Eyewitness Testimony and Lineups

Lineup Bias and Procedures

The Problem with Biased Lineups

  • A lineup was thrown out of court due to biases:
    • One photo was of different quality.
    • Highlighted "phenotypic Hispanic features" compared to "phenotypic African American features" in other photos.
    • The person in question was the only one wearing a black jacket.

Double-Blind Lineups

  • To avoid bias, lineups should be double-blind, similar to double-blind medical testing.
  • The person conducting the lineup should not know who the suspect is.
  • Example of suggestive behavior: "Are you sure that's the one? Why don't you take a look at number two?"

Dissertation Study on Lineup Administrator Knowledge

  • Participants watched a mock crime on TV where someone stole an iPod from a backpack.
  • Participants then watched a video on how to conduct a lineup.
  • The study varied whether lineup administrators knew who the suspect was.
    • Condition 1: Administrators were not told who the suspect was (double-blind).
    • Condition 2: Administrators were told person number two was the suspect (single-blind).
  • The actual suspect from the video was not in the lineup.

Results of the Study

  • Mistaken identifications occurred by chance.
  • Administrators knowing the suspect led to more mistaken identifications of an innocent person.
    • Almost 8% more identifications of the innocent suspect when the administrator knew who the suspect was.
    • Filler identifications decreased by about 10%.
  • Participants with good memory were not significantly affected.

Analysis of Switchers in Lineups

  • Secret videotaping and coding were used to analyze what was happening in the lineups.
  • 72 cases of "switchers" were observed (initially picking one person, then changing to another).
  • Y-axis: Number of switchers.
  • X-axis: Initial choice (bottom) and final choice (top).

Patterns in Switcher Behavior

  • Double-blind condition:
    • Switchers primarily shifted from filler to filler.
  • Single-blind condition (administrator knows the suspect):
    • Switchers were more likely to switch from a filler to the actual suspect.

Implications

  • In double-blind lineups, switches mostly occur between fillers (random noise).
  • When the administrator knows the suspect, switchers can "miraculously" identify the suspect.
  • Comparison of blue bar (double-blind) and black bar (single-blind) highlights the bias.