Identity: A crisis of confidence? Or is it resemblance? An exploration of the different approaches by which eyewitness evidence can be obtained from lineups

Author Identifiers

Dominic Jordan


Date of Award


Degree Type


Degree Name

Doctor of Philosophy


School of Arts and Humanities

First Advisor

Adrian Scott

Second Advisor

Donald Thomson

Third Advisor

Pamela Henry


Research has shown that eyewitness identification decisions are fallible and often mistaken. Although considerable attention has been afforded to identification decision accuracy and its improvement, mistaken identification decisions continue to contribute to costly errors at the evidentiary stage of the criminal justice system process (i.e., wrongful convictions). Several prominent researchers have suggested, by way of explanation, that the existing framework for obtaining eyewitness evidence from lineups, namely, identification, is inadequate. Indeed, the assumption that witnesses when presented with a lineup, can make reliable identification decisions (i.e., can reliably determine that a lineup member is the same unfamiliar person seen previously committing a crime) has become increasingly disputed. Instead, it has been suggested that witnesses, when presented with a lineup, use the perceptual or visual pattern of information available at encoding and retrieval to make inferences regarding the likely identity of a perpetrator. Guided by this rationale two alternatives, which have approached the eyewitness decision task as one which involves multiple match-based comparisons rather than a single categorical identification decision, have been considered - the non-categorical confidence and non-categorical similarity approaches.

The non-categorical confidence approach describes a process whereby witnesses express judgements, on a continuous scale, regarding how certain they are that each member of a lineup is the target. The non-categorical similarity approach describes a process whereby witnesses make judgements, on a continuous scale, regarding how similar each lineup member is to their memory of a target. Proponents have asserted that confidence and similarity can both facilitate recognition in lineup decision tasks by allowing witnesses to access probative memorial information (i.e., match-to-memory) free from distortive, non-memorial influences which impact categorical identifications.

While initial assessments seem to support this assertion, continued examination of alternative non-categorical approaches and their use in lineup-based recognition tasks is required. To date, attention afforded to understanding how non-categorical judgements are elicited or impacted under variable conditions remains limited. Further, the relationship between categorical identification decisions, non-categorical confidence judgements, and non-categorical similarity judgements in lineup-based decision tasks has not previously been examined. To address these gaps in the literature three experimental studies were conducted.

In each experimental study, participants completed four experimental lineup trials which involved: viewing a target person, completing a brief distractor task, and responding to the lineup. In responding to each lineup participants either made a categorical identification with an associated post-decisional confidence rating (Study 1, N = 343), or non-categorical judgements, on one of two 11-point scales, regarding how certain they were that each member of a photographic lineup was the target person (Study 2, N = 314) or how similar each member of a photographic lineup was to the target person (Study 3, N = 362). Variables in each of the three studies were identical, and included target appearance (no change, change, within-participant), lineup procedure (simultaneous, sequential, between-participant), target/replacement position (early, late, within-participant) and target presence (present, absent, within-participant). Results for each of the three studies are reported independently, however, in a separate chapter additional results are reported which compare data from Studies 1, 2, and 3.

In all three experimental studies, variables primarily influenced responses made from target present lineups. While the effects of procedure and position on participant responses varied across the three experimental studies, altering target appearance exerted a consistent, moderate to large effect on participant responses made using each of the three recognition approaches. As the appearance variable represents a direct manipulation of perceptual, and by extension memorial, signal strength this finding provides some evidence to suggest that non-categorical confidence and similarity judgements access and rely on the same information which underpins categorical identification. This finding was reinforced, in part, by more direct comparisons of the experimental data which revealed a high degree of correspondence in participant responses obtained using identification decisions, non-categorical confidence judgements, and non-categorical similarity judgements. Interestingly, however, direct comparisons also revealed some notable differences between mean measures of confidence and similarity. First, the numerical value of mean similarity exceeded that of mean confidence within all four lineups. Second, discrepancy, which measures the difference between the highest value judgement (i.e., max judgement) and second highest value judgement in a lineup trial, was significantly higher for confidence than similarity within all four lineups. The theoretical implications and practical considerations associated with the main findings and the methodological limitations associated with the experimental materials and research design are discussed. Additional research regarding alternative approaches to lineup-based decision tasks is required, and therefore recommendations for future research are made accordingly.

Access Note

Access to this thesis is embargoed until 24 09 2023.

Access to this thesis is restricted. Please see the Access Note below for access details.