Challenging the witness incompetence myth: Forensic interviews with individuals with intellectual disability

Irit Hershkowitz, University of Haifa

The current study followed investigative interviews with individuals with mild and moderate intellectual disability (ID) and observed both the types of prompts addressed to them and the nature of their responses. The sample comprised 200 alleged victims, in 4 equal and matched groups: individuals with mild ID and their mental-age (MA) matches, and individuals with moderate ID and their MA matches. All alleged victims were interviewed by trained investigators who followed the protocol of the National Institute of Child Health and Human Development, known to promote developmentally sensitive interviews. A descriptive analysis revealed that individuals with ID were faced with a substantial number of prompts, with recall prompts dominating the interviews. They replied to almost all prompts, addressed substantive issues following over 2/3 of the prompts, and provided new details following 1/3 of the prompts, but they tended to respond at lower rates than did their counterparts. The number of details was generally lower among individuals with ID. However, all but the moderate ID group produced more details following recall than recognition prompts. Performance varied according to the severity of ID, with the mild ID group outperforming the moderate ID group. The findings challenge the notion that individuals with ID cannot participate in criminal investigations or act as witnesses. The data suggest that when abuse is suspected, awareness of their weakness and strengths is likely to promote informed and effective interviewing, which may yield valuable forensic evidence.

abuse, eyewitness testimony, intellectual disabilities, forensic interviews

“Individuals with intellectual disability (ID) form a vulnerable population at increased risk of abuse and are thus more often required to participate in criminal investigations. However, little is known about the competency of individuals with ID in forensic contexts, although laboratory analogue studies have suggested that they can accurately report events they have experienced. The present study involves the first comparison of actual forensic interviews conducted with alleged victims diagnosed with ID to those conducted with typically developing (TD) alleged victims, matched for mental age (MA).” (p. 393)

“In the eyewitness testimony literature, individuals with mild and moderate ID are characterized by a range of functional problems considered to affect their competency as witnesses. Cognitive impairment is the central diagnostic feature of individuals with ID, thought to compromise their ability to provide meaningful and reliable eyewitness testimony. Deficits in information processing, working memory, and executive control, as well as poorer comprehension of concepts and events, may reduce the amount of information they encode during an event. When engaging in conversation about their experiences, individuals with ID generally show a narrower range of receptive and expressive skills. Receptive skills may be hindered by limitations in attention, information processing resources, and verbal short-term memory, all of which are likely to challenge their understanding and processing of communicative demands. Expressive skills are likely to be affected by less effective retrieval strategies; lower verbal and narrative abilities; and the tendency to please questioners, acquiesce to suggestions, and confirm external information, as well as by failure to understand the consequences of making a statement. Due to these cognitive limitations, many police officers, attorneys, and potential jurors doubt the reliability of individuals with ID and believe they cannot act as competent witnesses. Consequently, allegations are often not investigated, because these cases are less likely to result in successful prosecution and conviction.” (p. 393)

“When allegations do reach the investigation phase, police interviewers tend to act in counterproductive ways, failing to provide individuals with ID with proper conditions for reporting their experiences. Contrary to best practice recommendations, they often do not use free recall prompts, which have been found to encourage production of rich and accurate information. Rather, they tend to use types of questions that are known to compromise memory retrieval. […] Furthermore, interviewers may address individuals with ID with contaminating repeated questions, following which interviewees are likely to change their answers and produce contradictions, both in pretrial investigations and in court.” (pp. 393–394)

“Research findings have clearly shown that individuals with mild ID respond well to recall prompts and can provide equally accurate though less complete information in comparison to TD individuals. Even individuals with moderate ID can provide accurate, yet limited, amounts of information when addressed with recall prompts. Nevertheless, despite the importance of this research, it has involved mostly laboratory analogue procedures, and thus it is not necessarily indicative of performance in real-life investigations. The current study explores the criminal investigation of individuals, both adolescents and adults, with mild and moderate ID who were alleged victims of physical or sexual abuse.” (p. 394)

“The study explored the types of prompts addressed to individuals with ID over the course of their investigation, that is, recall versus recognition prompts, as well as the nature of their responses. The rate of replies in response to interviewers’ prompts, the rate of relevant responses addressing forensic issues, and the rate of responses including new forensic information were observed. In addition, the number of forensically relevant details provided in the average response was analyzed, while examining variation in detail according to the types of prompts. The investigation of the ID groups was compared to that of TD individuals matched for MA (hereafter referred to as TD matches). […] The interviews observed in the current study were conducted by trained investigators specializing in child investigations as well as investigations of individuals with ID. The investigators followed the protocol of the National Institute of Child Health and Human Development (NICHD), known to foster developmentally adapted interviews with TD children.” (p. 395)

“Following research on the NICHD protocol, interviewees with and without ID were expected to be offered more recall prompts rather than recognition ones, both in the transitional phase, aimed to identify target abusive events, and in the substantive phase, designed to explore the events in detail. In addition, all four groups were also expected to be more productive in response to the average recall prompt than to the average recognition one. Informed by previous research on interviewees with mild ID, we expected them to be as responsive (provide replies, informative responses, and responses containing new forensic detail) to the interviewers’ prompts as were their TD matches and to produce a similar number of details. However, the performance of those with moderate ID was expected to decline in comparison to both their TD matches and those with mild ID.” (p. 395)

“The sample consisted of 200 alleged abuse victims in four equal groups: 50 adolescents and adults diagnosed with mild ID (age range = 12–49, M = 20.05) paired with 50 TD matches (TD children of age range over 8 and under 12, M = 9.31) and 50 adolescents and adults diagnosed with moderate ID (age range = 12–58, M = 29.5) paired with 50 TD matches (TD children of age range over 4 and under 8, M = 5.05). All TD children attended regular schools, studied in regular classes, and had no known special needs. The individuals with ID were diagnosed by the Diagnosing Committee of the Ministry of Labor and Welfare (according to the Welfare Law (the treatment of individuals with intellectual and developmental disability), 1969, Israel), following the American Association on Intellectual and Developmental Disabilities (2010) recommendations and based on intellectual functioning and adaptive behavior. All participants (74 male, 126 female) were referred for investigation following suspicions of physical (n = 78) or sexual (n = 122) abuse and made allegations during the investigation. No other inclusion criteria were employed, and the interviews collected were the first ones available during data collection (2008 –2011) to meet the matching criteria.” (p. 395)

“The interviews were conducted in the centers of the Service for Child Investigations and Special Investigations in the four regions of Israel or at the interviewee’s residence. All investigations were performed according to the NICHD Investigative Interview Protocol. […] The NICHD protocol employs investigative practices found to facilitate reliable accounts from children about their experiences. It is fully structured and covers all phases of the investigative interview. In the introductory phase, interviewers introduce themselves, clarify the need to describe events in detail and to tell the truth, and explain the ground rules and expectations. […] In a transitional phase between the presubstantive and the substantive parts of the interview, a series of initially open-ended, nonsuggestive prompts are used to identify the target event(s) to be investigated. […] Once an allegation is made, the free recall phase begins with the main invitation. […] Follow-up open-ended prompts are then recommended, […] as are cued invitations aimed at eliciting uncontaminated accounts of the alleged incident(s) from free-recall memory. […] Only after open-ended questioning has been exhausted do interviewers proceed to directive questions. […] If crucial details are still missing at the end of the interview, interviewers may ask limited option-posing questions. […] Suggestive prompts that communicate what responses are expected […] are strongly discouraged in all phases of the interview. […] Videotape recordings of the interviews were transcribed, and transcriptions were reviewed to ensure their completeness and accuracy.” (p. 396)

“Interviewees with mild ID were as likely as their controls and those with moderate ID to make spontaneous allegations, but interviewees with moderate ID were less likely to make spontaneous allegations than were their controls and required more prompting before they made an allegation. Generally, more recall rather than recognition prompts were provided to all groups, both in the transitional phase and throughout the interview. However, the lower functioning groups were faced with fewer prompts, and specifically fewer recall prompts, in the substantive investigation and responded with lower rates of verbal replies. Both the ID groups and the lower functioning groups provided fewer informative responses in general and fewer responses including new details. They also provided fewer details per prompt and, specifically, fewer details per recall prompt. Recall prompts were typically followed by more details than recognition prompts, although the relative advantage of recall prompting was absent for the ID low-functioning group (moderate ID).” (p. 398)

“In an exploration of the quality of interviewing in the observed sample, the data showed that recall prompts dominated the interviews, both in the transitional phase and in the substantive phase. […] Note that nearly one third of the alleged victims made an allegation spontaneously, with no need for questioning at all (although this proportion was lower for the moderate ID group). Thus, despite some variation, the allegations seem to have been generally raised with relatively appropriate means.” (p. 399)

“When exploring the criminal events, we found that both individuals with and without ID were addressed with a higher proportion of recall than recognition prompts. Recall prompts consisted of over half of all prompts and 75% of all information-request prompts in the substantive part, thereby confirming the relatively high quality of interviewing.” (p. 399)

“In an inspection of the interviewees’ responses in the substantive investigation (regardless of their content), the data illustrated that individuals with ID were able to answer a substantial number of questions seeking diverse types of information. Even individuals with moderate ID were given about 150 prompts on average and replied to almost all of them, indicating that despite their limitations, they showed basic communicative competency.[…] Analysis of the content of responses showed that interviewees with ID not only complied with communicative demands but also seemed to understand and process many of the requests addressed to them, often providing relevant information. Despite some differences from their TD controls, individuals with ID provided relevant and informative responses to over two thirds of the prompts they received.” (p. 399)

“Providing responses that include new forensic details, rather than just repeating previously mentioned information, is a key communicative factor for the successful progression of an investigation. Interviewees with ID provided such responses to over one third of the prompts and did not perform as well as did their respective matches. […] In the current study, all individuals with ID, including those in the mild ID group, did not perform as well as did their TD controls, and their production of details was compromised. This pattern does not follow the hypotheses or the pattern found in previous research. It should be noted that prior research employing MA matches was performed on children but not on adolescents or adults as in the current study and that the events reported were not of a criminal nature. The relatively low performance of individuals with mild ID can perhaps be attributed to the stressful nature of the reported events (sexual or physical assault), as well as the stressful situation of a criminal investigation. It is also possible that because the individuals in the mild ID groups were older than were their matches, they were more aware of the potential consequences of a criminal investigation, which may have hindered their performance. Thus, contrary to expectations, their greater maturity and life experience may have acted to decrease rather than increase their performance during criminal investigations. In addition, individuals with ID tend to be subject to more severe abuse, by closely related figures, involving threats or use of force, and are more often silenced, factors that could have potentially limited their ability and/or motivation to provide elaborated descriptions of the abuse.” (pp. 399–400)

“Although production was lower for the ID groups (collapsed) than for their TD matches, their recall prompts were still more effective than were recognition ones, yielding larger amounts of information that were likely to be more accurate […] The current findings confirm the superiority of recall prompts for individuals with ID, suggesting that the latter employ memory retrieval mechanisms similar to those employed by TD individuals, although at a lower capacity. However, follow-up univariate analyses showed that the superiority of recall prompts was evident for all groups except the moderate ID group. Thus, not only did individuals with moderate ID produce fewer details (less than half the number of details) in comparison to their TD matches, but their production was insensitive to retrieval strategy (recall vs. recognition). This seems to represent a substantial difference in memory retrieval mechanisms among individuals with moderate ID, which may explain their poorer outcomes.” (p. 400)

“The current study yields important insight into the capabilities and limitations of individuals with ID, suggesting that they can participate in interviews and provide forensic information about experiences of abuse. Thus, misconceptions about individuals with ID among forensic professionals were challenged by the empirical data, and dismissal of their witnessing skills was shown to be unfounded. Instead, the data support adopting an empowering approach toward ID individuals throughout the legal process and suggest that involving them as witnesses may enable more rigorous investigations of crimes against them and help prevent further crimes.” (p. 400)

“The data suggest that use of developmentally sensitive investigative strategies, similar to those designed for TD children, can be appropriate when individuals with ID are interviewed. […] The current study supports [the conclusion that the NICHD protocol can be effective in the investigation of children with ID] based on a more ecologically valid exploration in a forensic context and further suggests that the NICHD protocol can be successfully used in interviews of adolescents or adults with ID of both mild and moderate levels.” (p. 401)

“The study also points to weaknesses in communication and memory performance of individuals with ID, of which investigators should be aware to set realistic expectations of these witnesses. Most important, investigators should be aware of their lower production of details and provide them with more cognitive support for recall instead of shifting to potentially harmful recognition prompts. […] Employing appropriate cueing strategies may perhaps support free recall among individuals with ID and improve their retrieval.” (p. 401)

“By utilizing ID individuals’ strengths while adapting to their limitations, forensic interviewers can help them become valuable informants in criminal investigations. Because a large variation among individuals with ID was evident and they tended to function at different levels of skill, a well-informed evaluation of memory, verbal, and communicative skills of the individuals referred to investigation is necessary to conduct developmentally adjusted forensic interviews.” (p. 401)

“Because this was a field study of individuals describing criminal events that were obviously not documented, it was impossible to assess the accuracy of the information reported. However, the number of recall details can be indicative of the quality of the statements, because extant research has shown that information elicited with recall strategies is generally accurate (over 80% accuracy) among individuals with ID.” (p. 400)

“Another limitation of the current study was the lack of CA-matched groups, in addition to MA matches. In an attempt to avoid variations in investigative agencies, in interviewers’ training, and especially in the investigative guidelines employed, we selected all groups from the investigation service for children and individuals with ID. Although this service investigates ID individuals of all ages, it does not serve TD individuals over 14 years of age. Thus, CA matching was not possible. Consequently, the individuals with ID were substantially older than were their TD matches, which may have affected both the interviewers’ interventions and the interviewees’ performance. In addition, the fact that all participants made allegations limits generalizability of the findings. It is possible that participants denying suspicions of abuse would have been addressed with a different composition of prompts and would have displayed lower cooperation and engagement or lower eyewitness skills.” (p. 400)

“Another limitation was the lack of distinction between open-ended and directive prompts. These prompts were collapsed during coding, which prevented exploration of their distinct effects in the investigation of individuals with ID. […] Future research may elucidate the role of open-ended versus directive utterances in real-life investigations.” (p. 400)

“Finally, in the current study, the forensic statements provided by individuals with ID were analyzed only in terms of their volume, whereas their clarity and organization were ignored. Examination of the narrative coherence of individuals with ID can be of special relevance to the forensic system.” (p. 400)

As always, please join the discussion below if you have thoughts or comments to add!