rodney Ellis, who serves as the chair of the national Innocence Project, asked that the panel investigate and recommend sequential lineups as a better alternative to simultaneous lineups.

“When combined with a ‘blind’ administrator, presenting lineup members one-by-one … rather than all at once… has been proven to significantly increase the overall accuracy of eyewitness identifications,” Ellis wrote. “The Innocence Project can provide all interested parties with substantial information on the scientific research about sequential presentation and the experiences of jurisdictions that are effectively using the practice, including studies, protocols, law enforcement references and other information.”

The National Institute of Justice acknowledges the different judgment calls that eyewitnesses make during the two processes but calls the overall research “inconclusive.”

“For sequential lineups, witnesses must exercise ‘absolute judgment,’ comparing each photograph or person only to their memory of what the offender looked like. In simultaneous lineups, witnesses must use ‘relative judgment’ to compare lineup photographs or members to each other,” the National Institute of Justice states. “Some research has indicated that a sequential lineup … produces fewer false identifications as well as fewer true identifications.”

Policy-wise, that means that good, true identifications that helped put real criminals behind bars may not have ever been obtained or used in some cases.

Scientifically, both Ellis and the National Institute of Justice contribute part of the controversy to a 2006 field study conducted on behalf of the Illinois State Police. While the study “found a higher rate of correct identifications in simultaneous non-blind lineups and a higher rate of false identifications in sequential double-blind lineups,” according to the National Institute of Justice, “questions have been raised about the methodology used in the study.” (Ellis wrote that the study was “severely flawed, non-scientific” and “neither written nor worthy of peer-review by scientists.”)

A 2006 policy evaluation by The University of Texas at El Paso’s Eyewitness Identification Research Laboratory that took into account numerous studies on the subject concluded that “it is important to be careful that the pressing needs of public service (in this case, the criminal justice system) do not move us to potential premature advocacy of any line of research and the application of that research. We believe that the recent promotion of the [sequential lineups] in research journals, conferences and policy arenas may be a case in that point. On the basis of our analysis, we argue that the research base for [sequential lineups] may not be sufficiently developed from a methodological or theoretical point of view to currently advocate for its implementation to the exclusion of other procedures.”

Edmonds compared the scientific and practical controversy surrounding whether to show one photograph at a time or six photographs at a time to global warming.
“No matter how many people say, ‘oh, everybody agrees there is global warming,’ there are people out there saying, ‘oh, nope, I don’t agree,’” he said. “And in fact there are still professors and other people conducting studies on it, so if they are still doing studies, that clearly shows the subject isn’t settled.”

Kerss said law enforcement agencies want to use the best investigation methods available, but added that it should ultimately be up to the courts to decide what works best in the justice system based on the best available scientific research — when it is complete.

“To the officer on the case, well, what difference does it make how you show them the photograph? It’s easy to adjust something that has no cost to it,” he said. “If it serves no other purpose, (Senate Bill 117) may serve to help better educate officers and agencies.”