Psychologists recently carried out hundreds of mock police lineups to determine the most successful procedure for identification of criminals. Inaccurate identification can obviously lead to the prosecution of innocent people while guilty ones go free.

The National Academy of Sciences (NAS) recently wrote a report titled ‘Identifying the Culprit’ about eyewitness evidence in order  to improve the way in which police officers collect reports from witnesses, strengthen the value of eyewitness reports in court, and improve the scientific foundation underpinning eyewitness identification. One of the papers was co-authored by Dr. Heather Flowe, from the University of Leicester's School of Psychology called ‘Receiver Operating Characteristics Analysis of Eyewitness Memory: Comparing the Diagnostic Accuracy of Simultaneous Versus Sequential Lineups’. That paper sought to determine whether police lineups were more successful at obtaining accurate results if the lineup members were viewed simultaneously (as a group) or sequentially (one-by-one).

Using a technique called Receiver Operating Characteristics (ROC), they concluded that the simultaneous procedure was better at discriminating between the guilty and the innocent.

Flowe said, “For decades, research suggested that sequential procedures reduced innocent suspect identifications, without having an appreciable effect on the rate at which the guilty are identified. This resulted in many jurisdictions in the US adopting sequential procedures. However, we showed in our paper that this previous research was using the wrong performance metric. Simultaneous procedures may actually be better.”

The team carried out hundreds of staged lineups in an attempt to determine which method – group or singular – yielded the most accurate results. Flowe said, “We carried out our work with participant witnesses in California, and then we replicated it, recruiting participant witnesses online throughout the world. “They watched a crime simulation, and then attempted to identify the culprit from a lineup.

“We varied whether or not the perpetrator was actually present in the lineup, and whether the lineup was simultaneous or sequential. “We found that people were better able to distinguish guilty from innocent suspects in a simultaneous lineup. “Several labs around the world have replicated our findings. As a consequence of our work, sequential lineups can no longer be recommended.

“The NAS report reflects this dramatic shift in scientific opinion. We also introduced ROC analysis, a new performance metric for evaluating lineup identification performance, and this is rapidly changing how lineup identification researchers conduct their studies.”

The NAS has highlighted a number of issues surrounding eyewitness reports based on the ROC paper as well as a number of other sources.

It found that memory is not always a faithful record of eyewitness's experiences and was affected by a number of factors including their emotional state, the environment, and the way the brain processes and stores memories.

Factors such as dim lighting, brief viewing times, large distances, duress, elevated emotion, and the presence of distracting elements such as weapons all play a part in how the witnesses interprets the crime.

The study also explains that the way the brain stores information can also affect the accuracy of the witness's memories. It said that human memories are ever changing – forgotten, reconstructed, updated and distorted. Therefore, caution must be taken when relying on eyewitness identifications court.

Flowe said, "Our research underscores the usefulness of asking eyewitnesses to evaluate their identification confidence at the time of the identification.

“Confidence is a valuable measure of how good an eyewitness’ identification is, but only if it is measured at the time of the identification. Eyewitness confidence is not a good index of an eyewitness’ memory strength if it is taken after the fact. The confidence we have in the accuracy of our memories is malleable. It can be shaped by social influences.“

Speaking at a NAS conference earlier this month, Professor Thomas Albright, director the Vision Centre Laboratory at the Salk Institute for Biological Studies, in California, suggested 5 main improvements that police and other law enforcement personnel should follow when obtaining the witness statements.

1) Provide personnel with training about vision and memory and the variables that affect them. For example, training which minimises the contamination of eyewitness memory – asking opened rather than leading questions.

2) Officers involved in the case should not be involved in the organization of the lineup and should not know the placement of the subject (criminal). Meaning they cannot inadvertently influence the witness.

3) Create a standard set of easily understandable instructions used for engaging a witness in an ID procedure. For example, telling the eyewitness that the perpetrator may or may not be in the lineup.

4) Officers should document the level of confidence of the eyewitness when he or she identifies the suspect. Prof Albright said that expressions of confidence which take place in the courtroom often deviate substantially from the lineup itself.

5) Eyewitness procedures should be videoed and kept as a permanent record of the identification.