memory in forensic contexts

Importance of memory

• critical in many forensic settings including interviewing and suspect identification

• Misunderstanding of how memory works can leads to wrongful conviction (garrett 2017)

• Innocence project found eyewitness misidentification accounts for 63% cases and 72% evidence supporting conviction came from eyewitness evidence. 192 cases in the US have been exonerated through DNA.

• Testimony of at least one eyewitness was involved in 90% 40 proven miscarriages of justice

• Ronald cotton spent 10 years in prison for rape he didn’t commit based on one victim that identified him out of a line up

Factors affecting encoding

• external: duration, exposure, frequency, physical factors such as illumination, obstructions, distance, salience and weapons etc

• Internal: divided attention, personal relevance, mood, emotion, stress, dissociative symptoms, distortions, psychological issues, substance effects and traumatic injury

• Perspective

• Bias

• Perception

Memory types

• procedural

• Semantic

• Episodic: made up of lots of pieces of related information e.g what happened, person descriptions and related information such as emotions, thoughts what we saw/ heard/ read but people can and do make mistakes

Multi-store model of memory - Atkinson and shiffrin 1968

• sensory input moves into sensory memory (short retention, large capacity) and can be transferred into short term/ working memory (15-30 second retention, limited capacity) if attended to

• Working memory can move to long term memory if rehearsed and can flick between the two stores whether you are retrieving information of storing it again.

• Long term memory has infinite retention and capacity but can be forgotten over time via decay/ retrieval failure

Working memory model (baddeley and hitch)

central executive is made up of the visuospatial sketchpad (visual semantics), episodic buffer (episodic LTM) and phonological loop (language/ sounds) and all interact with one another

Stages of memory

• encoding: processing/ acquiring information = formation

• Storage: maintaining information = maintenance

• Retrieval: recalling information = recovery

• If something impacts any of these stages → changes in memory

Yerkes Dodson law - arousal and performance

• cues attended to is related to degree of arousal

• Tunnel effect = central details remembered over peripheral

• As arousal increases, attention narrows and fewer environmental stimuli are attended to

• Simple task = focused attention, flashbulb memory and fear conditioning

• Difficult task = impairment of divided attention, working memory, decision making and multitasking

Stress and attention - encoding

• stress associated with worse recall (valentine and mesout 2009)

• Observers have better recall than those directly at risk in simulated armed police operations (hope et al 2016)

• Witnesses to a real crime showed a positive relationship between stress and accurate recall even after 5 months (yuille and cutshall 1986)

Attention and stereotyping - encoding

• people attend to the unusual items e.g. geese/ guns and tend to forget other details when a weapon is present. Affected by retention interval, exposure duration and threat but unaffected by setting (fawcett et al 2013)

• Tend to misremember schema incongruent information about crimes - that don’t match what you expect (tuckey and brewer 2003)

Non-deceptive factors that effect encoding

• stress

• Emotion

• Age

• Disease

• Disability

• Neurodiversity

• Poverty

• Mood

• Mental health

Intoxication - encoding

• alcohol usually associated with fewer details and tendency to focus on central features

• Vredeveldt et al 2018 found Amsterdam ‘coffee shop’ users remembered fewer details about a mock crime video when recruited leaving

Emotion - encoding

• deffenbacher found 10 studies showed increased/ stable accuracy with higher arousal levels but 11 showed the opposite and high levels of stress impaired accuracy and recall on perpetrator identification and recall of details of crime

Exposure duration - encoding

• Shapiro and penrod 1986 found hit rate increases with increases in time spent viewing a target however, increased exposure to faces increased false alarm rates

• Read 1995 suggested that longer exposure may bias peoples assumptions about what they ought to remember. Target present line ups → accuracy with longer exposure but target absent → increased false alarm

• Memon, hope and bull 2003 found longer exposure → accurate identification and rejection and inflated witness confidence

Memory retrieval

• process of accessing information from memory storage

• Can suffer retroactive and proactive interference

Retroactive interference - memory retrieval

• occurs when new information disrupts or ‘contaminates’ the information to be remembered - misinformation effect

• Can occur post incident from speaking to other witnesses, police, emergency services, reporters, family etc

• Recall can also suffer interference through poor questioning (leading questions)

• Suggestibility - influence of others expectations

• Misattribution - information attributed to an incorrect source

Proactive interference - memory recall

occurs when old information/ memories interfere with new information → taints in their recall of a recent event

Barriers to memory retrieval

• distraction at the time of encoding

• Trauma

• Age

• Lack of concentrated effort at recall

• Passage of time

Context reinstatement - memory retrieval

• tulving and Thomson 1973 - most efficient when the context of the original event is reinstated at the time of recall

• Prior to inviting an episodic recall from a witness, their recall can be assisted by mentally recreating the external environment and their cognitive and emotional states that existed at the time of encoding

• Breivik killed 69 on an island and police spent weeks interviewing and decided to take him back to the scene to help him recall/ show his account.

tools for effective recall in interview

• report everything

• Change perspective

• Change order

• Mental reinstatement of context - senses

• Category clustering recall

• Recreate circumstances associated with event (emotional, cognitive and physiological

• Can be traumatising for victims to re-live events like this

Free recall model

• initial recall, free recall/ context sketch, focused recall/ probing and review

• Experience and judgement will guide you on how long to spend at each stage, and which units meet situation requirements/ need revisiting

• Allows them to feel in control of the narrative

• Research has shown allowing interviewees to recall details in category clusters can help increase information yield e.g. person, action, location, object, conversation and sound/ sensory details

Brain structures that handle memory

• cortex = short and LTM, semantic and episodic memory

• Cerebellum = procedural memory

• Medial temporal lobe and hippocampus = LTM semantic and episodic memory

• Amygdala = emotional memory

Episodic memory

• recalling events in ones personal history such as details about people and objects present and locations of specific events

• Memory of personal experiences

• What happened, where and when

• Context dependent

• Susceptible to corruption and falsification (loftus and pickrell 1995)

• Emotional state at the time of encoding can impact later recall (kensinger and corkin 2003)

Repeated events

• can lead to issues in recall as events blur together

• E.g. child sexual abuse, domestic violence and criminal harassment

• Several similar events makes detailed memory for specific incidents, times more generalised

• Long time delays reduce success of conviction

Repeated recall

• likely to happen in forensic contexts

• Variability across interviews may be due to omission and metacognitive factors

• Contradictions can occur in repeated reports of single events and loss of availability of original response - are they recalling the original event or their previous recall?

• Longer delays → increased effect

• Can be dealt with be allowing witnesses to re-read statements

• People may repeat recall by talking with friends/ family, police call operator, initial reporting officer, secondary investigator, discussions, evidence in court etc

Incidental vs intentional encoding

• incidental = experienced and then later recalled rather than ‘learned’

• Among recalled following intentional encoding greater

• Greater binding of contextual and item information in intentional encoding

• Generic memory = details of repeated events e.g. abuse are incidental and compounded by lack of distinctiveness

Central vs peripheral details

• what the witness considers central is likely to be the details that give the event meaning

• From a forensic POV, central details would be details that help catch and convict a perpetrator e.g. descriptions

• Not always compatible

Perception

• people can be looking at the same thing but interpret them differently

• Eyes can deceive you and every stimulus is perceived in its most simplest form

• Continuations = the way in which an object appears may compel you to look elsewhere

• Closure = when an object is incomplete we will try to make it more recognisable ourselves and fill in the gaps

• Proximity = when elements are placed close together we may perceive it as one group

Delay

• initial interviews to crimes are often very quick but trial may be some months/ years later

• Some crimes may not be immediately recognised so even the initial interview can be delayed

• ‘Over learned’ events are reliable after time but forensic memories are not generally learned (incidental)

• Memory may not be rehearsed over time → decay

Gist vs verbatim - storage

• fine grain detail = specific details from memory that would fit with verbatim memory but liable to rapid decay and may need to be elicited via cued recall. Vulnerable to leading/ misleading questions

• Coarse grain detail = general details that will form gist memory and be more stable over time/ long lasting, most likely elicited via free recall but vulnerable to schema-consistent distortion

Meta memory processes - monitoring

• reality/ source monitoring relate to our ability to differentiate between potential sources of information - imagined/ real

• Distinctiveness heuristic = if an event isn’t distinctive, it may be unreported

• When processes fail to reject false information, incorrect information may be given → rejection of true information

• We might have a clear memory of something but we may not know where we got it from

Reconstructive memory - retrieval

• links to constructive memory and schemas

• Can be affected by perspective, imagination, beliefs, knowledge, self - esteem etc

• Schemas and memory can change since the first recall due to anything such as speaking with friends etc → reconstructive when you recall an event over a few times, each time you are recalling the recall then the original memory

Issues with memory - howitt

• wells et al 2007 - rates of incorrectly identifying a filler as the suspect in real police line ups range from 20-24% =15.2 same level of risk for innocents

• Granhag et al 2013 - witnesses to the murder of a Swedish foreign minister were not accurate - 35% details were correct, 24% partly correct → high percentage of misleading information

• Van koppen and lochun 1997 - correspondence between witness and police descriptions of robberies → sex and height being most common detail mentioned (not very helpful) but witness descriptions were relatively sparse, only mentioning 8/43 details on average with some features having low agreement. Longer statements, distance between offender, duration, feelings of threat and delays predicted accuracy.

Later intrusions to memory - loftus and palmer 1974

• ppts witnessed images of a car accident and later asked a series of questions in which key information was embedded

• Estimates of speed were affected by wording in questioning e.g. ‘smashed’ → 1/3 faster vs ‘contacted’ = encouraged faster estimates but not necessarily altering the memory

CCTV evidence - howitt

• doesn’t necessarily lead to crime reduction - welsh and Farrington 2002

• Brown 1995 - can reduce crime short term but → crime displacement

• Scott et al 2016 - found drug addicts admitted they stopped injecting in toilets and moved to streets instead

• Familiarity of the person in a low quality video → easier recognition

• Kemp et al 1997 - difficult to identify if two different images are the same person if it’s a stranger e.g. 29% correctly identified the woman from a video of an incident