Factors Affecting Eyewitness Testimony

• Eyewitness testimony:

  • eyewitness testimony is when an observer of a crime's recollection of evens is used as evidence in a criminal trial
  • memory is not like a 'video tape' that can be replayed exactly. It is reconstructive
  • this means that we 'reconstruct' an account of events and this can be influenced by our prior knowledge and expectations and it may also be influenced by things that we are told after the events

  factors that affect eyewitness testimony include;

  • anxiety
  • misleading information (leading questions and post-event discussion)
• Johnson and Scott (1976)- procedure:

  • research shows that anxiety has a negative impact on the accuracy of eyewitness testimony
  • Johnson and Scott (1976) conducted research into the effect of anxiety on eyewitness testimony
  • participants were invited to a psychology experiment and asked to wait in a waiting room. A nearby receptionist excused herself to run an errand, leaving the participant alone

  participants were then subjected to one of two conditions

  • low anxiety
  • high anxiety

  the participants were then shown a set of 50 photos and had to select the man that had left the room
• Johnson and Scott (1976) - low anxiety condition:

  • participants overheard a conversation from the laboratory about equipment failure
  • a man left the room holding a pen with his hands covered in grease
• Johnson and Scott (1976) - high anxiety condition:

  • participants overheard a heated exchange about equipment failure, followed by the sound of breaking glass and crashing chairs
  • a man ran from the room holding a bloodied knife
• Johnson and Scott (1976) - findings:

  • Johnson and Scott found participants in condition 1 (low anxiety) had a 49% success rate at identifying the man
  • participants in condition 2 (high anxiety) had a 33% success rate
• Johnson and Scott (1976) - conclusions:

  Johnson and Scott concluded that;

  • anxiety decrease the accuracy of eyewitness testimony
  • this happens because people focus on the weapon, rather than other details. This is known as the 'weapon focus effect' (proposed by Loftus)
• Yuille and Cutshall (1986) - procedure:

  • research showing that anxiety has a positive impact on the accuracy of eyewitness testimony
  • Yuille and Cutshall (1986) investigated the effect of anxiety in a real life shooting, in which one person was killed and another person was seriously wounded during a failed robbery of a Canadian gun store
  • Yuille and Cutshall interviewed 13 witnesses 4-5 months after they had been interviewed by the police
  • they also completed a retrospective 7 point questionnaire about how stressful they found the experience of witnessing the original shooting
• Yuille and Cutshall (1986) - findings:

  • it was found that all 13 witnesses gave eyewitness accounts that were highly consistent with the original interviews they gave the police
  • most importantly, they found that those who reported being more stressed by the events initially gave more detailed and more consistent accounts (88% accurate) of the original events, compared to those who were less stressed (75%)
  • this is in direct contrast to the conclusions found by Johnson and Scott
• Contrasting findings between Johnson and Scott and Yuille and Cutshall:

  • the contrasting results can be explained using the Yerkes-Dodson arousal curve
  • this curve suggests that too little anxiety leads to poor performance (recall), as does too much anxiety
  • instead, an optimal level of anxiety leads to the best performance (recall)
• Evaluation of anxiety affecting eyewitness testimony:

  S - Johnson and Scott's study was higher in control, whereas Yuille and Cutshall was higher in ecological validity

  W - Johnson and Scott measured surprise rather than anxiety - Pickel (1998)

  W - individual differences in responding to anxiety - Bothwell (1987)

  S - the positive effect of anxiety on eyewitness testimony can be supported by the Christianson and Hubinette bank robbery study
• Johnson and Scott's study was higher in control, whereas Yuille and Cutshall was higher in ecological validity (identify):

  • a major strength of the research by Johnson and Scott is that it is higher in control than the Yuille and Cutshall study
• Johnson and Scott's study was higher in control, whereas Yuille and Cutshall was higher in ecological validity (explain):

  • the Johnson and Scott study was a lab experiment with highly standardised procedures, meaning the experiment was tightly controlled. For example, the participants were each shown the same 50 photos of men to choose from
  • however, this study was low in ecological validity because it took place in artificial surroundings in a lab experiment
  • in contrast, the Yuille and Cutshall study may be less controlled as it took place in the real world with little control over extraneous variables
  • however, the Yuille and Cutshall study had much higher ecological validity because it used real life crimes
• Johnson and Scott's study was higher in control, whereas Yuille and Cutshall was higher in ecological validity (conclusion):

  • this is important as it suggests that the Johnson and Scott study is more internally valid but we are less able to generalise to real crimes, whereas the Yuille and Cutshall study is more realistic but more affected by potential extraneous variables
  • however, both contribute to our understanding of memory
• Johnson and Scott measured surprise, rather than anxiety - Pickel 1998 (identify):

  • a problem with the Johnson and Scott study is that it measures the effect of surprise, rather than anxiety
• Johnson and Scott measured surprise, rather than anxiety - Pickel 1998 (explain):

  • Pickel (1998) found that participants who had watched a video of a man approach a hair salon with a handgun (highly unusual, high threat) were no worse at identifying the same man approach with a whole raw chicken (highly unusual, low threat)
  • this suggests that the low accuracy of eyewitness testimony is down to the object being surprising, rather than threatening
• Johnson and Scott measured surprise, rather than anxiety - Pickel 1998 (conclusion):

  • this is important as it suggests that the Johnson and Scott study wasn't measuring the effect of anxiety on eyewitness testimony
  • instead it is measuring the effect of surprise
  • however, this has low ecological validity as the participants were watching a video of the event
• Individual differences in responding to anxiety - Bothwell 1987 (identify):

  • one problem with the research is that it cannot account for individual differences in how anxiety affects eyewitness testimony
• Individual differences in responding to anxiety - Bothwell 1987 (explain):

  • Bothwell et al (1987) assessed participants on their levels of neuroticism before testing their eyewitness testimony abilities
  • they found that participants labelled as neurotic (became anxious quickly) showed decreasing levels of accuracy as stressed increased, whereas those labelled as stable showed increasing levels of accuracy as stress increased
• Individual differences in responding to anxiety - Bothwell 1987 (conclusion):

  • this is important as it suggests that we cannot generalise the findings of anxiety research to everyone
  • anxiety affects different people in different ways and this needs to be taken into account when applying these findings to real eyewitness testimonies
• The positive effect of anxiety on eyewitness testimony - supported by Christianson and Hubinette's bank robbery study (identify):

  • the positive effect of anxiety in eyewitness testimony can be supported by the Christianson and Hubinette bank robbery study
• The positive effect of anxiety on eyewitness testimony - supported by Christianson and Hubinette's bank robbery study (explain):

  • Christianson and Hubinette interviewed 110 witnesses of 22 real life bank robberies and found that those who had been directly threatened or subjected to violence during the robberies were significantly more accurate in their eyewitness accounts than those who were just bystanders
  • this was consistent up to 15 months after the events. This suggests that in the real world anxiety increases the accuracy of eyewitness testimony
• The positive effect of anxiety on eyewitness testimony - supported by Christianson and Hubinette's bank robbery study (conclusion):

  • this provides support for the Yuille and Cutshall finding that eyewitness testimony is improved by higher levels of anxiety adding greater validity to this conclusion
• Misleading information:

  misleading information can be split into two subtopics;

  • the effect of leading questions on eyewitness testimony
  • the effect of post-event discussion on eyewitness testimony
• Loftus and Palmer (1974): the effect of leading questions on eyewitness testimony - procedure and results:

  • Loftus and Palmer (1974) conducted research into the effect of leading questions on eyewitness testimony, carrying out two experiments
  • experiment 1 - 45 undergraduate students were shown a video of two cars colliding. They were asked a series of questions about what they had seen. The critical question was "how fast were the cars going when they ___ into each other?"

  Loftus and Palmer found the average estimates;

  • smashed - 41mph
  • collided - 39 mph
  • bumped - 38mph
  • hit - 34 mph
  • contacted - 32 mph

  Loftus and Palmer concluded that leading questions can have a significant impact on the accuracy of eyewitness testimony
• Loftus and Palmer (1974): the effect of leading questions on eyewitness testimony - procedure and results:

  • experiment 2 - in a very similar setup, 150 different participants were shown another video of a car crash and were split into 3 groups
  • group 1 - their question was "how fast were the cars going when they hit each other?"
  • group 2 - "how fast were the cars going when they smashed into each other?"
  • group 3 - no questions were asked out the vehicles speeds

  one week later the participants were asked if they had seen any broken glass (there was no broken glass in the film). Participants in group 2 ('smashed') were more likely to report seeing broken glass than those in group 1 or group 3. This provided further support that leading questions can have a significant impact on the accuracy of eyewitness testimony
• Gabbert et al (2003): the effect of post-event discussion on eyewitness testimony - procedure:

  • post-event discussion is when people discuss what they have seen after the event
  • Gabbert et al (2003) wanted to investigate the effect of post-event discussion on the accuracy of eyewitness testimony
  • participants were split into two groups and then watched one of two videos. Video 1 - a girl entered an office to return a book and was seen stealing a wallet. Video 2 - the same video, but instead of seeing the girl steal the wallet there was a close-up of the book title
  • the participants were then paired together and allowed to discuss what they had seen (believed they had watched the same video). They were then asked about what they had seen individually
• Gabbert et al (2003): the effect of post-event discussion on eyewitness testimony - findings and conclusion:

  • it was found that 74% of participants reported seeing something that they had definitely not seen

  this means that;

  • people who sae video 1 reported seeing the title of the book
  • people who saw video 2 reported seeing the girl steal the wallet

  Gabbert concluded that post-event discussion can lead to contamination of (and decreased accuracy) of eyewitness testimony
• Evaluation of the effect of misleading information on eyewitness testimony:

  W - both studies are low in ecological validity

  W - both studies have sampling issues

  W - contrasting research - Vredebeldt's theatre study

  S - positive practical applications - improving EWT
• Both studies are low in ecological validity (identify):

  • a problem with the research into misleading information is that it lacks ecological validity
• Both studies are low in ecological validity (explain):

  • both the Loftus and Palmer and the Gabbert et al study used a video of an incident to test eyewitness testimony
  • this is a problem because it is not the same as witnessing an incident in the real world and so people's memories may be different in an artificial situation
  • for example, watching a video of an incident may cause less anxiety than one in the real world meaning that the eyewitness testimony is more or less accurate
• Both studies are low in ecological validity (conclusion):

  • this is important as it suggests that we do not know how misleading information affects eyewitness testimony in the real world
• Both studies have sampling issues (identify):

  • a problem with the research into misleading information is that it lacks population validity
• Both studies have sampling issues (explain):

  • Both the Loftus and Palmer and Gabbert et al study used a biased sample of participants in their research
  • Loftus and Palmer used university students in an opportunity sample. Gabbert used 60 students and 60 older adults
  • this is bad because it means that the sample is going to be skewed more towards younger people. Younger people tend to have better memories and so may make better eyewitnesses. Furthermore, they use a large number of students. University students are likely to have better memories than the general population
• Both studies have sampling issues (conclusion):

  • this is important as it suggests that research into misleading questions only applies to students and not the general population
  • this lowers the population validity of the findings in the general population
  • these effects may actually be significantly worse
• Contrasting research - Vredeveldt's theatre study (identify):

  • a problem with Gabbert's research is that there have been contradictory findings in the real world
• Contrasting research - Vredeveldt's theatre study (explain):

  • Vredeveldt (2015) interviewed couples leaving the theatre after watching a violent play and found that when the couples were interviewed together and had time to discuss they made significantly fewer errors (10 on average) than when they were interviewed alone (14.6 on average)
  • this contradicts the findings found by Gabbert et al in their lab experiment
• Contrasting research - Vredeveldt's theatre study (conclusion):

  • this is important as it suggests that Gabbert's findings cannot be generalised to the real world and real life post-event discussion may have a positive impact on the accuracy of eyewitness testimony
• Positive practical applications - improving eyewitness testimony (identify):

  • one strength of research into misleading information is that there are positive practical applications to improve eyewitness testimony
• Positive practical applications - improving eyewitness testimony (explain):

  • research from Loftus and Palmer has shown that leading questions can have a negative impact on the accuracy of eyewitness testimony
  • police forces are now trained in cognitive interview techniques which aims to reduce the use of leading questions
  • furthermore, Gabbert's research has shown how post-event discussion can contaminate eyewitness testimony. Police forces are now trained to separate eyewitnesses as soon as possible to prevent post-event discussion
• Positive practical applications - improving eyewitness testimony (conclusion):

  • this is important as it suggests that the research can be used in the real world to improve the accuracy of eyewitness testimonies and improve the criminal justice system