The goal of the article was to integrate the various techniques of eyewitness identifications: lineup instructions, interviewing techniques, and construction and presentation of lineups.

Experiment 1

In experiment 1, participants viewed a video reenactment of an armed robbery and then attempted to identify the robber. Variables of interest included:

Disguise of the robber

Weapon presence

Lineup type : robber-present or robber-absent

Presentation mode: Simultaneous or Sequential

Lineup instructions: cautioning instructions or no cautioning instructions

Timing of prejudgment questionnaire

Context interview: guided interview or time to reminisce

Lineup context: moving lineups or videotaped stills

The ‘context interview’ variable is important because it utilizes the effects of Tulving’s encoding specificity principle, which is that contextual features may aid in memory retrieval if those contextual features were encoded with the to-be-remembered stimulus. Thus, the authors of the present article claim that inclusion of context features during the identification may help to enhance accuracy of the robber.

These context features include the context interview (providing the witness with contextual information that is extrinsic from the robber) and lineup cues (which provide the witness with contextual information that is intrinsic to the robber).

The ‘presentation mode’ variable has shown to be important by Lindsay & Wells. They claim that the use of simultaneous lineups is flawed because it allows witnesses to use a relative judgment of the members of the lineup, thus leading to a larger number of false alarms.

Results

Of the 175 participants tested, the overall proportion correct was .74.

By doing a hierarchical regression with the main effects entered on the first step and the interaction terms entered on the second step, marginally significant (p< .07) effects were found for disguise and presentation mode (with correct identifications more often in the low disguise and sequential presentation conditions). None of the other main effects were significant.

The one interaction that was significant was that between lineup cues and presentation mode.

This finding suggests that both lineup cues, as well as sequential presentation improve identification accuracy, but that they are not additive.

To add more evidence that sequential presentations are superior, the false-identification rates were compared. In the simultaneous condition, the false-identification rate was .39, but in the sequential condition, it was only .19.

Further, to assess the confidence-accuracy correlation, it was found that lineup cues had a significant effect on this correlation, as well as sequential vs. simultaneous presentation. The use of the context interview had a trivial effect on the confidence-accuracy correlation.

Experiment 2

The variables of interest in experiment 2 were robber-absent or robber-present lineups, simultaneous or sequential presentation of lineup, and three levels of instructions (low, neutral, and high). The low-level instructions emphasized the point that if this was an actual identification, we would not want to let a guilty criminal back onto the streets. The high-level instructions emphasized the point that we would not want to convict an innocent person.

Results

The overall proportion of correct judgments for this experiment was .56.

The absence of the robber in the photo spread led participants to make more correct judgments than when the robber was present.

With sequential presentation, the high-level instructions (not wanting to convict an innocent person) led to a significant reduction in false identifications. (The other correlations were trivial).

The authors conducted a combined probability analysis on their two studies and Lindsay & Wells, (1985) to determine the statistical reliability of the interaction of Presentation Mode X Presence of Robber. A combined probability was obtained ( a z of 2.61) which indicates that across these three studies, the positive effect of sequential presentation on reducing the number of false identifications significantly outweighs any losses in the number of correct identifications.

General Discussion

The authors found that enriching the lineup cues enhances identification accuracy. They also found that presenting lineups sequentially reduces false identifications without hindering correct identifications.

Their results suggest that using lineup procedures with voice samples, showing lineup members from three-quarter poses, and showing lineup members walking into and out of the observation room should enhance accuracy. In addition, sequential presentation should be used, since it is no more costly than simultaneous presentation and yields far superior results.