3. Cognitive Interview PDF

Preview

Summary

Models, accuracy and critical research ...

Description

3. The Cognitive Interview BACKGROUND CONTEXT: PACE • PACE refers to the Police and Criminal Evidence Act of 1984 – governs police interviewing practice • Previous to this act, there were no formal rules or guidance for police interviewing of suspects and eyewitnesses. • Previous to PACE, Irving (1980) found the following police tactics, widely used in England: • manipulating the suspect’s self-esteem • minimising the seriousness of the offence • pretending that the police were in possession of more evidence than was, in fact, the case • pointing out the futility of denial • advising that it was in the interviewee’s best interests to confess. PACE: CODES of PRACTICE C & H Detention, treatment and questioning of persons by police officers. • Prior to interview: interviewee must be reminded of entitlement to free legal advice. • Interviewers must ask whether interviewees confirm or deny any earlier statement or significant silence and whether they want to add anything. • No interviewer may try to obtain answers by use of oppression (e.g. interviewer cannot indicate subsequent action taken by police unless an answer to a direct question). • The interview must cease when: – All relevant questions have been put. – Other available evidence has been taken into account. – It is considered that there is sufficient evidence for a conviction. Interview Records • An accurate written record must be taken at time of each interview, or as soon as practicable afterwards (including place, time, breaks, names of those present). • Written records must be timed and signed. Unless impracticable the interviewee (and solicitor if present) must be given an opportunity to read (or have read to them) the record and indicate whether it is correct or inaccurate. • Refusals to sign must be recorded. Juveniles, mentally disordered and mentally vulnerable • Cannot be interviewed as a suspect unless accompanied by an appropriate adult. • Appropriate adults must be informed that they are not simply an observer, they must – advise the interviewee,

• • •

– observe whether the interview is conducted properly and – facilitate communication with the interviewee. Requirement for appropriate adult may be waived if officer of superintendent rank or above considers this will cause delay with significant consequences. A record must be keep of the grounds for the above. Notes: Because of risk of unreliable evidence it is important to obtain corroboration of any facts admitted whenever possible.

Interviews in police stations • In any period of 24 hours a detainee must be allowed a continuous period of at least 8 hours for rest. • Before being interviewed an assessment must be made of whether the detainee is fit to be interviewed. • Suspects may refuse to answer questions but their consent is not necessary to interview them. • Interviewees shall not be required to stand. • Short refreshment breaks should be provided at 2 hour intervals. • A record should be kept e.g. of reasons it was not practicable to use an interview room, and decisions to delay a break in the interview. Are you free to leave the police station? If you are at a police station, either as a suspect or possible witness, but have not been arrested, you are free to leave at any time. How long you can be held in custody? • The police can hold you for up to 24 hours before they have to charge you with a crime or release you. • They can apply to hold you for up to 36 or 96 hours if you’re suspected of a serious crime, e.g. murder. • You can be held without charge for up to 14 days if you’re arrested under the Terrorism Act. Interpreters • Appropriate arrangements must be made to provide interpreters for people who are deaf and/or do not understand English. • Interpreters must be given sufficient time to note each question and answer. • The interpreter must not be a police officer when interpretation is needed for the purposes of obtaining legal advice. Otherwise OK with agreement of interviewee or appropriate adult. PEACE PEACE is the national model for best interviewing practice. • The PEACE interviewing model was developed by the Home Office and the Association of Chief Police Officers in 1992. Considerable input from forensic psychologists. • The aim was to improve interviewing standards and improve the information obtained from interviews.

• •

It is used when interviewing witnesses as well as suspects and has helped to focus police work onto the collection of facts rather than confessions. PEACE stands for: – – – – –

Preparation and planning Engage and explain Account Closure Evaluate

•

P(reparation) means not just the content of questions (i.e. aims and objectives of interview; assessing what evidence is needed) but also their style, working out where the interview should take place, where the interviewee should sit, how long the interview should last, and so on – optimising the conditions to get at the truth. • E(ngage and explain) means making sure the interviewee is comfortable and able to go ahead with the interview. The interviewer should be aware of the needs expectations etc. of the interviewee, and explain the purpose and routines of the interview. If a tape recorder is being used, it should not be switched on until this is established. • A(ccount) is the stage in which the interviewees recollection of events is obtained without interruption. It is also where the interviewer clarifies and challenges, as necessary. It is here that procedures such as the cognitive interview can be employed. • C(losure) means bringing the interview to an end in a clear way, with the interviewee asked whether there is anything else they would like to say. The interviewee is also thanked before leaving. • E(valuation) means identifying what has really been learnt from the interview and fitting it into the rest of the information gathered. As part of data gathering, when police interview witnesses they should apply the mnemonic ADVOKATE from R v Turnbull (1975). They must explicitly question the witness on these matters. A amount of time for which the event was observed (How long? The longer the better) D distance of the event from the witness (What distance? The shorter the more accurate) V visibility (How good? The clearer the visibility at the time, the better) O obstructions between the witness and the event (Were there any? The fewer the better) K known as seen before (Was person known or familiar? The more familiar a person the better) A any reason to remember (Any special reason to remember? The more emotionally salient or novel the better) T time lapse between the event and the witness’s recall of it (How long? The shorter the better)

E errors or material discrepancies (Any in witness accounts? The more inaccurate parts of a witness’s testimony are shown to be, the less reliable the rest) Useful reference: Walsh, D. & Milne, R. (2007). Perceptions of benefit fraud in the UK: Giving PEACE a chance? Public Administration, 85, 525-540. Concluded: General support for the model, but insufficient time to prepare for investigations. Concern over inflexibility over model’s framework. THE COGNITIVE INTERVIEW In 1970s books on police science included nothing on witness interviewing and police in USA and UK received little if any formal training in interviewing. Because of the crucial role witnesses play, in early 1980s Ed Geiselman (UCLA) and Ron Fisher (IUF) set out to design a set of procedures that would enhance the quality and quantity of evidence given by witnesses. Geiselman et al. (1985/1987) argued that a successful technique for improving testimony should have 3 characteristics. 1) It should reliably increase accurate recall. 2) When the technique is used, confidence should be related to accuracy. 3) Should have practical, applied value; it should be possible to train police officers to use it. Forensic Interviewing: ACCURACY What is an increase in accurate recall? • An increase in accurate recall is an increase in correct responses without a disproportionate increase in errors. E.g. • First row is acceptable, here with an enhancement procedure we have an increase in correct responses and also an increase in incorrect responses. However, the proportion of correct and incorrect responses remains the same in both the enhancement and control condition, both at 90% - acceptable because we have a decent number of correct responses and a small number of incorrect responses. • The ideal situation is that in row 2, here with an enhancement procedure we have an increase in correct responses but the rate of incorrect responses don’t change, as a result if we look at the accuracy rate in the enhancement procedure it is 97% whereas in the control it is 90%, not only have we had an increase in correct responses but the accuracy rate has improved in enhancement too. • What we don’t want is the situation in the last column where we have an increase in correct responses but a disproportionate increase in incorrect responses as a result if we look at accuracy rates, the rate for enhancement procedure has actually declined from 90% in control to 33% in enhancement procedure.

Enhancement Corr Incorr

Control Corr Incorr

Accuracy Rate Enhance Control

30 30 30

3 1 20

10 10 10

1 1 1

90% 97% 33%

90% Good 90% V. Good 90% Bad

Geiselman and Fisher’s Approach Included four features: 1) 2) 3) 4)

reviewing the relevant cognitive literature, looking at real life forensic interviewing conditions conducting experiments reporting results back to police and other professionals.

Cognitive literature: Context reinstatement • Based on Tulving and Thomson's (1973) Encoding Specificity Principle items are encoded together with features of the context within which they occurred. • E.g.Memory can be improved by physically reinstating the context e.g. Godden and Baddeley (1975) using divers finding that divers remembered info better if they returned to the context they learnt I e.g. under water, Smith (1979) using police officers returning to the lecture theatre they learnt something in. • Teasdale and Foggerty (1979) found that emotional or mood cues were also useful in reinstating context. People remembered info better if they returned to the mood they were in at encoding. Cognitive literature: Mental Context Reinstatement • Led investigators to look at mental reinstatement of the context; Malpass and Devine (1981) devised a ‘guided imagery’ procedure. Participants are encouraged to recall peripheral and contextual details of an event e.g. what they were thinking, feeling at the time. • In the 1980’s some investigators (e.g. Sanders, 1984) found that physical reinstatement of the context had little effect on face recognition (identification). • However, others have continued to find that mental context reinstatement is successful on free recall and cue recall and also face recognition, mental context reinstatement may on the whole be superior to physical context reinstatement. (Cutler et al., 1987; Hammond Wagstaff and Cole, 2006; Wagstaff, Cole et al., 2007; Wagstaff et al., 2010; Ohman, Eriksson, & Granhag et al., 2013). Cognitive literature: Report Everything • Memory is stored as a series of coded representations, with numerous interconnected codes that may trigger cues or associations (Bower, 1967; Melton and Martin, 1972). The more details that are retrieved by a witness the more likely other memories will be cued.

• •

Also, in practice, witnesses may hold back information that they believe to be irrelevant or unimportant. To overcome this Geiselman et al. suggest that witnesses should be encouraged to report everything; not leave anything out and report anything that 'pops into your head'.

Cognitive literature: Recall in different orders • Some assume that episodic memories can be stored in the form of scripts that fit preconceptions about what should happen in a particular situation. These scripts limit recall by filtering out information inconsistent with the script and filling in gaps in script with errors (Schank and Abelson, 1977). • Forward recall encourages the use of scripts, but some evidence that recall better when items recalled in reverse order (Whitten and Leonard, 1981). • Hence recall may be improved through recalling in different orders. Cognitive literature: Change perspectives • Anderson and Pitchert (1978) found that participants could remember additional details of a narrative if they had to recall it from the perspective of another person (e.g. a house-buyer rather than a burglar). • Geiselman and Fisher decided to include change perspectives as a technique. However, the change perspectives technique is often based on a change of spatial perspective; e.g. thinking about it from the perspective of someone else, 'think about where he/she was', and 'what he/she would have seen'. Doesn’t seem to be effective like this but according to Anderson and Pitchert (1978), its clearly not how it was originally conceived. These elements went to form GEISELMAN AND FISHER’s Original Cognitive Interview 1984/1985 which consisted of Four Mnemonics 1.reinstate the context (recall all detail, feelings, reactions); 2. report everything (all detail no matter how unimportant it may seem); 3. recall in different orders (including reverse order and starting in middle); 4. change perspectives (view from the perspective of another person); COGNITIVE INTERVIEW: Problems with Standard Police Interview (Fisher et al., 1987) Fisher et al. (1987) went on to look at problems with interviewing in real life investigations (Florida Police). Main problems were: • Interruption. • Excessive use of question-answer format • Inappropriate Sequencing of Questions. • Failure to establish rapport · Other problems: Negative phrasing

Leading questions Inappropriate language Judgemental comments Failure to follow up on leads Auditory clues underemphasised.

George (1991) in the UK broadly supported these criticisms. Problems with real life forensic investigations: Fisher et al., 1987; George, 1991 •

•

•

• •

•

•

• •

Police interviewing is characterised by frequent interruption: Fisher et al.’ found there were, on average, three open-ended question requiring an extended answer per interview; during these the interviewer interrupted the eyewitness on average 11 times; typically witness was interrupted after only 7.5 seconds. Interruptions cause two problems: 1) They break the concentration of witnesses when they are trying to retrieve information; unfortunate as free recall typically produces the most accurate recall (Yuille & Cutshall, 1986). 2) After the witness has been interrupted several times he or she begins to expect interruption throughout the interview and shortens responses accordingly. This may exclude information, important to a forensic investigation. Excessive use of Question-answer format; Police interviewing relies heavily on the use of a question-answer format. Most questions are closed questions (e.g. ‘what colour was the attacker’s shirt?’). Closed questions give no opportunity the eyewitness to elaborate or extend an answer. The witness becomes passive. Fisher et al. comment, ‘It is difficult enough for the eyewitness to retrieve detailed events from memory when actively trying; it is virtually impossible when he remains passive’ (p.181). If the interviewer forgets to ask a certain question, no information in that area is recorded. Inappropriate Sequencing of Questions: Many questions asked by interviewers are in a seemingly arbitrary order that may impair eyewitness performance through shifting their retrieval efforts from one memory domain to another (e.g. visual to auditory). Alternating retrieval in this way has been shown to produce a 19% decrease in witnesses’ performance (Fisher & Price-Rouch, 1986). General knowledge questions:

•

•

• •

Other problems can be caused by 'general knowledge' questions, such as ‘why do you think he did that?’ or ‘Was he married?’. Shifting from the recall of crime details to general knowledge questions, then back to crime details can cause decreases in the witness's performance. Failure to establish rapport: Interviewers did not develop a rapport with witnesses; so witnesses felt less inclined to divulge sensitive information; Indeed, often interviewers would make judgmental comments e.g. ‘you shouldn’t have been carrying so much cash’, or ‘that was a silly thing to do’. Can make the eyewitness defensive or offended and less likely to cooperate.

Other problems: • Negative phrasing e.g. "you don't remember if?" May actively discourage the witness from trying hard to retrieve information. • Leading questions subtly suggest that a certain answer is required. Loftus (1975) found that leading questions may bias eyewitnesses later recollections of an event. • Inappropriate language such as using sentences or words beyond the comprehension of the eyewitness (e.g. ‘what was the index number of the vehicle?’). • Police officers often fail to follow up on leads that they are given. For instance, a suspect may be described as looking like ‘a gangster’ without any attempt being made to follow up the comment, to elicit why the eyewitness felt that the suspect looked like a 'gangster'. • Finally, auditory clues are often underemphasised. Officers rarely enquire about what a suspect may have said or if they had an accent; though witnesses may volunteer such information if given the chance. THE ENHANCED COGNITIVE INTERVIEW •

•

Led to Enhanced Cognitive Interview (Fisher, Geiselman & Raymond, 1987; Fisher, Geiselman, Raymond, Jurkevich, & Warhaftig, 1987; Fisher, Geiselman & Amador, 1989). reinstate the context (recall all detail, feelings, reactions); report everything (all detail with minimal interruption); recall in different orders (including chronological sequence); change perspectives (view from the perspective of another person);

plus: build rapport with the witness (develop trust) use witness compatible questioning (structure interview so that it is directed by the witness; witness can generate information on own initiative); focus attention (encourage witness to focus concentration). Other instructions in ECI commonly used include: OK to say don’t know; Don’t make up answers; Ask interviewer to explain question if necessary; Repeating a question doesn’t mean previous answer is wrong

Similar instructions were found to be useful generally; particularly when interviewing children e.g. Danby et al., (2015) Uninterrupted free recall phase followed by question phase (preferably open ended).

Experiments on the Cognitive Interview ADULTS Much support for the cognitive interview CI compared with Standard Interview SI (without cognitive interview) with adults. For example, –  Geiselman et al. (1984) CI did better than a SI with no increase in FAs. –  Geiselman et al. (1985) found 40% more correct information with CI. No increase in False alarms. –  Fisher et al., (1987) found CI better than SI, no increase in errors. –  Kohnken, Milne, Bull & Memon (1999): Meta-analysis of 32 experiments (1200 subjects, 2500 interviews). Increased recall of both correct and incorrect items with CI, but increase greater for correct items. However, the mean accuracy rate (proportion of correct details recalled. relative to total number of recalled details) was similar to the standard interview (84% for the CI and 82% for the standard interview). – Milne, Clare and Bull (1999): CI better, no increase in errors (adults with learning disabilities). –  Dornburg and McDaniel (2006): CI better no increase in errors (Elderly 65-87 year-olds). –  Stein and Memon (2006): CI better, no increase in errors (Study conducted on cleaning staff in in Brazil). – Wright and Holiday (2007): CI better, no increase in errors (Elderly 75-96 year-olds). – Paulo et al. (2015): CI better, no increase in errors. Witnesses overestimate errors. Experiments on the Cognitive Interv...