One month ago, Duke law professor James Coleman expressed his concern about the circumstances under which the lacrosse case accuser identified the three defendants. “According to the police account of the identification,” he noted, “the police officer who presided over the proceedings told the alleged victim at the outset that he wanted her to look at people the police had reason to believe attended the party. Thus, the police not only failed to include people they knew were not suspects among the photographs shown the woman, they told the witness in effect that there would be no such ‘fillers’ among the photographs she would see. This strongly suggests that the purpose of the identification process was to give the alleged victim an opportunity to pick three members of the lacrosse team who could be charged. Any three students would do; there could be no wrong choice.”
To what extent did this behavior typify law enforcement in North Carolina? Are procedures in the state normally applied in an arbitrary and capricious fashion? Did the reforms growing out of a major state initiative, the North Carolina Actual Innocence Commission, have much of an impact? Quite apart from this case, these questions seemed like interesting ones for an academic project, and I’m going to spin an article out of this for Criminology and Public Policy.
The recommended guidelines laid down by the Actual Innocence Commission (AIC) (on which, it should be noted, Professor Coleman served) grew out of a late 2002 initiative from former state Supreme Court Chief Justice Beverly Lake. Declining confidence in the integrity of the system motivated Lake: as one of his former law partners noted, “He is really conscientious about the public perception of the judicial branch.” Barely two weeks after Lake’s initial gathering met, a highly publicized, intensively researched, series in the N&O uncovered an extreme case of prosecutorial misconduct, which ended in a death sentence (subsequently overturned) against a man named Alan Gell.
With political cover provided by Lake, a conservative Republican, and a somewhat favorable local climate, in 2003 the AIC voted 25-0 in favor of a variety of recommendations regarding eyewitness photo IDs. The commission urged, among other items:
- The individual conducting the photo or live lineup should not know the identity of the actual suspect. This is called a double-blind procedure and addresses misidentifications resulting from unintentional influences from those conducting the identification procedure.
- Witnesses should be instructed that the suspect may or may not be in the lineup.
- A minimum of eight photos should be used in photo identification procedures.
At the time, Iowa State psychology professor Gary Wells, an expert in misidentification by crime witnesses, hailed North Carolina as “out there in front, way ahead of the curve on these changes.” He added that down the line, innocent people would benefit, though “they'll never know that they perhaps were one of the people saved from a hellacious experience by these procedures." In 2004, police from around the state attended a daylong conference in Cary, where they heard from New Jersey officers who had embraced a similar set of policies. One supervisor from New Jersey’s Union County admitted, "I thought, why do we need new procedures? But it works out very well. And it really makes a difference.” By adopting transparent, standardized procedures, AIC backers hoped to eliminate situations in which courts had to make difficult calls to exclude procedurally dubious IDs. As the AIC’s executive director noted in 2004, the change was “something that should come from basic law enforcement.”
Since Nifong himself ordered the Durham Police to confine the photo ID session to pictures of the lacrosse team, I assumed this degree of D.A. involvement was standard practice for North Carolina. So I e-mailed every district attorney in the state, asking if they used AIC guidelines in photo ID lineups that they oversaw. I noted that my interest had been sparked by the revelations in the lacrosse case. In one-line replies, two refused to answer my questions; several others didn’t respond. But from the rest (with one exception, to be noted below), I received what was for me a surprising consensus, for which this note was typical: “In North Carolina, DAs have little power to dictate to local law enforcement what investigative techniques should be used.” Most (though not all) said that they encouraged police departments in their jurisdiction to use AIC guidelines.
One D.A. told me, “In North Carolina to be a law enforcement officer, one must complete 488 hours of instruction at a host of community colleges or at the central North Carolina Justice Academy and pass a state board examination. The course of instruction includes proper identification procedures—proper meaning consistent with Constitutional and statutory requirement (i.e. not unduly suggestive).” He suggested that I contact the Justice Academy for more details.
The Justice Academy’s representative reported that the “Basic Law Enforcement Training and Mandated In-Service Training curricula . . . differ from the Actual Innocence Commission in that the CJ [Criminal Justice] Commission recommends the use of six photographs in a photographic identification process rather than the eight recommended by the AI Commission.“ Like those of the AIC, the CJ guidelines require that “witnesses should be instructed that the suspect may or may not be in the lineup.” The CJ guidelines are slightly more permissive on who should conduct the session: they urge the “double-blind” approach, but note that departments who lack available personnel can waive this requirement—though they “should be prepared to articulate in court why.”
That said, the representative informed me, “The Justice Academy does not train officers to follow the guidelines, but the Academy does train officers on the use of the guidelines.” Police departments themselves, in consultation with local district attorneys, decide whether to follow the AIC guidelines, the CJ guidelines, or no guidelines at all.
Accordingly, I got in touch with North Carolina police departments. I contacted departments from five cities of more than 100,000 people in the 2000 census (Charlotte, Fayetteville, Greensboro, Raleigh, and Winston-Salem); five cities with populations of between 50,000 and 100,000 (Cary, Gastonia, Jacksonville, Rocky Mount, and Wilmington); and five cities of less than 50,000 people (Burlington, Chapel Hill, Hickory, Spencer, and Tarboro). I e-mailed each department to ask if it adhered to the AIC guidelines; and, if not, which, if any, procedure it employed. Follow-up calls went to departments that didn’t respond to my e-mail request. Every law enforcement officer to whom I spoke communicated a genuine desire to ensure that the procedure followed was fair, and answered every question I had. For those whose primary exposure to the North Carolina criminal justice system has come through watching Nifong’s behavior, this finding is reassuring.
Of the 15 departments, eight follow the AIC guidelines in their entirety. (A ninth, Winston-Salem, is currently revising its guidelines along the lines recommended by the AIC, after some local controversy over whether to follow the commission's recommendations.) Typical e-mail responses to my question of whether the department followed the AIC recommendations: “Yes—and every member of the agency was trained in (early) 2005 on the protocol.” Another: “Yes we do, and we do not steer from what the guidelines are.” And a third: “YES!”
A representative from Burlington’s department described for me its procedures, which seem to provide a model for conforming to all AIC guidelines. In the central North Carolina city, which is around one-quarter Durham’s size, all witnesses are shown arrays of eight photos—one of a suspect, seven of which are computer-generated from the department’s pre-existing database to resemble the suspect’s basic parameters. The person conducting the session, who might even be a department secretary, has no idea of which of the eight photos contains the suspect. Witnesses can ask to look again at a photo, but not until all eight photos have been shown; and if they re-check, the witness has to re-examine at all eight photos.
Similarly, Chapel Hill’s procedure rigorously follows AIC guidelines, even in cases with many possible suspects. According to its official document on eyewitness investigations,
1.) Line-ups shall consist of one known suspect and seven filler photos/images.
2.) If there are multiple suspects in a case, a different line-up shall be created for each suspect. Filler photos/images can not be used in more than one line-up.
3.) Filler photos/images are selected based upon shared characteristics with the suspect photo . . .
In displaying the line-up,
1.) Verbal instructions [which include the following statement: “The person who committed the crime may or may not be included”] shall be given to the witness by the person conducting the line-up.
2.) The photos/images will be shown to the witness sequentially and in a predetermined order. No more than one photo at a time should be seen by the witness. If the witness wishes to see the line-up a second time, the photos will be presented in the same order as in the first presentation.
. . . 4.) Whenever possible, the individual presenting the line-up should not be aware of the suspect’s identity.
5.) The witness should receive no feedback from the person conducting the line-up.
Four other departments conform to the CJ guidelines. In Gastonia, Greensboro, Fayetteville, and Jacksonville, the official policy calls for five "filler” photos for every suspect shown. (The departments vary on whether an uninvolved person will conduct the photo session, and also on whether the photos will be shown sequentially or all on one page; all have either formal or informal policies in which the person conducting the session opens with a statement that the lineup might or might not include the suspect.) Rocky Mount’s department, meanwhile, uses only five fillers rather than seven, but in all other respects conforms to the AIC guidelines.
The southeastern corner of the state features a unique hybrid structure. The largest city in this area, Wilmington, uses a “six pack,” which contains five “filler” photos and one of a suspect. Its police department—along with all others in North Carolina’s 5th Judicial District—runs its sessions according to a detailed form prepared by District Attorney Benjamin David.
To begin each interview, the officer informs the witness, “In a moment, I am going to show you a group of photographs. This group of photographs may or may not contain a picture of the person who committed a crime now being investigated.” The report produced from the ID session contains not only the specific results but considerable contextual material. Of the five areas in which officers record items, three deal with the background to the session: the witness’s opportunity to view the suspect during the crime itself; the witness’s degree of attention (i.e., the nature of the crime; the witness’s occupation, training, prior experience as victim; and the witness’s prior relationship, if any, to the defendant); and the accuracy of the witness’s principal description of the alleged assailants (i.e., when and to whom the description was first given; whether the witness identified an accent or distinctive marks, tattoos, limp, or other such feature on the suspect). During the photo array process, officers are required to observe the level of certainty (i.e., the time the witness viewed the lineup before making an identification; what the witness said or did when making an identification). Finally, the officer must record the length of time between the crime and the photo lineup. Obviously, a session that occurs a day or two after the crime will yield a far more reliable identification than one that takes place, say, three weeks after the incident.
Of all the procedures in the state—including those of the Actual Innocence Commission—that structured by D.A. David seems the most comprehensive. It combines the spirit of the AIC’s desired procedural protections for suspects with common-sense efficiency measures and the guaranteed recording of useful contextual information for law enforcement. Mistaken ID’s, of course, always can occur, but in the David system, they would seem highly unlikely; and proper ID’s would be more likely to yield convictions.
On the question of eyewitness IDs, then, North Carolina features an overwhelming consensus. And in the aftermath of the Gell case and the work of the AIC, that consensus is moving in favor of more, not less, fairness and transparency in the eyewitness ID process.
This procedures employed in the lacrosse case contradict from this pattern in almost every respect. As an N&O story observed, the policy followed in Durham before April 4, General Order 4077, reflected CJ guidelines. But in the first paragraph of Sergeant M.D. Gottlieb’s report of the lacrosse ID session, he wrote:
Mr. Nifong suggested we put together the mug shot type photographs [of the lacrosse players] into a group since we are under impression the players at the party are members of the Duke Lacrosse team and instead of doing a line up or a photographic array, we would merely ask the [alleged] victim to look at each picture and see if she recalled seeing the individuals at the party.
Gottlieb himself conducted the session—beginning, as Coleman noted, by informing the accuser that she would only be seeing “people we had reason to believe attended the party.” The accuser identified one person (Collin Finnerty) who she claimed attacked her in a variety of ways. Three others, she said, “looked” like people who might have attacked her. For reasons that remain unclear, Gottlieb treated each of these three partial IDs in very different ways—confirming the wisdom of policies like Burlington’s, where people who have no involvement with the case oversee the viewing and therefore are consistent throughout the process.
Gottlieb passed over the first partial ID quickly and without explanation. The second partial ID—of Dave Evans—led to the accuser saying Evans resembled her attacker, but the person who attacked her had a mustache (which Evans does not). Gottlieb then asked the accuser to give a “percentage-wise” estimate, something that he hadn’t asked of the previous partial ID. (The accuser said she was 90% sure.) The third partial ID—of Reade Seligmann—began with the accuser saying “he looked like one of the guys who assaulted me.” Then, under follow-up questioning from Gottlieb, the accuser claimed that she was 100% sure. As far as we know, this one statement is the only evidence against Seligmann, who was video-taped at a Wachovia ATM a mile away at the time of the alleged crime.
For the state: The AIC seems to have had a significant, and constructive, influence on fairness in eyewitness IDs—even though its guidelines are only recommendations. Many small departments have adopted the recommendations en toto. Among large departments, one (Winston-Salem) is moving in that direction, another (Raleigh) follows the basics of the AIC approach, and a third (Charlotte-Mecklenburg) is headed by a police chief who was himself a member of the AIC. Two of the smaller departments noted that the apparent preference of the North Carolina courts for AIC identifications had facilitated their decisions. All non-AIC departments surveyed adhere to CJ guidelines, except those in southeastern North Carolina, whose independent approach seems more conducive to the cause of justice than either the AIC or CJ policies.
For the lacrosse case: The manner in which Nifong orchestrated the photo ID deviates wildly from common North Carolina practice, in at least four ways:
1.) The basic debate in the state seems to revolve around following the AIC or CJ guidelines, or developing a hybrid (D.A. David). As far as I could determine, in no North Carolina jurisdiction is it the policy to confine eyewitness ID sessions to suspects in the case.
2.) Every police department that communicated with me has a standard policy—either formal or informal—of telling witnesses that the photo array might or might not include the suspect. As far as I could determine, in no North Carolina jurisdiction is it the practice, as was followed in this case, for the witness to be informed that the photo array would consist only of possible suspects.
3.) As far as I could determine, no North Carolina jurisdiction possesses a third identification procedure, one that can be used, to quote Gottlieb’s description of Nifong’s words, “instead of doing a line up or a photographic array.”
4.) Some district attorneys who communicated with me seemed to like the AIC guidelines; others weren’t particularly enthusiastic about them; others (like David) developed their own systems. But of those who said that they made recommendations to local law enforcement, all did so in favor of departments using fillers in photo ID sessions, not the reverse. And a district attorney setting policy for an eyewitness ID session seems to be unusual in and of itself.
The lacrosse session also departed from the statewide norm (although here no unanimity exists) in having someone intimately involved with the investigation oversee the photo ID session.
A notion to suppress the IDs in this case is pending.
[Originally published in Cliopatria.]