Task 4 Flashcards
Effect of stress on memory - believes
Marr, C., Otgaar, H., Sauerland, M., Quaedflieg,
C.W.E.M., & Hope, L. (2021). The effects of stress on eyewitness memory: A survey of memory experts and laypeople.
AIM
METHOD
RESULTS
LIMITATION
- examine the beliefs about the effect of stress on memory of laypeople, fundamental memory experts and eyewitness memory experts
- 35 in expert groups, 100 in the laypeople group
- survey, with statements about memory and it effects on stress (yes, no, i don’t know)
- high portions of both experts group and laypeople agreed that very high levels of stress impair the accuracy of eyewitness testimony
- most fundamental experts, but not eyewitness experts or laypeople thought that stress experienced during encoding can enhance memory
- -> Some topics quite some discrepancies between the expert –> could be used an indicator where more research or consensus is need
- -> Differences might also explain by different focus of the different fields (fundamental memory experts mostly look at the general mechanism of memory and the underlying neurological structure/ eyewitness experts are more focused on applied psychology)
- expert groups little underpowered
- closed-question (nuance lost)
- too technical for laypeople
Effect of stress on memory - differences in methodology
Marr, C., Sauerland, M., Otgaar, H., Quaedflieg, C.W.E.M., & Hope, L. (2021). The effects of acute stress on eyewitness memory: An integrative review for eyewitness researchers.
AIM
METHOD
RESULTS
(LIMITATION)
- examine the differences in finding of the two fields: eyewitness memory vs. fundamental memory
- systematic literature review
differences in findings:
- fundamental memory field: findings frequently demonstrate that encoding stress enhances memory recall performance and sometimes also benefits recognition performance
- eyewitness memory field: different finding either negative effect of encoding stress or null-finding
–> probably du to methodological differences!!!
!! Deffenbacher’s meta-analysis form 2004 showed a negative effect of stress (during encoding) on recognition and recall - BUT not well design studies!!
methodological differences:
stressor: type, check, timing
–> eyewitness: arousal, self-report, encoding and retrieval at the same time –> improvable!
–> fundamental: actual stressor (HPA-axis), blood pressure + cortisol, 24h apart so no stress at retrieval :)
stimulus:
–> eyewitness: relevant :)
–> fundamental: not relevant –> improvable!
participants:
–> eyewitness: few exclusion criteria
–> fundamental: many exclusion criteria
=> many methodological differences!
Effect of stress on memory
Marr, C., Quaedflieg, C.W.E.M., Otgaar, H., Hope, L., & Sauerland, M. (2021). Facing stress: No effect of acute stress at encoding or retrieval on face recognition memory.
AIM
METHODS
RESULTS
LIMITATION
- examine the effect of encoding stress on memory recall performance
- ALSO! include the suggestions to improve and combine the methodology of the fundamental and the eyewitness filed
- 119 participants (after power analysis was done)
- Experiment 1: 4 (condition: encoding stress vs. retrieval stress vs. stress at encoding and retrieval vs. no stress) x 2 (stress stage: acute stress vs. delayed) mixed design
- Experiment 2: 2 (condition: no encoding stress vs. encoding stress) x 2 (stress stage: acute vs. delayed) x 2 (stimulus type: faces vs. words) mixed design
- pp did the MAST or fake-MAST (“body” manipulation checks)
- then had to look at faces (or word list)
- 24h later came to do the recall test
- strong support for the Null-Hypothesis => stress did not have an effect (not encoding, nor retrieval) on the recall performance
- -> importance to do for well-powered experiments using contemporary methodology for a more complete understanding of the effect of acute stress on face recognition memory
- also not crime relevant stimuli
- future research could look at hormonal and gender differences
- generalizability is questioned because to avoid the effects of the own-race bias everyone was white
Recommendations for improving line-ups
Wells, G.L., Kovera, M.B., Douglass, A.B., Brewer, N., Meissner, C.A., & Wixted, J.T. (2020). Policy and procedure recommendations for the collection and preservation of eyewitness identification evidence.
- only results
1) Pre-line-up Interview Recommendation
- record it (–> recommendation 7)
- get a description, statement about familiarity/conditions/attention
- tell them to not discuss with other witnesses
2) Evidence-Based Suspicion Recommendation
- their should be other evidence then just the description or later identification of eyewitnesses
3) Double-Blind (or Equivalent) Recommendation
- best way is to have an blind administer otherwise you risk the “post-identification feedback- effect”
4) Line-up Fillers Recommendation
- at least 5 fillers
- only ever one suspect per line-up
- no one should stand out (especially not the suspect)
- description or suspect matched???
5) Pre-line-up Instructions Recommendation
- police should say nothing except…
(a) the line-up administrator does not know which person is the suspect and which persons are fillers
(b) the culprit might not be in the line-up at all, so the correct answer might be “not present” or “none of these”
(c) if they feel unable to make a decision they have the option of responding “don’t know”
(d) after making a decision they will be asked to state how confident they are in that decision
(e) the investigation will continue even if no identification is made
6) Immediate Confidence Statement Recommendation
7) siehe oben
8) Avoid Repeated Identifications Recommendation
- because of source monitoring error, the police would “give away” their hypothesis, mistakes tend to be repeated
9) Show-ups Recommendations
show-ups = no filler
- should be avoided!!!
- but happen in real life
=> reliability and integrity of eyewitness identification evidence is highly dependent on the procedures used by law enforcement for collecting and preserving the eyewitness evidence. These nine recommendations can advance the reliability and integrity of the evidence!
Estimator variables
= Variables that affect eyewitness accuracy but over which the system has no control
Cross-race identification
- Own race bias (ORB): the chance of a mistaken identification is 1.56x greater in other-race
Stress
- -> look at other Marr studies
- not one conclusion
Weapon focus
= the visual attention eyewitnesses give to a perpetrator’s weapon during the course of a crime
- not so important for the identification itself but for the statement
- mixed findings in field studies
- meta-analysis of lab studies the effect on description accuracy was larger (but its not the same)
Exposure duration
- very relevant! –> linear trend for exposure time was associated with improved performance
Disguise - Duh, bra!
Retention interval (Time) - longer delays led to fewer correct identifications and more false identifications
Intoxication
- less accurate/reliable
- more susceptible to false memory
System variable
= Variables that the criminal justice system has control over, such as how eyewitnesses are instructed before they view a line-up and methods of interviewing eyewitnesses
- pre-line-up interview (recorded, no leading questions)
- pre-line-up instructions (target might not be there)
- line-up procedure
> double-blind line-up: administrator should not know who the suspect is (other wise they might influence the witness for example through post-identification feedback)
> sequential vs. simultaneous (sequ: higher accuracy but lower response –> more conservative criterion)
> get confidence statement right away
> record the procedure - filler selection
> number: at least 5
> how selected? –> match the suspect vs. match the description (suspect-matched has some cons) - interaction at the station
> no further information give
> keep multiple witnesses apart
–> basically all the recommendations!!!
Post-diction variable
= Are neither system nor estimator variables in the traditional sense, because they are not presumed to causally affect the accuracy of eyewitnesses. These variables are measurable products that correlate with the accuracy of eyewitnesses in a non-causal manner.
- confidence -> is an indicator, but over-confidence is very much a thing!
- response latency -> if witness take longer time, indicator for inaccuracy (Brewer, & Weber, 2008) = quite decent predictor, especially if the line-up is in line with recent, scientific standards
- Self-reported decision process:
‘‘I compared the photos [in the line-up] to each other to narrow the choices’’ :(
‘I just recognized him, I cannot explain why” :)
Post-diction: confidence –> accuracy?
Wixted, J.T, Mickes, L., Dunn, J.C., Clark, S.E., & Wells, W. (2016). Estimating the reliability of eyewitness identifications from police line-ups.
(Wixted, J.T., & Wells, G.L. (2017). The relationship between eyewitness confidence and identification accuracy: A new synthesis.)
history aim method results limitation
- the perception of the relationship between confidence and accuracy in line-up performance is shaped by many examples of misidentification false conviction
- the most prominent case: misidentification of Ronald Cotton by Jennifer Thomson
- so representatives of the law and juries learned to mistrust confidence as a predictor of eyewitness identification
- this case study tries to examine the relation between accuracy and confidence in real life cases
- hang-up: in real life you do not have the ground truth
–> so, you have to use other evidence or convictions to get closer to the ground truth
(also no control though target absent conditions, because witnesses would be pissed according to Melanie) - they examined real life eyewitness identification data from the course of a year (nowadays: practice to always collect a confidence statement)
- actually confidence is a quite good predictor for accuracy WHEN the line-up is done properly according science-based standards (see recommendations)
- NO ground truth
- NO control target absent condition
Wells, G.L., & Olson, E.A. (2002). Eyewitness identification: Information gain from incriminating and exonerating behaviors.
(Warnung: bissel larifari)
- historically, police often only took into account the incriminating value of a suspect identification
- they did not take into account the exonerating value of a filler identification or the response that the perpetrator was “not there”
- especially filler identifications were seen as a sign that the witness “was not good”
- they did a information gain analysis (based on Bayesian reasoning) with old sample data sets
- -> “not there” or filler identification are as good or even better predictor for the suspect being not guilty as the identification is for the suspect being guilty
- -> this sheds new light on how the police should treat exonerating responding of line-up identification
- tutorial: this led to a new trend in the Netherlands (for example) that the police does not use line-ups very often anymore, because if there is a non-identification of the suspect they have to take that into account now
Clark, S.E., & Wells, G.L. (2008). On the diagnosticity of multiple-witness identifications.
…or how I call it “the article with the prosecutor’s fallacy”
(the gist of the article)
- the article is looking a the diagnosticity (not accuracy) of line-up identification of eyewitnesses
- -> the diagnosticty is the probability (p) of the suspect (S) being guilty (G) given the response (R) : p(S = G/R)
(which is might I add definitely the prosecutors fallacy and should actually sound like this: diagnosticity is the probability of the response (R) given the suspect (S) is guilty (G) : p(R/S=G))
- often in real life you have multiple witnesses of a crime, ergo you would also have multiple line-up responses
- if all the eyewitnesses had the same conditions and are independent
AND the conditions are ideal the probability that the suspect is guilty if every suspect identifies them is higher then the probability if they are just identified by a single suspect
–> also if only one witness identifies the suspect the probability goes down => showing that foil-identification and non-identification also have diagnostic value - the diagnostic value of certain responses changes if the conditions change (many examples in the article)
- if one suspect had a longer exposure their identification is “worth more” then a non-identification or a foil-identification by a witness with a short exposure => exposure time
- foil identifications had little or no diagnostic value for suspect-matched line-ups, but were diagnostic of innocence for description-matched line-ups => type of foil
- if the line-up is very biased (low similarity) suspect identification has no diagnostic value => foil selection
- also pressure to chose can minimize the diagnosticity of suspect identification => pre-line-up instructions
- pressure to choose the suspect minimizes the diagnosticty even more => (if not) double-blind
How to study line-up performance?
…see real life flash card
