Identifying the Perpetrator 1 & 2 (wk 5)

Question 1

Why is eyewitness identification important?

Answer 1

In the US 200 people per day become defendants after being identified bya witness in a lineup
in 75% of DNA exonerations at least one eyewitness made a false identification
So, “Eyewitness identification evidence is amongst the least reliable forms of evidence and yet is persuasive to juries” Wells et al 1988

Question 2

How to False Eyewitness Identifications Occur

Answer 2

Can occur in a range of stages of memory

1. Encoding = estimator variables
   = the variables that can impact encoding such as what happened during the crime
2. Storage = estimator variables
   =the variables that can impact storage such as a delay in time etc
3. Retrieval = system variable
   = the way in which we retrieve memory can influence its accurate such as suggestive questioning etc

Question 3

Types of variables that can attribute false ID’s

Answer 3

Estimator Variables
= variables that impact the encoding or storage of a memory

System Variables
= variables that impact the retrieval of a given memory
- tend to be able to be controlled for or modulated by crime investigators (ie. creating an interview that is not-suggestive etc

Both estimator and system variables can influence the accuracy of a memory trace

Question 4

Estimator Variables and eyewitness identification

Answer 4

= variables that impact the encoding or storage of a memory

• sometimes we just don’t have a good recollection of a perpetrator’s face and there’s a lot of factors that can influence our perception and memory of a face

Types of estimator variables

1. Stable Witness Factors; Stable factors about the witness that can influence their encoding (eg. age, ethnicity, own race bias etc). These are stable factors that don’t rapidly change.
   For example; children tend to encode faces based on superficial information like hairstyles, or we are less good at distinguishing between two faces of individuals who are from a different race to our own
2. Malleable Witness Factors = the changeable factors about a witness that may influence encoding and storage
   (eg. intoxication, tiredness etc)
3. Stable Target Factors = stable factors about the person we are witnessing (perpetrator). These can make the person distinctive or not.
   - facial distinctiveness is key here.
4. Malleable Target Factors = changeable factors about a perpetrators face (eg. disguises, hairstyle etc). A person will be harder to identify if between the event and the recognition test they try and change their appearance (hair as a superficial factor tends to be a key eyewitness descriptor)
5. Environmental Factors; the environmental factors during the crime or witnessed even such as light levels, visibility, presence of a weapon or crime seriousness etc can influence encoding) . This is mostly influencing salience
6. Post-Event Factors; delays, exposure to misinformation, verbal overshadowing etc can all influence our encoding and storage

Question 5

Race Bias; Hills and Pake 2013

Answer 5

In an eyetracking paradigm, it was found that different ethnicities will distinguish peoples faces by looking at different parts of the face (eg. amount of space between the eyes or amount of space between the nose and other features etc)

Is an example of our our own ethnicity can govern how we encode information (stable witness factor)

Question 6

Verbal Overshadowing

Answer 6

= The effect whereby people are required to describe someone verbally before a lineup is conduction. Which results in impaired identification.

Believed to occur as in the verbal description people pick apart facial features rather than considering the person holistically which is required for accurate lineup identifications (check the reading for more on this)

Question 7

Key Eyewitness Descriptors (Shepherd and Ellis 1996)

Answer 7

Malleable Target Factors aka the variable factors about a perpetrator can influence our encoding and subsequent ability to identify them in a lineup.

Certain factors tend to have a larger effect on our ability to identify people than others.

eg. Shepherd and Ellis
1. Hairstyle 27%
2. Eyes (14%)
3. Nose (14%)
4. Face Shape (13%)
This looked at facial features and the most salient details we encode based off of. Uniquely, the most malleable target factor aka hair seems to be one of the key features we use in encoding a face

Question 8

System Variables

Answer 8

System Variables
= variables that impact the retrieval of a given memory
- tend to be able to be controlled for or modulated by crime investigators (ie. creating an interview that is not-suggestive etc

The way in which eyewitness identification evidence is elicited can make things difficult for witnesses

Question 9

Methods for eyewitness identification (aka System Variables)

Answer 9

1. Verbal Descriptions
2. Facial Composites
3. Mugshot Books
4. Lineups
5. Instruction Bias
   6.Presentation Bias
6. Foil Bias
7. Investigator Bias

Question 10

Verbal Descriptions

Answer 10

Asks witnesses to describe the perpetrator.

Critique:
- These sorts of descriptions are often too brief to be forensically useful
- Descriptions often focus on external features and other features that are easily changed (ie hairstyles and clothes) - this effect tends to be worse in children
- When witnesses have initially provided a highly detailed description, investigators often place more weight on the subsequent visual identification (so lineup outcomes, rather than description made)
- despite this, there is no reliable association between thee nature of a verbal description and the subsequent visual identification accuracy

Question 11

Facial Composites

Answer 11

= the process of working with a witness to construct a visual representation of the perpetrator (creating a likeness)

• This could use manual processes like having a sketch artist work with the witness or a computer system (like composite image layering)

Question 12

Efficacy of facial composite methods

Answer 12

People are notoriously poor at constructing accurate facial composites even when these are of family members or celebrities and when there is a reference photo

The inefficiency may be due to facial composite formation processes being at odds with how we remember faces (being that they are feature-based ie. what did his eyes look like, rather than based on configural processes ie. the holistic feel)

Question 13

Identicate System of Facial Composites

Answer 13

This process rather than working with a sketch artist, has clear sheets layered ontop of each other to create a composite face. So, you can select from a range of images of ears, which ones look the most accurate and layer this with the most accurate looking forheard etc.
This rather than asking witnesses to recreate features of the face, they simply need to recognise facial features, however, could this be a source of post-event misinformation
These identicate systems can be manual or computerised

Question 14

Why are Facial Composites poor methods for perpetrator identification

Answer 14

On the whole, we are bad at remembering unfamiliar people, so producing an identicate image of their face is inherently difficult as there isn’t a strong memory of their face (even people we are familiar with, or celebrities are hard to create composite images of)

• the main critique of facial composites is that they require a feature based recognition of faces we have seen. Whereas our memory uses configural or holistic processing methods (so we look at the face as a whole).

Question 15

Whole-Face Processing Methods

Answer 15

Evofit - is a software that helps form configural based composite images.
In this a verbal description is entered into the software, then a series of images are presented that align with this and witnesses are asked to pick the face out of the selection that is the mot similar to the perpetrator and these get more and more specific to form an image

Question 16

Caricatures as a tool for forming facial composites

Answer 16

This works on the idea that our memory for faces is based upon the unique features of a face, that differ from the ‘average’ face we have formed in our mind.
Caricatures amplify these points of differences of a face. So, on the one hand could be good to portray the important features remembered, but also could be limited by artistic variation (as every artist will draw a face differently)

Question 17

Mugshot books for perpetrator identification

Answer 17

Mugshot books ask witnesses to look through albums of faces of individuals already in police systems (known offenders) and asks them to see if the perpetrator is in one of those books

These aren’t great methods as
1. Mugshot books can serve as a source of misinformation and infiltrate the original memory trace. as such mugshot books can contaminate subsequent identification performance (ie. in a lineup) as people recognise the face from the mugshot, rather than from the crime
2. The knowledge that photos are all of known offenders can decrease the identification threshold (people are less rigorous in their identification, as if they do pick the wrong person, it is still a ‘bad’ person who will be convicted – they aren’t at risk of mis-identifying an innocent person

Question 18

Ronald Cotton and Jennifer Thompson

Answer 18

Jennifer Thompson was sexually assaulted after a home-break-in. She was determined to remember the face of the assailant so was a very confident witness.
A composite sketch was made when Jennifer went to the Police and Ronald Cotton was believed to match this. On interview he accidentally gave a false alibi (got his weekends muddled) so was arrested and placed in a lineup.
In the lineup Jennifer identified Ronald as the assailant and he was imprisoned. Ronald filed for a retail but was found guilty a second time - in this second trial, Ronald argued that Bobby Pool, a man he met in prison who was convicted of sex crimes was responsible but Jennifers testimony where she said it was Ronald not Bobby placed him back in jail.
Finally Ronald was set free with a DNA exoneration
After his release, Ronald and Jennifer became friends

Question 19

Lineup Combination Types

Answer 19

In all lineups, a suspect in placed in and amongst a set of foils (or confederates).

1. Target Present Lineup
   The suspect is in fact the guilty party
2. Target Absent Lineup
   The suspect is innocent and the guilty party (aka the target) is not in the lineup

Question 20

Source of Error in Lineups

Answer 20

In all lineups, Police have some evidence or reason to place the suspect into the lineup.
In target present lineups, a witness will either select a foil (in which case they are a poor witness and nobody will be convicted) or the witness will select the suspect (in which case police proceed with a conviction allowing justice to be served)
In target absent lineups, a witness will either select a foil (in which case they are a poor witness and nobody will be convicted) or the witness will select the innocent suspect, in which an innocent person will be at risk of being convicted (as police don’t have reason to believe its not them)

Question 21

Role of Relative Judgement in Police Lineups

Answer 21

When witnessses are presented with a lineup, they make a relative judgement. So rather than selecting a person because they are the perpetrator, they may select a person as they look the most similar to the perpetrator out of the people present

Children are particularly susceptible to this error, which may be due to them feeling pressured to pick a person (aka instruction bias)

Question 22

Instruction Bias

Answer 22

When a lineup is presented it could be a target present or target absent lineup and in real life contexts we don’t know which is which but there is hope that the lineup will be a target present one

Generally speaking, witnesses are reluctant to reject the lineup altogether (ie say the target is not present). As a result, there can be a fair number of false alarms due to target absent lineups eliciting an errored identification being made.

Instructions (ie. don’t pick someone if the perpetrator is not there) can reduce the tendency but not eliminate it, as often witnesses assume police wouldn’t present the lineup if a suspect wasn’t in it. As such instructions may be taken as a grain of salt thing

Additionally, there is the risk of relative judgement (people thinking they have picked the perpetrator when really they’ve picked the person who looks most similar out of the presented lineup).

Question 23

Memon, Hope and Gabbert 2002 study of Lineups

Answer 23

Participants were given a post-lineup questionaire.
90% of the participants advised that they expected the target to be present, even though there were clear instructions to advise they might not be
95% of these participants were able to recall the instructions (so they hadn’t forgotten that this was mentioned)
47% of incorrect witnesses wold make the same decision in ‘real life’

Question 24

Zajac and Karegeorge 2009 Study of Children and Lineups

Answer 24

• sampled 8 - 11 yr olds
• they did a visit to the Police station during which a confederate event occurred in which a confederate asked for the keys from a cop whilst he was giving a speech
• The next day a varied lineup was conducted. These were either target present (for the confederate) or target absent
• Within the TP or TA trials, participants would either do a control lineup manipulation (where they are asked to do a lineup under normal conditions) or a wildcard manipulation in which there was the standard images or a silhouette to select if the target was absent.

Results;
in control trials; as expected in TA trials there were False Alarms (in which someone was selected who was not the target)
In wildcard trials; correct performance on the target-present trials weren’t affected (so just as many correct identifications occurred) and in target absent trials fewer false alarms occurred. In fact, in wildcard conditions, target absent and target present accuracy did not differ

Question 25

Presentation Bias

Answer 25

There are two main ways we can present suspects to eyewitnesses for identification. These apply to both live lineups and photospreads

1. Simultaneous lineups (uses relative judgement) – all people are presented at the same time
2. Sequential lineup (uses absolute judgement) – people are presented one at a time.

Sequential lineups result in fewer identifications overall (so more often will at the end of all the photos have not identified anyone) but importantly produces fewer false alarms (which is good)

There’s 2 key reasons for this;
1. In sequential lineups, witnesses don’t know how many images in total will be presented
2. As soon as an identification is made, no more photos are presented

Question 26

Foil Bias

Answer 26

Looks at how investigators select FOILS for their investigation.

• in NZ foils are selected to be similar to the suspect
• alternatively, foils could be selected to be similar to a participants verbal description (even if that differs from the suspect)

A Backfire effect
- occurs when foils are too dissimilar to the suspect, which may increase the change of an innocent suspect being wrongly identified (eg. the suspect is a black man whilst foils are white women - the black man even if he is innocent is more likely to be selected).

Conversely, foils who are too similar to the suspect may confuse the witness and increase the risk of a foil being identified

Question 27

What are the options in a lineup for optimal FOIL similarity?

Answer 27

1. FOILS could be selected to match the witnesses verbal description
2. FOILD could be selected to match the suspect

Question 28

Wells et al 2000 - study of FOIL selection in lineups

Answer 28

Serves as evidence for foils needing to be chosen on the witnesses verbal description

“Research and theory indicate that the fillers should fit the verbal description of the perpetrator (which is given by the eyewitness prior to viewing the lineup), but additional similarity should not be sought. Physical similarity between the suspect and the fillers beyond the level of the description provides no additional protection to the innocent suspect and can actually harm the eyewitness’s ability to identify the perpetrator”

Question 29

Clark and Tunnicliff 2001 - FOIL Selection

Answer 29

Shows participants a simulated crime.
Participants were asked to give a verbal description of the perpetrator.
Then participants underwent one of 3 lineups;
1. target present - description -matched foils
2. target absent - description matched foils
3. target absent - suspect matched foils

Findings;
In the target absent (aka suspect is innocent) lineups, the suspect made identification errors more often (in error) when foils were suspect matched rather than description matched
Overall, the accuracy (ability to reject the lineup as say target is not present) was about the same.

This suggests that in accordance with Wells et al 2000 that foils should be description matched (not suspect matched)

Question 30

Investigator Bias

Answer 30

Looks at the influence of investigator cues on eyewitness decisions

Could include
- verbal behaviour (ie. take another look to be sure)
- Facial gestures (nodding, growing etc)
- body movements (nodding or head shaking)

Question 31

Phillips et al 1999 - study of investigator bias

Answer 31

Presented a series of target absent lineups under one of two conditions
1. single blind condition (investigator is aware of who the suspect vs foils are)
2. double blind condition (neither witness or investigator know who the suspect is)

Findings showed that fewer people picked the suspect (who is innocent and should be rejected) when they did a sequential lineup and the investigator was blind. This suggests that investigator knowledge has a bigger effect in sequential lineups than simultaneous ones (suggesting there is more investigator influence in single blind sequential lineups)

In a lineup (sequential or simultaneous) an investigators knowledge of suspect placement, can increase false identification rates when the suspect is innocent. This influence is greater in sequential lineups

Question 32

Postdiction Variables

Answer 32

= measurable products that correlate with identification accuracy in a non-causal manner
Ie. certain factors (system or estimator variables) may help or hinder accuracy but don’t cause it. eg. if a person is far away and it’s dark, it’ll cause your inaccuracy but if you’re a confident witness (a postdiction variable) you may be more likely to give accurate resp

Question 33

Types of Postdiction Variables

Answer 33

1. confidence
2. response latency
3. self- reported decision processes

Question 34

Confidence as a Post-diction Vaariable

Answer 34

The relationship between confidence and accuracy is controversial. But simple correlations tend to find either no relation or a small positive one.

Confidence–Accuracy calibration;
Suggests there perfect calibration would show a linear relationship between confidence and accuracy (100% confidence = 100% accuracy) but this is not the case, (ie. 100% confidence produces 85% accuracy)

Problems with Confidence-Accuracy Relationships ;
1. Confidence tends to get stronger over time and over subsequent identifications (ie. in the lineup.. I think it is him –> in court… I know it is him)
2. Postivie feedback can artificially inflate confidence (eg policeman saying yup I thought it was him too)
3. Postive feedback also inflates retrospective reports of condience and viewing conditions - especially for people who are mistaken

Question 35

Response Latency as a Postdiction Variable

Answer 35

Accurate witnesses may make their decisions more quickly that inaccurate ones

However, there’s not much scientific specifity in fast or slow responses - ie. fast or slow compared to what?
Dunning and Peretta 2002 suggest 10 - 12 seconds

Subsequently research suggests this number may vary depending upon viewing conditions

Question 36

Self-Reported Decision-Process

Answer 36

How do witnesses report their identification processes?
Are they prone to the same kinds of distortion as confidence?