Good Intentions – and Lawyers – Are Not Enough

The Innocence Scholarship Column July 24, 2023

The last column, Hit – Smashed – Broken Glass, highlighted a 1974 article by Loftus and Palmer that led to her work on eyewitness issues and began to reorient eyewitness research. Today’s column addresses an article by Gary Wells, published four years later in 1978. It became a foundation for applied eyewitness research. Rather than reporting experimental results, Wells’ article reconceptualized the variables studied by eyewitness researchers. He observed that eyewitness experiments to that time had not really advanced the goal of applied research – to design error-reducing lineup procedures. His insight was quickly recognized by experienced psychological scientists as a game changer.

Before describing Wells’ idea, let’s consider a preliminary question that might occur to some readers. Do accurate lineups have to be based on psychological laboratory research? Can’t professional, unbiased police conduct accurate lineups using fair and “common sense” procedures? In fact, such non-scientific approaches were advanced by lawyers, police, and even the US Supreme Court for decades before it became clear that accurate lineups had to be based on scientific principles.

The reasons were explained in the first eyewitness white paper, (see White Papers 1) published in 1998, twenty years after Wells’ 1978 article galvanized eyewitness researchers. Their cumulative work ultimately led to a few basic lineup principles and procedures that reduced the probability of honestly mistaken eyewitness identification. Gary Wells, incidentally, was the lead author and the prime motivator of that white paper.

Eyewitness identification errors were documented since the early twentieth century. Beginning in the 1950s, many legal scholars, prosecutors, and police officials (in Canada and the UK as well as the US) were concerned about inaccurate criminal identifications. In law journals and in official documents they published many common-sense and fair-minded recommendations, guidelines, rules, and checklists.

Here are some examples. Witnesses should not be allowed to view suspects in handcuffs. Police must share lineup results with defense attorneys. A prosecutor and defense attorney should both be present at a police lineup. Lineup “fillers” should be of the same race and similar in age to the suspect. Only one witness at a time should be allowed in the identification room. To the extent possible, a witness should not view a photograph of the suspect prior to the lineup. Gestures and comments by lineup participants should be done “uniformly.” The witnesses should state their identification in writing rather than verbally. Avoid showups (one-suspect lineups).

Beginning in the 1960s, the Supreme Court decided that defense lawyers must be present in post-indictment lineups. The Court also held that suggestive lineups violate a defendant’s due process rights, but allowed the introduction of suggestive lineups if the identification is reliable, considering the totality of the circumstances. One so-called reliability factor was the confidence of the eyewitness.

There were problems with all these well-intentioned suggestions. The guidelines were incomplete and inconsistent. They could be multiplied into so many rules as to be unworkable in practice. Some rules were too vague; others were too specific or detailed. Most important, however, is that the rules were not based on empirical research findings. Even those that seemed fair were not proven to reduce error. Some, like the Supreme Court’s ruling about witness confidence was proven by research to have little relation to accuracy.

Into this stew, came Gary Wells’ insight. His article, Applied Eyewitness-Testimony Research: System Variables and Estimator Variables, examined the variables studied by researchers that could affect eyewitness accuracy. All the variables could be controlled in the laboratory. But not all the variables could be controlled in criminal justice practice. In the lab, the race of the perpetrator and the witness, for example, could be controlled. In actual criminal cases, it could not be controlled. The police took witnesses and perpetrators as they came. Wells labeled such factors as "estimator variables". They included all manner of factors such as crime seriousness, the degree of the witnesses’ attention, and the like.

Other variables that could affect the accuracy of identifications could be controlled by the police. They mostly related to the construction and conduct of the lineup. These system variables included lineup instructions, the number of lineup fillers, whether the lineup administrator knew the identity of the perpetrator/suspect or not, and others.

Members of the eyewitness research community recognized that by conducting experiments controlling system variables they could eventually discover, with scientific validity, which methods of conducting lineups could reduce the likelihood of eyewitness error. As noted in a prior column, White Papers 2, they did so. Today the principles for conducting more accurate lineups exist. Lineups based on scientific research cannot totally eliminate honest lineup misidentifications. But reducing the probability of eyewitness error is a real advance.

Copyright © 2023 Marvin Zalman