ccc bishop.pdf

7/27/2019 CCC Bishop.pdf

1/14

J. Child Psychol. Psychiat. Vol. 39, No. 6, pp. 879891, 1998Cambridge University Press

' 1998 Association for Child Psychology and PsychiatryPrinted in Great Britain. All rights reserved

00219630\98 $15.00j0.00

Development of the Childrens Communication Checklist (CCC):A Method for Assessing Qualitative Aspects of Communicative

Impairment in Children

D. V. M. Bishop

MRC Cognition and Brain Sciences Unit, Cambridge, U.K.

The Childrens Communication Checklist (CCC) was developed to assess aspects ofcommunicative impairmentthat are not adequately evaluated by contemporary standardisedlanguage tests. These are predominantly pragmatic abnormalities seen in social com-munication, although other qualitative aspects of speech and language were also included.Some items covering social relationships and restricted interests were incorporated, so thatthe relationship between pragmatic difficulties and other characteristics of pervasivedevelopmental disorders could be explored. Checklist ratings were obtained for 76 childrenaged 7 to 9 years, all of whom had received special education for language impairment. In 71cases,2 raters (usually a teacher and speech-language therapist) independently completed the

checklist, making it possible to establish inter-rater reliability. From an initial pool of 93items, 70 items, grouped into 9 scales, were retained. Five of the subscales were concernedwith pragmatic aspects of communication. A composite pragmatic impairment scale formedfrom these subscales had inter-rater reliability and internal consistency of around n80. Thiscomposite discriminated between children with a school diagnosis of semantic-pragmaticdisorder and those with other types of specific language impairment (SLI). The majority ofchildren with pragmatic language impairments did not have any evidence of restrictedinterests or significant difficulties in the domains of social relationships.

Keywords: Specific language impairment, pragmatics, assessment, ratings, PDD.

Abbreviations: CCC: Childrens Communication Checklist; CLIC: Checklist forLanguage Impaired Children; IRR: inter-rater reliability; PDDNOS: Pervasive Develop-mental Disorder Not Otherwise Specified; PLI: pragmatic language impairment ; SLI:specific language impairment; SP: semantic-pragmatic.

Introduction

Specific language impairment (SLI) is diagnosed whena child has significant problems in mastering language forno known reason. Given that language is a complexfunction and this is a diagnosis largely by exclusion, it isperhaps not surprising that a fairly wide range ofproblems is encompassed by this category. Some childrenhave major comprehension deficits, others have diffi-culties predominantly with expression. Syntax and\or

phonology usually pose particular problems, but in somecases these are relatively intact, and there are peculiaritiesin the content and use of language. Although mostprofessionals concerned with SLI would agree thatchildren in this category are heterogeneous, it has proved

Requests for reprints to: D. V. M. Bishop, MRC Cognitionand Brain Sciences Unit, 15 Chaucer Road, Cambridge, CB22EF, U.K. (E-mail: dorothy.bishop!mrc-apu.cam.ac.uk).Copies of the Childrens Communication Checklist can beprovided at cost for research studies. Please contact the authorfor further details.

remarkably difficult to arrive at any consensus as to howSLI should be subclassified.

One approach to subclassification is based on clinicalobservation. Rapin and Allen (1983) and Bishop andRosenbloom (1987) independently produced clinicalaccounts distinguishing between a range of disorders,including those where the main problems are withlanguage form (syntax and phonology) and those wherethe principal difficulties are with language content anduse (semantics and pragmatics). Rapin listed the fol-

lowing features of semantic-pragmatic disorder" :Verbosity with comprehension deficits for connectedspeech. Word finding deficits and atypical word choicesfrequent. Phonology and syntax unimpaired. Inadequateconversational skills: speaking aloud to no one inparticular, poor maintenance of topic, answering besidesthe point of a question. (Rapin, 1996, p. 646). Thisdescription contrasts sharply with textbook accounts of

" Rapin (1996) refers to semantic-pragmatic deficit disorder,and Bishop and Rosenbloom (1987) to semantic-pragmaticdisorder. The latter term will be used henceforth in this paper.

879


2/14

880 D. V. M. BISHOP

typical SLI, where sentence structure is immature, speechmay be unintelligible, and grammatical morphemesomitted. It has been suggested that the pattern oflanguage difficulties seen in semantic-pragmatic disordermay reflect a distinctive neurological location of under-lying dysfunction (Shields, 1991). Questions have alsobeen raised about the relationship between semantic-pragmatic disorder and pervasive developmental disorder

(PDD). Semantic-pragmatic disorder tends to co-occurwith autistic features, although it is also seen in language-impaired children who do not meet diagnostic criteria forautism (e.g. Bishop & Rosenbloom, 1987; Rapin, 1996).Bishop (1989a) suggested that the distinction betweenautistic disorder, Aspergers syndrome, and semantic-pragmatic disorder may be more a matter of degree ratherthan a sharp divide.

Clinical descriptions of subtypes are a vital startingpoint in developing a classification, but a great deal ofwork needs to be done in validating the proposedcategorisation. If a category is to be useful, it must bedefined in a way that is reliable and valid, and, as researchin psychiatry has amply demonstrated, syndromes based

on behavioural characteristics pose major problems(Cantwell, Russell, Mattison, & Will, 1979). Multivariatestatistical analyses have been used in the past to findnatural groupings in the SLI population, but in generalthese studies do not lead to the same types of sub-classification as exist in the clinical literature (Aram &Nation, 1975; Wilson & Risucci, 1986; Wolfus, Mosko-vitch, & Kinsbourne, 1980). This is unsurprising when weconsider that the data typically entered into such analysescome from standardised language tests. Most languagetests measure complexity of expressive language form,verbal memory, or comprehension of vocabulary or ofsentences of increasing length and complexity. Many ofthe impairments that are regarded as diagnostically

important in the clinical literature are not identified onsuch measures. For instance, several of the features thattypify semantic-pragmatic disorder, such as verbosity,over-literal responding to questions, or problems inunderstanding connected discourse, would not bedetected by a conventional language test battery.

Conti-Ramsden, Crutchley, and Botting (1997) foundthat a cluster analysis of data from children with SLIyielded a coherent pattern of results when test resultswere supplemented with teachers opinions as to whetherthe child had difficulties in the areas of articulation,phonology, syntax and\or morphology, semanticsand\or pragmatics. They noted that psychometric testswere particularly poor at identifying semantic-pragmatic

problems, and concluded that there was an urgent needfor better methods of reliable assessment of this area ofcommunicative functioning. In 1984, Roth and Spekmannoted that most pragmatic assessment procedures wereinformal and lacked information about reliability andvalidity. Although some progress has been made sincethat time, the fact remains that assessment of childrensdifficulties in this domain remains considerably moreproblematic than assessment of other levels of languagefunctioning. Furthermore, as McTear and Conti-Ramsden (1992) remarked, such procedures as do existare often developed on the basis of what is known aboutnormal development, and may not be appropriate for

pinpointing the types of problem that arise for language-impaired children.

Thereare three methods thatcan be used in assessment;(1) tests that elicit the specific behaviours of interest; (2)observational methods for identifying the critical featuresin naturalistic contexts (e.g. conversational analysis); (3)ratings of behaviour by someone familiar with the child.

Standardised tests. There is a handful of standardised

assessments of pragmatic aspects of language in children;however, none is entirely satisfactory for evaluating therange of pragmatic abnormalities described in clinicalaccounts. The Test of Pragmatic Language (Phelps-Teraski & Phelps-Gunn, 1992) requires the child toformulate appropriate utterances in relation to picturedsituations. One limitation of this approach is that childrenwith semantic-pragmatic difficulties can perform muchbetter when given clear instructions in relation to aconcrete context than they do in naturalistic settings(Bishop & Adams, 1991). The Test of Pragmatic Skills(Shulman, 1985) and the protocol devised by Klecan-Aker and Swank (1988) assess the childs use of differentcommunicative functions in a standardised but more

natural setting, but they quantify pragmatic skill inrelation to developmental norms, rather than enablingone to pinpoint qualitative abnormalities in communi-cation.

Observation in naturalistic contexts. Observationalprocedures have been developed at both the macro- andmicro-level of analysis. Macro-analytic methods are thosewhere a stretch of conversation or other language isreviewed and rated globally on various dimensions ofpragmatic function. Prutting and Kirchner (1987)demonstrated that trained raters could achieve acceptablelevels of inter-rater reliability with this kind of approach.However, making a global judgement of whether prag-matic behaviours are impaired or adequate becomes

problematic when dealing with children who have fairlysubtle difficulties. If we identify just one or two instancesof a particular abnormal feature in a 10-minute stretch ofconversation, it can be difficult to know how typical thisis of the child, and how much significance to attach to thebehaviour. Micro-analytic methods, which involvecoding each utterance in a conversation (e.g. Bishop &Adams, 1989), allow one to achieve a more gradedquantification of pragmatic difficulties, but they areenormously time-consuming and require extensive train-ing in order to achieve adequate levels of inter-raterreliability. Nor do they escape the difficulties inherent ingeneralising from one brief section of conversation thatmay not be representative of the childs typical behaviour.

For example, Bishop and Adams (1989) confirmed thatover-literal responding to questions is more common inchildren with a diagnosis of semantic-pragmatic disorderthan in other language-impaired children. However, thisis nevertheless a rare phenomenon, and a large amount ofconversation needs to be coded if one is to detect it. Thisis time-consuming and one remains unclear whether, ifthe behaviour is not observed, this is because it does notoccur, or whether it does so but only rarely, so that theprobability of encountering examples in a stretch ofconversation remains fairly low.

Ratings. Ratings of a childs typical behaviourhave the disadvantage that they are particularly prone to


3/14

881CHILDRENS COMMUNICATION CHECKLIST

subjective interpretation and may be influenced bystereotyped concepts of language disorder. On the otherhand, ratings have three distinct advantages over othermethods. First, they take relatively little time. Second, ifthey are completed by someone who knows the child welland has observed behaviour over a prolonged period,they are more likely to give a representative account oftypical behaviour without being unduly influenced by

day-to-day fluctuations. Third, they allow one to assessbehaviours that are difficult to elicit in test situations andthat are rare in occurrence. This approach to assessmenthas been adopted by Dewart and Summers (1988) in theirPragmatics Profile, but their method was not suitable forthis study because it focuses on preschool children and ispredominantly geared toward assessing the normal de-velopmental progression of pragmatic skills. Further-more, it is intended to offer a framework for interviewingparents about childrens pragmatic skills rather thanacting as a formal assessment, and no information aboutreliability and validity are provided.

The aim of this study was to develop and evaluate achecklist for rating features of language impairment that

are given importance in clinical accounts but that are notreadily amenable to conventional forms of assessment.The interest was not so much in assessing children inrelation to normal development, but in obtaining a morequalitative picture of the profile of communicativedifficulties in children with SLI, including some be-haviours that may seldom be encountered in the course ofnormal development. The specific questions that wereconsidered were (1) can such behaviour be reliablyassessed; and (2) do checklist data provide any validationfor a distinct syndrome of semantic-pragmatic disorderwithin the SLI population?

The checklist that will be described here builds on aseries of studies extending over several years. The overall

approach is an iterative one: we start with hypothesesderived from clinical accounts of pragmatic impairmentsin children, and devise items to measure the relevantbehaviours. We then test these items, both in terms oftheir reliability and their validity, as indexed by how wellthey distinguish groups who are thought, on globalclinical impression, to have different types of disorder.Based on the results of such an exercise, a new set of itemsis developed, and the exercise repeated until we arrive atan instrument with acceptable reliability and validity. Aninitial version of the checklist was designed and data werecollected on 34 language-impaired children attendingspecial schools, in the course of a detailed study ofchildrens conversational behaviour (see Bishop, Chan,

Adams, Hartley, & Weir, in press, for details of this pilotchecklist). Raters who completed this checklist wereencouraged to offer comments and suggestions forimprovement, and on the basis of their feedback, the poolof items was revised to give a new Checklist for LanguageImpaired Children (CLIC), consisting of 23 items, eachwith multiple-choice format. Each choice consisted of astatement describing behaviour in a given domain. Forinstance, one item asked: How likely is the child toinitiate conversation with others?, andoffered the options(a) Seldom or never initiates conversation; (b) Tends notto initiate conversation but will do so in some settings,e.g. with younger children or family members; (c) Will

initiate conversation normally when this is appropriate ;(d) Talkative, but not excessively so; (e) Verbose; willtalk incessantly if given the opportunity.

The inter-rater reliability of the CLIC was assessed in astudy using ratings by teachers and speech-languagetherapists for 158 pupils from 3 residential schools forchildren with SLI. Some results from this initial studywere encouraging, but the reliability of several of the

items was disappointingly low, and there was a relativelylarge amount of missing data, which reflected the ratersreluctance to check any of the five options offered for anitem if they felt it did not correspond closely to the child.Some results from CLIC are presented by Bishop (inpress) but, given the low reliability of some items, it wasdecided that this should be treated as an exploratory pilotstudy, and further development was needed. It wasdecided, therefore, to proceed with a revised version,entitled the Childrens Communication Checklist (CCC),which is the focus of the present report.

It is important to emphasise that the checklist was notintended to identify children with SLI from within thegeneral population. Rather, the aim was to see whether,

within a population that is already identified as language-impaired, professionals who are familiar with the childrencan agree about the extent to which particular com-municative behaviours apply in a given case and, if theydo, whether there is any evidence for a distinct subgroupof children with difficulties predominantly affecting sem-antics and pragmatics.

Method

Sample

The children involved in this study were a subset of thosetaking part in a survey of children attending language units(i.e. special language-based classrooms) for children with SLI

(Conti-Ramsden et al., 1997). Most children selected forplacement in a language unit will have undergone a formalassessment leading to a statement of special educational needs,and will be deemed to require intensive speech-languagetherapy. Although criteria vary across the country, most unitsexclude children who fall below the average range on nonverbalability, those who have a permanent bilateral hearing loss, orthose who have major physical, psychiatric or behaviouraldifficulties. Conti-Ramsden et al. selected a random sample ofapproximately half the 7-year-olds attending language units inEngland for more detailed study, giving a cohort of 242participants (186 boys and 56 girls). As reported by Conti-Ramsden et al., these children were given a detailed individualassessment, and their teachers and speech-language therapistswere interviewed about their speech-language difficulties.

Approximately 12 months after children had been indi-vidually assessed, school staff were invited to participate in thisvalidation study of the Childrens Communication Checklist.Validation of the CCC had not been part of the original surveythat staff had agreed to participate in, and because they hadalready given up time helping with the main study, they werenot pressurised to take part. Staff from 52 language units agreedto participate, giving checklist data for 76 pupils (32% of thewhole cohort, comprising 44% of those children who were stillattending a language unit, and 17% of those who had left).Language data obtained by Conti-Ramsden et al. (1997) werecompared for those members of the cohort for whom checklistswere completed versus the remainder, and no significantdifferences were found, confirming the representativeness of thesubset of children studied here. The mean age of children at the


4/14

882 D. V. M. BISHOP

time of checklist completion was 8n25 years (range 7n55 to 9n83years).

Diagnostic Information

Diagnostic information was gathered from the school staff,including a rating of whether children were thought to havesemantic-pragmatic disorder. For each child in the study, adiagnosis sheet was completed by a member of the school staff,

as shown in Appendix 1, using any information available in thechilds records. The categories that were included were notmutually exclusive, and the respondent was encouraged tocheck as many diagnoses as seemed to apply. The purpose ofgathering this information was not to provide a gold standardagainst which to evaluate the checklist: there was no attempt toadopt uniform diagnostic criteria across the different schools inthis study and it was not possible to assess the inter-raterreliability of this diagnosis. Nevertheless, comparison betweenchecklist scores and global diagnosis should give a preliminaryindication of whether the checklist was likely to be useful.Minimally, we would expect it to be able to reveal differentqualitative patterns of impairment in children given differentdiagnostic labels.

Procedure

Checklists were completed by members of staff who hadknown the child for at least 3 months. Wherever possible (71children), two staff members independently completed thechecklist, and these cases were used to establish inter-raterreliability. In 76 % of cases, the first rater was a speech languagepathologist and the second was a teacher, but some checklistswere completed by other professionals such as classroomassistants. There were a further seven cases where only one ratercompleted a checklist, and these checklists were included whenanalysing internal consistency of checklist subscales. Themedian time that raters had known the children was 18 months(range from 5 months to 96 months).

Instructions on the front of the checklist included questions

about the identity and profession of the person completing thechecklist, and the length of time they had known the child.Raters were also asked to indicate the kind of educationalprovision the child was receiving, and whether the child hadever had a permanent hearing loss diagnosed, whether the childhad a permanent physical handicap or chronic illness, andwhether English was the main language spoken at home.Instructions for completing items in the body of the checklistwere as follows:

This checklist contains a series of statements describingaspects of childrens behaviour. For each statement, youare asked to judge whether the statement DOES NOTAPPLY, APPLIES SOMEWHAT or DEFINITELYAPPLIES. Please tick ONE box per item, choosing theresponse that, in your judgement, best describes the child

named above. Do not leave any items blank. If you areunable to answer the question, please tick the box labelledUnable to judge. PLEASE FILL IN THE CHECK-LIST ON YOUR OWN, DO NOT DISCUSS YOURANSWERS WITH ANYONE ELSE. The checklist can-not capture every childs behaviour perfectly, so do notworry if you feel that none of the response alternatives isexactly appropriate; tick the one you think comes closest,and, if necessary, add an explanatory comment.

We added the final sentence to the instructions for two reasons.First, past experience had taught us that teachers and therapistswere more willing to use a multiple-choice format if they feltthey had the opportunity to add verbal comments to theirresponses. Second, we had found in pilot work that such

comments were useful in drawing attention to possible ambi-guities in item wording, or aspects of communication that wehad overlooked. Explanatory comments were noted but did notinfluence the scoring.

Development of the Checklist

Creation of initial item pool. As noted above, checklist itemswent through a protracted course of development over several

years; the pool of items used in the Checklist for LanguageImpaired Children (CLIC) was used as the basis for developingthe Childrens Communication Checklist (CCC). The main aimwas to devise a checklist that could evaluate communicativebehaviours that were difficult to assess using conventionalpsychological tests but which had been given prominence inclinical descriptions of childrens language disorders. Theseincluded items assessing use of prosody, word-finding diffi-culties, production and comprehension of discourse, appro-priate use of language, and nonverbal communication (Rapin &Allen, 1983, 1987; Bishop & Rosenbloom, 1987). Some itemswere motivated by findings from a previous study of childrensconversation, which identified characteristics that led to a

judgement of inappropriacy (Adams & Bishop, 1989 ; Bishop &Adams, 1989). These included verbosity, failure to use context

in comprehending utterances, socially inappropriate utterances,tendency to provide too much or too little information for thelistener, and use of stereotyped language. In addition, itemsassessing intelligibility of speech and complexity of spokengrammar were included in the checklist, as these are regarded asimportant in differentiating subtypes of language impairment.Because the diagnostic relationship between autistic disorderand language impairment, especially semantic-pragmatic dis-order, has been a matter of debate, we also wanted to includeitems assessing social relationships with peers and adults, andunusual and restricted interests.Two additional areas, attentionand gross motor skills, were also included, as these are bothdomains where children with SLI are known to have difficulties,and it was of interest to see whether they would act as anexternal criterion that would differentiate between subgroups of

children selected on the basis of their communicative charac-teristics.One of the problems with the CLIC had been that it used a

five-choice item format, and raters had appeared reluctant tocheck any of the five options if they did not feel it closelymatched the child. It was decided therefore to proceed with arevised version, CCC, using a larger pool of items with a simplerresponse format, so that, in effect, each response optionpreviously offered in the five-choice format of CLIC waspresented in isolation, with the rater being asked to checkwhether it definitely applied, applied somewhat, or did notapply. The further option unable to judge was also offered,but raters were instructed to use this only if they had noopportunity to observe the behaviour in question. At this stage,some further items were dropped or added, on the basis offeedback from raters who had participated in the CLIC study.This gave a final pool of 93 items, grouped into 14 broaddomains. To provide raterswith theopportunity of commentingon strengths as well as weaknesses, and to avoid ratersdeveloping a response set, some items asked about positivecharacteristics of the child.

Preliminary evaluation of subscales in terms of internalconsistency and inter-rater reliability. Each of the 14 domainsof functioning was treated as a subscale, and a mean item scorewas computed by awarding 2 points for each item wheredefinitely applies was checked, and 1 point for appliessomewhat. Scores for the subscale were then totalled, with apositive sign given to items describing a childs strengths, and anegative sign to those describing a weakness (see Appendix 2),and the total was added to 30 to ensure that all subscale scores


5/14


were positive, for ease of computation. Scores for items thatwere omitted (1n57% of all responses) or coded unable to

judge (1n45% of all responses) were prorated provided theydid not constitute more than one fifth of the items in a subscale.(The majority of unable to judge responses occurred onnonlinguistic items rated by speech-language therapists).

Two measures of reliability were considered: internal con-sistency (coefficient , Cronbach, 1951) gives an indication ofhowfar theitems in a subscale differentiate in the same direction

as the whole scale. This was computed for both sets of ratings.Inter-rater reliability (IRR, measured in terms of Pearsonsintraclass correlation, r) indicates how well the two sets ofratings agreed. An iterative procedure was adopted for opti-mising the subscales. The total score for the subscale wascomputed (onthe basis of a prioriassignationof items accordingto item content), and the correlation of each item with the totalscale was inspected. Items were then progressively deleted (onthe basis of low intercorrelation with total scale) until internalconsistency was optimised, except that no item was dropped ifthis led to a drop in IRR. (Items that were discarded in thisexercise were subsequently considered for reassignment toanother subscale, see below.)

Discarding unsatisfactory subscales. Any subscale where or IRR fell below n61, or where the subscale consisted of less

than four items, was discarded, and items from that scale wereconsidered for reassignment. Five of the original subscales(prosody, lexical semantics, conversational engagement, at-tention, and motor skills) were discarded at this point.

Reassignment of discarded items. Discarded items were nextexamined in terms of correlations with remaining subscales.Items that correlated significantly with one of the retainedsubscales were considered for incorporation in that subscale,and were retained if this led to increased without decreasingIRR.

The final set of items, grouped into subscales, is shown inAppendix 2, together with data on the internal consistency andIRR of each subscale.

ResultsValidation: Relationship of School DiagnosticInformation to CCC Subscales

As noted earlier, the diagnostic information availablefor each child could not be regarded as a gold stan-

Table 1Mean Checklist Subscale Scores (Rater A) for Language-impaired Children Subdivided According to School Diagnosis

SubscalesPossible

rangeActualrange

SP plus(Nl8a)

SP pure(Nl 14)

SLI(Nl 37b) F p

A. Speech 1638 1736 30n1 (4n49) 30n6x

(4n65) 26n5y

(5n05) 4n34 n018B. Syntax 2432 2432 30n3 (1n58) 29n2 (2n12) 28n7 (2n24) 1n86 n165C. Inappropriate initiation 1830 2030 25n0

x(3n29) 25n9 (3n57) 28n0

y(2n37) 5n48 n007

D. Coherence 2036 2035 23n6x

(3n11) 25n5x

(2n38) 28n7y

(4n44) 7n34 n002E. Stereotyped conversation 1430 1530 21n4

x(4n87) 23n9

x(3n65) 27n5

y(2n04) 17n84 n0001

F. Use of context 1632 1932 22n3x

(4n27) 24n2x

(3n58) 28n5y

(2n90) 16n86 n0001G. Rapport 1834 1834 25n0

x(4n72) 29n1

y(3n36) 31n0

y(2n87) 11n54 n0001

H. Social relationships 1434 2034 25n3x

(4n00) 28n9 (4n19) 29n9y

(3n56) 4n25 n020I. Interests 2034 2534 28n3

x(2n07) 30n8 (2n29) 31n8

y(2n08) 7n19 n002

Pragmatic composite(subscales C to G)

86162 100158 119n6x

(12n68) 129n8x

(8n79) 143n2y

(10n39) 21n18 n0001

a Nl7 for social, Nl 6 for interests.b Nl36 for use of context, Nl33 for social and interests.Means with different subscripts are significantly different on Scheffe! test at n05 level.

dard, as no uniform diagnostic criteria had beenapplied. Nevertheless, we would anticipate that thechecklist should be able to differentiate children who wereregarded as having a semantic-pragmatic disorder fromthose who were not. In evaluating this hypothesis,children who were noted as having any permanenthearing loss (Nl 2), nonverbal IQ below 80 (Nl6),definite autistic disorder (Nl3), physical handicap (Nl

1), or bilingual home background (Nl4) were excluded.One further child for whom no diagnostic informationwas provided was also excluded. The remaining 59children were divided, on the basis of school diagnosticinformation, into three broad groups :

(1) Semantic-pragmatic, pure (SP pure): (Nl14) chil-dren who were identified as having possible ordefinite semantic-pragmatic disorder, but whowere not regarded as having autistic disorder,autistic features, or Aspergers syndrome

(2) Semantic-pragmatic, plus (SP plus): (Nl8) chil-dren who were identified as having possible ordefinite semantic-pragmatic disorder, but who

were also thought to have possible or definiteautistic features, autistic spectrum disorder, orAspergers syndrome. (Note that children whowere coded as having definite autistic disorder havebeen excluded from consideration.)

(3) Other SLI: (Nl37) children who were identifiedas having some other kind of communicationdisorder (developmental language disorder, de-velopmental verbal dyspraxia, language delay),and who were not regarded as having eithersemantic-pragmatic disorder or any kind of per-vasive developmental disorder (including autisticfeatures, autistic spectrum disorder, or Aspergerssyndrome).

Scores for these three subgroups on the nine CLICsubscales are shown in Table 1. The only measure onwhich the three groups of children did not differ was thesyntax subscale. On the speech scale, children with SLIwere significantly impaired relative to the SP pure group.


6/14

884 D. V. M. BISHOP

16

12

8

4

0

SLI

SP pure

SP plus

106-111 112-117 118-123 124-129 130-135 136-141 142-147 148-153 154-159

Mean pragmatic composite

Frequency

Figure 1. Distribution of scores on pragmatic composite in relation to school diagnostic information.

Table 2Intercorrelations Between CCC Subscales: Correlations Above the Diagonal Are for Rater A (Nl59) and ThoseBelow the Diagonal for Rater B (Nl51)

A B C D E F G H I

A. Speech n289 kn224 kn123 kn303 kn202 kn136 n005 kn027B. Syntax n501* kn450* n354* kn355* n013 n186 kn007 kn213C. Inappropriate initiation kn242 kn339 n151 737* 598* n172 n354* n547*D. Coherence n164 n530* n064 n348* n496* n529* n178 n256

E. Stereotyped conversationkn

230kn

277n480*

n314

n729*

n426*

n502*

n653*F. Use of context kn087 n030 n452* n508* n632* n618* n659* n596*

G. Rapport kn065 n044 n283 n470* n558* n505* n551* n589*H. Social relationships kn082 kn202 n451* n194 n453* n528* n513* n650*I. Interests kn234 kn322 n264 n027 n490* n322 n565* n452*

* Denotes a correlation that is significant at n01 level.

On all other scales, the tendency was for the SP plusgroup to obtain lowest scores, the SLI group to obtainhighest scores, with the SP pure group intermediate. Thedifferences between SLI and SP plus groups was sig-nificant on all these subscales. The difference between theSP pure and SLI groups reached significance on the

following scales: coherence, stereotyped conversation,and use of context. The SP pure and SP plus groupdiffered significantly on rapport.

Next, a composite scale was formed by summingsubtests C to G, all of which are concerned with pragmaticaspects of communication. This gave IRR of n80, withinternal consistency of n867 for rater A and n797 forrater B. The distribution of composite scores is shown inFig. 1. (To maximise sample size, the plot shows the meanscore from the two raters, with a single raters score beingused where only one set of ratings was available.) Themajority of children with a composite score greater than132 were categorised as other SLI, whereas the majority

of those with a composite score lower than this wereclassified as either SP pure or SP plus.

Intercorrelations between SubscalesThe sample size is too small in relation to the number

of variables to give reliable output from a multivariateprocedure such as factor analysis, but exploration of thepattern of intercorrelations can give a preliminary in-dication of how the variables interrelate. The inter-correlations between the CCC subscales are shown inTable 2, with rater A shown above the diagonal and raterB below the diagonal. In general the two sets of ratingsshowed a similar pattern of intercorrelations betweenscales, giving one confidence that these are stable, and notthe result of chance fluctuations. Overall, correlationsbetween the two indices of language structure (A: speech


7/14


Table 3Mean Language Test Scores (Centiles) and Nonverbal IQ (Matrices) for Language-impaired Children Subdivided According to School Diagnosis

SP plus(Nl 8)

SP pure(Nl 14)

SLI(Nl35a) F p

Naming vocabularyb 45n9 (26n23) 48n9 (23n94) 41n4 (28n84) 0n39 n673Grammatic closurec 27n2 (8n91) 31n5 (9n86) 28n4 (8n18) 0n82 n445Bus story informationd 29n7 (34n94) 26n6 (25n07) 14n4 (16n69) 2n54 n088

Goldman-Fristoee 66n9x

(32n36) 57n2 (38n34) 31n2y

(26n77) 6n15 n004Test for Reception

of Grammarf28n1 (22n86) 23n6 (25n94) 25n1 (22n94) 0n09 n910

Ravens matrices IQ 102n0 (11n81) 102n3 (14n39) 98n8 (11n73) 0n49 n611

a Nl 33 for Goldman-Fristoe and Grammatic Closure.b Subtest from British Ability Scales (Elliott, Murray, & Pearson, 1983).c Subtest from Illinois Test of Psycholinguistic Abilities (Paraskevopoulos & Kirk, 1969).d A test of narrative recall (Renfrew, 1991).e Articulation test (Goldman & Fristoe, 1986).f Multiple-choice grammatic comprehension test (Bishop, 1989b).Means with different subscripts are significantly different on Scheffe! test at n05 level.

and B: syntax) were positive, though the correlation fell

short of significance in the first set of ratings. Scale A(speech) had nonsignificant correlations with the re-maining scales. For both sets of ratings, there was asignificant positive correlation between scale B (syntax)and scale D (coherence). Each scale contributing to thepragmatic composite was positively intercorrelated withat least two of the other scales contributing to thecomposite. Two of the scales contributing to the prag-matic composite (E: stereotyped conversation, and G:rapport) were reliably correlated with the nonlinguisticscales H (social) and I (interests), which were alsopositively correlated with one another.

Relationship of School Diagnostic Information toLanguage Test Scores

One rationale for developing a checklist had been thatconventional assessments are insensitive to pragmaticimpairments. To evaluate this assumption, the same threesubgroups of children were compared on the formalspeech-language tests that had been administered byConti-Ramsden et al. (1997). The reduced numbers ineach group reflect missing data. Results are shown inTable 3. The only test on which groups differed sig-nificantly was the Goldman-Fristoe articulation test, onwhich there was a trend for the two SP groups toobtain higher scores than the SLI group, with the

difference reaching significance for the SP plus group. Onthe Bus story test, there was a nonsignificant trend forbetter performance by the SP groups. In a supplementaryanalysis, children with full test data were classifiedaccording to whether they scored above the 10th centileon all of the 5 speech-language measures. The numbers ofchildren doing this well were 10 out of 35 (29%) of theSLI subgroup, 8 out of 14 (57%) of the SP pure subgroup,and 2 out of 8 (25 %) of the SP plus subgroup. Althoughthese proportions do not differ significantly [#(1,57)l4n00, pln135], the trend is for fewer of the SP puresubgroup to have significant impairments on conven-tional tests.

Relationship of the CCC Pragmatic Composite to

Language Test and IQ ScoresThe CCC pragmatic composite was correlated with the

test measures shown in Table 2 for the 50 children onwhom there were complete data. The n01 level ofsignificance was adopted to allow for the inflated prob-ability of a significant finding by chance when sixcorrelations are computed. Only one correlation reachedsignificance: that between the pragmatic composite andthe Goldman-Fristoe articulation test [r(48)lkn349].Thus, as is also suggested by other analyses, children whoobtain a low score on the pragmatic composite tend tohave relatively good articulation.

Discussion

This study has shown that pragmatic aspects ofcommunicative difficulty in 7- to 9-year-old children canbe rated reliably by teachers and speech-languagetherapists. A substantial proportion of children with SLIdo have significant problems in this domain, and they arenot detected by standardised tests. The question arises asto how one should refer to those children who havesignificant pragmatic difficulties and who do not meetcriteria for autistic disorder. The term semantic-prag-matic disorder is widely used in the U.K., but isunpopular in many quarters because the label impliesthat impairments in semantics and pragmatics will alwaysgo together. An attempt was made in CCC, and in its

forerunner, CLIC, to include items assessing semantic aswell as pragmatic functioning, using items based onclinical accounts of word-finding difficulties, use of vagueterminology (e.g. thing rather than a specific noun), orparaphasic errors. However, these did not form a co-herent scale, and did not show strong correlations withthe items assessing pragmatic functioning (see Appendix2). Furthermore, as can be seen in Table 3, a test ofNaming Vocabulary did not differentiate children with adiagnosis of semantic-pragmatic disorder from otherchildren with SLI. Although there is clear evidence fromclinical descriptions and case reports that semanticimpairments can go hand in hand with pragmatic


8/14

886 D. V. M. BISHOP

difficulties (e.g. Sahle! n & Nettelbladt, 1993), the as-sociation does not appear to be a necessary one. Takentogether, these results suggest the semantic part of thesemantic-pragmatic disorder label is inappropriate,and that a term such as Pragmatic Language Impairment(PLI) is preferable.

Another result that was not in line with clinicalaccounts was the lack of any significant differences

between the school diagnostic groups on the syntax scaleof the CCC (scale B). This contrasts with clinical accountsof semantic-pragmatic disorder, which postulate thatchildren with this type of language impairment haveunusually complex syntax. The lack of difference could beinfluenced by poor sensitivity: there was a nonsignificanttrend in the predicted direction, and scale B is the shortestscale of the CCC, with a restricted range. However, thelack of significant group differences on the two languagetests that are concerned with syntax (i.e. Grammaticclosure and Test for Reception of Grammar) suggests thisis not the whole story. There are (at least) two ex-planations that can be proposed for this finding. The firstis that pragmatic difficulties may simply be more obvious

when they occur in the context of good structurallanguage skills. If a child talks in very simplifiedsentences,we may focus more on the structural limitations, anddisregard, or make allowances for, pragmatically in-appropriate behaviours. On this view, there is no realdissociation between syntax and pragmatics: rather, weare simply more aware of those cases where there is amismatch between these domains. A second possibility isthat the profile of impairment seen in PLI may vary withage. Bishop and Rosenbloom (1987) noted that manychildren with semantic-pragmatic disorder had verydelayed language in the preschool period, and may notstart speaking in sentences until the age of 5 or 6 years. Itis possible that the picture of pragmatic limitations

associated with complex and fluent expressive languagebecomes increasingly apparent as children grow intomiddle childhood, and that the children studied here werestill too young to show such a striking profile. Accordingto this view, there is an inverse relation between syntacticand pragmatic impairment, but only in older children.

The question arises as to whether children withpragmatic difficulties should be regarded as a distinctsubgroup within the SLI population. This study cannotprovide a definitive answer, but, in providing a methodfor identifying pragmatic difficulties with acceptablereliability, it lays the ground for future research on thisquestion. Cantwell and Rutter (1994) reviewed variousmethods for validating categories and suggested that the

most satisfactory approach is to find some externalcriterion that is associated with the category of interest.So if, for instance, we could demonstrate that childrenwith PLI had a different profile on nonlinguistic cognitivemeasures, this would help validate the notion thatpragmatic impairment characterises a qualitatively dis-tinct subgroup of children. We had hoped to useinformation on childrens attentional problems andmotor competence as external criteria, but the items thatwere included were not sufficiently reliable to be retained.There is, however, a suggestion in the literature that theremight be a distinctive neurological etiology for PLI:Woodhouse et al. (1996) reported that children with a

diagnosis of semantic-pragmatic disorder were signifi-cantly more likely to have unusually large heads whencompared with other children with more typical SLI. Inthis regard, they resemble children with high-functioningautism (Piven et al., 1995).

This brings us on to the question of the relationshipbetweenPLI and autistic disorder. Much of the debate onthis question has implicitly accepted the categorical

distinction between specific developmental disorders(such as SLI) and pervasive developmental disorders(such as autistic disorder), which is given prominence inthe fourth revision of the Diagnostic and StatisticalManual (DSM-IV) of the APA (American PsychiatricAssociation, 1994). Arguments have revolved around thequestion of whether semantic-pragmatic disorder is

just another name for high-functioning autism (ListerBrook & Bowler, 1992). The data reported here fromCCC lead one to question the emphasis on differentialdiagnosis between language disorder and autistic dis-order, and rather to follow the logic of Rapin and Allen(1987), who regarded language impairment and autisticdisorder as different levels of description, rather than

alternative diagnoses. None of the children in our studyhad a diagnosis of autistic disorder, although some,referred to as SP plus, were described as having possibleor definite autistic features, autistic spectrum disorder, orAspergers syndrome. However, the majority of childrenwho received a school diagnosis of definite or possiblesemantic-pragmatic disorder were not regarded as havingany autistic symptomatology, and children in this SPpure group did not differ from other children with SLI interms of ratings of social relationships or unusualinterests, although they were clearly differentiated interms of their communicative behaviour. Overall, theevidence is compatible with the view that while pragmaticdifficulties often occur as part of the autistic triad of

impairments in social interaction, social communication,and imaginative development (Wing, 1989), they may beseen in children who would not meet diagnostic criteriafor autistic disorder, in whom communication problemsare the most obvious area of difficulty. This is consistentwith evidence from a study of relatives of individualsaffected with autistic disorder (Bolton et al., 1994), whichused a standardised interview to assess functioning in thethree domains of communication, social interaction, andstereotyped behaviour. They reported that although veryfew relatives met criteria for autistic disorder, many ofthem had impairments in just one or two of thesedomains. If we combine the notion of three correlated butdissociable domains of impairment with the concep-

tualisation of PLI as a disorder that is intermediatebetween autistic disorder and SLI (cf. Bishop, 1989a), thisgives a picture of diagnostic possibilities as shown inFig. 2.

The category of Pervasive Developmental DisorderNot Otherwise Specified (PDDNOS) was introduced inDSM-IV to provide a diagnostic label for cases wherethere is a severe and pervasive impairment in thedevelopment of reciprocal social interaction, verbal andnonverbal communication skills or the developmental ofstereotyped behaviour, interests and activities, includ-ing cases where there is late age at onset, or atypical orsubthreshold symptomatology. Because this diagnosis


9/14


Figure 2. Pragmatic language impairment (PLI) in relation toSLI and autistic disorder. The full triad of impairments inlanguage, social interaction, and stereotyped behaviour is seen

in autistic disorder, but there may be dissociation between theseareas of impairment.

requires impairment in only one of the autistic triaddomains, and because it includes subthreshold symp-tomatology, it could be said to apply to a child who hadrelatively pure pragmatic difficulties. However, the prob-lem with this category is that it is very broad and couldalso include children with verydifferentsymptom profiles,e.g. one whose only problems were in the domain ofstereotyped behaviour, and with normal communication.If ones main aim in making a diagnosis is to identifychildren with similar characteristics who might benefit

from similar kinds of intervention, then PDDNOS seemsto be far too broad a category to be useful.

The checklist looks promising as a tool for bothresearch and clinical practice, but limitations of this studymust be noted. As noted in the Introduction, checklistratings are a quick and efficient way of gatheringinformation, but they run the risk of subjective bias. Wedeliberately collected information only from teachers andothers who knew the children well (for at least 3 months),because it was felt that accurate identification of prag-matic difficulties might require prolonged contact, but,although this should reduce the degree of random error,it could increase the risk of systematic bias arising if thestaff member is familiar with the childs diagnostic label

and has a stereotyped conception of the disorder. We dotherefore need studies that cross-validate checklist ratingsagainst other kinds of information, such as the micro-analysis of conversational behaviour developed byBishop et al. (in press). It would also be of interest to seewhether reliable ratings could be obtained on the basis ofa much shorter acquaintance with the child, or, con-versely, whether the checklist could be used as means forobtaining parental report of qualitative aspects of com-municative behaviour. Another limitation of the studyconcerns the sample. We restricted consideration only to7- to 9-year-olds receiving special educational provisionfor SLI, and we did not have systematic data on

psychiatric diagnoses using a standardised procedure.There would be considerable interest in using CCC withchildren who had been diagnosed according to a standarddiagnostic algorithm that takes into account early de-velopmental history, such as the Autism DiagnosticInterviewRevised (Lord, Rutter, & LeCouteur, 1994). Itwould also be informative to compare and contrastcommunicative difficulties in other clinical groups where

pragmatic impairments have been described, such asthose with Fragile X syndrome (Sudhalter, Scarborough,& Cohen, 1991), Williams syndrome (Karmiloff-Smith,Klima, Bellugi, Grant, & Baron-Cohen, 1995; Rapin,1996), hydrocephalus (Cromer, 1994 ; Rapin, 1996),Turners syndrome (Skuse et al., 1997), childhood schizo-phrenia (Caplan, 1996), and Attention Deficit Hyper-activity Disorder (Tannock, Fine, Heintz, & Schachar,1995). In future work, we plan to extend use of CCC toother ages anddiagnostic groups, andto consider whetherit is adequately reliable when used by parents.

AcknowledgementsThis study would nothave been possible

without the generous assistance of Gina Conti-Ramsden,Nicola Botting, and Alison Crutchley from the University ofManchester (supported by grant AT251 OD from the NuffieldFoundation). Robin Chapman gave valuable feedback on anearly draft of this paper. I would also like to thank ClaireSalmond for assistance with data coding, the teachers,therapists, and other professionals who made this study possibleby giving their time to complete checklists, and two anonymousreviewers for constructive criticism of an earlier version of thepaper.

References

Adams, C., & Bishop, D. V. M. (1989). Conversational charac-teristics of children with semantic-pragmatic disorder. I.

Exchange structure, turntaking, repairs and cohesion. BritishJournal of Disorders of Communication, 24, 211239.American Psychiatric Association. (1994). Diagnostic and stat-

istical manual of mental disorders (4th edn.). Washington,DC: Author.

Aram, D. M., & Nation, J. E. (1975). Patterns of languagebehaviour in children with developmental language disorders.Journal of Speech and Hearing Disorders, 18, 229241.

Bishop, D. V. M. (1989a). Autism, Aspergers syndrome andsemantic-pragmatic disorder: Where are the boundaries ?British Journal of Disorders of Communication, 24, 107121.

Bishop, D. V. M. (1989b). Test for Reception of Grammar (2ndedn.). Age and Cognitive Performance Research Centre,University of Manchester, M13 9PL : Author.

Bishop, D. V. M. (in press). Whats so special about Aspergersyndrome? The need for further exploration of the border-lands of autism. In A. Klin, F. R. Volkmar, & S. S. Sparrow(Eds.), Asperger syndrome. New York: Guilford Press.

Bishop, D. V. M., & Adams, C. (1989). Conversational charac-teristics of children with semantic-pragmatic disorder. II.What features lead to a judgement of inappropriacy? BritishJournal of Disorders of Communication, 24, 241263.

Bishop, D. V. M., & Adams, C. (1991). What do referentialcommunication tasks measure ? A study of children withspecific language impairment. Applied Psycholinguistics, 12,199215.

Bishop, D. V. M., Chan, J., Adams, C., Hartley, J., & Weir, F.(in press). Evidence of disproportionate pragmatic difficultiesin a subset of children with specific language impairment.Development and Psychopathology.


10/14

888 D. V. M. BISHOP

Bishop, D. V. M., & Rosenbloom, L. (1987). Classification ofchildhood language disorders. In W. Yule & M. Rutter(Eds.), Language development and disorders. Clinics in De-velopmental Medicine (double issue), nos. 1012 (pp. 1641)London: Mac Keith Press.

Bolton, P., MacDonald, H., Pickles, A., Rios, P., Goode, S.,Crowson, M., Bailey, A., & Rutter, M. (1994). A case-controlfamily history study of autism. Journal of Child Psychologyand Psychiatry, 35, 877900.

Cantwell, D., Russell, A., Mattison, R., & Will, L. (1979). Acomparison of DSM-II and DSM-III in the diagnosis ofchildhood psychiatric disorders. Archives of General Psy-chiatry, 36, 12081222.

Cantwell, D., & Rutter, M. (1994). Classification: Conceptualissues and substantive findings. In M. Rutter, E. Taylor, & L.Hersov (Eds.), Child and adolescent psychiatry: Modernapproaches (pp. 321). Oxford: Blackwell Scientific.

Caplan, R. (1996). Discourse deficits in childhood schizo-phrenia. In J. H. Beitchman, N. J. Cohen, M. M. Kon-stantareas, & R. Tannock (Eds.), Language, learning, andbehaviour disorders (pp. 156177). Cambridge: CambridgeUniversity Press.

Conti-Ramsden, G., Crutchley, A., & Botting, N. (1997). Theextent to which psychometric tests differentiate subgroups of

children with SLI. Journal of Speech, Language, and HearingResearch, 40, 765777.Cromer, R. (1994). A case study of dissociations between

language and cognition. In H. Tager-Flusberg (Ed.), Con-straints on language acquisition: Studies of atypical children(pp. 141153). Hove, U.K. : Lawrence Erlbaum.

Cronbach, L. J. (1951). Coefficient alpha and the internalstructure of tests. Psychometrika, 16, 297334.

Dewart, H., & Summers, S. (1988). The pragmatics profile ofearly communication skills. Windsor, U.K.: NFER-Nelson.

Elliott, C., Murray, D. J., & Pearson, L. S. (1983). BritishAbility Scales. Windsor, U.K.: NFER-Nelson.

Goldman, R., & Fristoe, M. (1986). Goldman-Fristoe Test ofArticulation. Circle Pines, MN: American Guidance Service.

Karmiloff-Smith, A., Klima, E., Bellugi, U., Grant, J., &Baron-Cohen, S. (1995). Is there a social module? Language,face processing and theory of mind in individuals withWilliams syndrome. Journal of Cognitive Neuroscience, 7,196208.

Klecan-Aker, J. S., & Swank, P. R. (1988). The use of apragmatic protocol with normal preschool children. Journalof Communication Disorders, 21, 85102.

Lister Brook, S., & Bowler, D. (1992). Autism by anothername? Semantic and pragmatic impairments in children.Journal of Autism and Developmental Disorders, 22, 6182.

Lord, C., Rutter, M., & LeCouteur, A. (1994). AutismDiagnostic InterviewRevised: A revised version of a di-agnostic interview for caregivers of individuals with possiblepervasive developmental disorders. Journal of Autism andDevelopmental Disorders, 24, 659685.

McTear, M., & Conti-Ramsden, G. (1992). Pragmatic disability

in children. London : Whurr.Paraskevopoulos, J. N., & Kirk, S. A. (1969). The development

and psychometric characteristics of the Revised Illinois Test ofPsycholinguistic Abilities. Urbana, IL: University of IllinoisPress.

Phelps-Terasaki, D., & Phelps-Gunn, T. (1992). Test of Prag-matic Language. Austin, TX: Pro-Ed.

Piven, J., Arndt, S., Bailey, J., Havercamp, S., Andreasen,N. C., & Palmer, P. (1995). An MRI study of brain size inautism. American Journal of Psychiatry, 152, 11451149.

Prutting, C., & Kirchner, D. M. (1987). A clinical appraisal ofthe pragmatic aspects of language. Journal of Speech andHearing Disorders, 51, 105119.

Rapin, I. (1996). Developmental language disorders: A clinicalupdate. Journal of Child Psychology and Psychiatry, 37,643655.

Rapin, I, & Allen, D. (1983). Developmental language dis-orders: Nosologic considerations. In U. Kirk (Ed.), Neuro-

psychology of language, reading and spelling (pp. 155184).New York: Academic Press.

Rapin, I, & Allen, D. (1987). Developmental dysphasia andautism in preschool children: Characteristics and subtypes.In Proceedings of the first international symposium on specificspeechand language disordersin children (pp. 2035).London:AFASIC.

Renfrew, C. (1991). The Bus Story : A test of continuous speech.Bicester, U.K. : Winslow Press.

Roth, F., & Spekman, N. J. (1984). Assessing the pragmaticabilities of children: Part 2. Guidelines, considerations, andspecific evaluation procedures. Journal of Speech and HearingDisorders, 49, 1217.

Sahle!

n, B., & Nettlebladt, U. (1993). Context and compre-hension: A neurolinguistic and interactional approach to theunderstanding of semantic-pragmatic disorder. EuropeanJournal of Disorders of Communication, 28, 117140.

Shields, J. (1991). Semantic-pragmatic disorder: A right hemi-sphere syndrome? British Journal of Disorders of Communi-cation, 26, 383392.

Shulman, B. B. (1985). Test of pragmatic skills. Tucson, AZ:Communication Skill Builders.

Skuse, D. H., James, R. S., Bishop, D. V. M., Coppin, B.,Dalton, P., Aamodt-Leeper, G., Bacarese-Hamilton, M.,Cresswell, C., McGurk, R., & Jacobs, P. (1997). Evidencefrom Turner syndrome of an imprinted X-linked locusaffecting cognitive function. Nature, 387, 705708.

Sudhalter, V., Scarborough, H. S., & Cohen, I. L. (1991).Syntactic delay and pragmatic deviance in the language of

fragile X males. American Journal of Medical Genetics, 38,493497.

Tannock, R., Fine, J., Heintz, T., & Schachar, R. J. (1995). Alinguistic approach detects stimulant effects in two childrenwith attention-deficit hyperactivity disorder. Journal of Childand Adolescent Psychopharmacology, 5, 177189.

Wilson, B. C., & Risucci, D. A. (1986). A model for clinical-quantitative classification generation 1: Application tolanguage disordered preschool children. Brain and Language,27, 281310.

Wing, L. (1989). The diagnosis of autism. In C. Gillberg (Ed.),Diagnosis and treatment of autism (pp. 522). New York:Plenum Press.

Wolfus, B., Moscovitch, M., & Kinsbourne, M. (1980). Sub-groups of developmental language impairment. Brain and

Language, 10, 152171.Woodhouse, W., Bailey, A., Rutter, M., Bolton, P., Baird, G.,

& LeCouteur, A. (1996). Head circumference in autism andother pervasive developmental disorders. Journal of ChildPsychology and Psychiatry, 37, 665671.

Manuscript accepted 16 December 1997


11/14


Appendix 1

Diagnostic Information

Listed below are a number of diagnostic terms. We would like to know if any of these terms have been thought to apply to the child forwhom you are completing the checklist.

The child might have been given more than one diagnosis. Please tick any of the terms below that have been used to describe this childsdifficulties:

possible definite

developmental language disorderOR

developmental dysphasiaOR

specific language impairment

delayed language

developmental dyspraxia

articulation disorderOR

phonological disorder/impairment

dyslexia

ORspecific reading retardation

semantic-pragmatic disorder

autistic disorderOR

infantile autism

autistic featuresOR

autistic spectrum disorder

Aspergers syndrome/disorder

pervasive developmental disorderOR

PDDNOS

attention deficit disorderOR

ADHD

clumsy childOR

developmental inco-ordination disorder

mental handicap/retardationOR

learning disability (not specific)

other (please specify)


12/14

890 D. V. M. BISHOP

Appendix 2

Items from the Childrens Communication Checklist(CCC)

Note: For girls, an alternative form was used with ap-propriate pronoun gender. Response options, and correspond-ing scores were: does not apply (0), applies somewhat (1),

definitely applies (2), and unable to judge. Ajsign precedingthe item number denotes that scores were positive; all otheritems were negative. * denotes items that were originallyassigned to a different subscale, but were reassigned on the basisof internal consistency analyses.

I. Items Retained in the Final Version of theChecklist

A. Speech output: Intelligibility and fluency. (A)ln863, (B)ln838, IRRln749.

j1. people can understand virtually everything he says2. people have trouble in understanding much of what he

saysj3. seldom makes any errors in producing speech sounds

4. mispronounces one or two speech sounds but is notdifficult to understand; e.g. may say th for s orw for r

5. production of speech soundsseems immature, like that ofa younger child, e.g. he says things like: tat for cat,or chimbley for chimney, or bokkle for bottle

6. he seems unable to produce several sounds; e.g. mighthave difficulty in saying k or s, so that cat andsat are both pronounced as tat

7. leaves off beginnings or ends of words, or omits entiresyllables (e.g. bella for umbrella)

8. it is much harder to understand him when he is talking insentences, rather than just producing single words

j9. speech is extremely rapid10. seems to have difficulty in constructing the whole of what

he wants to say: makes false starts, and repeats wholewords and phrases; e.g., might say can I-can I-can-canI have an-have an icecream

j11. speech is clearly articulated and fluent

B. Syntax. (A)ln739, (B)ln779, IRRln663.

12. speech is mostly two to three word phrases such as megot ball or give dolly

j13. can produce long and complicated sentences such as:When we went to the park I had a go on the swings;I saw this man standing on the corner

14. tends to leave out words and grammatical endings,producing sentences such as: I find two dog; John gothere yesterday She got a bag

15. sometimes makes errors on pronouns, e.g. saying sherather than he or vice versa

C. Inappropriate initiation. (A)ln790, (B)ln801, IRRln684.

16. talks to anyone and everyone17. talks too much18. keeps telling people things that they know already19. talks to himself20. talks repetitively about things that no-one is interested in21. asks questions although he knows the answers

D. Coherence. (A)ln860, (B)ln835, IRRln619.

22. it is sometimes hard to make sense of what he is sayingbecause it seems illogical or disconnected

j23. conversation with him can be enjoyable and interesting

j24. can give an easy-to-follow account of a past event such asa birthday party or holidayj25. can talk clearly about what he plans to do in the future

(e.g. tomorrow or next week)26. would have difficulty in explaining to a younger child

how to play a simple game such as snap27. has difficulty in telling a story, or describing what he has

done, in an orderly sequence of events28. uses terms like he or it without making it clearwhat

he is talking about29. doesnt seem to realise the need to explain what he is

talking about to someone who doesnt share his ex-periences; for instance, might talk about Johnnywithout explaining who he is

E. Stereotyped conversation.

(A)ln867, (B)ln865, IRRln681.

30*. pronounces words in an over-precise manner: accent maysound rather affected or put-on, as if child is mimickinga TV personality rather than talking like those aroundhim

31*. makes frequent use of expressions such as by the way ,actually, you know what?, as a matter of fact,well, you know or of course

32. will suddenly change the topic of conversation33. often turns the conversation to a favourite theme, rather

than following what the other person wants to talk about34. conversation with him tends to go off in unexpected

directions35. includes over-precise information in his talk, e.g. will give

the exact time or date of an event. For instance, whenasked when did you go on holiday may say 13th July1995 rather than in the summer

36. hasfavourite phrases, sentencesor longer sequences whichhe will use a great deal, sometimes in inappropriatesituations

37*. sometimes seems to say things that he does not fullyunderstand

F. Use of conversational context. (A)ln844, (B)ln737, IRRln747.

38*. tends to repeat back what others have just said39. his ability to communicate clearly seems to vary a great

deal from one situation to another40. takesin just one or two words in a sentence, and sooften

misinterprets what has been saidj41. can understand sarcasm (e.g. will be amused rather than

confused when someone says isnt it a lovely day!when it is pouring with rain)

42. tends to be over-literal, sometimes with (uninten-tionally) humorous results. For instance, a child whowas asked Do you find it hard to get up in themorning replied No. You just put one leg out of thebed and then the other and stand up. Another childwho was told watch your hands when using scissors,proceeded to stare at his fingers

43. gets into trouble because he doesnt always understandthe rules for polite behaviour, and is regarded by othersas rude or strange

44. may say things which are tactless or socially inap-propriate


13/14


45*. treatseveryonethe same way, regardless of social status:e.g. might talk to the head teacher the same way as toanother child

G. Conversational rapport. (A)ln868, (B)ln871, IRRln828.

46*. ignores conversational overtures from others (e.g. ifasked what are you making? the child just continuesworking as if nothing had happened)

47*. seldom or never starts up a conversation; does notvolunteer information about what has happened

48. doesnt seem to read facial expressions or tone of voiceadequately and may not realise when other people areupset or angry

49. poor at using facial expression or gestures to convey hisfeelings; he may look blank when angry, or smile whenanxious

j50. makes good use of gestures to get his meaning across51. seldom or never looks at the person he is talking to :

seems to actively avoid eye contact52. tends to look away from the person he is talking to :

seems inattentive or preoccupiedj53. smiles appropriately when talking to people

H. Social relationships. (A)ln799, (B)ln882, IRRln691.

j54. is popular with other childrenj55. has one or two good friends

56. tends to be babied, teased or bullied by other children57. is deliberately aggressive to other children58. may hurt or upset other children unintentionally59. a loner: neglected by other children, but not disliked60. perceived as odd by other children and actively avoided61. has difficulty making relations with others because of

anxiety62. with familiar adults, he seems inattentive, distant or

preoccupied63. overly keen to interact with adults, lacking the inhibition

that most children show with strangers

I. Interests. (A)ln840, (B)ln725, IRRln653.

64*. uses sophisticated or unusual words; e.g. if asked foranimal names might say aardvark or tapir

65*. has a large store of factual information : e.g. may knowthe names of all the capitals of the world, or the namesof many varieties of dinosaurs

66. has one or more over-riding specific interests (e.g.computers, dinosaurs), and will prefer doing activitiesinvolving this to anything else

j67. enjoys watching TV programmes intended for childrenof his age

68. seems to have no interests: prefers to do nothing

j69. prefers to do things with other children rather than onhis own

70. prefers to be with adults rather than other children

II. Items Dropped from CCC Because of PoorReliability or Internal Consistency (Nl23)

$ speech is slow and laboured

$ the beginning of words are repeated or prolonged (a kind ofstammer)$ speech is monotonous or unmelodious, rather like a robot

speaking$ speech melody is over-expressive and exaggerated; as if he is

an actor speaking a script$ has difficulty adjusting loudness of speech to a specific

context: may talk too loudly or too softly (e.g. whisperingwhen far away from someone, or talking very loudly whenclose up)

$ often pauses to grope for a word, although he knows it$ uses over-general terms such as thing , rather than a more

specific word$ confuses words of similar meaning: e.g., might say dog

for fox, or screwdriver for hammer$

confuses words of similar sound: e.g. might say telephonefor television or magician for musician$ seems unsure of the exact pronunciation of some long words,

so might, for instance say vegebable rather thanvegetable or trellistope rather than telescope

$ tends to use actions rather than words in response to aquestion, e.g. if asked: what are you making ? would showwhat he was doing rather than saying a boat

$ answers readily when asked a question (even though theanswer may be wrong)

$ no problems in starting conversation with very familiarpeople, but reluctant to talk to children or adults if he doesnot know them well

$ produces meaningless words or phrases that sound ratherlike a foreign language, and which cannot be accounted forin terms of poor articulation

$ understands more than he can say$ tends to stare at the person he is talking to$ inhibited with strangers, but will warm up as he gets to know

someone$ highly distractible; difficult to keep him engaged on one task

because his attention is grabbed by any noise or movementthat occurs

$ can concentrate very well on something that interests him$ unco-ordinated in activities such as sports, riding a bike,

dancing$ movements are graceful and well co-ordinated$ messy when eating or drinking$ handwriting is neat


14/14

ccc bishop.pdf

Documents