The association between investigative interviewers' knowledge of question type and adherence to best‐practice interviewing

It is well established that not all investigative interviewers adhere to ‘best‐practice’ interview guidelines (i.e., the use of open‐ended questions) when interviewing child witnesses about abuse. However, little research has examined the sub skills associated with open question usage. In this article, we examined the association between investigative interviewers' ability to identify various types of questions and adherence to open‐ended questions in a standardized mock interview. Study 1 , incorporating 27 trainee police interviewers, revealed positive associations between open‐ended question usage and two tasks; a recognition task where trainees used a structured protocol to guide their response and a recall task where they generated examples of open‐ended questions from memory. In Study 2 , incorporating a more heterogeneous sample of 40 professionals and a different training format and range of tests, positive relationships between interviewers' identification of questions and adherence to best‐practice interviewing was consistently revealed. A measure of interviewer knowledge about what constitutes best‐practice investigative (as opposed to knowledge of question types) showed no association with interviewer performance. The implications of these findings for interviewer training programs are discussed.     

A field study of adult witness interviewing practices in a Canadian police organization

Purpose. The current study examined witness interviewing practices in a Canadian police organization. The effect of interviewer, interviewee, and interview characteristics on those practices was also examined. Method. Ninety witness interviews from a Canadian police organization were coded for the following interviewing practices: types of questions asked (i.e. open‐ended, probing, closed‐ended, clarification, multiple, leading, opinion/statement, and re‐asked), the number of interruptions, percentage of words spoken by interviewer, type of pre‐interview instructions (consequential vs. generic), and whether or not a free narrative was requested (and when requested during the interview). Characteristics pertaining to the interviewer (e.g. primary interviewer's age), interviewee (e.g. witness gender), and interview (e.g. crime type) were also coded. Results. Results showed that closed‐ended and probing questions were the most widely asked questions, and that open‐ended questions were asked relatively infrequently. It was also found that the 80–20 talking rule was violated in 89% of the interviews, interviewers rarely interrupted the witnesses, and free narratives were requested often. Overall, the effect of interviewer, interview, or interviewee characteristics on interviewing practices was minimal. Conclusions. The finding that scientifically prescribed interviewing practices are employed rarely by Canadian police officers highlights a need for increased professional interviewing training. The finding that practices are largely unaffected by personal and situational factors suggests that such training would be equally beneficial for all types of interviewers, interviewees, and contexts.     

The relationship between job status,interviewing experience,gender, and police officers' adherence to open‐ended questions

Purpose. The current study examined whether several factors related to the job and demographic profile of police officers are associated with adherence to best‐practice guidelines when interviewing children. Method. One hundred and seventy‐eight police officers completed a standardized (simulated) interview regarding an allegation of abuse by a 5‐year‐old child. Immediately prior to this interview, details were obtained from the officers' regarding their job status, gender, interview experience, the timing and nature of prior training/supervision, and experience outside the policing profession with young children. Results. The results showed that timing of training was the only factor that related to interview performance. The proportion of open‐ended questions among participants who completed their interviewer training course less than 1 month prior to the simulated interview was better than those who completed the training earlier. Interestingly, the performance of the latter group was identical to that of a group of participants who had not yet received any formal interview training. The implications of the findings are discussed, along with directions for future research.     

The role of confirmation bias in suspect interviews: A systematic evaluation

Purpose. The three studies presented in this paper systematically examined the effect of expectations of guilt on interviewer questioning style, confession, denial rates, and suspects' verbal behaviour during interview. Method. Undergraduate students were recruited to participate in the three studies. In Study 1, 61 participants formulated questions that they wanted to ask a suspect to determine whether or not they cheated on a task. Prior to formulating their questions, participants were led to believe that the suspect was likely to be guilty or innocent. In Study 2, 45 ‘innocent’ and ‘guilty’ participants were accused of cheating on a task and were interviewed with either guilt‐presumptive questions or neutral questions. In the final study, 58 participants listened to a selection of audiotaped interviews from Study 2. They then rated various aspects of the suspects' verbal behaviour. Results. As hypothesized expectations of guilt resulted in the formulation of more guilt‐presumptive questions even when participants were free to generate their own questions (Study 1). A significant association was found between suspect guilt/innocence and whether they confessed or denied, although there was no association between questioning style and confession or denial rates (Study 2). However, as expected, ratings of independent observers who listened to tape recordings of the suspect interviews indicated an influence of questioning style on the suspects' verbal behaviour whereby a self‐fulfilling prophecy effect occurred (Study 3). Conclusions. These results indicate that expectations of guilt can indeed have an effect on questioning style and that this in‐turn can lead to a self‐fulfilling prophecy effect.     

The Effects of Interviewer Gender and Child Gender on the Informativeness of Alleged Child Sexual Abuse Victims in Forensic Interviews

Most students of forensic interviews have focused on the interrogatory techniques used to elicit information from alleged child abuse victims. We asked how the gender of the interviewer and the gender of the child affected this process. Forensic investigators in three countries used either the NICHD structured interview protocol or local standard interview practices to interview 672 alleged victims who ranged in age from 4 to 14 years. Analyses of the interviews showed significant effects of gender on both the interviewers' behavior and the amount of information provided by children. Female interviewers asked boys more invitations, as well as absolutely and proportionally more suggestive questions, than they did girls, whereas male interviewers interviewed boys and girls similarly. Children's responses varied depending on their gender and age, the gender of the interviewer, and the type of question asked. Girls of all ages provided more information in response to directive questions posed by female rather than male interviewers whereas boys did not respond differently to male and female interviewers. The oldest girls provided more information in response to option-posing questions posed by male interviewers. More information was provided by the younger children in response to suggestive prompts from interviewers of the opposite gender. The gender-of-interviewer effects were attenuated in protocol-guided interviews.     

Examination of the consistency of interviewer performance across three distinct interview contexts

Abstract

The current study examined the consistency of investigative interviewers' performance (n=31) across three distinct interview paradigms: (a) a mock interview where an adult actor played the role of a child recalling abuse, (b) a mock interview where a school child recalled an innocuous event that was staged at the child's school, and (c) a field interview where the interviewer elicited a statement of abuse from a child. Performance was measured by calculating the proportion of open-ended and leading questions, and by eliciting expert ratings of the presence of a range of problem behaviours commonly exhibited by interviewers. Overall, the performance of individual interviewers was relatively stable across the tasks. Heterogeneity in stability, however, differed according to the type of question and the nature of the event being examined. In particular, the mock interview paradigm where the adult acted the role of an alleged child abuse victim produced a measure of performance that was more similar to the field interview than the interview where a school child recalled an innocuous event. The implications of the findings for trainers, and directions for future research, are discussed.     

Direct mixed and open questions in child interviewing: An analog study

Purpose. This study evaluated the relative value of direct questions, open‐ended questions, and mixed questions (including cued invitations) in eliciting accurate statements from young children about a known episode. Methods. Children, 25 aged 5–7 years, and 25 aged 10–12 years, were interviewed concerning a standardized experience the previous week. Direct, mixed, and open‐ended question sequences were counterbalanced between participants. Transcribed videos were assessed for quantity and quality of statements. Results. Mixed questions tended to fall in the middle, between direct and open questions on length of statement, and number of errors. For the younger children, direct questions, but not mixed questions, increased the number of errors of commission in subsequent open question sequences. Cued invitations, where the direct portion of the mixed question pair reiterated information previously elicited from the child, produced fewer errors of omission than mixed questions without adding errors of commission. Conclusions. Mixed questions, especially cued invitations, were found to be a useful alternative for obtaining specific content when open questions alone were not sufficient. Support for conducting the interview in an ‘inverted pyramid’ or ‘funnel’ fashion was found in an analog study where ground truth was firmly established.     

An examination of the questioning styles of police officers and caregivers when interviewing children with intellectual disabilities

Purpose. This research provided a detailed analysis of the types of questions and verbal strategies used by police officers and caregivers when interviewing children with intellectual disabilities about events. Method. Twenty eight children aged 9 to 13 years with a mild or moderate intellectual disability participated in a staged event at their school. Each child was then interviewed on separate occasions by the child's primary caregiver and by a police officer who was authorized to conduct investigative interviews with children. Results. While the approach used by the police officers was broadly consistent with best‐practice recommendations (i.e. their interviews contained few leading, coercive or negative strategies), they frequently interrupted the child's account and used relatively few minimal encouragers and other strategies designed to keep the child talking. The caregivers used a high proportion of direct, leading and coercive strategies to elicit information from their children. Even when caregivers used open‐ended questions, their children provided less event‐related information than they did to the police interviewers. Conclusion. The quality of evidence obtained from children with intellectual disabilities is likely to be dependent (albeit in part) on the degree to which police interviewers adhere to best‐practice guidelines, as well as the children's general experience with an open‐ended style of communication.     

When Does Quality Count?: Perceptions of Hearsay Testimony About Child Sexual Abuse Interviews

This study assessed how the quality of a sexual abuse investigative interview with a child and the age of the child influence jurors reactions to either the original interview with the child or to testimony by an adult hearsay witness (the interviewer). Participants (N = 360) were randomly assigned to 1 of 12 conditions in a 2 (type of testimony:hearsay testimony vs. child interview) × (interview quality: poor, typical, or good) × (age of the child: 4 years old vs. 10 years old) factorial design. Participants reached individual verdicts, answered a series of questions, and then deliberated in a group with five other participants. As predicted, jurors in the child interview conditions were more likely to find the defendant guilty if they read the good interview than if they read either the poor or the typical interview, but in the hearsay conditions verdicts did not significantly differ by interview quality. These findings suggest that there is a significant loss of information when the testimony of a hearsay witness is used in place of the actual interview with the child, and call into question the appropriateness of admitting hearsay testimony by interviewers.     

Getting the story from child witnesses: exploring the application of a story grammar framework

Abstract

Investigative interviews with children about alleged abuse were analysed to determine the degree to which the child's responses adhered to a story grammar framework, and whether the presence of story grammar elements was associated with interviewers’ adherence to best-practice (i.e. open-ended) questioning. The sample included 51 interviews with child witnesses from across Australia. The interviews were administered by a police officer with children (37 girls and 14 boys) aged 3–16 years (M age = 103.82 months, SD = 34.21 months). The interviewers’ questions were categorised as open-ended or specific and the children's responses were classified as a story grammar element, context/background information, or ‘don't know’ responses. The majority of interviewer questions were specific in nature and the majority of children's responses were context/background details. Open-ended questions were more successful in eliciting story grammar from children. Of the story grammar elements, the interviewers’ specific questions usually targeted setting and attempt details. These findings suggest that improvement in the narrative coherence of children's reports of abusive events can potentially be achieved by increasing interviewers’ use of open-ended questions.     

The effect of repeated questioning on children's accuracy and consistency in eyewitness testimony

Purpose. In police interviews children may be asked the same question many times. We investigated how the number of repetitions and the interval between those repetitions affected the accuracy and consistency of children's responses. Methods. 156 children aged 4–9 years watched a staged event and were interviewed individually 1 week later. Children were asked eight open‐ended questions, which were each repeated a further four times (making a total of forty questions). Half these open‐ended questions could be answered from information in the event, and half were unanswerable (so children should have said ‘don't know’ in response to these questions). The questions were repeated in gist form. The interval between an initial question and its repetitions was varied by use of other questions and twenty non‐repeated filler questions. The intervals between repetitions were immediate repetition, repetition after a delay of three intervening questions, after a delay of six intervening questions, and after ten or more intervening questions. Results. Over a quarter of children's responses to repeated questions changed, usually resulting in a decline in accuracy, particularly after the first repetition. Subsequently, the number of repetitions and delay interval had little effect on responses to answerable questions although accuracy to unanswerable questions continued to decline. Conclusions. Question repetition had a negative affect on children's consistency and accuracy. For unanswerable questions in particular, the more often a question was repeated the more likely children were to invent a response.     

Exploring the Disclosure of Forensic Evidence in Police Interviews with Suspects

Despite many years of empirical research focusing on investigative interviewing and detecting deception, very little research attention has been paid to the various types of evidence which feature in police interviews with suspects. In particular, the use of forensic evidence in the context of police interviews has not been previously considered, although in recent years the availability of various types of forensic analyses has dramatically increased. In the current study 398 experienced police interviewers from various countries completed a questionnaire about their experience of using various types of forensic evidence in interviews with suspects, as well as their perceptions regarding the strength of various sources of forensic information and how this may affect their interviewing strategy. The results indicated that although the participants have forensic evidence available in a large proportion of their interviews with suspects, the vast majority of police interviewers have received no training about how to interpret or use such forensic information. However, the perceived strength of forensic evidence was reported by some participants to affect their interview strategy and specifically the timing of the disclosure of such evidence during an interview. These findings are discussed with reference to police training and interview techniques, and suggestions for further research are offered.     

Simulations of child sexual abuse interviews using avatars paired with feedback improves interview quality

We tested whether simulated child sexual abuse (CSA) interviews with computer-generated child avatars could improve interview quality. Feedback was provided not only on question types, as in previous research, but also on whether the conclusions drawn by the interviewers were correct. Twenty-one psychology students (average age M = 24.5) interviewed four different avatars which had a simulated story of either abuse or non-abuse. The participants were randomly divided into two groups: one received feedback on question types and conclusions after each simulated interview and the other one did not receive any feedback. Avatars revealed pre-defined ‘memories’ as a function of algorithms formulated based on previous empirical research on children's suggestibility. The feedback group used more open-ended and fewer closed questions. They also made more correct conclusions and found more correct details in the last two interviews compared to the no-feedback group. Feedback on both the question types and conclusions in simulated CSA interviews with avatars can improve the quality of investigative interviews in only one hour. The implications for training practice were discussed.     

Age trends in eye-witness suggestibility and compliance

Abstract

This paper addresses two issues: whether there is a developmental trend in suggestibility to misleading post-event information, and whether suggestibility can be reduced by use of part of the cognitive interview. Twenty participants from each of three age groups (four-to-five year olds, eight-to-nine year olds and adults) watched a filmed event, and half of the participants in each age group were subsequently asked to recall everything they had seen using a method derived from the cognitive interview procedure. Following this, all participants were asked questions about the filmed event, some of which incorporated misleading information. Twenty four hours later the witnesses were interviewed again, this time critical questions were included about the truth of the presuppositions introduced in the initial questionnaire. It was found that although the eight-to-nine year olds were more suggestible than adults, the apparent greater suggestibility of very young children (four-to-five years) could potentially be explained in terms of heightened compliance to the perceived demands of the interviewer. The 'be complete' part of the cognitive interview only produced an improvement in performance for the eight-to-nine year olds.     

Is That What I Said?: Witnesses' Responses to Interviewer Modifications

Modifications occur when interviewers contradict statements made by witnesses or imply that witnesses provided information that they (interviewers) did not provide. Because of their suggestive nature, modifications threaten the reliability of investigative interviews. This study investigated developmental differences in witnesses' responses to modifications during interviews as well as in inclusion of modified misinformation in subsequent answers. Preschool, elementary school, and college students were interviewed about a video presentation. In the experimental conditions, the interviewer contradicted information about the video provided by the participants. Participants then answered two sets of follow-up questions: one immediately following the interview and another 6–8 days later. Results indicated that participants were more likely to ignore modifications than to correct or agree with them. Adult participants were most likely to disagree with modifications. Preschoolers were most likely to incorporate modified misinformation into subsequent answers. Implications of these findings for investigative interviews are discussed.     

LET'S GO TO THE VIDEOTAPE: WHY THE FORENSIC INTERVIEWS OF CHILDREN IN CHILD PROTECTIVE CASES SHOULD BE VIDEO RECORDED*

Child sexual abuse is often difficult to prove due to a lack of physical evidence. In many instances, the only evidence is a child's statement given during a forensic interview. Forensic interviews are conducted to assess whether the child has been abused, to develop a plan to protect a child's safety pending trial, and to provide further investigative leads. Note taking by the interviewer is currently the primary method for documenting what takes place in a forensic interview. Research shows that this form of documentation is problematic because interviewers tend to omit abuse‐related details in their notes. This Note suggests that federal law should require that forensic interviews of children in child sexual abuse cases be video recorded. State law can provide for a policy of video recording even in the absence of a federal law mandate. Video recording would better preserve the child's statements thereby improving the reliability of the information that is obtained during forensic interviews. The child's demeanor would also be fully captured on video as opposed to getting lost in an interviewer's notes.     

Weekly community interviews with high-risk participants: operational issues

To address several key questions in social science research, repeated interviews of individuals drawn from difficult populations are required. This article describes an approach for addressing the challenges associated with longitudinal interview studies, including locating research participants, obtaining reliable and valid interview data over time, and retaining participants across the course of the study. We applied this approach to conduct a study designed to identify changeable risk factors for violence among high-risk people with mental illness. To successfully conduct weekly interviews of these individuals in the community across a 6-month period, we developed a flexible and personalized interview format; carefully selected, trained, and supervised staff; and developed incentives to maximize participant retention. Each of these three steps is discussed as a guide for future longitudinal studies that involve interviewing difficult populations.