Achieving Best Evidence: A Comparison of Three Interview Strategies for Investigative
Interviews in a Forensic Sample with Mild Learning Disabilities
Calderstones NHS Trust
Calderstones NHS Trust
Recent legislation and guidance with respect to vulnerable witnesses has very far
reaching implications for people with a learning disability who become potential
witnesses in criminal proceedings. The foundations of these changes were laid 17
years ago when attempts were made to combine aspects of child protection and
criminal investigations involving child abuse. Between the first draft of "working
together" and the present legislation there have been numerous significant milestones.
Perhaps one of the most significant was the Pigot report published in 1989 (Pigot
1989). This recommended legislative change and the drafting of guidance with
respect to conducting videotaped interviews with child witnesses in anticipation of
these videotapes being played in criminal proceedings.
The Pigot report recommended that children should be able to make a recorded initial statement which
would constitute a their 'evidence in chief' and also a subsequent video recorded interview in which
aspects of their statement were questioned. This would constitute the cross-examination in court. In
fact only the first part became law and the guidance was published in 1992 in the form of V a
memorandum of good practice.
This partial implementation of the Pigot report enabled professionals to appreciate the strength and
potential of videoed evidence, but also problems encountered. One such problem was the fact that
cross-examination continued to occur 'live' in court. This meant that the video was played to the court
and then the child was asked questions on the content of the video. These questions were often, leading
confusing and distressing. Prosecution lawyers, reluctant to have a child suddenly plunged into live
cross-examination, adopted the practice of electing to have a live evidence- in chief to 'warm them up'
before cross-examination. This generated the somewhat bizarre situation in which child witnesses
endured both live cross-examination and examination in chief (as any other witness) and their original
statement was presented to the court in the form of a video. Contrary to the hopes and expectations of
professionals working with children the reforms did little to reduce inappropriate pressure and distress
on young witnesses. It was also patently clear were that the careful conduct of memorandum
interviewers was often in stark contrast to the largely unregulated and pressurised questioning that took
place in court.
The second shortcoming that became evident was that the benefits of memorandum type interviews
were not extended to older children and adults with a learning disability. Learning disabled people are
much more likely to be victims of offences against the person, but did not enjoy any special measures
to enable them to give evidence against their alleged abusers. This irony was not lost on many
professionals working with people with learning disabilities. This included the police who, faced with a
learning disabled witness were only too aware of facilities and procedures in existence for children that
could help them elicit evidence of a crime, but they were not permitted to use them.
These two shortcomings, among others (see Sanders et al, 1996) eventually lead to a committee report
on the evidence of vulnerable witnesses in 1998 (Home Office, 1998), which made over 70
recommendations for changes in provision and procedure, from the start of an investigation to after a
trial. In the following year, a youth justice and Criminal evidence Bill was introduced, which
extended the existing provisions available to children and applied them to some adults. It also enabled a
full implementation of the original Pigot proposals, namely that both evidence in chief and cross-
examination could be conducted during videoed interviews, thus drastically limiting the need for a
witness to endure the ordeal of giving live evidence in court.
In 2002 substantial extension and revision of the memorandum of good practice was published entitled
'Achieving Best Evidence'. This dramatically changes a number of aspects of the existing procedures,
however the basic gist of the new provisions is summarised:
Examination in chief – The procedure in the trial where, normally, the lawyer representing
the side who has called the witness takes that person through his or her evidence........ The
Youth Justice and Criminal Evidence Act 1999 allows a video recording of an interview with
an eligible witness to be played as the witness‟s evidence in chief. When such a recording is
admitted, the witness is not normally examined in chief by the lawyer at the trial. Depending
on the matters raised in cross-examination, the party who called the witness in the first place
may choose to conduct a further examination in chief, or re-examination, as it is called. Thus,
for example, where the prosecution calls a woman to give evidence that she has been raped by
two men, she will give evidence in chief on behalf of the prosecution, and will be open to
cross-examination on behalf of both defendants, with the prosecution having the option to re-
examine. Where cross-examination is pre-recorded ...... re-examination will take place at the
Achieving Best Evidence, Appendix A p131
Thus both examination in chief and cross-examination can be videoed. The adults benefiting from the
new provisions ie. „eligible‟ witnesses are described as "vulnerable or intimidated witnesses" , and
explicitly includes people with learning disability. The definition of learning disability has been based
on two of the three strands required to categorise a person as „Mentally Impaired‟ within the Mental
Health Act, that is, the presence of „significant impairment of intelligence and social functioning‟.
There is at present a considerable commitment to training professionals in issues relating to identifying
witnesses who may have emerging disability and Achieving Best Evidence itself includes the
These are indications only and buy themselves do not necessarily indicate that the witness has
a learning disability:
A slow and/or confused response to questions
Difficulty in understanding simple questions
Difficulties/inability with reading and writing
And unclear concept of time and place
Difficulty in remembering personal details or events"
Whilst identifying witnesses who may have a learning disability is something of a challenge, it is as
nothing compared with the challenging face is trained interviewers who wished to formally interview
people with a learning disability as potential witnesses in criminal proceedings. Although development
of cognition and the motion vary dramatically between children, there are at least useful outlines of
child development that enable an a priori a judgement be made about the likely level of ability to be
evidenced by a prospective interviewee. Not only is there an extensive literature on development will
psychology, many professionals have practical experience of children in their own family and therefore
have established internal models of what to expect of a 10 year-old child compared with say a four
year-old child. The same is certainly not true with respect to people with learning disability, and thick it
is extremely unlikely that those who are responsible for the conduct of interviews have the necessary
skills and experience to accommodate the diversity of ability and disability they are likely to encounter.
A number of studies and papers have identified some of the characteristics evidenced by interviewees
who have a learning disability. These include impaired receptive communication; impaired expressive
communication; higher interrogative suggestibility and/or compliance; greater vulnerability to
sanctions against communication imposed on the witness by abusers (threats, promises and bribery
etc.); difficulties perspective taking; idiosyncratic perceptions of the interview process; greater
emotional lability; impaired autobiographical memory; and finally, higher rates of comorbidity for a
range of mental health problems that may themselves make interviews even more difficult (see
Glasgow, 1993 and Milne & Bull 2001).
However, identifying potential difficulties is a far cry from accommodating the strengths and needs of
a particular individual in an appropriate investigative interview. In recognition of the potential need for
more sophisticated interview techniques, Achieving Best Evidence states:
"there are a number of specialised interview techniques which have been developed for
interviewing children and these may be acceptable to the courts as an alternative to the Mecca
method recommended in this guidance, provided evidential considerations are borne in
"in a systematic approach to gathering evidence or SAGE interview the child is encouraged
over a number of separate sessions did talk about significant persons and places in the child's
life and his or her attitude towards them. Systematic comparison of the child's responses
enables the trained interviewer to identify areas of particular concern which can then be
explored more thoroughly using a open-ended questions (Wilson and Powell, 2001)"
SAGE was originally developed for children with special needs (Roberts & Glasgow 1993). It has
since been incorporated into a computer assisted interview package called MacInterview. This
development project was funded by the Department of Health for approximately 10 years and this
included extensive testing on both young and learning disabled children (Calam et al 2000). It has also
been piloted on some learning disabled adults, and has been used as an investigative tool on behalf of
the police during inquiries relating to serious crimes.
MacInterview is divided into modules which can be accessed in different orders depending on the
needs of the particular case. The modules are:
This allows the interviewee to choose and name a figure to represent themselves
This introduces to the interviewee and emotions palate which later enables them to
communicate their feelings with respect to a range of experiences and significant others
Emotions and scenes
This module enable the interviewer to practise the use of the emotions palate and to begin to
broach with the interview we salient emotional experiences
This enables the interviewee to select drawings representative of significant places in their
lives, for example where they live, college, places they visit etc.
Here the interviewee can populate care settings with significant figures about which the
interviewee can subsequently express their feelings and thoughts.
This enables interviewees to describe physical experiences involving pain or other sensations,
locate them on a body plan, identify accompanying emotions and indicates what caused them.
Developing a research programme paradigm to evaluate the effectiveness of such a complex and
sophisticated tool is a considerable challenge. One major problem is that victims of offences have
typically experienced physical and/or emotional trauma. It is the cause of this trauma that is the focus
of questioning, and interviewers are generally 'blind' to these. It is of course unethical to inflict physical
or emotional trauma on participants in research, and it is therefore necessary to seek either
experimental or naturalistic analogues of such trauma. Ideally one should be able to find out post hoc
and details of the cause of the trauma and thus evaluate the accuracy and completeness of the account
emerging at interview.
Calderstones hospital offers a forensic and challenging behaviour service to approximately 200
individuals. As is often the case in such services 'critical incidents' are not uncommon and in some of
which patients suffer physical harm which arises as a result of an accident, the deliberate act of another
patient care or deliberate self-harm.
The study reported here used standard critical incident reporting systems to identify individuals who
had suffered an injury of some significance.
It was decided to include a injuries reporting reported to have arisen as a result of both class „C‟ and
class „D‟ incidents. These are described as follows:
“These are incidents that seriously affect or have the potential to seriously affect the health or
the psychological well-being of the individual involved. They include, but are not limited to,
errors of medication,sexual impropriety, sexual, racial, or gender harassment, accidental
injuries, assaults and acts of deliberate self harm”
Class „D‟ =
“Incidents which result in no injury or minor injury or in very minor injury and do not involve
any blameworthiness on the part of the staff. They include but are not limited to, minor
accidental injury, fights between clients without weapons, and deliberate self harm”
Incidents in either category which did not lead to actual physical injury were not included in the study.
The interviewer was blind to the nature and cause of injuries and had received only basic training in
three interview procedures.
The phased approach
This is the procedure that was published in the original memorandum of good practice and was based
upon the work of Professor Yule and the 'step wise' approach to forensic interviewing. Space prohibits
a detailed account of this approach but the main phases are:
1 rapport building
2 free narrative
The computer assisted investigative interview described above. The relevant MacInterview module,
„Somatic Experiences‟ offers a graphic representations of body layout, pain type and severity, and
emotional response. Figure 1 illustrates a completed 'pain episode' as generated and described by an
interviewee. It shows two pain representations (one large and one smaller) placed on a body layout.
The emotional salience is represented by a face from the palette at the top of the picture. The verbal
description of the episode becomes its label, appearing below the image.
Figure 1: The product of the MacInterview Somatic Experiences Module
This is not an established interview strategy, but was a condition devised in order to investigate
whether any differences between the phased approach and MacInterview might be attributable to the
use of a computer per se. There are numerous studies which indicate greater levels of disclosure using
computer-based clinical assessment, and it may be that the same is the case in forensic investigative
interviewing. In his condition the interviewer used the same techniques that are implemented in
MacInterview but under took them using a paper, line drawings and a pencil.
Forty-five individual cases were identified, of which 24 were excluded from the study for the reasons
described in figure 2.
Of the remaining 21, 13 were male and 8 were female. The mean IQ was 66.7 (S D 52 - 75). The mean
age was 32 (S D 8.7) and the mean delay in days between the incident and interview was 15 (SD 4.7).
Figure 2: Attrition of identified cases of physical harm in class C or D incidents
45 individual cases
6 unsettled or interview
4 "Too aggressive"
4 Unavailability of
2 Staff reported no 2 cases police
1 Interview tape 1 Too long a time elapsed
inadvertently erased since incident
1 deemed unable to
21 interviews analysed
Three main quantitative measures were taken (with a more detailed qualitative discourse analysis under
1. interview length
This was simply the total duration of the interview in minutes. Each interview was continued
for as long as the interviewee was prepared to, and as long as further information was
2 Interview facet score
Records and narratives can often used fully be subdivided according to 'facets'. These
represent domains of information that combine to fully represent an event or circumstance.
The facets identified in this study were as follows:
The interviewer scored one point for each facet which emerged at interview and was
subsequently determined to be correct by checking with independent records.
3. Positive comments regarding the interview (Post Interview)
After all interviews were completed the interviewer invited the interviewee to give feedback
on the process. Each positive utterance was scored one point and contributed to an overall
Given the small number of participants, and the fact that the data were not normally distributed, the
non-parametric equivalent of one-way analysis of variance, i.e. Kruskall Wallace, was used to analyse
the data. Analysis indicated that there was a significant difference between the interview strategies on
all three measures (Figures 3, 4 & 5 below). With respect to the interview facets score (figure 4) the
significant difference was not between the MacInterview and the Phased Approach, but between
MacInterview and „Manual Macinterview‟.
Figure 3: Critical incidents by cause of injury and severity of incident (‘C’, ‘D’ or Unclassified
C D U/C To ta ls
Se lf In jury 4 5 0 9
Accident 0 0 5 5
Ph ysical assau lt 1 4 2 7
To ta ls 5 9 7 21
Figure 4: Mean duration of interviews
p = .002
MacIntervie Verbal Phased
Figure 5: Facet scores of interviews
Interview Method Facet Scores
MacIntervie Verbal Phased
Figure 6: Comparison of positive comments made by interviewees after interview
Interview Method Positive Comments
MacIntervie Verbal Phased
It would be premature to conclude that MacInterview can offer a structured supported interview
procedure in all cases where the witness has a learning disability. However, the data from this study are
very promising. They show that a relatively inexperienced interviewee can elicit a very considerable
amount of accurate information using the procedure. Further, interviewees are highly positive about the
interview process when the computer is used. In many respects this is consistent with research on
clinical use of computers in which there is strong evidence of a high degree of acceptability and greater
disclosure of sensitive information when computerised procedures are used.
Interpreting the fact that the Macinterview procedure lasted longest is somewhat difficult to interpret.
A very simplistic view might be that the short interview is preferable. On the other hand a longer
interview which is also a more positive experience may generate the best quality evidence. In this study
it was certainly associated with the highest facet scores.
Although the difference between the acceptability of a phased approach and the manual MacInterview
approach were not significant, it is interesting and that on all three measures the manual MacInterview
scored lowest. One possibility is that using the computer engages interviewees in a unique way and
adds a 'special ingredient' to the interview process. It is also possible that the task of manually drawing
figures and pain representations adds a „task overhead‟ to the interview, which impairs the process.
That is, rather than simply clicking an image to indicate feelings or pain location, a greater cognitive
burden was imposed on interviewees who were required to generate the same information in a more
complex way. This might have little significance for many interviewees, but for those with a disability
it might be of crucial significance.
MacInterview continues to be researched and refined, and a new cross platform version („Zara‟) will be
available at the end of 2003. It has already been used to generate evidence in civil proceedings
involving both learning disabled and non-learning disabled children. It has also been used to assist the
investigation of two murder enquiries, again involving child witnesses. The data emerging from this
study strongly suggests that it has a potential to assist with the very considerable forensic and
psychological challenges associated with conducting investigative interviews with adults who have a
Calam, RM, Cox, AD, Glasgow, DV, Jimmieson, P and Groth Larsen, S (2000) Assessment
and therapy with children: Can computers help? Child Clinical Psychology and Psychiatry,
Glasgow, D (1993) Factors associated with learning difficulties demanding special care during sexual
abuse investigations and interviews Newsletter of National Association for the Prevention of Sexual
Abuse of Adults & Children with Learning Disabilities
Home Office, Department of Health, Crown Prosecution Service, (2001) Achieving Best Evidence in
Criminal Proceedings: Guidance for Vulnerable or Intimidated Witnesses, including Children, London
Home Office (1998) 'Speaking up for Justice London HMSO
Home Office, Department of Health. (1992) Memorandum of Good Practice on Video Recorded
Interviews With Child Witnesses for Criminal Proceedings. London: HMSO
Department of Health, Home Office Department for Education and Employment (1999) Working
Together to Safeguard Children, London HMSO
Milne,R. and Bull, R. (2001). Interviewing witnesses with learning disbalities for legal purposes: A
review. British Journal of Learning Disabilities, 29, 93-97
Pigot et al. (1989). Judge Thomas Pigot QC, Report of the Advisory Group on video-recorded
evidence. London: HMSO
Roberts H & Glasgow, D (1993) SAGE: A Systematic Approach to Gathering Evidence from
Children Division of Criminological & Legal Psychology Occasional Papers No. 22
Sanders A, Creaton J, Bird S and Weber L (1996) Witnesses with learning disabilities Home Office
Research & Statistics Directorate Research Findings No. 44
The Use of the PCL-R in Forensic Populations with Learning Disability
Rampton Hospital, Nottinghamshire Healthcare NHS Trust
Although the PCL- R is widely used for assessing forensic populations in general, there has
been no published research relating to the reliability and validity of such assessments with
offenders with a learning disability. The problems with applying the PCL-R with this
population are discussed. The results of a pilot study analysing data collected for clinical
purposes on a high security LD admission ward are described, and recommendations for
further development and research are made.
In recent years the Psychopathy Checklist-Revised (PCL-R, Hare 1991) has been increasingly
employed in forensic settings. Its establishment as a robust predictor of recidivism, and
particularly violent recidivism (eg Hemphill et al 1998), has led to its inclusion in several
major risk assessment tools (eg HCR-20 (Webster et al 1999); VRAG (Quinsey et al 1998)).
There is therefore reason to believe that the PCL-R is a potentially useful tool in the
assessment of offenders with learning disabilities (ie those with low intellectual functioning
and significant impairment in adaptive functioning). However, although several of the large
sample studies of PCL-R with prisoners and forensic psychiatric patients will have included
offenders with borderline to mild learning disabilities, there has been no published research
relating specifically to the applicability, validity and reliability of the PCL-R with such
Use of the PCL-R with Offenders with Learning Disability
Clinical experience with patients with borderline, mild and moderate learning disability (LD)
would suggest that scoring PCL-R assessments with this population could be less
straightforward than with those patients functioning in the average intellectual range. There
are a number of reasons why this may be the case.
The first problem is that the item descriptions in the PCL-R manual relate to the way the item
presents in people of average intelligence. It is probable that some of the key features of
psychopathy (particularly those interpersonal features encompassed by Factor 1) may present
differently in people with LD. For example, the item description for Conning/Manipulative
(Item 5) states that “behaviours include criminal activities, such as collecting social assistance
under many different names, passing bad cheques and setting up phoney businesses …
making use of loopholes in the law …[and] convictions for fraud and embezzlement”. While
people with LD are undoubtedly capable of manipulating others, this manipulation is likely to
take a less broad and criminally sophisticated form than that described in the PCL-R manual.
For LD offenders there may need to be a different emphasis: for example on evidence that the
person being assessed attempts to use people around them, such as family, staff in institutions
or other, less able, patients. Similar arguments apply in relation to the item descriptions for
several other PCL-R items, for example Glibness/Superficial Charm (Item 1) and
Grandiose (Item 2), and specific examples of how these features manifest themselves in
people with LD would facilitate item scoring.
A second potential problem with scoring the PCL-R with this group is that the lives of people
with LD often restrict opportunities to demonstrate some of the behaviours relevant to the
PCL-R. They may have experienced long-term institutional care, have fewer opportunities
for sexual relationships and live-in relationships, and do not have the same employment
opportunities as more able people. These factors may affect scores on Item 9 (Parasitic
Lifestyle), Item 13 (Lack of Realistic Long-Term Goals) and Item 17 (Many Short-Term
Marital Relationships) amongst others. In cases where opportunity has been severely
restricted, it is therefore arguable whether such items should be omitted. Reduced
opportunities combined with limited abilities also narrow the range of types of offences
committed by LD offenders: fraud, robbery and firearm convictions, for example, are very
rare in this population. It is also important to note that that LD offenders are frequently not
processed through the criminal justice system at all. Offenders with LD who commit violent
acts while living in institutions, whether or not detained under the Mental Health Act, are
often not charged even when serious offences have taken place. These factors will affect
scoring on the Criminal Versatility item (Item 20) which is based largely on charges and
A third area of difficulty arises in relation to the clinical features of some people with LD
(and syndromes associated with LD) and their relevance to some PCL-R items. Autism, for
example, is a condition associated both with learning disability and with a lack of social and
emotional reciprocity. Many people with autistic spectrum disorders might score positively
on Item 7 (Shallow Affect) which describes “an individual who appears unable to experience
a normal range and depth of emotion” and Item 8 (Callous/Lack of Empathy). Some
clinicians might argue that if the characteristic can be explained by the autism diagnosis, that
this should be taken into account in the scoring. However PCL-R training is clear on the
point that there is no clinical override for any long-term condition, and if the feature (e.g. lack
of empathy) is present, then it may be relevant to the PCL-R score (and risk) regardless of it‟s
aetiology. More generally, emotional and developmental delay may affect characteristics
such as reliability and ability to take responsibility for one‟s actions, which is pertinent to
Items 15 (Irresponsibility) and 16 (Takes Responsibility). For these items it may be most
appropriate to make comparison with a non-offending LD population when considering the
evidence, and to code the item as definitely present only if the feature is outside the norm for
Finally, a problem arises with the difficulties of administering the standard PCL-R interview
with people with moderate (and even mild) intellectual disabilities. Their degree of
comprehension, concentration and recall all create potential difficulties. Some people with
greater impairment may even be impossible to interview in a structured way. In these cases it
would be necessary to score the PCL-R from file information, which could, in turn, be
incomplete because few writers will have had informative interviews with the patient.
Despite the problems outlined above, for the past four years the author has used the PCL-R as
part of a broader assessment of patients admitting to the Learning Disability Service of
Rampton Hospital, which is the National Centre for High Secure Learning Disability
Services. The second part of this paper describes the basic findings of a small-scale pilot
study which analysed this data collected for clinical purposes, and points the way for further
research in this area.
Description of The Pilot Study
The sample comprised 30 patients admitted to (or residing in) a male LD Service admission
ward between August 1998 and April 2002. 21 were admitted to the ward from outside the
hospital, and represented over 80% of all new admissions to the service during the study
period. The remaining 9 were either existing patients on the admission ward or were
transferred to the admission ward from other wards for re-assessment.
The hospital admission criteria require the patients to have a diagnosed Mental Disorder (as
described in the Mental Health Act), and to represent a “grave and immediate danger to the
public”. In addition, they need to be clinically assessed as having a learning disability. The
patients in the LD Service may have MHA classifications of Psychopathic Disorder, Mental
Illness or Mental Impairment, and some patients are dually or triply classified. At the time of
the study there were an average of 80-90 patients in total in the LD service. The sample was
comparable to the population in the service on most characteristics (including IQ), but
because of the focus on new admissions does have a shorter current length of stay.
The PCL-R was administered as part of a broader psychological assessment process by an
experienced psychologist trained in the use of the PCL-R, and with established inter-rater
reliability (with non LD patients). The scores were based on full interview and file review in
PCL-R Score N Mean Range
Total 30 22.5 7.2 9-37
Factor 1 30 8.9 3.4 4-15
Factor 2 30 11.9 3.7 3-17
Table 1: PCL-R scores in Learning Disability Service sample.
MENTAL HEALTH SERVICE* LD SERVICE
PD Classification MI Classification All Classifications
Mean PCL-R Score N = 40 N = 30 N = 30
Total (SD) 20.1 (8.1) 18.3 (8.9) 22.5 (7.2)
Factor 1 (SD) 8.2 (3.5) 6.9 (3.7) 8.9 (3.4)
Factor 2 (SD) 9.7 (4.5) 9.4 (4.9) 11.9 (3.7)
* Hughes et al 2000
Table 2: PCL-R scores for different samples of forensic psychiatric patients within Rampton
The PCL-R scores for the present Rampton Hospital Learning Disability Service sample are
presented in Table 1. The mean PCL-R total score was 22.5, with individual scores ranging
between 9 and 37.
In Table 2 these figures have been compared with a sample of patients detained in Rampton
Hospital who were not diagnosed as having a learning disability (Hughes et al 2000). The
patients in that study were admitted to the Mental Health Service and were divided into 2
groups: those with a MHA classification of Psychopathic Disorder (n = 40) and those with
one of Mental Illness (n = 30). The mean total PCL-R score and Factor 1 score in the LD
sample are slightly higher than both the Mental Health Service samples. The Factor 2 score is
more than 2 points higher than both samples.
In terms of the prevalence of “high” scorers, 5 (16.7%) of the current LD sample scored 30 or
above (the original suggested cut off for determining high levels of psychopathy (Hare
1991)), and 10 (33.3%) scored 25 or above (the current agreed British cut off). This
compares with 15% scoring over 30 (30% over 25) in the Rampton sample referred to above
(Hughes et al 2000), and 10.7% over 30 (27.7%) in the large forensic psychiatric sample
described by Hare (1991) in the PCL-R manual.
The results for this sample of the current Learning Disability Service patients at a high secure
hospital are broadly comparable with patients without a learning disability in the same
hospital. The results are also comparable to those in large forensic psychiatric samples in
North America (eg Hare 1991). However the higher Factor 2 in this sample scores may point
to higher levels of antisocial lifestyle characteristics and impulsive behaviour in learning
disabled forensic psychiatric populations as compared to those without learning disability.
The level of psychopathy appears higher than in a representative UK prison sample where the
mean total was 16.5 (Hare, Thornton, Clark and Gann 2000). However, this is unsurprising
given the admission criteria for the hospital, which include a significant level of
It was the author‟s experience that judgements regarding item scores with the LD sample
described here, were indeed more complex than for assessment of offenders with IQs in the
normal range. This became more problematic when assessing patients in the moderate LD
range (ie IQ below 60). The difficulties are due in part to the absence of clear coding
instructions and examples relevant to LD patients, development of which would seem
appropriate in view of some of the problems outlined in the introduction. Some of these
problems could lead to relatively high numbers of omitted items, and hence more pro-rated
total scores (12 of the 30 scores in the current study involved pro-rating). Moreover such
difficulties may lead to lower inter-rater reliability, and it is recommended that future studies
should employ dual ratings.
The patients with more severe disability were experienced as more difficult to interview, and
there were at least three cases encountered during the assessment period where PCL-R
assessments were not conducted for this reason. For example, one patient had a deteriorating
neurological condition and had very poor speech, comprehension and concentration. Another
had English as a second language in addition to a moderate learning disability, had very
limited comprehension, and was not able to sit still long enough to be interviewed. It may
therefore be that patients with greater impairment are under represented in this sample.
Concerns regarding the validity of the instrument for this latter group led to the author being
cautious in utilising such assessments for clinical risk assessment purposes. Further research
may well conclude that that the PCL-R is an assessment that cannot usefully be employed
with those with moderate LD.
This pilot study suggests that the PCL-R is an assessment tool that can be conducted with a
borderline to mild LD population. However, further research and development in a number
of areas would seems to be appropriate before it can be confidently described as valid and
reliable with this group of offenders:
1. Given the problems outlined in the introduction, it is possible that there would be less
agreement between raters in relation to assessment of LD populations, as compared to
other offender populations. The inter-rater reliability of the PCL-R, specifically with
LD populations, therefore needs to be established.
2. In order to facilitate inter-rater reliability, general coding guidelines relating to using
the tool with people with LD, and item descriptions incorporating examples relating
to this population have been developed (contact author for details). The existence of
characteristics relevant to the PCL-R may need to be considered in relation to what is
„normal‟ for the LD population. It could be argued that raters may need particular
experience with rating LD offenders before embarking on independent assessments
with this group.
3. Normative data on the PCL-R for offenders with LD needs to be established. In
particular, the relationship between Factor 1 and Factor 2 scores (and scores on the
new factor structure to be described in the forthcoming second edition of the PCL-R
manual) should be explored in relation to non-LD samples.
4. The predictive validity of the measure for this population also needs to be established
via reconviction/re-offending studies, before it can be legitimately employed as part
of risk assessments for LD offenders.
Hare, R.D (1991). The Hare Psychopathy Checklist – Revised. Multi Health Systems.
Hare, R.D, Clark, D. Grann, M. Thornton, D. (2000). Psychopathy and the predicitive
validity of the PCL-R: an international prespective. Behavioural Sciences and the Law,
Hemphill J.F, Hare R.D, Wong S. (1998). Psychopathy and recidivism: A review. Legal
and Criminological Psychology, 3, 141-172.
Hughes, G.V, Sinclair S., Braham, L. (2000) Psychopathy ratings in male forensic
psychiatric patients. In Forensic Psychology and the Law. European Association of
Psychology and Law. Krakov.
Quinsey, V.L., Harris, G.T, Rice M.E, Cormier C.A (1998). Violent offenders :
Appraising and Managing Risk. American Psychological Association. Washington DC.
Webster, C.D, Douglas K.S, Eaves, D. Hart, S. D (1997). HCR-20: Assessing Risk for
Violence. Simon Fraser University, BC.
A version of this paper has appeared in The British Journal of Forensic Practice Volume 5 .
Issue 1. February 2000.