Eyewitness Testimony

What are estimator variables?

variables that affect EW accuracy but are not under the ctrl of the justice system - but might be used to estimate accuracy

inc event characteristics, witness characteristics and testimony characteristics

as a psychologist, you can only educate people sa jurors about

What do you need to be an accurate eyewitness?

ppl must complete 3 stages of memory processing - acquisition, storage and retrieval

What are system variables?

under the control of the justice system (Wells, 1978)

What are some examples of event characteristics?

disguise, viewing conditions, greater stress, weapon focus and own race bias

How does disguise affect EWT?

accurate identification is difficult if during the witnessed event the person wears sunglasses (hockley et al 1999) knit caps/anything covering their hairline and esp if they wear both (Mansour 2012)

configuration of face may change when one is intoxicated

What type of witnesses are common?

intoxicated

73% of officers report interviewing at least 1 within a month period

13% of prosecuted cases have at least 1

What did Schreiber et al, 2011 do?

pps drink either alcohol/placebo and then witness a staged convo

found no dfr in person description - drunk and sober witnesses desc ppl equally accurately

same amt of detail overall but sober witnesses focused more on physical details and intoxicated focused more on subjective impressions sa seeming angry or nervous

was a highly controlled lab setting - alc didnt drastically harm descriptive accuracy

What did real bar studies show about alcohol?

pps were grouped into low vs high intoxication

found no dfr in correct identifications in a line up - high intoxication did not reduce whether they picked the right suspect

highly intox witnesses were worse on accuracy and provided fewer details

real bar settings show more impairment than lab settings (Altman 2018)

Why do lab and bar findings differ?

lab - v controlled, pps only consume alcohol and lvls are kept low to moderate - upto legal driving limits - 0.08% so effects appear mild

bar studies - ppl drink much more - 0.13-0.29% BAC and uncontrolled so they may have mixed drinks and no food or other substances

in real world ppl are much drunker so impairment is stronger

What else are a lot of witnesses under the influence of?

18% - cannabis which has been found to impair memory to word lists (D’Souza, 2004)

What was the Amsterdam Coffee Shop Studies?

looked at ppl in amsterdam coffee shops where cannabis use is common and tested their memory after witnessing a mock crime and then doing a line up

pps reported how many joints they smoked and how intoxicated they felt

ppl more intoxicated gave fewer accurate details abt what happened in the crime scenario - gave less accurate descriptions of suspect

no dfr in correct identifications in a lineup - even tho recall was worse, intoxicated pps werent worse at actually picking correct person from a line up

(Vredeveldt, 2018)

What is worth considering?

combined effects of event and witness characteristics

What are people getting better at?

recognising other ppl, if they are the same race as they are (Brigham et al., 2007)

What is the own race bias/cross race effect?

that ppl are generally better at recognising faces of their own racial grp compared to faces from other racial grps

first identified by Malpass and Kravits (1969)

What is contact hypothesis?

(Brigham and Malpass, 1985)

proposes that the more contact you have with people of another race, the better you become at recognising faces from that race

What evidence is there for CH?

research examined whether white fans who freq watch black players wld recognise black faces better - found the effect existed but was not strong (Li, Dunning and Malpass, 1998)

across many studies, the impact of contact was v weak, explaining only 2% of that variance - means contact helps a little but not enough to fully expl the bias (Meissner and Brigham 2001)

What do many studies on CH rely on?

subjective estimates of contact sa I think i have lots of contact with x group which may not be accurate

What is the evolutionary exp about CRB?

bias may have deeper biological roots - our ancestors needed to quickly and accurately recognise members of own grp, for cooperation and survival and to distinguish outsiders who might be threats

better recognition of in grp faces may have evolved bc it helped maintain stable social grps

out grp faces were less essential to differentiate at the indv lvl so recognition accuracy didnt dvlp as strongly

What effect does age have?

own age bias - ppl have difficulty discriminating bt ppl of diff age grp

and are better recognising faces of own age grp bc they have more daily contact with them

Rhodes & Anastasi 2012 - uni students were sig better at recognising faces of ppl their own age compared to adults in 50s - agr dfr reduces recognition accuracy

Which age performs better in EW performance?

younger and middle aged adults

cld be due to cdr’s dvlping memory systems and older adults having memory decline

What did Pozzulo et al 1998 find?

when culprit not in line up - younger and middle aged adults make less false identifications than cdr/ elderly witnesses

when culprit in line up - no sig dfr bt age grps

age dfrs matter most when situation is ambiguous

What are most crimes events of?

high stress and pps experience high lvl of neg emotions

most studies show that stress decreases both accuracy and detailedness of memories

under some special conditions however, certain detail become sharper (Deffenbacher et al 2004)

What did the London Dungeon study show?

pps in state of high anxiety are more likely to make wrong identifications (Valentine and Mesout 2008)

How do real world conditions affect EWT accuracy?

24% of violent crimes occur bt 8pm and midnight , with a peak at 9pm (Us Dept of Justice, 2016) - means many crimes occur in low light conditions which neg affects how well witnesses can observe and later identify perpetrators

Yarmley 1986 - best conditions is daytime and beginning of twilight, witnesses can see facial fts more clearly- worst is night time and end of twilight, 13% correct id - lighting is one of the strongest env infl on EW accuracy

the longer the witness sees perp, more accurate the ID bc longer exposure allows for better encoding of fts into memory but theres no consensus on how long is enough

What was the US army survival camp studies?

Morgan et al (2004) studied real military personnel undergoing POW-style training inv isolation, food, sleep dep and stressful ints

high stress drastically red EW accuracy

earlier lab studies found no stress effect but these used mild stress

intense stress sig harms EW memory

What can a criminal have which makes them more difficult to identify?

if they have a gun, sword or knife vs a check, money or pen

eye tracking shows they look at gun longer than face and fixate on it more

What is the traditional explanation of weapon focus?

(Steblay , 1992)

a weapon acts as an attention magnet so bc it is threatening, witnesses focus on it and leave less attention for the perps face

What is expectedness within the weapon focus context?

(fawcett et al 2016)

wf stronger when weapon is unexpected in context

pps more likely to recognise susp holding racket at tennis court than gun at tennis court & gun at shooting range than tacket at shooting range

What is novelty within the wf context?

(erickson, 2014)

even harmless but novel objs can draw attention the same way a weapon does - sugg effect is not only abt fear but also abt attention being captured by unusual items

What can have a direct impact on changing system variables?

psychology can

all SVs can be estimator Vs but not EVs can be SVs

What is the relationship of confidence, speed and accuracy in EWT?

confidence alone does not mean accuracy

fast decisions are more likely to be accurate, esp when paired with high initial confidence (under 6 seconds) - sauerland et al, 2009

fast choices reflect automatic recognition and slow decisions indicates uncertainty

c and se is easy to manipulate and greatly infl by factors which it can and shld change

What are the guidelines about confidence?

there are gl which req police to record a witness’ confidence at exact time of identification bc conf can change after the fact

research shows that giving witnesses positive feedback after they make an identification can artificially boost their confidence sa ‘ good you identified the suspect’ - wells and bradfield 1998

confi can change even w/o verbal feedback sa facial expressions and tone (w+b, 1998)

effect seen in both lab studies but also actual EWs in Uk - B, W and olson, 2002

gls - record conf immediately, avoid giving fb, treat later conf as less reliable

What was Smalarz and Wells 2014 2 phase experiment?

examined how fb given to witnesses affects wits own bv and conf and jurors judgements of wit accuracy

pps watched vid of crime and completed line up identification

randomly given fb or not and then cross examined with video taped interview

phase 2 - new pps acted as jurors and watched ce videos and asked to judge how accurate witnesses were

shows how fb can distort wts confidence and testimonies and how jurors can be misled by these changes even though the witness’ actual accuracy hasnt improved

What was the Loftus carr accident studies?

asked how fast car as going when it smashed or hit the car

leading questions, with minor changes in wording can also change ppls memories

What are egs of SVs?

line up construction/fairness

line up administration

testimony administration

What are the AP-LS EW identification recommendations? (1980)

outline best practices for collecting and preserving EW evidence to improve accuracy and fairness

match fillers to witness’ desc

use sequential administration - view lineup one at a time to reduce relative judgemnt errors

provide pre-id instructions - tell witness the culprit may or may not be present

use double blind admin - admin must not know who suspect is

assess and record conf at moment of id

What new AP-LS EW Identification guidelines were included in the 2020 update?

video and audio record entire procedure

evidence-based susp selection - must be based on reliable evidence

pre-lineup interview - record wits memory and description b4 seeing any lineup

avoid repeat identifications - prevent same wit from identifying same suspect multiple times which inflates confidence artificially

What were the 2020 pre identification instructions?

2020 - witness must be clearly informed that perp may or may not be in lineup, its ok to not choose anyone, idk is acceptable response and they will be asked abt their certainty at the end - instructions reduce pressure and prevent forced choices

What does research show abt 2020 pre ID instructions?

research shows witnesses who receive these instructions are less likely to make a selection - adopt higher decision standard, less likely to correctly identify perp bc they only choose when v sure and less likely to falsely id someone innocent

What are alternatives to double blind admin?

blinded admin methods

envelope/folder shuffle method - photos placed in identical folders and shuffled so administrator cant see order or which folder contains susp

self administered env - wit opens env and views lineup on own

online/laptop presentation - lineups shown on computer, system record wits resps automatically and administrator cant see selection

What are pre-lineup factors?

infl EW accuracy b4 lineup even occurs, during time witness is retaining memory