Compared to typically developing children, children and youths with disabilities are more likely to be abused (Crosse et al. 1993; Sullivan and Knutson 1998, 2000; Westcott and Jones 1999). Children and youths with intellectual disabilities (IDs) who are alleged victims of abuse also report more severe forms of sexual abuse than typically developing children (Hershkowitz, Lamb, and Horowitz 2007). Despite a paucity of experimental research showing their abilities as eyewitnesses they are nonetheless viewed as less reliable informants (Henry and Gudjonsson 2007). Courts seem to be reluctant to accept these children's testimonies because of their age and their intellectual disability (Henry and Gudjonsson 2003). In addition, they are not interviewed properly in real forensic interviews, which can imply that they are not given the opportunity to report their experiences as accurately and coherently as possible (Cederborg and Lamb 2008a).
Courts seldom ask for expert guidance when assessing their credibility, and as a result courts can make decisions about credibility largely in ignorance of these children and youths’ capabilities, behaviour, and limitations. Regardless of interviewer behaviour and insufficient knowledge about their capabilities, courts argue that credible accounts from children and youths with IDs should have the same characteristics as those from alleged victims without intellectual disabilities (Cederborg and Lamb 2006). When these children are not understood correctly there is a risk that they are excluded from a proper legal trial (Cederborg and Gumpert 2010).
Knowledge from research can develop professionals’ understanding of capacities and difficulties of children and youths with IDs. So far, we do not know enough about the quality of details given by them in real forensic interviews, however, and this was the focus of the present study.
Research from laboratory studies has shown that children with mild IDs are less able to develop their reports compared to typically developing (TD) children of the same chronological age but that they do not differ when reporting accurate information from free recall or misleading questions. In addition, children with mild IDs do not differ in relation to suggestibility compared to TD children of the same chronological age. However, children with moderate IDs provide less information than both TD children and children with mild IDs. They are also more suggestible, although their responses to free recall questions tend to be accurate (Henry and Gudjonsson 2003). This means that the severity of disability influences how children and youths with IDs perform, but overall, the accuracy of their accounts has been described as comparable to that of mental-age-matched TD peers (Fowler 1998; Henry and Gudjonsson 1999; Iarocci and Burack 1998; Michel et al. 2000; Zigler 1969). However, children and youths with IDs may change their responses when option-posing and suggestive questions are repeated in real forensic interviews (Cederborg et al. 2009).
In sum, children and youths with IDs (including those with autism spectrum disorder (ASDs)) may have problems remembering the events in question, they may acquiesce to suggestions, and have difficulties communicating their experiences thus making them unable to provide coherent and detailed reports of their experiences (Cederborg and Lamb 2006). When interviewing alleged witnesses who have a variety of IDs, police officers should therefore give priority to strategies that will help possible victims provide the most accurate and complete information possible (Cederborg, Gumpert, and Abbad Larsson 2009; Home Office 2002; Jones 2003). Interviewers should use open question types whenever possible because they maximize accurate recall from both typical and intellectually disabled witnesses. Hence, police officers should start with open questions and then proceed to more specific questions as needed (Gordon and Schroeder 1995; Poole and Lamb 1998), bearing in mind that responses from people with IDs may become less accurate when they are asked focused questions (Henry and Gudjonsson 2003; Kebell et al. 2004).
So far, we do not know enough about how children and youths with memory and communicative difficulties are able to describe their experiences of abuse in real forensic interviews. Consequently, this was the focus of the present study. We explored the quality of central and peripheral details in interviews with 32 children and youths with IDs about their abuse experiences in relation to the quality of question asked. We first made a quantitative analysis to identify all the police officers’ question types, invitations, directives, option-posing, and suggestive utterances. Second, all the children's elicited details were coded into subcategories of central and peripheral information, and third, we analyzed how they responded to possible influencing prompts; that is, option-posing and suggestive questions.
Prosecutors from all 39 Swedish districts were asked by letter and phone to send as much information as possible about cases processed during the last five years in which disabled children and youths’ were allegedly victimized. ‘Disability’ was not systematically recorded, so case selection depended on the prosecutors’ and police officers’ memories and the sample was thus selective rather that representative. In total, we obtained files involving 69 children and youths. Our focus in this study was on those participants who were said to have intellectual disabilities including those with autism spectrum disorder. We examined the police officers’ first formal investigative interviews with 32 children and youths with IDs. One youth was involved in two different cases with two various perpetrators. Therefore the study involves 33 different crime cases. When analyzing the child and youth's elicited details in relation to type of questions, the focus was on two groups of question types. The first group was open questions, i.e. invitations that encourage free recall responses, and directive questions refocusing the child's attention on details or aspects of the alleged incident that the child already mentioned. The second group comprised focused questions, i.e. option-posing questions that focus attention on details or aspects of the alleged incident that the child has not previously mentioned, as well as suggestive questions in which the interviewer strongly communicates what response is expected.
In order to gain insight into each witness's possible reporting capabilities, circumstances and experiences, we first conducted an inductive review of all the documents (the transcribed interviews, documents from the police investigations and the court files) in each case. Information about the different participants’ test results and capacities was seldom obtained formally during the investigation and the courts were often given this information third-hand (Cederborg and Lamb 2006). As a result, the sample involved children and youths with a diverse array of disabilities.
From the limited information available, it was found that:
- 1. Twenty-two of the 33 cases involved children and youths who were developmentally delayed; nine were assessed with mild IDs (one youth was involved in two different cases); and 13 with unspecified degrees of ID.
- 2. Four others were reported to have autism spectrum disorder (one with Asperger syndrome).
- 3. Seven had been diagnosed with ID (two mild and five unspecified) combined with ASDs (one with Asperger syndrome).
The interviews involved 24 females and nine males whose chronological ages were between 5.3 and 22 years (M=12.9 years) when the last incident of abuse was believed to have occurred, and between 5.4 and 23.7 years of age when subsequently interviewed (M=13.2 years).
Thirty-one of the participants were thought to be exposed to abuse for the first time when their chronological age was less than 18 years. One case involved, however, a girl older than 18 years of age when first being abused. She was described as having severe developmental delays. Because of the participants’ IDs and presumed memory limitations they have been referred to as children and youths throughout this paper.
Most of the children and youths were suspected victims of sexual abuse. Most of the suspected perpetrators were well known or familiar to the children and youths (see Table 1).
|Type of crime||Relationship to perpetrator|
|Diagnosis||Sexual abuse||Physical abuse||Relative||Immediate family||Familiar||Unfamiliar|
|Intellectual disabilities/Autism spectrum disorder||5||2||3||4|
|Autism spectrum disorder||3||1||4††||1|
|Notes: There were 32 witnesses and 33 interviews.|
|†One victim mentioned two immediate family members as perpetrators and so was interviewed twice.|
|††One victim mentioned two immediate family members.|
All case material was given to the first author by the prosecutors and police officers in accordance with the provisions of the Official Secrets Act in Sweden. Personal details and references to places that might permit identification were removed to ensure that none of the victims could be recognized. When the study was conducted, Swedish researchers were not required to have their studies reviewed by human subjects’ protection committees, but the present study was reviewed and approved by the official at Linköping University, Sweden, responsible for monitoring research being conducted by University staff. This official ensured that the study was designed and implemented in accordance with the Declaration of Helsinki (World Medical Association 1975) regarding research on human beings.
After checking the transcribed interviews from video recordings to ensure their completeness and accuracy, the first author reviewed the transcripts and identified each interviewer utterance that was an invitation, directive, option-posing, or suggestive, using the categories developed by Lamb et al. (1996, 2007) and another researcher independently coded 10 of these interviews. Inter-rater reliability was 95%. Differences were resolved through discussion.
Invitations: Utterances, including questions, statements, or imperatives, prompting free recall responses from the child.
Directive: These refocus the child's attention on details or aspects of the alleged incident that the child has already mentioned, providing a category for requesting additional information.
Option-posing: These were utterances were those that focused the witness's attention on details or aspects of the alleged incident that the witness had not previously mentioned, asking the witness to affirm, negate, or select an investigator-given option using recognition memory processes.
Suggestive: These were utterances were used in such a way that the interviewer strongly communicated what response was expected or assumed details that had not been revealed by the witness.
All the details elicited from the children were then coded using a developed version of Lamb et al. (1996). The central details were categorized in three different subgroups and peripheral information in two different subgroups.
Central details: Description of the crime, identification of the suspect, time and place, the suspects’ actions and temptations, the victims’ actions and perceptions during the abuse were coded into separate categories.
Peripheral details: Descriptions of the crime not involving the abuse event and description of victims’ ‘state’ during the abuse as well as emotions and thoughts attributed to the suspects or possible other witnesses were also coded.
Repeated details were coded only once.
All substantial event information elicited using option-posing and suggestive prompts were categorized using two different combined response categories:
Agreement: The child or youth accepted an option proposed or detail suggested by the investigator and may have elaborated upon it.
Disagreement: The child or youth did not accept an option proposed or detail suggested by the investigator and may have proposed an alternative option.
All 33 transcripts, in the second and third step of analysis, were coded by the second author. Twenty percent of the transcripts were randomly selected and independently coded by the first author. Inter-rater reliability was 95%. Disagreements were resolved through discussion.
The types of questions used in the interviews
The 33 interviews of children and youths in this study contained a total of 4027 questions pertaining to alleged abuse with a mean of 122 questions per interview. There were equal numbers of open questions 50% (5% invitations and 45% directives) asked as there were focused questions 50% (43% option-posing and 7% suggestive).
The numbers of central and peripheral details elicited
Table 2 shows the mean numbers of details that were elicited and were analyzed using a repeated measure ANOVA and post-hoc t-tests. Similar numbers of central details (M = 73.21, SD = 44.98) were elicited as peripheral details (M = 58.96, SD = 40.61), F(1, 96) = 3.00, p = .09, η2=.08. Most of the central details reported were descriptions of the crime including the time and place as well information about the identity of the suspect (M=63.78, SD = 40.18) compared with details about the suspects actions (M = 4.27, SD = 5.36) and the victims actions (M = 5.15, SD = 5.35), ts(32) > 8.63, ps<.001, which did not differ. The peripheral details disclosed were predominantly about the context of the abuse events (M = 48.00, SD=34.78) compared with emotional details (M = 10.96, SD = 8.75), t(32) = 6.99, p>.001.
|Central details||Peripheral details|
|Crime||Suspects' actions||Victims' actions||Context||Emotions||Total|
|M (SD, %)||M (SD, %)||M (SD, %)||M (SD, %)||M (SD, %)||M (SD, %)|
|Invitations||5.78 (14.28, 4%)||0.24 (0.56, 0%)||0.24 (0.93, 0%)||5.36 (10.26, 4%)||0.81 (1.50, 0.5%)||12.45 (23.31, 9.5%)|
|Directives||27.15 (19.00, 21%)||1.66 (3.02, 1%)||1.93 (2.12, 1%)||21.69 (22.44, 17%)||5.75 (5.17, 4%)||58.21 (41.21, 44%)|
|Option-posing||27.33 (16.00, 21%)||2.18 (3.14, 2%)||2.39 (3.62, 2%)||18.42 (14.45, 14%)||3.75 (4.01, 3%)||54.09 (26.71, 41%)|
|Suggestive||116 (3.51, 3%)||0.18 (0.52, 0%)||0.57 (0.96, 0%)||2.51 (2.74, 2%)||0.63 (1.45, 0.5%)||7.42 (6.57, 5.5%)|
|Total||63.78 (40.18, 49%)||4.27 (5.63, 3%)||5.15 (5.35, 4%)||48.00 (34.78, 36%)||10.96 (8.75, 8%)||132.18 (71.53, 100%)|
There was also a significant difference in the numbers of details elicited for the different question types, F(3, 96) = 41.38, p < .001, η2=.56. A similar number of details were elicited by directives (M = 58.21, SD = 41.21) compared to option-posing questions (M = 54.09, SD = 26.71). More details were elicited by the directives and option-posing questions compared to invitations (M = 12.45, SD = 23.31), ts(32) > 6.69, p<.001, and suggestive questions (M = 7.42, SD = 6.57), ts(32) > 7.18,<p.001. As can be seen in Table 2, 42% of the details elicited by directives and option-posing questions were central details about the crime (21% each).
The effectiveness of each of the questions types
Although there were few invitations used in the interviews their efficacy can be seen by examining the numbers of details they elicited each time they were used in the interviews. Table 3 shows the numbers of details that were elicited as a function of the type of questions asked. An ANOVA revealed a significant effect of question type, F(1, 32) = 3.38, p<.05 and no other effects (p>.6). On average invitations lead to slightly longer responses compared to the other question types, although pair-wise post-hoc tests did not reveal reliable differences between the invitations and other question types (ps>.22). This means that the children and youths with ID did not disclose more details when invitations were asked compared to the other questions.
|Open questions||Focused questions|
|Central||0.95 (1.27)||0.60 (0.38)||0.64 (0.39)||0.52 (0.49)|
|Peripheral||1.30 (3.40)||0.60 (0.55)||0.52 (0.72)||0.44 (0.46)|
Answers to option-posing and suggestive questions
Because of the problems associated with asking focused questions we examined how the children and youths with ID answered option-posing and suggestive questions (Table 4). There were more agreements than disagreements to option posing questions t(32) = 4.52, p<.001, whether they regarded central, t(32) = 3.46, p<.002, or peripheral information, t(32) = 3.57, p<.001. Agreeing to suggestive questions predominated, t(32) = 3.67, p<.001, whether they regarded central, t(32) = 2.79, p<.009, or peripheral information, t(32) = 2.80, p<.009.
|M (SD, %)||M (SD, %)||M (SD, %)||M (SD, %)||M (SD, %)|
|Central||19.60 (13.62, 32%)||12.30 (8.82, 20%)||2.84 (2.84, 5%)||1.42 (2.06, 2%)||38.63 (18.32, 59%)|
|Peripheral||13.96 (9.71, 23%)||8.24 (8.22, 13%)||2.24 (2.68, 4%)||0.90 (1.80, 1%)||22.87 (15.58, 41%)|
|Total||33.54 (17.32, 55%)||20.54 (13.88, 33%)||5.09 (4.34, 8%)||2.33 (3.46, 4%)||61.51 (29.22, 100%)|
So far, laboratory studies have shown that children and youths with ID, including those with intellectual and communicative disabilities, can provide important and accurate information about their experiences if interviewed appropriately (Lamb et al. 2008). Knowledge about how these children and youths are able to report their experiences in real forensic interviews is, however, limited.
In this study we therefore wanted to contribute with further knowledge about the quality of details elicited from children and youths with IDs in real life interviews. We expected that the informants could be asked many contaminating questions and that invitations encouraging free recall information were few (Cederborg and Lamb 2008a; Cederborg et al. 2009). Consequently, the quality and amount of their information had to be understood in relation to the quality of question asked.
As expected, we found that few invitations were asked but that these free recall questions did not elicit a larger amount of central and peripheral information than other question types. This is inconsistent with findings in studies of typically developing children (Lamb et al. 2003; Orbach and Lamb 2000; Sternberg et al. 2001), but is consistent with the finding from laboratory studies that children and youths with IDs report less information to open questions compared to chronological age matched peers (Henry and Gudjonsson 2003). This finding must be interpreted with caution, however. Very few invitations were used in the interviews so the children and youths did not have much ‘experience’ or ‘practice’ at providing longer answers. Instead, most of the questions asked only required short answers and in this context it is perhaps not surprising that the open-ended invitations were not as effective as we would have predicted. On the other hand, this finding also indicates that children and youths’ with IDs may have difficulties in developing their details from open-ended invitations. Instead, they may need to be asked open directive questions to elicit details of abuse experiences.
From the perspective of possible influences on details elicited, it is worrying that option-posing and suggestive questions elicited almost half of the details given. The police officers’ extensive use of option-posing questions may have hindered these witnesses in reporting their abuse experiences in detail. Moreover, the details elicited from these very focused questions may be limited and influenced by the police officers because the interviewees agreed to the options and suggestions provided by the interviewers most of the time. From the data available for this study it is not possible to determine why the participants gave affirmative answers to options and suggestions. It could be that the police officers hinted at the right details but it is also possible that they did not trust their own memory and agreed because they did not want to contradict information provided by authority figures’ opinions. Whatever the reason for agreeing to options and suggestions, the content of the elicited details from these risky questions may be called in question by legal professionals later in the investigation because of the way in which they were elicited.
It was encouraging that directive questions elicited a similar number of details compared to option-posing prompts. These types of questions were also frequently used by the police officers. Directive prompts encourage the respondents to reveal more specific details about events they have previously mentioned. These prompts can also help elicit a great deal of information that is more accurate than information elicited using option-posing and suggestive prompts (Lamb et al. 1996). However, such questions nonetheless provide interviewees with limited opportunities to provide information (Henry and Gudjonsson 2003), and they can also elicit inaccurate information (Dent 1986; Perlman et al. 1994; Kebbell and Hatton 1999; Henry and Gudjonsson 2003; Kebbell et al. 2004). Compared to invitations, directive questions may have limited these eyewitnesses’ opportunities to report important information, but the details elicited from these open prompts may be more accurate than those elicited from focused prompts (option-posing and suggestive).
Children and youths with ID may have varying abilities to remember and communicate their experiences. They may have poorer memory and be at higher risk of suggestibility compared to their TD peers (Gudjonsson and Henry 2003). Even children and youths with impaired communicative ability, poor memory capacity, and impaired ability to cope with uncertainty or understand the purpose of the interview are nonetheless able to answer open questions and provide information about their experiences, especially when directive questions were asked (Cederborg and Lamb 2008a). The present study has strengthened this understanding in terms of their abilities and capacities to provide central and peripheral information above all in response to directive questions.
Irrespective of the fact that the children and youths may have had individual differences in reporting capacity as well as varied motivation to report their experiences, these findings mean that police officers are not using enough invitations before using any other question type. They need to provide children and youths with IDs greater opportunity to report details without the influence of suggestive and option posing questions. This study also indicates that directive questions, if asked in an open manner, may facilitate the elicitation of uninfluenced central and peripheral information about abuse experiences. Uninfluenced central details about the actual abuse event as well as peripheral details surrounding the abuse are of great importance when courts are to assess these children and youths’ experiences. As long as police officers continue to rely on option posing prompts, there is a risk that the courts, without distinction, confirm their preconceptions of these victims as being incompetent in providing detailed reports. Disproportional use of option-posing and suggestive prompts affects the perceived credibility of children and youths with intellectual disabilities (Cederborg and Lamb 2006; Tubb et al. 1999).
This study is limited by the fact that the sample was retrospective and ‘diagnosis’ of intellectual disability was made from examination of the case notes as opposed to a clinical examination. Moreover, there are a wide range of disabilities that were included in the sample and a wide age range among the alleged victims. These research conditions are not ideal compared to those of controlled experiments with children with intellectual disabilities. On the other hand, what goes on in real life interviews where police officers need to interview children and youths about abuse experiences is not always comparable to findings from laboratory studies. More research is clearly needed, on both experimental and real life interviews, to develop the knowledge we have about this very vulnerable group.
This study highlights, however, that police officers need to recognise that children and youths with IDs do not necessarily need their options or suggestions of central and peripheral details concerning the investigated crime event. In addition, these children and youths can be vulnerable to implicit or explicit suggestions built into the questions asked. This may be due to their restricted memory capacities, but they may also have a lack of self-confidence, and thereby, may feel more obliged to agree to or follow what is suggested to them.
We further found that children and youths with intellectual disabilities may be able to report qualitatively strong and important information about their abuse experiences without the ‘help’ from possible contaminating question types. The practical implications are that police officers who conduct real forensic interviews need to develop their understanding of how to interview these children and youths. They need to provide them with more opportunities to report information in response to open-ended invitations. If these children and youths are interviewed in line with recommended best practices (e.g. Lamb et al. 2008) they would have the opportunity to report information as accurately as possible. If they are not developing their responses from open-ended invitations this study indicates that they should be asked open directive questions, which may facilitate the elicitation of both central and peripheral information. The latter finding should be the focus of future research efforts because we need to know more about how these children and youths can be offered opportunities to report uninfluenced details about abuse experiences.