Eyewitness Testimony

It may seem that eyewitness testimony is an effective means of identifying and convicting criminals. Still, as we will see in this article, eyewitness testimony is not the most reliable method of identifying criminals.

In 2011, for example, DNA testing led to the release of 51-year-old Cornelius Depree. He spent 30 years in prison after being convicted of raping and robbing a 26-year-old woman. He was convicted because the victim identified him as the culprit through eyewitness testimony, but DNA evidence cleared his name.

How does this misidentification occur? And what is the relationship between eyewitness testimony and memory biases? Eyewitness testimony is an important area of research in psychology. Let us look at some factors affecting eyewitness testimony and look at some eyewitness testimony examples.

Eyewitness testimony, Flaticon

Factors affecting eyewitness testimony

There are several factors affecting eyewitness testimony. Let’s cover them one at a time.

Confirmation bias

Lindholm and Christianson (1998) examined the concept of in-group/out-group status’ and its effects on eyewitness testimony. The researchers showed Swedish and immigrant students videos where a simulated robbery occurred, and the culprit seriously injured a cashier.

Misleading information

Loftus and Palmer (1974) suggested that misleading information given after the event can distort eyewitness memories through leading questions (implying an answer). Their two experiments demonstrate the concept of misleading information.

Experiment 1

Aim: To investigate the effects of leading verbs on eyewitness accounts of a car accident. They hypothesised that misleading information in the form of a leading question could change the accuracy of eyewitness testimony.Participants: 45 students participated in the experiment.Procedure: They divided the participants into five groups. Each group watched seven clips of traffic accidents. After watching each clip, participants described what happened as if they were eyewitnesses. They were then asked several questions about what they had seen, including one crucial question. This was, ‘How fast were the cars going when they ___ into each other?’. Each group had to answer this question with a different verb:

• Smashed.

• Collided.

• Hit.

• Bumped.

• Contacted.

Findings: The more extreme the verb, the faster participants estimated the car’s speed. The average speed estimates for each of the verbs were:

• Smashed: 40.5 mph.

• Collided: 39.3 mph.

• Hit: 34 mph.

• Bumped: 38.1 mph.

• Contacted: 31.8 mph.

The researchers concluded that there could be two explanations for their findings:

1. Response bias influenced the participants’ answers – they were unsure of what estimate to give for speed, so their choice of verb influenced their answers. For example, for the group labelled ‘smashed’, the verb ‘smashed’ caused them to report a higher speed. However, their memories were not distorted.

2. The misleading question (with the choice of a verb) caused a change in the participants’ memories. For example, the verb ‘smashed’ caused participants in the ‘smashed’ group to remember the accident as being more severe than it was. If this memory bias is the case, the researchers could assume that participants remembered other features not initially shown in the clips. Therefore, they conducted a second experiment to test for memory distortion.

Eyewitness testimony experiment, Flaticon

Experiment 2

Participants: 150 students participated in the experiment.Procedure: Participants watched a movie about a car accident and were given a questionnaire about the movie they had just seen. The crucial question was about the speed of the cars.Fifty participants had to answer: ‘How fast were the cars going when they smashed into each other?Fifty participants had to answer: ‘Approximately how fast were the cars going when they collided?Researchers did not ask the other fifty any questions about the speed of the cars (control group).

After one week, the participants had to complete a questionnaire again. The crucial question was: ‘Did you see any broken glass?’ (there was no broken glass in the video).

Findings: 32% of participants in the ‘smashed’ condition reported seeing broken glass, compared with 14% in the ‘hit’ condition and 12% in the control condition.Loftus and Palmer’s (1974) study shows that suggestive questions can easily distort memory. Information given to someone after an event can mix with the original memory.

Evaluation

Strengths:

• Because this was a laboratory study, researchers could control confounding variables, which improved internal validity.

• The results of this study impact real-life interrogation tactics by demonstrating the effects of leading questions.

Weaknesses:

• The study only interviewed university students, so the results cannot be generalised to other populations.

• The study has low ecological validity because it was conducted in a laboratory. Participants experienced fewer emotions while watching the videos than in an actual situation. Questions might not distort your memories due to emotions in a real situation.

• The questions were leading (using arousing or emotional verbs instead of simple, less arousing verbs), which suggests what the study was investigating. Participants may then have figured out what was being investigated in the study and responded accordingly, affecting the validity and reliability of the experiment.

Post-event discussion

Gabbert et al. (2003) studied the effect of post-event conversations on the accuracy of eyewitness testimony.

Post-event discussion, Pixabay

The simulated crime involved a girl walking into an empty university room to return a book. Two videos were shot of this event, but from different perspectives: video A showed details that were not in video B, and vice versa. Only in video B did the girl commit the crime and took 10 pounds from a wallet.Procedure: Researchers divided participants into two groups, the control group (tested individually) and the co-witness group (tested in pairs). In the individual group, one half watched video A and the other half watched video B. In the co-witness group, each person in the pair saw a different video, although they were told they saw the same video. Afterward, participants in the co-witness group could talk to their partners about what they had seen.Afterward, all participants were tested individually with a questionnaire about their memory of the crime.

Findings: 71% of the co-witness group remembered information they had not seen in the video. Moreover, in the co-witness group, 60% of the participants who had not seen the girl commit a crime said that the girl was guilty. Thus, although they had not seen the crime themselves, they thought the girl was guilty after discussing it with their partner who had seen it. There were no differences in memory distortion between old and young.The results show that post-event discussions can distort memory. The information participants in the co-witness groups had heard from others influenced their memory. This information was then incorporated into their original memory.

Strengths:

• Two populations participated in this study, university students and older adults, and there were no differences in memory distortion between young and old. Therefore, this study has good generalisability that memory distortions affect young and old equally.

• The study can be applied to police work to train police officers not to judge the statements of multiple witnesses as more accurate just because they have the same information.

Weaknesses:

• Because this was a laboratory experiment, ecological validity is low. Participants knew they were taking part in an experiment and thus may have paid more attention to the videos. In real-life situations, people may be exposed to less information.

• We cannot be sure that the memory distortion was due to post-event information. It could be due to conformity effects (informational influence).

Anxiety

In real life, eyewitness testimony is often used when witnesses recall anxious situations, such as a violent crime scene. How does anxiety affect memory? Psychologists have researched the effects of anxiety on eyewitness memory.Let us take a look at some of these studies.

Johnson and Scott (1976)

Procedure: Participants were invited into a laboratory and had to wait in the reception area. The receptionist there excused herself to run an important errand and went into an adjacent room. At this point, participants experienced one of two conditions. These were:

1. ‘No weapon’ condition: Participants heard a conversation about an equipment failure. Then a man left the room and walked past the participants with greasy hands and a pen in his hand.

2. ‘Weapon’ condition: Participants heard a hostile exchange of words, the sound of breaking glass and overturned chairs. Afterwards, a man with a bloody knife ran out of the room.

Both groups of participants were shown 50 photographs and asked to identify the man.Findings: The ‘no weapon’ condition group correctly identified the man 49% of the time, compared to only 33% for the ‘weapon’ condition group. The participants who saw the knife felt higher levels of fear and focused more on the weapon than on the man’s face. This phenomenon is known as the weapon focus effect.

Strengths:

• There is scientific evidence for the gun focus effect. In a study by Loftus et al. (1987), participants were asked to observe a person either pointing a gun at them or handing a check to a cashier and subsequently receiving cash. Researchers recorded participants’ eye movements. Participants made more eye fixations and looked longer at the gun than at the check. In addition, participants’ memory was worse in the gun condition than in the check condition.

• In their meta-analysis, Fawcett et al. (2013) found that focusing on the weapon harmed eyewitness memory.

• The study has high ecological validity because participants did not know it was a study and should have responded authentically.

Weaknesses:

• There are ethical problems, as participants were deceived about the nature of the study and exposed to a man with a bloody knife.

Yuille and Cutshall (1986)

Yuille and Cutshall (1986) studied eyewitness memory of a real-life crime.

Procedure: Twenty-one witnesses saw a shooting in which one person was killed, and two were seriously injured. The police interviewed them all. Four to five months later, 13 of the witnesses agreed to participate in this research study and were interviewed about what they saw. The eyewitness accounts they provided to the police and the research team were analysed.Findings: Participants were highly accurate in their descriptions, and after five months, researchers noted only a slight change in recall accuracy. In addition, they resisted leading questions, and the level of anxiety they felt at the time of the crime did not appear to affect their recollection. These results suggest that weapon focus and fear do not affect eyewitness memory accuracy in real life.

Strengths:

• It was a real-life event with much higher ecological validity than laboratory experiments.

Weaknesses:

• This study was only one study with 13 participants. Therefore, it is difficult to generalise the results or to say that the results of this study indicate that the results of the laboratory experiment are wrong.

• Ethical issues related to asking participants to recall traumatic events.

• Because this was a field study, there was no way to control for outside variables, such as that participants may have read information about the crime before the interview, which contributed to their recall accuracy.

• It is difficult to determine whether proximity to the crime or the stress of witnessing the crime contributed to recall accuracy.

The Yerkes-Dodson Law states that there is a relationship between stress and performance. Stress increases performance, but only up to a point. After that, excessive stress harms performance. Applied to eyewitness memory, this means that anxiety can increase memory performance, but only up to a certain point.Deffenbacher (1983) examined 21 studies of eyewitness testimony and found that low levels of anxiety lead to a poor recall, whereas higher levels of anxiety increase recall, but only up to a point. If a witness feels too much fear in connection with a crime, memory decreases.