P sychlgy F actsheets www.curriculum-press.c.uk Number 12 Eyewitness Testimny This Factsheet summarises memry research int eyewitness testimny. It cvers the tpics f recnstructive memry (by Bartlett) and research int the effect f language n eyewitness testimny (e.g., rle f leading questins by Lftus). It als cvers the effect f emtin n recall and ways f imprving eyewitness testimny. The main fcus f this Factsheet is n what; research int recnstructive memry tells us abut the accuracy f memry; psychlgical research tells us abut the accuracy f eyewitness testimny. What is eyewitness testimny? An eyewitness is a persn wh has seen a crime r event. Their testimny is their spken r written statement r evidence. Basically it is their recall r memry f a crime/event. Frm nw n, we will call eyewitness testimny EWT. Why is EWT a critical issue? It is a critical issue because an eyewitness recall tends t be unreliable (inaccurate). Hwever, the peple in the jury believe that the eyewitness testimny is accurate and reliable and will cnvict a persn based n this. The prblem is that peple s memry f an event is very ften inaccurate. EWT is an example f where memry research applies t everyday life. An example f unreliable eyewitness testimny: In 1979, in the USA, the plice arrested a Cathlic priest fr a series f armed rbberies. Seven eyewitnesses identified the priest. He wuld have been cnvicted if the real rbber had nt cnfessed t the crimes. The priest was shrter, bald and 14 years lder than the rbber (wh had a full head f hair). The rbber (Rnald Cluser) The priest (Father Pagan) Psychlgical research shws that peple s memry f events is wrse than they think. One research study shwed that peple were bad at identifying a thief they had been in a rm with yet the jury believed them! Wells, Liepe and Ostrm (1979) asked participants t wait in a rm befre a study began. A cnfederate (the thief) f the experimenters entered the rm, picked up a calculatr and put it in her purse. The participants then had t identify the thief frm six phtgraphs. Only 58% were crrect even thugh they had been in the rm with them. The experimenters then set up a mck trial where participants gave their evidence. 80% f the jury believed them. This shws that EWT is unreliable but many jurrs are cnfident that it is accurate. Why might it be unreliable? We can explain the unreliability f EWT by; 1. The recnstructive nature f memry (Bartlett s research); 2. The effects f language n memry (leading questins used by Lftus). We will lk at each f these in turn. Exam Hint: In the sectins t fllw What is it? describes the research. This is material fr an AO1 questin. The strengths and weaknesses part evaluates the research. This is material fr an AO2 questin. 1
1. Recnstructive memry! What is it? Bartlett (1932) argued that peple d nt just recrd memries and passively play them back like a tape recrder. Instead they actively try t make sense f infrmatin by fitting it in with what they already knw (he called this effrt after meaning ). He suggested that we interpret and recall what we see accrding t what we expect t see and what we assume is nrmal. (e.g., we may recall seeing a stapler in an ffice because it is smething we expect t be in an ffice). That is, we rebuild r recnstruct ur memry by using ur existing knwledge (called schemata). Hwever, this makes ur memry inaccurate because we distrt what really happened. If we see smething that des nt fit in with ur expectatins, then we may alter ur memry. Fr example, if given the testimny I ran up t the burglar alarm, many peple assume that the burglar alarm had been rung (Harris, 1978). Bartlett tested the idea f recnstructive memry by asking participants t listen t stries and then recall them later. He did this by testing the participants n a number f ccasins, frm 15 minutes t 10 years later. This is a serial reprductin methd f testing recall. He tested English participants by using an unfamiliar stry frm Nrth America, called The War f the Ghsts. The stry prduced cnflict between its cntents and the participant s wn knwledge. He examined whether the participant wuld impse their wn schema n the stry. Bartlett fund when participants recalled the stry, they made it mre like an English ne and ver time they cntinued t distrt it. In general, he fund that; recall became shrter recall was distrted by the participant s culture (they left ut unfamiliar details and ratinalised the stry t make it mre nrmal t their wn culture). the participants changed phrases t be mre similar t their wn language (e.g., cane was changed t bat). Extract frm The War f the Ghsts (Bartlett, 1932): One night tw yung men frm Egulac went dwn t the river t hunt seals, and while they were there it became fggy and calm. They heard war cries and thught, Maybe this is a warparty. They escaped t the shre and hid behind a lg. Nw canes came up and they heard the nise f paddles and saw ne cane cming up t them.! Strengths and weaknesses f Bartlett s research Bartlett s schema thery demnstrates the active retrieval f memries. This research is eclgically valid as it uses stries instead f artificial nnsense syllables. Mre recent studies which shw the recnstructive effect f schemata n memry supprt Bartlett s research. Fr example, evidence by Brewer and Treyens (1981) supprts Bartlett s research. They asked participants t wait in an ffice fr 35 secnds. After, they fund that recall fr bjects expected t be in an ffice (e.g., a desk) was gd whereas recall fr bjects nt expected t be in an ffice (e.g., a pair f pliers) was nt gd. S their ffice schema affected their recall. Bartlett s research is criticised fr ging t far in claiming that memry is usually inaccurate. Sme memries are very accurate. His research is criticised fr lack f experimental cntrl. S, Bartlett shwed that ne reasn why recall might be inaccurate is because we recnstruct memry, fitting it in with ur expectatins.. Exam Hint: Sme exam questins ask yu what psychlgical research tells us abut EWT. Be careful when talking abut Bartlett s research. Bartlett did nt directly research EWT he examined the recnstructive nature f memry. His research des tell us abut why memry f an event might be inaccurate. 2
2. Language n memry Elizabeth Lftus research! What is it? Lftus and clleagues have directly examined EWT. They have examined hw language can affect the recall f eyewitnesses, especially leading questins. Leading questins are nes which imply smething that may nt have happened/been there (e.g., Did yu see the man? suggests that a man was present althugh there may nt have been). Research by Lftus and Palmer (1974) Aims: T examine the effect f leading questins n the accuracy f the estimate f car speed in an accident. Prcedure: Participants saw a film f a traffic accident. The film lasted frm 5 t 30 secnds. The participants then had t recall what they had seen and then answer specific questins. Imprtantly, there was ne critical questin in which the wrds were changed. The questin was Abut hw fast were the cars ging when they hit each ther? Different sets f participants heard different wrds. They heard hit, smashed, cllided, bumped r cntacted. The estimated speed was used as a measure f the effect f the different wrds. One week later the participants were asked Did yu see any brken glass? Findings: The wrding f the questin did affect the participants judgement f speed. Participants guessed that the cars were ging fastest fr the wrd smashed, then secnd fastest fr cllided, then bumped, hit and cntacted. Fr example, they estimated a speed f 40.8 mph fr smashed but 34.0 mph fr hit. Als, 32% f the participants wh heard smashed said that they had seen brken glass whereas nly 14% f the participants wh heard hit said that they had seen brken glass. Cnclusin: The change f ne wrd had a large effect n the participants recall. The language used did affect the recall f eyewitnesses. Research by Lftus and Zanni (1975) Participants saw a shrt film f a car accident. Sme participants were asked Did yu see a brken headlight?, thers were asked Did yu see the brken headlight?. There was n brken headlight in the film (althugh the questin implies that there was). They fund that, fr the wrding the brken headlight 17% f participants said that they had seen it cmpared t 7% fr the wrding a brken headlight. S, ne wrd can lead peple t add infrmatin t their recall that they had nt witnessed. S, Lftus and clleagues shw that EWT can be inaccurate as peple can add infrmatin t their recall (e.g., brken headlight) r distrt their recall (speed estimates).! Strengths and weaknesses f Lftus research Lftus and clleagues research supprts the idea that language easily distrts memry and has implicatins fr EWT in curt. A strength f this research is that they have cntrlled the studies well (t stp ther irrelevant factrs affecting the results). The research can be criticised fr lacking eclgical validity as the participants nly saw a vide clip whereas real-life accidents are likely t be mre emtinal. EWT f a real-life event can be very accurate (Yuille and Cutshall, 1986). This type f research tends t fcus n minr details (e.g., headlights) in the event, nt majr nes (e.g., car clur), and it may be easier t distrt these minr nes. [By the way, this effect f wrding n recall is an example f interference (see the Theries f Frgetting Factsheet).] Exam Hint: Yu shuld knw the experiment by Lftus and Palmer (1974) very well! 3
Emtin and recall EWT culd als be unreliable due t high emtin experienced by the eyewitness at the time f the crime. There are tw ways that emtin can affect recall by repressin r as flashbulb memries. Repressin Freud (1915) suggested that repressin is where we purpsefully frget memries which cause us anxiety by repressing them (pushing them int uncnsciusness). S we may repress the memry f, fr example, a car accident. This makes it harder t recall. Sme research supprts repressin. Peple wh are repressrs (based n persnality tests) take lnger t recall negative childhd memries than ther persnality types (Myers and Brewin, 1994). It is difficult t get experimental evidence fr repressin because the experiments wuld need t create anxiety-prducing situatins. This is nt ethical. Evidence fr repressin cmes frm case studies f individuals and s the results may nt generalise t the whle ppulatin. Flashbulb memries Flashbulb memries are vivid, lng-lasting memries f highly emtinal events (Brwn and Kulik, 1977). Flashbulb memries are mst likely t ccur when the event is surprising and has cnsequences fr the persn s wn life. They imprve recall fr an event. Flashbulb memries may nt be as accurate as first thught. They are subject t frgetting just like ther memries. Imprving the reliability f EWT Psychlgical research can imprve the reliability f EWT: Expert witness Psychlgists can be an expert witness. An expert witness is smene wh ges int curt t explain t the jury that EWT may be unreliable and inaccurate. This shuld prevent them wrngly cnvicting smene n the basis f a EWT. Cgnitive Interview Psychlgists have used research int memry and EWT t develp a way f interviewing eyewitnesses that helps them t recall mre infrmatin and mre accurately. It is a cgnitive interview (Geiselman and clleagues, 1985). The plice in the UK use cgnitive interviews. A cgnitive interview reduces the number f interruptins and stps the use f leading questins when an eyewitness recalls the event t the plice. A cgnitive interview has different stages that fllw a set rder, fr example; Free recall ccurs first. It is where the persn recalls what they saw. Change rder is where the persn recalls the event starting frm the thing that happened last and wrking backwards t the beginning. Questins are asked last. Each time the persn answers a questin they use cntext reinstatement where they imagine themselves back in the situatin at the time f the crime. This uses cntext-dependent recall (see the Theries f Frgetting Factsheet) and imprves recall. Geiselman and clleagues fund that recall was better and mre accurate with the cgnitive interview cmpared t the standard plice interview. Cgnitive interview eyewitnesses recalled 41.1% crrect facts abut an event cmpared t 29.4% fr a standard plice interview. Exam Hint: If an exam questin asks what psychlgical research tells us abut the unreliability f EWT, yu shuld fcus n the wrk f Elizabeth Lftus wh did directly research EWT. Glssary Cnfederate: a persn wh is in n the experiment but is acting as anther participant in the experiment. Cntrl: stpping factrs that are irrelevant t the experiment affecting the results. Eclgically valid: the extent t which the findings f the research apply t everyday life. Schemata: Schemata are the plural f schema. A schema is a framewrk f knwledge that we have. Schemata represent all kinds f knwledge (e.g., yu will have a schema fr what happens in restaurants). Example exam questin Describe a study f recnstructive memry and give ne criticism f the study. (6 marks) T answer this yu culd talk abut Bartlett s study f recnstructive memry. Explain that he used a serial reprductin methd t test recall f a stry called The War f the Ghsts. The stry did nt fit the English participants cultural expectatins. He fund that when they reprduced the stry, their recall became shrter and mre cnventinal. It was als ratinalised. He fund evidence fr peple recnstructing stries using their schemata. A criticism f Bartlett s research is that it ges t far in claiming that memry is usually inaccurate. Sme memries are very accurate. (Yu can als use a psitive criticism such as the research being eclgically valid). Alternatively yu culd answer this questin by talking abut the study by Brewer and Treyens (1981). Acknwledgements: This was researched and written by Amanda Albn. The Curriculum Press, Bank Huse, 105 King Street, Wellingtn, Shrpshire, TF1 1NU. 4
Wrksheet: Eyewitness Testimny Name 1. Why is eyewitness testimny an imprtant issue in Psychlgy? 2. What did Bartlett (1932) examine in his research? 3. In Bartlett s study, peple recalled a stry n a number f ccasins. In what ways did they change the stry? 4. What did Lftus examine in her research? 5. What did Lftus discver abut the effect f language n recall? 6. In which ways can emtin affect recall? 7. Describe ne way in which the reliability f eyewitness testimny can be imprved. 5