EWTs + cognitive interviewing ALL

Eyewitness Testimonies

What is an EWT?

• the evidence provided by people who witnessed a particular event or crime

• It relies on recall from memory

Eyewitness Testimonies

what might an EWT include?

• descriptions of criminals (eg: hair colour, height)

• descriptions of the crime scenes (eg: time, date, location)

Eyewitness Testimonies

What is the accuracy of EWTs?

• often inaccurate

• this has important implications when it comes to police interviews

Eyewitness Testimonies

What do many cognitive psychologists focus on?

working out what factors affect accuracy of EWTs + how accuracy can be improved in interviews

Eyewitness Testimonies

What is the first stage of EWT memory?

• The witness encodes into LTM details of event + people involved

• Encoding may be only partial + distorted, esp bcs most crimes happen v quickly, often at night, + sometimes accompanied by rapid, complex and often violent action

Eyewitness Testimonies

What is the second stage of EWT memory?

• Witness retains the info for period of time

• Memories may be lost/modified during retention (most forgetting happens within first few minutes of a retention interval)

• other activities b/w encoding + retrieval may interfere w/ the memory

Eyewitness Testimonies

What is the third stage of EWT memory?

• Witness retrieves the memory from storage

• The presence/absence of retrieval cues/the nature of the questioning may significantly affect accuracy of recall

Eyewitness Testimonies

What are ways EWTS can be affected?

• misleading information

• post event discussion

• anxiety

Misleading information

What are the types of misleading information?

• Leading Qs - when Q’s wording implies answer (investigated by Loftus + Palmer (1974) + + Yuille and Cutshall (1986))

• Post-event discussion

• Repeat interviewing

Misleading information

How many experiments did Loftus and Palmer conduct investigating misleading information?

2

Misleading information

Loftus & Palmer Experiment #1 (1974) Procedure:

• 45 Ps watched seven films of different car accidents

• Were given a series of Qs → one was the “critical Q” (the one about speed estimates of the car) - a leading Q → IV = verb used, DV = average speed estimate

Misleading information

What were the 5 verbs used in Loftus and Palmer #1 (1974), and what was the average speed estimate? + a mnemonic to remember

Some Cars Bring Home Children

1. Smashed - 40mph

2. Collided - 39mph

3. Bumped - 38mph

4. Hit - 34mph

5. Contacted - 31mph

Misleading information

Loftus & Palmer Experiment #1 (1974) conclusion:

Two potential explanations for the results:

1. Response Bias Explanation - Ps were unsure → leading Q activated schema → guided answer (this does not necessarily suggest the memory was manipulated)

2. Substitution Explanation - leading Q caused Ps to recollect the crash differently → affecting memory of the speed

Misleading information

Why did Loftus and Palmer conduct the second experiment?

To get more evidence for substitution explanation, as would suggest memory can be manipulated

Misleading information

Loftus & Palmer Experiment #2 (1974) Procedure:

• 150 different Ps → one film of car accident → used only smashed and hit in critical Qs

• Control: no Q about speed

• Week later asked: did you see the broken glass?

Misleading information

Loftus & Palmer Experiment #2 (1974) findings:

• Both experimental conditions = the higher speed estimate = the higher the recollection of broken glass

• 'Smashed' (16/50) → more likely to experience confabulation (saying ‘yes’ to Q about glass) vs control (6/50) + 'hit' (7/50)

• Leading Qs changed memory Ps had of video

Misleading information

Loftus & Palmer Experiment #2 (1974) conclusion:

• Substitution explanation (schemas alter memories)

• Confabulation - type of memory error where gaps in memory = unconsciously filled w/ fabricated/misinterpreted/distorted info

Misleading information

Why are post event discussion not good?

Memory of event may be altered/contaminated by discussing events w/ others and/or being questioned multiple times

Misleading information

What is a study that investigates post-event discussions (P-ED) effects on EWT accuracy?

Gabbert et al (2003):

• 120 Ps watched video of girl entering an office, looking through some drawers, + examining a wallet → half saw her take money, half didn’t

• Control: no discussion

• Experimental: P-ED in pairs (saw diff versions) → told saw same video

• Individually answered questionnaires → P-ED: 71% recalled info they hadn’t seen

• Gabbert concluded P-EDs = dangerous, bcs can severely affect accuracy of EWTs

Misleading information

What are the post-event discussion (P-ED) explanations? Which is more likely?

• Memory Contamination - Memories become distorted due to P-EDs → information + misinformation = combined

• Memory Conformity (conformity effect) - co-witnesses reach consensus of what actually happened (memories conform) → may be to fit in/gain social approval (NSI), OR bcs think they’re wrong, + others = right (ISI)

Skagerberg + Wright (2008): memory contamination = more likely than conformity bcs Ps often reported blend of info from discussion + video clip they saw

Misleading information

Evaluation: contradicting evidence to Loftus + Palmer (1974)

Yuille and Cutshall (1986): interviewed 13 witnesses of real armed robbery in Vancouver 4mnths after → asked ‘did you see a broken headlight?’ OR ‘did you see the broken headlight?’ → Ps = v reliable/detailed (accuracy = 79%-84% - verified w/ forensic evidence) → 10/13 recalled accurately no broken headlight → contradicts Loftus + Palmer, + has higher mundane realism/eco val bcs real event → suggests L+P’s results due to artificial task bcs watching video = NOT as emo arousing as real-life events (esp bcs Ps safety = threatened) → affecting recall

Misleading information

How does repeat interviewing effect EWT accuracy?

• Each time an eyewitness is interviewed there’s a chance interviewer comments will become incorporated into their recollection of events

• If interviews use leading Qs, may also effect it → alter the memory

• Have the biggest impact when children are being interviewed

Misleading information

Evaluation: Practical application on EWTs

Supporting research has had a GREAT INFLUENCE on how EWs are interviewed → eg: cognitive interview: don’t use leading Qs, interviewer = encouraged to ask v. few, if any Qs → SO has been shown to have positive real-life applications

Misleading information

Evaluation: P-ED supported by real life evidence

1995, Oklahoma bombing → one witness claimed saw murderer, w/ an accomplice → initially no other witnesses could describe accomplice, BUT later they too claimed to recall them → eventually first witness realised their recollection was wrong → support’s idea that P-ED can negatively impact accuracy of EWTs → gives eco val

Misleading information

Why might the 1995, Oklahoma bombing witnesses make the same mistake as another witness? explain in terms of the conformity effect

• Maybe part of ISI, because they thought they were wrong, and the other person is right, and so doubt themselves, which may have lead to internalisation

• Also NSI is less likely because the context is serious, so they are less likely to be motivated by need for approval

Misleading information

Evaluation: Potential for demand characteristics:

experimental design of misleading info experiments may cause demand characteristics → results skewed bcs of Ps expectations about purpose → leading Qs could give Ps clues about nature of experiment (eg: may have realised experiment = about susceptibility to leading Qs/effect of P-ED/repeat interviews) → so Ps change behaviour to prove/disprove what they think aim is → issue bcs reduce validity + reliability of experiment

Misleading information

Evaluation: issue w/ Loftus + Palmer (1974) methodology

Asking about speed of car = difficult bcs alr estimate (some may be more/less accurate perception in first place bcs do/don’t drive etc → P variables) and then gets distorted: asking to recall details of event they experienced can be more reliable and less likely to be distorted → so Loftus + Palmer (1974)’s methodology = vulnerable to extraneous variables

Misleading information

Evaluation: issue w/ Yuille + Cutshall (1986) + 1995 Oklahoma bombing

No control variables bcs quasi-experiment/one off case study → difficult to know level of rehearsal/P-ED of each P → (for Y+C) could be those who agreed to study spent the most time thinking/reading about the case so why less vulnerable to distortion → (for both) not replicable, + hard to generalise bcs v specific/unique circumstances → SO need more research into real EWTs to make confident conclusions about influence of leading Qs

Anxiety

What is anxiety?

What is it’s effect on EWTs?

• Physiological response to environmental stressor: increased heart rate, blood pressure, breathing rate ect

• Can have either positive or negative effect on accuracy of EWT

Anxiety

What is the weapon focus effect?

Loftus et al (1987): The presence of a weapon draws witness’s attention + impairs ability to identify other key details (eg: person’s face)

Anxiety

What studies shows a positive effect of anxiety on EWT accuracy?

Christianson & Hubinette (1993) - meta analysis:

• Ps = 110 witnesses of real life bank robberies (some onlookers, some bank employees)

• Ps = surprisingly accurate recall of robbers' clothing + behaviour (accuracy still evident 15 mnths later)

In their follow up:

• Interviewed 58 witnesses to bank robberies in Sweden (victims (high anxiety) or bystanders (low anxiety)) 4-15 mnths later

• All = >75% accurate recall of details robberies

• Most anxious victims = more accurate

ISSUE: culture bias

Anxiety

AO1 + AO3

What study found negative effects of anxiety on EWTs?

Johnson and Scott (1976)

• Ps thought were taking part in lab experiment

• While in ‘waiting room’ (deception), experiment took place w/o them knowing

  1. Low anxiety condition: Ps overheard casual convo in next room → man walked out past Ps w/ pen + grease on his hands

  2. High anxiety condition: Ps overheard heated argument in next room → man walked out holding a knife covered in blood

• Ps then had to identify the man from a set of 50 photos

• 1 had 16% more accuracy than 2 → suggests anxiety has negative effect on EWT + supports weapon focus effect

Anxiety

what do psychologists believe about the different amounts of anxiety/arousal increases? (inverted U theory)

small increases in anxiety and arousal may increase the accuracy of memory, but high levels have a negative effect on accuracy

Anxiety

example of how different amounts of of anxiety/arousal increases affect on accuracy:

In violent crimes (where anxiety and arousal are likely to be high), the witness may focus on central details (eg: a weapon) and neglect other details (eg: what the criminal was wearing)

Anxiety

What study suggests that the weapon focus effect (WFE) is due to surprise, not anxiety?

Pickel (1998):

• Showed Ps video thief entering hairdresser’s w/ several variations including: scissors (high threat, low surprise), raw chicken (low threat, high surprise)

• EWT accuracy sig. worse in high surprise condition vs high threat

• suggests WFE is due to surprise rather than anxiety

Anxiety

What is the inverted U theory/Yerkes-Dodson Law model?

• Relationship between emotional arousal and performance looks like an 'inverted U' (see graph).

• Deffenbacher (1983) used this to explain EWT

  1. Lower levels of anxiety produce lower levels of recall accuracy

  2. Becomes more accurate as level of anxiety increases

  3. The optimal anxiety level gives maximum accuracy

  4. If anxiety goes beyond this point → sharp decline in the accuracy of memory

Anxiety

Evaluation: limitation of Yuille + Cutshall (1986)

No control variables bcs quasi-experiment → difficult to know level of rehearsal/P-ED of each P → could be those who agreed to study spent the most time thinking/reading about the case so why less vulnerable to distortion → not replicable, + hard to generalise bcs v specific/unique circumstances → SO need more research into real EWTs to make confident conclusions about influence of leading Qs

Anxiety

Evaluation: evidence that anxiety increases accuracy/contradicts weapon focus effect

Yuille and Cutshall (1986): interviewed 13 witnesses of real armed robbery in Vancouver 4mnths after → Ps = v reliable/detailed (Ps who reported highest levels of stress = most accurate (~88% vs ~75% of less-stressed group) → weapon = NO affect) → 10/13 recalled accurately no broken headlight (leading Q had not effect) → supports idea that anxiety can increase EWT accuracy + contradicts weapon focus effect

Anxiety

Evaluation: strength of Yuille and Cutshall (1986)

high mundane realism/eco val bcs real event → more emo arousing than artificial task (esp bcs Ps safety = threatened) → easier to generalise to real life than lab experiments

Anxiety

Evaluation: alternative explanation

Alt of impact on accuracy of EWT = leading Qs → Response Bias Explanation - when unsure of answer → leading Q activated schema → guided answer (this does not necessarily suggest the memory was manipulated) or Substitution Explanation - leading Q caused people to recollect event differently → affecting memory → supported by Loftus and Palmer (1974) → suggests leading Qs can also affect EWT accuracy

Evaluation: inverted U explanation oversimplistic

Anxiety = difficult to define + measure bcs has many elements: cognitive, behavioural, emotional + physical → inverted U explanation only acknowledges physiological (physical) arousal’s affect on performance (bio reductionist) → fails to account for other factors (eg: effect of emotional experience of witnessing crime) on accuracy of EWT

Anxiety

Evaluation: unethical

Creating anxiety (eg: Johnson + Scott (1976)) = potentially unethical bcs breaks ‘protection from harm’ duty → So real life studies = beneficial: psychologists interview people who have alr witnessed event, so, anxiety = naturally occurring, not created → ethical issues don't challenge findings of studies, but do raise questions about ethical misconduct

Cognitive interview (CI)

What are the 4 cognitive interview components?

1. Report everything

2. Mental reinstatement of original context

3. Changing the order

4. Changing the perspective

Cognitive interview

How does reporting everything help a CI?

Small details the witness sees as irrelevant may be important + trigger other memories

Cognitive interview

How does mental reinstatement of original context help a CI?

Witness returns to original crime scene 'in their mind' → based on concept of context dependent forgetting + cues from this context may trigger recall

Cognitive interview

How does changing the order help a CI?

• Prevents people from using their expectations of how the event must have happened instead of the actual event

• Prevents dishonesty (harder to remember rehearsed story out of order)

Cognitive interview

How does changing the perspective help a CI?

• Witness recalls incident from other people's pov

• Prevents influence of expectations + schemas on recall

Cognitive interview

What does an enhanced cognitive interview entail?

13 basic skills for the interviewer, eg:

• Establishing rapport

• Active listening

• Avoiding interrupting + leading questions

Cognitive interview

(AO1 or AO3)

Evaluation: Supporting research of CI

Geiselman et al (1984): Ps saw film which simulated violent crime + questioned 2 days later → CI = higher number of correct items recalled, no difference in number of incorrect items/confabulation → Geiselman et al (1985) found same results when also compared to interview under hypnosis → supports higher effectiveness of CI

Cognitive interview

Evaluation: mundane realism

Geiselman’s studies lack mundane realism (how well a study mimics the real world) → bcs isn't everyday occurrence to watch a film and then be interviewed + bcs less emotional arousal from video than real witnessing may effect results bcs stakes lower → SO limits generalisability

Cognitive interview

Evaluation: strength of supporting research

Geiselman’s studies have ecological validity → field studies: took place in interview rooms (where real CIs would take place) → bcs environment = same as real life, giving it ecological validity → strength Ps may have displayed more representative behaviour than if lab study

Cognitive interview

Evaluation: practical application w/ children

Research found children recalled more details, less likely to be influenced by leading Qs, and no increase in inaccurate/false info reported w/ CIs → strength of CIs bcs can be practically applied to use w/ child witnesses, as can be hard to interview children after crimes (eg: violent crimes → scared) + more susceptible to leading Qs (so interviewer must be V careful w/ wording) →using CI can aid interviewers to get higher volume of accurate/detail info from child witnesses

Cognitive interview

Evaluation: slightly different CI techniques

Studies on effectiveness of CI inevitably use slightly different techniques → diff researchers may use variations of CI/enhanced CI, + police have evolved own methods → means it’s difficult to draw conclusions about CI in general bcs wide differences in application + use