alt 162 final cards

Memory is ...

-reconstructive

-While tapping our memories, we filter or fill in missing pieces of information to make our recall more coherent

Misinformation effect:

incorporating misleading information into one's memory of an event

Hit versus smashed what asked about speed

-Then asked about broken glass at the scene

-Smashed versus hit -- smashed reported glass more

-Elizabeth Loftus experiment?

-Change of word changes speed estimate -- and what they remembered could potentially be used in court

Recovered Memory Epidemic

-Unfolded in 1990's

-In therapy, adult victims recover memories of having been abused during childhood

-Charges brought against incredulous parents

-Guilty verdict

The Courage to Heal

-Published in 1988

-Guide for women of survivors of sexual abuse

-Made assersitions as if they were scientifically true -- influential with therapists

Theory behind psychotherapy for repressed memories

-Certain memories are too painful to bear

-As a defense mechanism, those memories are repressed

-Repressed memories create psychology disorders

Therapeutic Methods

-Assigned reading of self-help books

-Dream analysis

-Hypnosis

-Guided imagery

Wade, Garry and Lindsay experiment:

-Using false photographs to create false childhood memories

-Digitized photos of 3 true events childhood -- took photos of family members and pasted them into false photo of hot air balloon

-Interviewed family member three separate times about those 4 events (tell every detail)

-Everyone remembered the true events -- 30% remembered false balloon ride

-After 3 interviews almost 50% remembered false memory

Linda Williams Study

-1994 -- recall of childhood trauma

-Participants: 129 adult women who had been evaluated in early 1970's for possible sexual abuse

-Method: participants were interviewed about possible traumatic experiences during childhood

-40/129 did not describe an episode of sexaul abuse

-long periods with no memory of abuse should not be regarded as evidence that the abuse did not occur

Critique of Williams study

-Unwillingness to report childhood sexual trauma to an unfamiliar interviewers does not mean that the trauma was forgotten

-Some of the abuse took place when the women were as young as 10 months (infantile amnesia)

-Some of the abuse was such that the child might not have recognized it as such

What the False Memory debate taught us about memory

-False memories are surprisingly easy to implant

-What makes them so east to implant is that memories are reconstructive

Proponents (recovered memories are real)

-15% of adult women who report memory for childhood abuse also report having forgotten about that at some point

-None of these women recovered their memories in therapy

Opponents (recovered memories are false)

-False memories are extremely easy to implant, so it does not make sense to convict anyone without corroborating evidence

Since the 1990s, DNA testing has overturned more than

350 wrongful convictions of the innocent

Of the 350 wrongful convictions, how many were due to eyewitness misidentifications?

72%

The Case of Ronald Cotton

-In 1984, a college student named Jennifer Thompson was raped

-Shortly thereafter she picked Ronald Cotton out of a photo lineup (said she was certain)

- Cotton was sentenced to life in prison plus 54 years, and he served almost 11 years in jail before being exonerated by DNA testing

The National Registry of Exonerations

-20 or so per year (DNA exonerations)

-But there are a lot of other ways to be exonerated

-DNA isn't the only way to get them out

-Purgeory and false accusations and Official Misconduct

Research has shown that sequential lineups

reduce misidentifications and increase accuracy compared to traditional simultaneous lineups

Lineup Construction and Lineup Fairness

-A lineup should contain only 1 suspect

-The suspect should match the physical description provided by the eyewitness

-The police should have some additional reason for believing that the suspect may have committed the crime

-All of the fillers (aka "foils") should also match the physical description provided by the eyewitness (i.e., the suspect should not stand out)

If a non-witness can identify the suspect from the description alone then

the lineup is unfair

Simultaneous lineup

show pictures simultaneously, all at once

Sequential Lineup

one picture at a time

Mock-Crime Laboratory Studies

-Each participant (n = 200) watches a simulated crime (e.g., a video of a young man stealing a laptop)

-Followed by a lineup memory test:

-Half (n = 100) are then tested using a target-present lineup

-The other half (n = 100) are tested using a target-absent lineup

Lindsay & Wells (1985)

- Simultaneous versus Sequential lineups

- Huge drop in False ID rate with sequential lineup

2013 Survey of US Police Departments

-30% adopted sequential procedure

-Huge impact of psychological research

Absolute vs. Relative Judgments

1. Simultaneous lineup

-Eyewitnesses feel "pressure to choose" and therefore rely on a relative judgment (choose the most familiar lineup member)

-Alternative way of thinking: eyewitness use a lower standard for making an ID

2. Sequential lineup

-Eyewitnesses feel less pressure to choose, so they rely on an absolute judgment (choose a lineup member only if the match to memory is good enough)

-Alternative way of thinking: eyewitness use a higher standard for making an ID

Reduce "pressure to choose" by inducing a more

conservative standard

Conservative approach means

fewer ID's (both false and correct)

A higher ROC curve indicates

greater discriminability

The Concept of Response Bias

-Do not make an ID if you are just guessing

-Do not make an ID unless you are reasonably sure

-Do not make an ID unless you are very sure

-Do not make an ID unless you are absolutely certain

More than One Way to Perform ROC Analysis

-Manipulate response bias across conditions, using instructions to vary the degree of confidence the witness should have before making an ID from the lineup

-Collect confidence ratings from eyewitnesses who make an ID and vary the requisite level of confidence after the fact

To compute the most "liberal" ROC point,

count all suspect IDs (including guesses)

To compute the most "conservative" ROC point,

count only suspect IDs made with the highest level of confidence

Simultaneous has ______ ROC than sequential

higher

Why are Simultaneous Lineups Diagnostically Superior (Wixted & Mickes (2014, Psychological Review)

- proposed a diagnostic feature-detection model

-Basic idea: simultaneous lineups make it easier to appreciate the existence of shared (and therefore non-diagnostic) features

-The enhanced awareness of non-diagnostic feature makes it easier for eyewitnesses to tell the difference between innocent and guilty suspects

Diagnostic feature-detection hypothesis (Wixted & Mickes, 2014)

-Innocent and guilty suspects share facial features - namely, features that correspond to the suspect description

-The common features are largely non-diagnostic and should therefore be discounted (i.e., ignored)

-Simultaneous lineups immediately teach the eyewitness which facial features should be discounted, but sequential lineups do not

Relationship between eyewitness confidence and accuracy

-MISCONCEPTION: Research shows that even under perfect conditions, usually confidence doesn't tell you shit

-This misconception is based on the use of the wrong statistic to measure the relationship between confidence and accuracy

Correlation vs. Confidence-Accuracy Characteristic (CAC Analysis)

-Correlation turns out to be the wrong measure to use

-A better approach is to simply plot accuracy as a function of confidence

-Take low confidence and calculate % correct and high confidence and % correct and plot it

-When witnesses say they have high confidence, they are usually correct

Behrman and Davey (2001)

-Analyzed confidence ratings made by real eyewitnesses to IDs made from real lineups in Sacramento

-Witnessed had 3 response options: I'm sure, I'm not sure, I don't recognize

-"I am sure": 92% correct (11 of 12)

-"I think": 47% correct (7 of 13)

Houston Police Department Field Study (Wixted, Mickes, Dunn, Clark & Wells (2015))

showed that false filler identifications had a relationship with low confidence

The diagnosticity ratio wrongly suggested that

sequential lineups are superior to simultaneous lineups (but ROC analysis shows that the opposite is true)