Psychiatry, Psychology and Law

ISSN: 1321-8719 (Print) 1934-1687 (Online) Journal homepage: http://www.tandfonline.com/loi/tppl20

The Effects of Distributed Learning on Enhanced Cognitive Interview Training

Constance Tamara Heidt, Katherine D. Arbuthnott & Heather L. Price

To cite this article: Constance Tamara Heidt, Katherine D. Arbuthnott & Heather L. Price (2016) The Effects of Distributed Learning on Enhanced Cognitive Interview Training, Psychiatry, Psychology and Law, 23:1, 47-61, DOI: 10.1080/13218719.2015.1032950

To link to this article: http://dx.doi.org/10.1080/13218719.2015.1032950

Published online: 23 Jun 2015.

Submit your article to this journal

Article views: 30

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=tppl20

Download by: [University of Regina] Date: 25 February 2016, At: 07:22 Psychiatry, Psychology and Law, 2016 Vol. 23, No. 1, 47À61, http://dx.doi.org/10.1080/13218719.2015.1032950

The Effects of Distributed Learning on Enhanced Cognitive Interview Training

Constance Tamara Heidta, Katherine D. Arbuthnottb and Heather L. Pricec aDepartment of Clinical Psychology, University of Saskatchewan, Saskatchewan, Canada; bDepartment of Psychology, Campion College, University of Regina, Saskatchewan, Canada; cDepartment of Psychology, University of Regina, Saskatchewan, Canada

Interview training for police officers is generally limited and, when it occurs, rarely translates into optimal interviews. Training ineffectiveness may be partly due to the structure of the training programme. In the present study, 60 participants received two hours of training on the Enhanced Cognitive Interview (ECI), in either a massed (one two-hour session) or spaced (two one-hour sessions) format. Following training, participants conducted an ECI. Advantages for spaced training were found in open-ended prompt use, perpetrator-specific details elicited from open prompts, and the utilization of two critical ECI components. These results suggest that a simple alteration in training protocols could improve forensic interviewing skills. Key words: forensic interviewing; distributed learning; interview training; skill acquisition.

Introduction (police officers and social workers) following Eyewitness evidence is often a critical aspect an intensive one-week interview training pro- of a criminal investigation, yet police inter- gramme with those who had not received the view training programmes generally occur training. Despite an increase in knowledge, no over a short period of time, with an average behavioural differences were observed training time of two days (Clarke & Milne, between the trained and untrained interviewers 2001; Snook, Eastwood, Stinson, Tedeschini, in appropriate and inappropriate question &House,2010; Wells, Memon, & Penrod, types, and adherence to interview protocol. 2006). A recent study examining 170 Canadian The present study examines whether a simple police officer interview training programmes alteration in the structure of the interview training format would improve interview per-

Downloaded by [University of Regina] at 07:22 25 February 2016 found that fewer than half of the officers received formal interview training (Snook, formance following training. House, MacDonald, & Eastwood, 2012). Even The learning distribution, either massed when training has occurred, it is uncommon or spaced, of interviewing training pro- for the information to effect a significant grammes may be one factor contributing to change in interviewing behaviour (Aldridge & poor interviewing performance post-training. Cameron, 1999;Lambetal.,2000; Warren et Cognitive and educational research indicates al., 1999). For instance, Aldridge and Cameron that information learned over longer periods (1999) compared investigative interviewers of time, or temporally ‘spaced’, is better

Correspondence: Constance Tamara Heidt, Department of Clinical Psychology, University of Saskatche- wan, 9 Campus Drive Saskatoon, Saskatchewan S7N 5A5, Canada. Phone: C1 306-850-4114. Email: [email protected]

Ó 2015 The Australian and New Zealand Association of Psychiatry, Psychology and Law 48 C. T. Heidt et al.

retained than information received over a serial position effect indicates that information shorter period of time, or in a ‘massed’ fash- presented first and last is more likely to be ion (Carpenter, Cepeda, Rohrer, Kang, & recalled than information in the middle (Mur- Pashler, 2012; Cepeda, Pashler, Vul, Wixted, dock, 1962). Massed learning allows only one & Rohrer, 2006; Gluckman, Vlach, & first and last exposure to the information. Sandhofer, 2014). Most formal education sys- However, spaced learning has multiple oppor- tems are premised on the notion that optimal tunities of first and last exposure to the infor- learning occurs when information is spread mation and decreases the amount of poorly- over months, rather than condensed over days recalled middle material. Although unclear (Dunlosky, Rawson, Marsh, Nathan, & exactly what neuropsychological and/or cogni- Willingham, 2013; Pashler et al., 2007). One tive mechanisms are underlying the benefits of theory of why spaced learning is beneficial is distributed learning, it is likely to be an auto- that scheduling spaces in between instruction matic process independent from learner behav- allows a learner to review the material, which iour (Stickgold & Walker, 2007). increases the likelihood that the information When police officers participate in an will be better retained (Kornmeier & interview training programme it often occurs Sosic-Vasic, 2012). The mechanisms behind in a massed, rather than a spaced, format the advantage for spaced learning are (Dando, Wilcock, & Milne, 2008; Warren grounded in cognitive and neuropsychological et al., 1999). Interviews conducted by investi- research. Consolidation theory posits that dis- gative interviewers of children before, during, tributed learning allows breaks between and after two days of interview training instruction during which the memory traces revealed no difference in the pre- and post- are strengthened and learning consolidation interview behaviours (Rischke, Roberts, & occurs, which results in the improvement of Price, 2011). It was not until a second two- skills without practice (Litman & Davachi, day interview training session À which took 2008; Robertson, Pascual-Leone, & Miall, place two months later and comprised a 2004). Optimal consolidation of learned infor- review and classroom-based practice À that a mation may occur while sleeping, as a result significant improvement in performance was of the slow-wave sleep (SWS) and rapid eye observed (Rischke et al., 2011). Price and movement (REM) sleep support system aiding Roberts (2011) found an increased use of in synaptic consolidation (Diekelmann & open-ended prompts and more information Born, 2010; Stickgold & Walker, 2007). It is elicited following eight months of training, also suggested that spaced learning does not which consisted of both classroom compo- improve memory itself, but instead acts to nents and weekly feedback. A subsequent decelerate forgetting by reinforcing memory refresher session two months after the initial Downloaded by [University of Regina] at 07:22 25 February 2016 consolidation (Litman & Davachi, 2008). Cor- training resulted in continued improvements responding neuropsychological support for the in interviewing behaviour (Price & Roberts, consolidation theory indicates that, due to syn- 2011). Though the same material was cov- aptic plasticity, there is a strengthening of the ered in each training session, these results synapses that produces an increase in the com- suggest that the spaced format may facilitate munication between neurons (Kornmeier & interviewing skill acquisition. However, Sosic-Vasic, 2012). It is this process that instead of conceptualizing massed and spaced results in the benefits of spaced learning training within a competitive framework, it is (Kornmeier & Sosic-Vasic, 2012). probably more realistic to consider the types Another possibility is that the benefits of of training as two available options that are spaced learning can be attributed to the both being used, with the underlying question recency and primacy effects associated with as to which will better facilitate the transition the serial position curve (Murdock, 1962). The of knowledge into skilled interviewing. Effects of Distributed Learning 49

The benefits of spaced training to enhance by investigative interviewers on child learning and performance post-training have witnesses and alleged victims found that on been demonstrated outside the domain of average 40% were option-posing (e.g., ‘Were police interview training. Vlach and Sandhofer you inside or outside when it happened?’), (2012) found that when simple and complex 30% directive (e.g., ‘What colour was the science concepts were learned over the span shirt?’), and only 8% were open-ended invita- of four days, rather than in one day, children tions (e.g., ‘Tell me what happened’; Luther, aged five to seven years were significantly Snook, Barron, & Lamb, 2014). Improperly more likely to acquire and generalize the conducted interviews can further traumatize knowledge of simple science concepts. Simi- the interviewee and/or reduce the chances of larly, a study examining fifth-grade children in acquiring crucial investigative information a classroom setting found that when vocabu- (Fisher & Geiselman, 2010). Furthermore, an lary learning occurred one week apart, com- improperly conducted interview can ultimately pared to during the same day, children contribute to wrongful convictions (Snook recalled significantly more of the vocabulary et al., 2010). definitions five weeks later (Sobel, Cepeda, & Kapler, 2011). The benefits of spaced learning have also been observed in samples of individ- The Enhanced Cognitive Interview uals with moderate and severe brain injury Founded on empirically-based forensic inves- (Hillary et al., 2003). When to-be-learned tigative interview guidelines, the Enhanced words were not presented consecutively, but Cognitive Interview (ECI) is the gold stan- instead spaced apart using other words, partic- dard of investigative interviewing with adult ipants recognized significantly more words witnesses (Memon & Bull, 1991; Memon, (Hillary et al., 2003). A quasi-experimentally Meissnier, & Fraser, 2010). The ECI is a designed study found that when goal-focused structured interview that aims to enhance wit- coaching training occurred over a two-day ness memory and elicit unbiased information period with three weeks in between or once from witnesses using memory retrieval tactics per week for thirteen weeks that the coaching (i.e., outline interview, unlimited time, skills of the participants in both conditions sig- focused retrieval, report everything, context nificantly increased (Grant, 2007). However, mental reinstatement, picture activation, vary participants in the thirteen-week condition temporal recall order, changes perspectives were found to have significantly enhanced lev- technique, memory jogs, thanking the inter- els of emotional intelligence. These benefits of viewee) based on psychological research spaced learning have been observed in many (Fisher & Geiselman, 2010; Memon et al., different areas of learning (Carpenter et al., 2010). Because the amount of information Downloaded by [University of Regina] at 07:22 25 February 2016 2012). provided is often associated with the psycho- Special strategies are required in order to logical state of the interviewee, explaining enhance memory retrieval from witnesses the interview outline, emphasizing the lack of (Fisher & Geiselman, 2010;Milne,2004). a time restriction, and thanking the inter- Many relevant factors are outside police offi- viewee near the completion of the interview cers’ control, including the contextual condi- all serve to make the interviewee feel more at tions of the crime and the psychological state ease. Other methods which are used to obtain and cognitive/verbal abilities of the inter- maximum retrieval include directing the viewee (Fisher & Geiselman, 2010). However, interviewee to focus on retrieving the infor- the type of questions posed by interviewers is mation pertaining to the witnessed event, an well within the control of the interviewer and open-ended prompt to report everything, to is thus an important aspect of police training. reinstate the mental context of the inter- A recent study examining question types used viewee while the event was occurring, to 50 C. T. Heidt et al.

activate a more specific mental image of a Wright & Powell, 2006). In one study, even specific episode during the event, to report after a week of training, a sample of investiga- the event in different temporal sequencing, to tive interviews were found to use only 23% change perspectives when recalling the event, open-ended compared to 77% specific ques- and using memory jogs (e.g., did the person tions during the post-training interviews. Fur- remind you of anyone you know?). thermore, only 41% of the questions coded The ECI began as the Cognitive Interview were non-leading (Aldridge & Cameron, (CI) and was created in response to the need 1999). A review of 90 investigative interviews for an evidence-based investigative interview- performed by a Canadian police organization ing protocol (Fisher & Geiselman, 2010). also found that, of all questions asked during Over the last 25 years the CI has been the interviews, only 6% were open-ended explored in over 100 studies and has been (Snook & Keating, 2011). deemed a highly effective tool in obtaining Specific-closed questions inquire about a information (Fisher & Geiselman, 2010). The particular detail. These questions are CI has repeatedly been validated as extracting designed to elicit a much shorter response, 25 to 40% more correct information than con- often one or two words or a short phrase, and trol interviews (Fisher & Geiselman, 2010; generally start with ‘what’, ‘where’, ‘how’, Memon et al., 2010). In 1992 the CI was or ‘when’ (Milne, 2004). enhanced by Fisher and Geiselman (1992; i.e., the ECI) to include more social compo- nents and add a consistent interview structure. Inappropriate Question Types These alterations in structure did not reduce A leading or misleading question is one that the protocol’s efficacy (Memon et al., 2010). lacks impartiality. It can also include inquiries In modified form, the ECI still elicits signifi- relating to information that has yet to be men- cantly more forensically-relevant details than tioned by the interviewee. A leading question other structured interviews (Colomb, Ginet, steers the interviewee to the right response, as Wright, Demarchi, & Sadlet, 2013). defined by the interviewer, while a misleading question leads to the wrong response. Multi- ple questions involve asking more than one ECI Question Types question before the interviewee has had a A common theme in evidence-based inter- chance to answer. Forced-choice questions viewing protocols is to focus on open, non- offer only a limited number of response suggestive questions (Fisher & Geisleman, options and are most commonly found in the 2010). Consistent with this approach, the ECI form of yes/no questions, but can also include specifically describes both desirable and a multiple-choice format (Milne, 2004). Downloaded by [University of Regina] at 07:22 25 February 2016 undesirable questions in training, as outlined below (Milne, 2004). ECI Structure Appropriate Question Types Nine Enhanced Cognitive Interview Phases Questions beginning with ‘tell me’, ‘describe’, Nine phases comprise the ECI (for a sum- or that are formatted in a way which elicit mary refer to Table 1). As a highly structured from the interviewee as unrestricted and interview, the ECI requires strict adherence impartial a response as possible are open- to phase order and appropriate and inappro- ended prompts (Milne, 2004;Wright& priate question types (Fisher, Geiselman, & Powell, 2006). While the majority of questions Amador, 1989; Milne, 2004). Prematurely asked during the ECI should be open-ended, employing a phase, or the unintentional omis- this is often not the case in practice (Hughes- sion of one, could be damaging to an inter- Scholes & Powell, 2008; Luther et al., 2014; view (Milne, 2004). It is important that Effects of Distributed Learning 51

Table 1. The nine ECI phases.

Phase Purpose/components

1: Establish Rapport To make the interviewee as relaxed as possible 2: Explain Interview Explain the interview purpose, interview outline, and transferring control to Objectives interviewee Instructions are comprised of focused retrieval, no time restriction, and report everything 3: Initiating the Free Report Use an open-ended prompt to request the free report Context mental reinstatement, cue all senses, people, objects, layout, describe events free report initiation. Good interviewer behaviour, guggles/facilitators (‘okay’, ‘mmhmm’, ‘yeah’, etc.), and active listening strategies (echo probing, querying, and summarizing) 4: Questioning Appropriate questioning types (open-ended and specific-closed), inappropriate (forced-choice, multiple, and leading/misleading) 5: Varied and Extensive Utilizing memory facilitation techniques such as vary temporal recall order, Retrieval memory jogs, and change perspectives to obtain more information from the interviewee 6: Important Investigative Questions necessary for the investigation but not pertaining to the Questions information provided in the free report 7: Summary Summarize all the information obtained back to the interviewee 8: Closure Leave the interviewee in a positive state of mind 9: Evaluation Two types of evaluation that should occur, evaluation of obtained evidence and of interviewer

interviewers recognize how their behaviour quality of trainers, structure of training, level affects the interviewees’ ability to recall criti- of feedback, perceived relevance of material). cal information. Participating in an interview, The present study was designed to compare especially if the interviewee has been the effects of training using massed versus exposed to a traumatic experience, can be spaced learning conditions, which is just one overwhelming and anxiety-provoking (Fisher of the factors that could influence interview & Geiselman, 2010; Fisher, Geiselman, & skill transferral. Previous research has found Raymond, 1987). If the interviewee is not that when interview training is spaced, rather comfortable then he or she is less likely to than massed, correct procedures are more Downloaded by [University of Regina] at 07:22 25 February 2016 focus on maximum memory retrieval, which likely to be followed, and more open-ended translates into less information obtained from questions, and fewer closed questions are the interviewee (Milne, 2004). Highlighting asked (Rischke et al., 2011). The present both appropriate and inappropriate question study differs substantially from the Rischke types, the ECI focuses on acquiring as much et al. (2011) study, which investigated inter- accurate information as possible (Fisher & views with child witnesses, used the National Geiselman, 2010). Institute of Child Health and Human Development protocol (Lamb, Orbach, Hershkowitz, Esplin, & Horowitz, 2007), and The Current Study implemented a within-subjects design with a Hypotheses different interview spacing and presentation There are a number of reasons for the non- format. Participants in the present study were transferral of skills following training (e.g., trained in the proper way to conduct an ECI 52 C. T. Heidt et al.

in either a massed or spaced training session. take part in the ECI training. Interview partic- Training sessions were structured to resemble ipants received one course credit for their as much as possible the actual police inter- involvement in the study. view training format. Following the training session(s), participants were paired with a na€ıve interviewee (i.e., an undergraduate wit- Materials ness), who had previously watched a 10-min- ute video of a crime, and the pair participated Academic Motivation Scale in an audio-recorded ECI. It was hypothe- (Vallerand et al., 1992) sized that participants in the spaced condition The Academic Motivation Scale (AMS-28) is would use more appropriate (open-ended and a 28-item self-report measure designed to specific-closed) and fewer inappropriate assess learning motivation (e.g., ‘Why do you (multiple, leading/misleading, and forced- go to university?’). It has seven subscales choice) prompt types, would utilize a larger measuring three forms of extrinsic motivation number of ECI components, and would elicit (identified, introjected, and external regula- more accurate and fewer inaccurate details tion), three forms of intrinsic motivation (to than the massed condition. We also included know, toward accomplishment, and to experi- a measure of motivation in interviewers. ence stimulation), and amotivation. Each Because intrinsic motivation to learn has item is rated on a seven-point Likert scale been strongly associated with skill acquisi- ranging from 1 (Does not correspond at all) tion, it was hypothesized that individuals to 7 (Corresponds exactly). The AMS-28 has high in intrinsic academic motivation would been found to have a good level of internal pose more appropriate and fewer inappropri- consistency (a D .81) and an acceptable test- ate question types than individuals with lower retest reliability score (r D .79), while a factor intrinsic motivation (Quinones,~ 1995; Ryan & analysis has endorsed the seven-factor struc- Deci, 2000). It was also hypothesized that ture (Vallerand et al., 1992). motivation would interact with the temporal distribution manipulation condition and that intrinsic motivation combined with spaced Crime Video instruction would be associated with the most Participants viewed a real surveillance video appropriate question types and fewest inap- of a convenience store robbery (found on propriate question types. YouTube, titled ‘Kirtland Giant Armed Rob- bery 8-3-2012’) with a running time of 9 minutes and 22 seconds. The video presented two store clerks cleaning and several different

Downloaded by [University of Regina] at 07:22 25 February 2016 Method customers entering and exiting the store. An Participants armed robbery occurs, one of the store clerks There were two samples of participants: train- is threatened with a gun, and forced to pro- ing and interview participants. The 60 train- vide the money from the cash register. The ing participants who took part in the ECI perpetrator leaves, the police are called, and training were undergraduate students enrolled the remaining customers in the store commu- in a third-year Forensic Psychology course nicate with the store clerks. (46 women, 14 men). Participating in the ECI Although participants were not asked if training was a required component of the they had previously viewed the video, the Forensic Psychology class. The 60 interview video was specifically selected as a target participants were introductory psychology video in the present study for its low fre- students recruited from the psychology par- quency of views (<1200 at the time of the ticipant pool (46 women, 14 men) and did not study). Though they were not explicitly Effects of Distributed Learning 53

asked, no participant spontaneously reported session. Training participants were informed having previously viewing the video. that the interview participants had viewed a video of a crime unfolding, and were instructed to conduct an ECI in the way in which they had been trained to do during the Procedure session(s). After training, participants were Training Participants then randomly paired with the interview par- Training participants attended highly scripted ticipants (approximately 15À30 minutes fol- training sessions that were presented either in lowing the completion of training), provided one two-hour ECI training block (massed with a digital audio-recorder, and instructed condition) or two one-hour ECI training to find a quiet place to proceed with the inter- blocks distributed one week apart (spaced view. Debriefing and collection of the record- condition). Each training session was con- ers and questionnaires followed. ducted by the principal researcher, and con- Following the training session while the sisted of identical instruction and material interview participants watched the crime emphasis. A sign-up sheet was circulated video, training participants completed the Aca- around the Forensic Psychology class for par- demic Motivation Scale (AMS-28; Vallerand ticipants to choose which training session to et al., 1992). During the interview, training attend. Students were led to believe that the participants were permitted to refer to the training sessions were organized for their handouts received during the training session, scheduling convenience and, thus, were but it was suggested that constructing a quick- unaware of the spacing manipulation. Both glance reference sheet would be a convenient massed and spaced training sessions were option for ease of use during the interview. scheduled in the mornings and afternoons, and were interspersed over a three-week period each semester. Participation in the Interview Participants study took approximately three hours, with Interview participants watched a 10-minute two hours consisting of the ECI training and surveillance video depicting an armed rob- one hour allotted for conducting the ECI. All bery at a convenience store. Interview partici- nine phases of the ECI were covered during pants were told that they would be the training session, with training participants interviewed regarding what they had viewed paired to participate in mandatory 10-minute on the video. Immediately following the practice sessions following critical phases of video, the interview participants were paired the interview (i.e., phases 1 to 5). In the with a training participant and interviewed massed condition participants were taught all about what they could recall from the video. Downloaded by [University of Regina] at 07:22 25 February 2016 nine phases of the ECI over a two-hour period, whereas the participants in the spaced condition were taught the first three phases Results during the first week and the last six phases in second week. Spaced training participants Coding were given a short (approximately 5-minute) The interviews were transcribed and coded refresher session at the beginning of the sec- for the question types used by the training ond week of training consisting of a brief participants, utilization of critical ECI com- overview of the material covered the previous ponents, and accuracy. A total of 13 inter- week. To ensure similarity between the con- views (22% of the total sample) were coded ditions, massed training participants also by two trained researchers, one of whom was received the same refresher session at the the principal researcher, with an inter-coder same time during the interview training reliability of .88. Considering the acceptable 54 C. T. Heidt et al.

inter-coder reliability score, the remaining 47 Accuracy interviews were coded solely by the principal Critical details from the crime video and researcher. Coders were blind to the partici- incorrect details provided were coded and pant condition. used to determine interviewee accuracy. Overall, there were 33 critical details consist- ing of 16 perpetrator-specific and 17 other Question Types details. The critical details included informa- Utterances were coded as either appropriate tion pertaining to the setting of the crime and (open-ended and specific-closed) or inappro- the behaviour, appearance, clothing of the priate (multiple, leading/misleading, and store clerks, perpetrator, and witnesses (e.g., forced-choice). Utterances beginning with ‘the setting is a convenience store’, ‘the cus- ‘tell me’, ‘describe’, or formatted in a way tomer has pink hair’, ‘the perpetrator has that produced as unrestricted and impartial nylon over his face’). Incorrect information interviewee response as possible were coded was determined by calculating the number of as open-ended prompts, whereas questions incorrect details provided by the interviewee inquiring about a particular detail, often start- during the interview. ing with ‘what’, ‘where’, ‘how’, or ‘when’, were coded as specific-closed questions. If a question began with ‘tell me’ or ‘describe’ but was inquiring about a specific detail, it Question Types was coded as a specific-closed question. Separate univariate analyses of variance Although in the ECI procedure it is preferred (ANOVAs) were conducted for each appro- that specific-closed questions follow an open- priate (open-ended and specific-closed) and ended prompt, in the present study training inappropriate (multiple, leading/misleading, participants were not penalized for not fol- and forced-choice) question type to assess lowing this particular questioning sequence. whether the use of each question type varied Questions which lacked impartiality, by learning condition (see Table 2). Open- including inquiries pertaining to information ended prompts differed significantly, with not yet mentioned by the interviewee or ques- participants in the spaced condition (M D tioning that steered the interviewee to the 3.53) using more open-ended prompts than ‘right’ or ‘wrong’ response were coded as those in the massed condition (M D 2.27), F leading/misleading questions. Questions were (1, 58) D 4.07, p D .05. There were no signif- coded as multiple when more than one ques- icant differences between the massed and tion was asked before the interviewee had a spaced conditions in the use of specific- chance to respond. Questions offering a closed questions, F(1, 58) D 0.02, p D .88. Downloaded by [University of Regina] at 07:22 25 February 2016 restricted number of response options (e.g., yes/no questions) were coded as forced- choice questions. Table 2. Mean number of question types used in the massed and spaced conditions.

Massed Spaced Critical ECI Components Inter- Ten critical ECI components (outline inter- Question type MSDMSDrater view, unlimited time, focused retrieval, report Open-ended 2.27 1.64 3.53 3.03 0.83 everything, context mental reinstatement, Specific-closed 3.53 2.81 3.67 3.87 0.82 picture activation, vary temporal recall order, Multiple 0.93 1.34 0.47 0.63 0.87 changes perspectives technique, memory Leading/misleading 0.63 1.35 0.17 0.46 1.00 jogs, thanking the interviewee) were coded as Forced-choice 6.00 9.10 4.37 3.59 0.80 either present or absent. Effects of Distributed Learning 55

Table 3. Mean proportion of question types in the 1.93, p D .17; leading/misleading, F(1, 58) D massed and spaced conditions. 2.71, p D .11; forced-choice, F(1, 58) D 0.27, D Massed Spaced p .58.

Question type MSDMSD Interview Components Open-ended .20 .18 .25 .17 Specific-closed .22 .14 .23 .18 Separate Chi-squared tests were conducted to Multiple .06 .08 .03 .05 determine if there were differences between Leading/misleading .45 .20 .47 .19 the massed and spaced conditions in the pres- Forced-choice .03 .06 .02 .03 ence of the critical components of the ECI (see Table 4). Two of ten critical components were found to differ significantly. Participants in the spaced condition were more likely to There were also no significant differences request that the interviewee attempt focused between the two conditions in any of the retrieval, x2 (1, n D 60) D 5.71, p D .03, and inappropriate question types: multiple ques- used the vary temporal recall order retrieval tions, F(1, 58) D 2.99, p D .09; leading/ strategy more often than those in the massed misleading, F(1, 58) D 3.20, p D .08; forced- condition, x2 (1, n D 60) D 5.46, p D .04. A choice, F(1, 58) D 0.31, p D .58. univariate ANOVA was performed to assess To control for the overall number of ques- whether there was a difference in the total tions asked by the interviewers, the propor- number of interview components used within tions of each question type per interview the massed and spaced conditions. The result were also calculated and analysed using sepa- was not significant, F(1, 58) D 1.61, p D .21. rate ANOVAs (see Table 3). There were no Two separate univariate ANOVAs were significant differences between the massed conducted to determine if there were differen- and spaced conditions for either of the appro- ces in the use of active listening strategies priate question types: open-ended prompts, (echo probing, querying, and summarizing) F(1, 58) D 1.523, p D .22; specific-closed, or guggles/facilitators (‘okay’, ‘mmhmm’, F(1, 58) D 0.06, p D .81. Nor were there any ‘yeah’, etc.) between the massed and spaced significant differences found for any of the conditions. There were no significant differ- inappropriate question types between the ences between the massed and spaced condi- conditions: multiple questions, F(1, 58) D tions in the use of: active listening strategies,

Table 4. Critical interview components.

Downloaded by [University of Regina] at 07:22 25 February 2016 Massed Spaced Interview components Present Present x2 Cramer’s V Inter-rater

Outline interview 13 (43%) 17 (57%) 1.07 .13 1.00 Focused retrieval 7 (23%) 16 (53%) 5.71 .31 1.00 Unlimited time 26 (87%) 24 (80%) 0.48 .09 1.00 Report everything 28 (93%) 27 (90%) 0.22 .06 1.00 Context reinstatement 10 (33%) 13 (43%) 0.64 .10 1.00 Picture activation 6 (20%) 3 (10%) 1.18 .14 1.00 Vary recall order 21 (70%) 28 (93%) 5.46 .30 1.00 Change perspectives 11 (37%) 9 (30%) 0.30 .07 1.00 Memory jogs 7 (23%) 7 (23%) 0.00 .00 1.00 Thank interviewee 20 (67%) 23 (77%) 0.74 .11 1.00 56 C. T. Heidt et al.

F(1, 58) D 0.43, p D .52; guggles/facilitators, differences between the massed and spaced F(1, 58) D 0.08, p D .78. conditions on the number of inaccurate details (total and perpetrator-specific) elicited from the interviewees. Neither the total number of Accuracy critical details, F(1, 58) D 0.03, p D .86, Analysing the number of critical details nor perpetrator-specific details, F(1, 58) D obtained by the interviewer is a means for 0.38, p D .54, varied between conditions. determining how successful the interview has been. Acquiring critical details is, after all, the primary purpose of almost any interview. Motivation Two separate univariate ANOVAs were con- Motivation to learn has been strongly associ- ducted to assess differences between the ated with skill acquisition (Quinones,~ 1995; massed and spaced conditions in the number Ryan & Deci, 2000). It was hypothesized that of critical details (total and perpetrator-spe- individuals high in intrinsic academic motiva- cific) elicited from interview participants (see tion would use more appropriate and fewer Table 5). Neither the total number of critical inappropriate question types than those with details, F(1, 58) D 1.12, p D .30, nor the lower intrinsic motivation. Separate Pearson perpetrator-specific details, F(1, 58) D 0.01, product-moment correlations were conducted p D .92, varied between conditions. However, to assess whether high scores on intrinsic the difference between the numbers of perpe- motivation were associated with a higher trator-specific details obtained through open- number of appropriate questions asked. This ended prompts differed significantly between hypothesis was not supported; open-ended the massed (M D 1.10) and spaced (M D prompts, r(56) D .18, p D .17, and specific- 2.23) conditions, F(1, 58) D 3.88, p D .05. closed questions, r(56) D .04, p D .76, were Details acquired through open-ended prompts not associated with higher intrinsic motiva- are more reliable as they are unlikely to be tion scores. Similar non-significant results influenced by interviewer bias. were found for the hypothesis that high levels Another method of determining the suc- of intrinsic motivation would be inversely cess of an interview is by analysing the num- correlated to inappropriate question types: ber of incorrect details provided by the multiple, r(56) D¡.11, p D .41, leading/mis- interviewee. Obtaining incorrect information leading, r(56) D¡.20, p D .88, and forced- can be detrimental to an investigation and can choice, r(56) D¡.03, p D .83. There is noth- waste scarce resources. Two separate univari- ing which suggests that high intrinsic motiva- ate ANOVAs were conducted to assess tion scores were related to the use of fewer inappropriate question types in this sample. Downloaded by [University of Regina] at 07:22 25 February 2016 To determine whether motivation inter- Table 5. Mean number of critical details (total acted with the massed vs spaced learning and perpetrator-specific) elicited by question type condition, partial correlations of condition, by condition. controlling for intrinsic motivation score, Massed Spaced with appropriate questions and inappropriate questions were examined. These analyses Question type Total Perp. Total Perp. indicated no significant correlations with Open-ended 1.90 1.10 3.23 2.23 condition. Specific-closed 1.43 0.90 1.00 0.53 Multiple 0.13 0.13 0.07 0.00 Discussion Leading/misleading 0.07 0.07 0.00 0.00 Forced-choice 0.10 0.03 0.17 0.13 Forensic interview training for police officers is often limited, averaging around two days, Effects of Distributed Learning 57

which can impact the acquisition of witness were referenced numerous times during the evidence (Clarke & Milne, 2001; Snook training session for both conditions. Though et al., 2010; Wells et al., 2006). However, the proportion of open-ended prompts did not even when training has occurred it does not differ between experimental conditions, this always translate into properly conducted finding remains important because this larger interviews (Aldridge & Cameron, 1999; number of open-ended questions resulted in a Warren et al., 1999). Based on the theory of greater number of perpetrator-specific details distributed learning, skill acquisition is more reported in the spaced condition. In relation likely to occur when the information is to a criminal investigation the perpetrator- learned in a spaced, rather than a massed, for- specific details are arguably the most relevant mat (Cepeda et al., 2006). Although literature to the task of identifying the culprit and thus exists on factors that improve police inter- furthering the investigation. Although the viewing (e.g., Snook et al., 2010) and on the present study did not find any significant dif- benefits of spaced learning (e.g., Cepeda et ferences between the massed and spaced con- al., 2006), there is a dearth of research on the ditions in the use of specific-closed questions impact of spaced learning on interviewing or inappropriate question types, the signifi- practices. As previously mentioned, Rischke cant difference between open-ended prompt et al. (2011) observed improvements in inter- usage between the massed and spaced condi- viewing performance following a refresher tions is a promising finding. Open-ended session in which the same material was pre- prompts are well regarded for their ability to sented again. The present study was unique in acquire information that is as free from inter- that the interview material was learned over a viewer influence as possible, and infrequent one-week span, with a 5-minute refresher use of such prompts is often cited as a cause session in between, and different material of improperly conducted interviews presented in each session. If the structures of (Aldridge & Cameron, 1999; Milne, 2004). If police interview training programmes are not such a minimal manipulation used in the pres- empirically evaluated, it is likely that the field ent study has increased the number of open- will continue to observe poor transfer of ended prompts, this has the potential to sig- training to performance. Unlike much of the nificantly improve the effectiveness of inter- research on improving investigative inter- view training. viewing practices, the present study investi- Additional differences were observed gated the effect of a simple change in training between conditions in the use of varying tem- format, either massed or spaced instruction. poral recall order and focused retrieval tech- If such a simple alteration can influence inter- niques, which are key ECI components. view effectiveness, this could have important Considering the substantial amount of Downloaded by [University of Regina] at 07:22 25 February 2016 resource implications for increasing the suc- research which suggests that information cess of training programmes. acquired over a longer time period results in Participants in the spaced condition used an improvement in skill acquisition, it could a significantly greater number of open-ended be that two hours of interview training was prompts, and elicited more correct perpetrator not enough time to produce a noticeable details from such questions. Because partici- behaviour change in all elements of the ECI, pants in both conditions received the same so it remains to be seen whether longer train- ECI training, this is as a direct result of the ing periods would show additional differen- experimental manipulation of learning spac- ces. There is also the possibility that a longer ing. Many ECI components (i.e., vary tempo- duration of time between training and inter- ral recall order, picture probing, and memory viewing would have produced a larger differ- jogs) require the interviewer to use open- ence between conditions. Research suggests ended prompts and thus those question types that when information is retrieved after a 58 C. T. Heidt et al.

delay, rather than immediately, performance education system does by spacing new mate- is enhanced (Cepeda, Vul, Rohrer, Wixted, & rial over time. These differences in the tem- Pashler, 2008; Loaiza & McCabe, 2012). One poral spacing of learning and its effects on study that investigated the optimal gap found the subsequent benefits, coupled with the that there was no ideal amount of time scarcity of research regarding the benefits of between review and testing of material spaced learning on police interview training, (Cepeda et al., 2008). However, if testing makes this research a stepping stone to fur- occurs too soon after material is learned, the thering the field of police interview training. results will likely yield misleadingly high lev- With growing restrictions on training els of immediate mastery (Cepeda et al., budgets, there may be an interest in having 2008). Still, with only two hours of training select team members attend extensive train- there were significant and promising findings ing with the aim of having these trained mem- related to properly conducting the ECI. Con- bers then disseminate their knowledge to their sidering the time limitations on police inter- colleagues. In such a format, knowledge of view training, these findings are particularly how to best temporally distribute ‘in-house’ important because with only two hours of training sessions in the most practical and training there were several significant areas economical way is essential. In the context of of improvement, each of which may consid- the present research, it appears that spacing erably improve interviewing effectiveness. even very brief training or refresher sessions It was hypothesized that individuals high would be most advantageous. This strategy in intrinsic academic motivation would use could be used for any discipline seeking to more appropriate and fewer inappropriate best make use of limited training budgets. question types. This hypothesis was not sup- ported. but there was little variation in the motivation scores, which limited the ability Limitations to observe correlations. It is possible that the To assess skill acquisition properly it would lack of differences related to motivation is have been desirable to have a longer time simply an indication that spaced training interval between the training and the reten- would improve performance for all trainees, tion test (Donovan & Radosevich, 1999). In irrespective of their motivations to obtain this study, the training participants had training. 15À30 minutes to refresh themselves on the There are differences in the benefits earlier material. Having the training partici- gained by spacing learning and the subse- pants return after a longer interval would quent effects on skill acquisition and post- have been a more ecologically valid gauge of training behaviour (Carpenter et al., 2012). skill. Further, the short delay meant that the Downloaded by [University of Regina] at 07:22 25 February 2016 Carpenter et al. (2012) note that there is not a interview participants’ memory for the event ‘one-size-fits-all’ approach to the improve- was quite strong and that much of the video ments gained from spaced learning, and that was remembered for both conditions, which optimal time between sessions is based on could be an underlying cause of the relatively what type of information is required for mem- small effect sizes. ory retrieval. It is important from a research A further issue relates to the training of and applied point of view to examine the opti- university students, which may not be gener- mal amount of time and of spacing to deter- alizable to a population of police officers. For mine how police interviewing can be police officers, interview training is a skill enhanced by using spaced learning, as cur- necessary for a career in law enforcement, rently there is no empirical research investi- whereas for the university participants it was gating the impact of spacing interview a required assignment. Interest, involvement, training in the same format as the formal and motivation may not be the same between Effects of Distributed Learning 59

the two groups. However, the lack of effect of Carpenter, S. K., Cepeda, N. J., Rohrer, D., Kang, motivation suggests that these factors may S. H., & Pashler, H. (2012). Using spacing to not influence relative performance between enhance diverse forms of learning: Review of recent research and implications for instruc- the two learning conditions. These factors tion. Educational Psychology Review, 24, would have little impact on skill acquisition 369À378. if the neuropsychological and/or cognitive Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & mechanisms underlying the benefits of spaced Rohrer, D. (2006). Distributed practice in ver- learning are automatic. It is also crucial to bal recall tasks: A review and quantitative syn- thesis. Psychological Bulletin, 3, 354À380. explore if a more representative sample of Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & police officers would show similar benefits Pashler, H. (2008). Spacing effects in learning: with the spaced training format. The findings A temporal ridgeline of optimal retention. Psy- would not only contribute to the distributed chological Sciences, 19, 1096À1102. learning literature, but would also suggest Clarke, C., & Milne, R. (2001). National evalua- tion of the PEACE investigative interviewing whether police interviewing research should course. London: Home Office. focus on police officers as participants or if Colomb, C., Ginet, M., Wright, D., Demarchi, S., university students are an approximately & Sadlet, C. (2013). Back to the real: Efficacy equivalent population. and perception of a modified cognitive inter- view in the field. Applied Cognitive Psychol- ogy, 27, 574À583. Dando, C., Wilcock, R., & Milne, R. (2008). The Conclusion cognitive interview: Inexperienced police offi- Inferior interviewing practices have been cers’ perceptions of their witness/victim inter- viewing practices. Psychology, Crime & Law, implicated in causing failures within the jus- 17,59À70. tice system (FPT Heads of Prosecutions Diekelmann, S., & Born, J. (2010). The memory of Committee Working Group, 2011). It is sleep. Nature Reviews Neuroscience, 11, hypothesized that thousands of innocent indi- 114À126. viduals have been charged with and convicted Donovan, J. J., & Radosevich, D. J. (1999). A meta-analytic review of the distribution of of crimes they did not commit because of the practice effect: Now you see it, now you don’t. improper interviewing of witnesses (Snook Journal of Applied Psychology, 84, 795À805. et al., 2010). With only a simple alteration in Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, training format it was possible to observe sig- M. J., & Willingham, D. T. (2013). Improving nificant improvements in interviewing practi- students’ learning with effective learning tech- niques: Promising directions from cognitive ces. If distributed learning principles are and educational psychology. Psychological utilized then interviewing practices could Science in the Public Interest, 14,4À58. benefit, which could result in a substantial Fisher, R. P., & Geiselman, R. E. (1992). Memory improvement in skill acquisition and a subse- enhancing techniques for investigative inter- Downloaded by [University of Regina] at 07:22 25 February 2016 quent reduction in injustice. viewing: The cognitive interview. Springfield, IL: Charles C. Thomas. Fisher, R. P., & Geiselman, R. E. (2010). The cog- nitive interview method of conducting police Disclosure statement interviews: Eliciting extensive information No potential conflict of interest was reported by and promoting therapeutic jurisprudence. the authors. International Journal of Law and Psychiatry, 33, 321À328. Fisher, R. P., Geiselman, R. E., & Amador, M. (1989). Field test of the cognitive interview: References Enhancing the recollection of actual victims and witnesses of crime. Journal of Applied Aldridge, J., & Cameron, S. (1999). Interviewing À child witnesses: Questioning techniques and Psychology, 74, 722 727. the role of training. Applied Developmental Fisher, R. P., Geiselman, R. E., & Raymond, D. S. Science, 3, 136À147. (1987). Critical analysis of police interviewing 60 C. T. Heidt et al.

techniques. International Journal of Police Memon, A., & Bull, R. (1991). The cognitive Science and Management, 15, 177À185. interview: Its origins, empirical support, FPT Heads of Prosecutions Committee Working evaluation and practical implications. Journal Group, Department of Justice Canada. (2011). of Applied and Social Psychology, 1, The path to justice: Preventing wrongful con- 291À307. victions. Retrieved from http://www.ppsc- Memon, A., Meissner, C. A., & Fraser, J. (2010). sppc.gc.ca/eng/pub/ptj-spj/ptj-spj-eng.pdf The cognitive interview: A meta-analytic Gluckman, M., Vlach, H. A., & Sandhofer, C. M. review and study space analysis of the past 25 (2014). Spacing simultaneously promotes mul- years. Psychology, Public Policy, and Law, tiple forms of learning in children’s science 16, 340À372. curriculum. Applied Cognitive Psychology, 28, Milne, B. (2004). The enhanced cognitive inter- 266À273. view: A step-by-step guide. (Unpublished Grant, A. M. (2007). Enhancing coaching skills training guide) University of Portsmouth. and emotional intelligence through training. Murdock, B. B. (1962). The serial position effect Industrial and Commercial Training, 39, of free recall. Journal of Experimental Psy- 257À266. cholgy, 64, 482À488. Hillary, F. G., Schultheis, M. T., Challis, B. H., Pashler H., Bain P. M, Bottge B., Graesser A., Koe- Millis, S. R., Carnevale, G. J., Galshi, T., & dinger K., McDaniel M., Metcalfe J. (2007). De Luca, J. (2003). Spacing of repetitions Organizing instruction and study to improve improves learning and memory after moderate student learning (NCER 2007À2004).Wash- and severe TBI. Journal of Clinical and ington, DC: National Center for Education Experimental Neuropsychology, 25,49À58. Research, Institute of Education Sciences, U.S. Hughes-Scholes, C. H., & Powell, M. B. (2008). Department of Education. An examination of the types of leading ques- Price, H. L., & Roberts, K. P. (2011). The effects tions used by investigative interviewers of of an intensive training and feedback program children. Policing, 31, 210À225. on police and social workers’ investigative Kornmeier, J., & Sosic-Vasic, Z. (2012). Parallels interviews of children. Canadian Journal of between spacing effects during behavioral and Behavioural Science, 3, 235À244. cellular learning. Frontiers in Human Neuro- Quinones,~ M. A. (1995). Pretraining context science, 6,1À5. effects: Training assignment as feedback. Lamb, M. E., Orbach, Y., Hershkowitz, I., Esplin, Journal of Applied Psychology, 80, 226À236. P. W., & Horowitz, D. (2007). A structured Rischke, A. E., Roberts, K. P., & Price, H. L. forensic interview protocol improves the qual- (2011). Using spaced learning principles to ity and informativeness of investigative inter- translate knowledge into behavior: Evidence views with children: A review of research from investigative interviews of alleged child using the NICHD Investigative Interview Pro- abuse victims. Journal of Police and Criminal tocol. Child Abuse & Neglect, 31, 1201À1231. Psychology, 26,58À67. Lamb, M. E., Sternberg, K. J., Orbach, Y., Hersh- Robertson, E. M., Pascual-Leone, & Miall, R. C. kowitz, I., Horowitz, D., & Esplin, P. W. (2004). Current concepts in procedural consol- (2000). The effects of intensive training and idation. Nature Reviews Neuroscience, 5, ongoing supervision on the quality of investi- 576À582. gative interviews with alleged sex abuse vic- Ryan, R. M., & Deci, E. L. (2000). Intrinsic and Downloaded by [University of Regina] at 07:22 25 February 2016 tims. Applied Developmental Science, 6, extrinsic motivations: Classic definitions and 114À125. new directions. Contemporary Educational Litman, L., & Davachi, L. (2008). Distributed Psychology, 25,54À67. learning enhances relational memory consoli- Snook, B., Eastwood, J., Stinson, M., Tedeschini, dation. Learning & Memory, 15, 711À716. J., & House, J. C. (2010). Reforming investi- Loaiza, V. M., & McCabe, D. P. (2012). Tempora- gative interviewing in Canada. Canadian lÀcontextual processing in working memory: Journal of Criminology and Criminal Justice, Evidence from delayed cued recall and 52, 215À229. delayed free recall tests. Memory & Cognition, Snook, B., House, J. C., MacDonald, S., & East- 40, 191À203. wood, J. (2012). Police witness interview Luther, K., Snook, B., Barron, T., & Lamb, M. E. training, supervision, and feedback: A survey (2014). Child interviewing practices in of Canadian police officers. Canadian Journal Canada: A box score from field observations. of Criminology and Criminal Justice, 54, Journal of Police and Criminal Psychology, 363À372. (nv), 1À9. Effects of Distributed Learning 61

Snook, B., & Keating, K. (2011). A field study of children’s acquisition and generalization of adult witness interviewing practices in a Cana- science concepts. Child Development, 83, dian police organization. Legal and Crimi- 1137À1144. nological Psychology, 16, 160À172. Warren, A. R., Woodall, C. E., Thomas, M., Sobel, H. S., Cepeda, N. J., & Kapler. I. V. (2011). Nunno, M., Keeney, J. M., Larson, S. M., & Spacing effects in real-world classroom Stadfeld, J. A. (1999). Assessing the effective- vocabulary learning. Applied Cognitive Psy- ness of a training program for interviewing chology, 25, 763À767. child witnesses. Applied Developmental Sci- Stickgold, R., & Walker, M. P. (2007). Sleep- ence, 3, 128À135. dependent memory consolidation and reconso- Wells, G. L., Memon, A., & Penrod, S. D. (2006). lidation. Sleep Medicine, 8, 331À343. Eyewitness evidence: Improving its probative Vallerand, R. J., Pelletier, L. G., Blais, M. R., value. Psychological Science in the Public Briere, N. M., Senecal, C., & Vallieres, E. F. Interest, 7,45À75. (1992). The academic motivation scale: A Wright, R., & Powell, M. (2006). Investigative measure of intrinsic, extrinsic, and amotiva- interviewers’ perceptions of their difficulty in tion in education. Educational and Psycholog- adhering to open-ended questions with child ical Measurement, 52, 1003À1017. witnesses. International Journal of Police Sci- Vlach, H. A., & Sandhofer, C. M. (2012). Distrib- ence and Management, 8, 316À325. uted learning over time: The spacing effect in Downloaded by [University of Regina] at 07:22 25 February 2016