Eye Witness Flashcards
Introduction
Eye Witness: 75% of eyewitness testimony errors involved in 75% of DNA exoneration cases.
Eye witness may be only evidence available in absence of other forensic evidence.
Eyewitness Testimony Is Error Prone: Buckhout, (1980) Recorded crime on TV, 2000 people rang in. But, 1800 made wrong ID.
Similarity of fillers in line ups:
Too highly similar = problematic – eyewitnesses experience difficulty in identifying the suspect.
But, moderately similar = higher identification and fewer false positives (Fitzgerald et al 2015).
Face composites: can harm line up identification performance (Wells et al 2016). The composite building process can harm builder’s memory for the face. Reduction in chance of later identifying the original face.
Combining face composites: yields improvements in face likeness (Bruce et al 2003). A morph of all four composites created a better likeness than individual composites. In a line up, 4 morphs performed better (more correct choices and fewer false positives).
Evofit composites: holistic evolutionary face imaging technique (holistic processing) Higher suspect identification.
Eye Witness Interviewing:
Enhanced Cognitive Interview: Several Stages of Training: Very few police offices engage in it
83% of interviews in the UK are not done with the cognitive interview – because it is time consuming and the pressures of resources are too great.
Amount of time under observation Distance from suspect Visibility Obstructions the view of the witness Known or seen before Any special reason for remembering the suspect Time lapse since witness saw suspect Error or discrepancy between the witness’s description in their first and subsequent accounts.
Postdiction Variables:
Individual differences:
Facial recognition Wells et al, (2006): • Immediate, longer delays • Fewer corrects IDS (51% vs 61%) • Increase False IDs (32% vs 24%)] • Increased Delay Longer response time
Time estimation:
exposure to the event and the Face is correlated with face recognition performance (Memon et al 2003).
More exposure to target: (45s vs 12s) = increase identification.
But people vary hugely in time estimates, some being poor judgers of time (16 to 496% of actual duration), but consistent across time estimation tasks. This can create a problem when lab experiments are moved into interviews.
• Is it possible to determine reliable from the non-reliable?
Self-report decision processes
Self-report measures of face recognition skill linked with identification accuracy for culprit (Olsson & Juslin, 1999).
• Subjective, which can cause problems
How much awareness do we really have?
Bindemann et al 2014 – not very, tend to generalise ‘good; ability to recognise faces with unfamiliar faces.
Response Latency (how long to make an identification)
Event Details (Ebbesen & Rienick 1998)
Post event suggestion & misinformation
Choice Blindness
- Immediate vs 4-week delay
- Reduction in number of recalled facts
- % error consistent
Post event suggestion & misinformation:
86% of witness discuss their memory with co-witnesses
• Causes conformity (Gabbert et al 2004)
• Can cause misinformation (Wright et al 2000)
• Couples vs strangers (French et al 2008
Choice Blindness:
• Difficulty detecting manipulation of choice they made (Sagana et al 2014).
• ‘This is the person you selected’
• 39% - 68% of manipulations remained undetected
• Many people thing they would able to detect this manipulation
We know a lot about what factor impair eye witness reliability:
Estimator variables:
Eyewitness: (emotional state, Intoxication “Alcohol myopia theory” suggest that alcohol increases focus on central detail (Evans et al, 2009).
Perpetrator Variables: Disguises (Busey & Tunniclif, 1999)
Study phase: Participants viewed a staged robbery for ½ of the participants the robber wore knit pullover cap.
Test Phase: Identified perpetrator from video line-up 45% no hat group vs 27% hat group.
Situation Variables: exposure duration (Memon et al 2003) Distance (Lindsay et al 2008).
Affecting Multiple Stages
Perception: Design to attend to and encode stimulus properties (Poor encoding).
Encoding: Stereotypes and Prejudice – Remember more information that is consistent with our scripts (Holst and Pezdek, 1992).
Storage: Misleading Information – Representation with additional information & Regenerate (Visualisation).
Choice Blindness: difficult detecting manipulation of a choice they made (Sagana et al 2014).
39 – 68% of manipulations remain undetected. But many people think they would be able to detect this manipulation.
Retrieval: Demand Characteristics (Pressure to give an answer).
