theeducationaluses oe the wisc-iii - don pugh interests/iq testing/the... · wisc-iii, gexxj...

8
THE PLACE OF INTELLIGENCE TESTING he notion of intelligence is readily seen in the concept of the person-environment fit . In simple terms, individuals who are well matched to their environment will be able to maximise the quality and productivity of their existence, while poorly matched individuals will engage in behaviours which produce a low return and may even threaten their lives . Individuals who Fave a full understanding of their envir ; nment are well placed to make insightful decisions about their actions . Their choices will be informed by their knowledge and their ability to process this knowledge. The product of these thinking processes will allow them to make predictions about the impact of their proposed behaviour both on themselves and on the environment . The result of competently "thinking through" their behavioural options should lead to a choice of "intelligent" hehaviours that produce the best possible outcomes for the individual . One role of education is to teach students the skills to generate Page 13 O R EFE R E ED P A PE R S O THE EDUCATIONAL USES OE THE WISC-III DR . DARYL GREAVES ` ABSTRACT In practice, the WISC-III is the foundation oja thcRnuRh assessment vja child uiith a learning difficulty . The interpretation of iu scores provides suggestions fcrr further confirmatory assessmenu and suggesu the nature of educaticmal interventicms . Grt the other hand, while there is general agreement the WISC-111 provides a sound assessment of general intelligence, there is little support jrom psychometric analyses of both the WISC-R and the WISC-III for the use of factor scores and profiles such as ACID in the interpretation of the scores . This paper supports the continued interpretationof WISC-III scores u~ provide hypotheses regarding the nature of learning difficulties experienced by children . When interpreting the results frcnn a WISC-111 or a similar assessment, it is essential tv keep the following caveat in mind, that "those who are responsible for interpreting the results of intelligence testing must be careful to distinguish between test scores m IQs on the one hand and intelligence on the other" (Wechsler, 1991, p . 3) . "intelligent" hehaviours to maximise the individual's gains from their interactions with the environment . In this context, IQ scores are small samples of cognitive ahilities which our culture asserts are the foundations of intelligent behaviour . Alongside cognitive ahilities, other ahilities and talents are also necessary for an individual to achieve in our society . These include non-intellective capacities of self-discipline, motivation, a goal focus, anxiety control, and physical health . When considered in the context of the person-environment fit, "intellectual ability is only one aspect of intelligence" (Wechsler, 1991, p . 3) . Success in the sch<x>l environment is clearly related to intellectual ability . Numerous studies over decades have shown high significant correlations between IQ scores and academic achievement (Sattler, 1992) . These data support the contention that IQ scores are predictive of the academic achievement of a student . It is also acknowledged that different patterns of intellectual ahilities are found in children . Some are more talented in language usage while others may he more able in perceptual-motor tasks . AICSTRAL1Afi IOI~RnAL Oi If.~RNING DISABILITtEC These patterns are often predictive their performances in the different aspects of the curriculum . Converse :, when attempting to explain a lack o' academic achievement IQ scores car he used as a diagnostic tool to identir a lack of knowledge or particular cognitive processing difficulties . As the WISC-I11 provides predictive data, an assessment woul~ include the student's educational assessment, the teacher's ohservatior the cultural background, and the child's general behaviour. These additional data may provide confirmation of the predictions arisir from the WISC-ltt scores or it may show the IQ scores were not g<x~d predictors of the educational assessment scores . In this second instance, other non-intellective factors such as motivation or personality are needed to explain the student's level of academic achievement . In the diagnosis of learning difficulties, the interactions of all relevant factors are considered . The remainder of this paper will explore the contribution of the score of the WISC-111 to an educational assessment of a student. An interpretation of the W1SC-IIf cognitive assessment is based on a Voi . 2 No . 2 tune IY

Upload: danglien

Post on 05-Oct-2018

217 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

THE PLACE OF INTELLIGENCETESTING

he notion of intelligence isreadily seen in the concept ofthe person-environment fit . In

simple terms, individuals who are wellmatched to their environment will beable to maximise the quality andproductivity of their existence, whilepoorly matched individuals willengage in behaviours which produce alow return and may even threatentheir lives . Individuals who Fave a fullunderstanding of their envir; nmentare well placed to make insightfuldecisions about their actions. Theirchoices will be informed by theirknowledge and their ability to processthis knowledge. The product of thesethinking processes will allow them tomake predictions about the impact oftheir proposed behaviour both onthemselves and on the environment.The result of competently "thinkingthrough" their behavioural optionsshould lead to a choice of "intelligent"hehaviours that produce the bestpossible outcomes for the individual .

One role of education is to teachstudents the skills to generate

Page 13

O R E F E R E E D P A P E R S O

THE EDUCATIONAL USES OE THEWISC-IIIDR. DARYL GREAVES

` ABSTRACTIn practice, the WISC-III is the foundation oja thcRnuRh assessment vja child uiith alearning difficulty . The interpretation of iu scores provides suggestions fcrr further

confirmatory assessmenu and suggesu the nature ofeducaticmal interventicms . Grt the otherhand, while there is general agreement the WISC-111 provides a sound assessment of generalintelligence, there is little support jrom psychometric analyses of both the WISC-R and theWISC-III for the use of factor scores and profiles such as ACID in the interpretation of the

scores . This paper supports the continued interpretationof WISC-III scores u~ providehypotheses regarding the nature of learning difficulties experienced by children .

When interpreting the results frcnn a WISC-111 or a similar assessment, it is essential tvkeep the following caveat in mind, that "those who are responsible for interpreting the resultsof intelligence testing must be careful to distinguish between test scores m IQs on the one

hand and intelligence on the other" (Wechsler, 1991, p . 3) .

"intelligent" hehaviours to maximisethe individual's gains from theirinteractions with the environment . Inthis context, IQ scores are smallsamples of cognitive ahilities whichour culture asserts are the foundationsof intelligent behaviour. Alongsidecognitive ahilities, other ahilities andtalents are also necessary for anindividual to achieve in our society.These include non-intellectivecapacities of self-discipline,motivation, a goal focus, anxietycontrol, and physical health . Whenconsidered in the context of theperson-environment fit, "intellectualability is only one aspect ofintelligence" (Wechsler, 1991, p. 3) .

Success in the sch<x>lenvironment is clearly related tointellectual ability . Numerous studiesover decades have shown highsignificant correlations between IQscores and academic achievement(Sattler, 1992) . These data supportthe contention that IQ scores arepredictive of the academicachievement of a student. It is alsoacknowledged that different patternsof intellectual ahilities are found inchildren . Some are more talented inlanguage usage while others may hemore able in perceptual-motor tasks .

AICSTRAL1Afi IOI~RnAL Oi If.~RNING DISABILITtEC

These patterns are often predictivetheir performances in the differentaspects of the curriculum . Converse : ,when attempting to explain a lack o'academic achievement IQ scores carhe used as a diagnostic tool to identira lack of knowledge or particularcognitive processing difficulties .

As the WISC-I11 providespredictive data, an assessment woul~include the student's educationalassessment, the teacher's ohservatiorthe cultural background, and thechild's general behaviour. Theseadditional data may provideconfirmation of the predictions arisirfrom the WISC-ltt scores or it mayshow the IQ scores were not g<x~dpredictors of the educationalassessment scores . In this secondinstance, other non-intellectivefactors such as motivation orpersonality are needed to explain thestudent's level of academicachievement. In the diagnosis oflearning difficulties, the interactionsof all relevant factors are considered .

The remainder of this paper willexplore the contribution of the scoreof the WISC-111 to an educationalassessment of a student. Aninterpretation of the W1SC-IIfcognitive assessment is based on a

Voi . 2 No . 2 tune IY

Page 2: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

great number of research studies .While these studies have providedguidelines, they have not produceddata to support definitiveinterpretations . The followinginformation is congruent with thegeneral consensus regarding theinterpretation of the WISC-lll scores .

INTERPRETING DIFFERENCESBETWEEN SCORESA cognitive assessment provides onepart of the data to discriminatebetween those with learningdisabilities, visual or auditorydifficulties, slow learners, and thosewho come from a non-Englishspeaking background .

The diagnosis of a learningdifficulty is based mainly on therelationship between the cognitiveassessment and the educationalassessment. A diagnosis of a slowlearner, for instance, would show amatch between the cognitiveassessment and an educationalassessment . In this case, the child mayhe a couple of years behind his or herpeers on the educational assessmentand would have an IQ score belowaverage .

In some cases however, adiscrepancy appears l,etween the IQscores and the educational scores(Lerner, 1997). A clear example ofthis can he found in the definition ofdyslexia . Dyslexia is generallyacknowledged when a child is shownto have average or above averageintellectual abilities, along with areading score that is significantlybelow that expected for a child of hisor her age . Lerner suggests that thediscrepancy between intellectualpotential and achievement must hesevere and notes that "a one-yeardiscrepancy at the second grade levelis more severe than a one-yeardiscrepancy at the eleventh-gradelevel" (p . 14) . As part of thedefinition dyslexia, the discrepancybetween the IQ score and the readingscore is not explained by cultural,educaticmal, physical, or emotionalfactors .

The logic behind the use of adiscrepancy score is that the sccxe onan intelligence test provides areas<mahly valid prediction of thechild's expected academic attainment .A discrepancy between the IQ scoreand academic attainment suggests that

Vol.2 IVn.2 June 1997

O R E F E R E E D P A P E R S 4

A diagnosis of a slowlearner, for instance,would show a matchbetween the co~r~.itiveassessment and. aneducationalassessment . In thiscase, the child maybe a couple of~'earsbehind his or herpeers on theeducationalassessment andwould have an IQscore below average .

other factors are needed to explain thedifference . The successfulidentification of this factor or factorsmay provide the basis for anintervention .

THE MECHANICS OF WISC-IIISCORESDifferent types of scores are derivedfrom the suhtests of the WISC-lll . Inthe first instance, each suhtestproduces a raw score which isconverted into a Scaled Score . scaledScores are used in all furthercalculations . A suhtest Scaled Scoreallows a comparison to be made of thechild's score with those of his or herchronologically-aged peers. A score ofone places the child in the lowest onepercent of his or her peers ; a score often places the child at the middle ofthe range, while a score of nineteenplaces the child in the top onepercent .

Scaled Scores from specifiedsuhtests are added together to createan IQ or index score (different typeswill he described below) . In turn, theIQ or index score is translated into apercentile. The 1Q score basedpercentile gives an estimate of thestanding of the child in relation to hisor her peers . The Performance ScaleIQ score of 1 Z0, for example, providesa percentile rank of 91 while a

A1lSTRALIAM JOIiRfYAL OF LF1RMIfVG DISARIIJTI~

Performance Scale IQ score of 110provides a percentile rank of 75 . Theexact percentile ranks of 91 and 75(for example) are in fact somewhatmisleading . The percentile rank isreally an estimate of the student'srelationship to his peers as the IQscore which determined the percentilerank may well contain somemeasurement error_ This possiblemeasurement error is dealt with by theuse of a confidence interval . AllWISC-111 IQ scc,res have their ownconfidence intervals .

As it is impossible to determinethe amount of error in each score, it ispostulated that the true score for thechild could well fall in the intervalbetween a score of x and y . When theconfidence interval is applied to ascore of 100 on the Verbal Scale IQ, itis concluded with 95% confidencethat the child's true score fallsbetween 95 and 105. In other words,given the probability of error arisingin measuring cognitive ability, onecan he 959f, confident that the truescore falls between 95 and 105 .

The confidence interval providesa forceful reminder that eachmeasurement of cognitive ability(each suhtest) is likely to containsome measurement error . Thesemeasurement errors can arise from avariety of sources . It may be that nnthe day of testing the child's score wasunderestimated as the child was tired,unwell, or stressed . Alternatively, thescore may have been over estimatedbecause the child had been taughtexcellent problem solving strategies orthe child had made a number of"lucky guesses" . Given thesearguments, it must he rememberedthat the apparent precision of thepercentile ranking of a child's IQscore is c,nly an estimate' .

To firm up the estimate providedby the IQ scores of the student'scognitive abilities, informationregarding the child's performance froma number of independent sources isconsidered . Parent's observations oftheir child are important along withthe teacher's assessment of the child'scognitive ability . Standardised andinformal educational assessments mayalso he helpful . Data from all of thesesources needs to he kept in mind whenthe scores from the WISC-Ill areinterpreted .

Page 14

Page 3: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

WISC-111 ASSESSMENT OFINTELLECTUAL ABILITIESThe WISC-I11 consists of thirteensubtests, each claiming to measure adistinctive facet of intellectual ability .Five of these are used to calculate theVerbal IQ Scale Score and five tocalculate the Performance 1Q ScaleScore . The Verbal IQ scale containsthe subtests of Information,Similarities, Arithmetic, Vocabulary,and Comprehension . ThePerformance IQ scale consists ofPicture Completion, Coding, PictureArrangement, Block Design, andObject Assembly . These 10 subtestsare combined to give the total IQscore . As both the Verbal Scale andthe Performance Scale are measures ofthe student's overall cognitive ability,a significant difference between themrequires interpretation . A differencetnay show that a student is able to usetheir intellect more successfully in onemode more than the other mode .Students who are hearing impaired orare from non-English speakingbackgrounds, for instance, may notshow their true ability on the VerbalScales . Other hypotheses can be testedfor explanatory factors such as,achievement motivation, hemisphericspecialisation, spatial ability, motorproblems, time pressure, academicproblems, and lonj;-tertn memoryword retrieval .

INTERPRETATIONS USINGSUBTESTSIn addition to comparisons of verbaland performance scores found in theWISC-III, gexxj use can he made ofthe individual subtests . While theresults of these subtests are onlyestimates of intellectual functioning,they can be used to generatehypotheses regarding a student'sperformance in particular learningareas . Some subtests arepsychometrically stronger than others .Information, Arithmetic, andComprehension among others aremore easily interpreted than PictureArrangement and Object Assembly(Sattler, 1992) . Interpretation of thesubtest scores must be compatible withthe overall picture presented by thestudent's results on the other subtests .

Two types of comparisons can heused to interpret individual subtests .The student's subtest Scaled Scores

Page IS

O R E F E R E E D P A P E R S 4

can be compared with the norm fur achild of his or her age (normative) . Inthis comparison, the following set ofdescriptors can be used : far belowaverage (a score of 1 to 4) ; belowaverage (5 to 7) ; average (8 to 12) ;ahove average (13 to 15) ; andexceptional strength (16-19) . Theseccmd type of comparison attempts te,identify a specific pattern of subtestscores which provide the basis fur aninterpretation . This is an examinationof intro-individual strengths andweaknesses .

An interpretation on intra-individual strengths and weaknesses ishaled nn the student's own averagescore rather than the normativeaverage which is a Scaled Score of 10 .It may he, for example, that a studenthas an average score of 8 for the verbalsubtests of Information, Similarities,Arithmetic, Vocabulary,Comprehension, and Digit Span . Thesame student may have an averagescore of 7 on the Performance subtestsof Picture Completion, Coding,Picture Arrangement, Blcxk Design,Object Assembly, Symbol Search, andMazes . In this case, scores nn theVerbal subtests with a score wellahove 8 are considered strengths . On aPerformance subtest, scores well below7 would he considered a comparativeweakness . In essence, theinterpretation is based on subtestsaxes that deviate from the student'sown average score . This method ofinterpreting the WISC-Ill can hedescribed as comparing each Verbalsubtest Scaled Score to the student'saverage Verbal Scaled Score andfollowing the same prcxess with thePerformance Scaled Scores . Thefollowing type of statement may result .'The student performed poorly but hascomparative strengths in the subtestsof . . . .

USING FACTOR SCORES TOINTERPRET THE WISC-IIILex,king for strengths or weaknessesusing patterns or factor analyticgroupings of subtests has peensomewhat contentious . Bannatyne(1974) for instance, proposed patternsbased on spatial abilities,conceptualisation, sequencing, andacquired knowledge.'The pattern ofsubtests in acquired knowledge wasInformation, Arithmetic, andVexahulary . While some statistical

AItSTRALIAN ~OIiRNAI, OF I'.IR"IING DISARIIJTIFS

support was found to support thevalidity of these patterns, replicationof the analyses on other samples didncx provide good confirmation of thepatterns . The recent psychometricwork on the four factor solution of theWISC-Ill demonstrates that theinterpretation of the two factors ofPnx:essing Speed and the Freedomfrom Distractibility are in need offurther development (Kamphaus,1993 ).

While the ~uhtests are measures c,ithe facets of intellectual ability,statistical analyses of W ISC-R andWISC-lll scores have peenundertaken in an attempt to identifythe cognitive abilities underpinningthese facets . Recent factor analyses c,fthe WISC-lll have shown that allsubtests except Mazes are helpful inidentifying underlying cognitivefacets. These analyses show a generalfactor of intelligence composed ofcognitive components called VerbalComprehension, PerceptualOrganisation, Processing Speed, and ,;weak factor called Freedom fromDistractibility . The use of the VerbalComprehension and PerceptualOrganisation factors to interpret astudent's scores provides a reliablestatement of the student's cognitiveabilities . The Processing Speed andFreedom from Distractibility arepsychometrically less robust than theother two factors and prc,vide moretentative estimates of a student'sabilities .

The Verbal Comprehension Scc,reis derived from the subtests ofInformation, Similarities, Vocabulary,and Comprehension . Children whoscore well on Verbal Comprehensionshow that they have been successfulverbal learners and that they canapply these verbal skills to understandnew situations . The total of thesubtests comprising this factor is thepest predictor of academicperformance when compared with thescores from the other three factorscores .

Perceptual Organisation consistsof the subtests of Picture Completion,Picture Arrangement, Blcxk Design,and Object Assembly . PerceptualOrganisation represents a ncmverhalskill in competently interpreting andanalysing visually perceived material .

The ability identified by Freedcxnfrom Uistractihility is difficult tc,

Vol . 2 No.2lunc 147 ;

Page 4: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

determine and needs to he interpreteDon a case by case basis. As it iscomposed of the suhtests ofArithmetic and Digit Sean, it wouldseem to describe a language basednumerical facility using auditory shcmt-term memory .

Pnxessing Speed consists of thetimed suhtests of Ceiling and SymbolSearch . It is a measure of the student'sspeed in proxessing nonverbalinformation along with concentrationand eye-hand c<xmdination .

These four factor scores providediagnostic information to assiststudents experiencing learningdifficulties . They can provideinformation on issues such as languagecompetence, attention span, theability to make fine discriminations,the ability n~ prcxess different types ofcognitive tasks, and to deal with visualproblems .

In addition, a comparison of thescores on these four factors canprovide an indication of the student'srelative strengths and weaknesses . Theidentification of a strength hasimplications for the most appropriatemcxle to use in the Delivery ofinstruction . A relatively lowPrcxessing Speed suggests that thevisual presentation of instructionalmaterial needs to he slowed down,that is, the student will need moretime feet reading eu copying from thehoarD in class . A relatively low scemecm Freedom from Distractibilitysuggests that the key elements <~f theconcepts neeD to he highlighted,rehearseD, anD presenteDdiagrammatically to cope with thechilD's IimiteD short-term memorycapacity anD his or her ability to thinkahcuo relationships .

Mc,rr specifically, the value of thePrexrssing Speed factor is that itreflects Psyehexneuem and mentalprocessing speed in solving nonverbalproblems . It is possible that a studentcewlD have an aptitude in solvingw,nverhal problems (PerceptualOrganisation same), but sage pcxmlyon the Prc,cessing Speed factor as aresult mf a lack of pcxm memorccxmdination, reflectiveness, ercc~tnpulsiveness (Kaufman, 1994). ADiscrepancy W this type is suggestive ofADIiD especially if academicPerfc,rmance is DepresseD . A majoreliscrrpaney between a PerceptualOrganisatiom sa,re anD a Prcxessing

Val . 2 Nn.2lunc 1497

O R E F E R E E D P A P E R S O

Speed score casts doubt on theaccuracy of the Perfcmmance 1Q scoreas is illustrated below.

The use of the four factors in aninterpretation of the W1SC-III can hehelpful as the following case shows . AVietnamese student, who was hcmn inVictoria, provided the followingpercentile ranks: VerbalComprehension, 5; Freedom fmmDistractibility, 79 ; PerceptualOrganisation, 77 ; and PrcxessingSpeed, 27 . As with any interpretation,such discrepancies need to he carefullyexamined .

As noted above, the factors ofVerbal Comprehension and Freedom

In the learningdisability literature, aconsensus emergedthat an ACID profilecommonly appearedfor children whoexperienced learningdisabilities and itoften appeared inchildren with anAttention DeficitHyperactivityDisorder (Cordoni,O'Donnell, Ramaniah,Kurtz, ~ Rosenshein,1981 ; Salvia,

ajar,Meenakshi, £~ Salvia,1988; Bauman, 1990).

from Distractibility contain all of theverbal suhtests . Freedom fmmDistractibility contains ArithmeticanD Digit Span anD provided thepercentile rank c 77 . This percentileis in stark contrast to the VerbalG~mprehension percentile rank of 5.It is Probable that the NESLhackgre,unD c~f the chilD explains thePcxm result on the VerbalG~mprehensicm suhtests ofInformation, Similarities, V<xahulary,anD C<,mprehension . A's both verbalfactors are gcxx3 predictors of academic

AI!STRALIAN ~OI~RNAI. OF LF,IRNING DISARILITIFS

achievement (RoiD, Prifitera, &Weiss, 1993) and both pn,viDe areasemahle Prediction mf generalintellectual ability (Blahs &Wallhrnwn, 1996), a choice has to hemade between them . Given the NESLbackground, the Freedom frc,mDistractibility Percentile rank c,f 77woulD he tentatively accepteD . TheFreedom fmm Distractibility sc<,reindicates that the student is likely tcihave quite adequate intellectualfunctioning . This.hyPothesis wc,uldneed further testing with arecommendation for an intelligencetest in Vietnamese and a Diagncuticlanguage test .

Both factors c~f PerceptualOrganisation and Prcxessing SpeeDcontain all of the performancesuhtests . Both provide predicticms ofacademic achievement (Raid et al .,1993) and of general intellectualability (Blahs & Wallhrown, 1996).These two nonverbal predicte,rs c,fintellectual ability are ncm as reliableas those based on the two Verbalfactors. The Vietnamese student'sPercentile rank of 77 em PerceptualOrganisation is markedly differentfrom his rank of 27 on PrcxessingSpeed (Ccxling anD Symbol Search) .During the Ccxling and Symhc,lSearch suhtests, the student wemkedvery slowly and double checkeD allanswers to ensure a perfect resP<mse . Itcc~ulD he argued that his taskapproached on these two suhtests DiDnot give a reliable estimate of his trueability which is better reflecteD in thePerceptual Organisation rank of 77 .

When taken together theFreedom fmm Distractibility anD thePerceptual Organisation factorssuggest that the Vietnamese student isintellectually quite capable, but thiswuulD not he evident in academictasks requiring the use cif the Englishlanguage . The value of the four factemsfeet interpretation in this case isobvious because an interpretatiombased solely un student's Full Scale IQscore of 90 (25th percentile rank)would have suggested a slew learnerwhen an examination of the facets ofPerceptual Organisatiom (77) andFreedom from Distractibility (77)suggest that he may well he a capablestudent. This hypothesis woulD neeDto he confirmeD with at least twc~further pieces c>f evidence .

Pagc 16

Page 5: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

TAB(E 1

WISC-111 results highlighting subtest profile scores for the identification of learning disabilities

PROFILE ANALYSIS OF THEWISC-111In addition to factor scores teeing usedin the interpretation of the WISC-Ill,clinical experience has also identifiedsome patterns of subtest scores whichappear to be associated with specifictypes of learning difficulties ordisabilities . The use of these patternsis known as profile analysis . In 1981,Kaufman reported earlier research thathad identified the ACIU profile . Thisprofile he described as comparativelylow scores in the suhtests ofArithmetic, Coding, Infcxmation, andDigit Span .

Further research was undertakento determine if consistent subtestprofiles such as ACIU could he used toidentify different types of disability . Inthe learning disability literature, aconsensus emerged that an ACIUprofile commonly appeared forchildren who experienced learningdisabilities and it often appeared inchildren with an Attention DeficitHyperactivity Disorder (Cordoni,O'Donnell, Ramaniah, Kurtz, &Rosenshein, 1981 ; Salvia, Gajar,Meenakshi, & Salvia, 1988 ; Bauman,1990) . The identification of studentsin such groupings is quite important asthe educational interventions andrecommendations for them aresometimes different to those withlearning difficulties (Spafford, 1989) .

Page 11

4 R E F E R E E D P A P E R S O

The main method to determinethe presence of an ACID profile hasbeen to compare the ACID subtestscores with the remaining subscales . Ifthe ACID subtest scores are equal to,or less than the other subtest scores,then an ACID profile is said to exist(Ward, Ward, Hatt, Young, &Mollner, 1995) .

Preliminary research has foundevidence of a modified ACID profilefor girls . Spafford (1989) identified anAVID profile in dyslexic girls . In thiscase, the suhtests were Arithmetic,Vcxabulary, 1_nformation, and LOigitSpan . The data repcxted in Table 1were recently obtained from a sevenyear old boy who was referred to aclinic fetr an assessment . These datawill he used to demonstrate profileanalysis and interpretation ..

An examination of the profile inTable I suggests that this child has alearning disability as the ACIUsubtest scores are equal to, or lowerthawthe other suhtests with theexception of Picture Arrangement .Thus, it can not be classified as anACIU profile even though a clinicianwould use this profile as a basis togenerate an hypothesisthat this childmay he LD . As noted above, theACIU profile was developed in the eraof the WISC-R, that is, before theintroduction of the Symbol Searchsubtest to the WISC-Ilf . On the basis

AIISIRALIAfI )UI~Rf1AL OF I1 .IRNINC DISARIl1T1F.S

of their research, Ward et al . (1995)suggest that not only should theACID profile be examined, but thatnew profile he considered . This newprofile includes Symbol Search alongwith ~cxling, Arithmetic, and DigitSpan and is called SCAU.

The SCAR profile is determinedby the following formula . The ScaledScores of Symbol Search, Coding,Arithmetic, and Digit Span he addedand then subtracted from the total ofthe Scaled Scores of PictureComl+letion, Picture Arrangement,Block Design, and Object Assembly . Ifthe difference is greater than ninethen the SCAR profile is deemed toexist . When this formula is applied tc,the data in Table 1, a difference of tenresults .

While both formulas for ACIDand SCAU appear to provideconfidence in the prediction of LU,caution needs to he exercised .Research on the ACID profile hasshown it to he statistically moreprevalent in LU samples, it is,however psychometrically weak as itonly identifies a small percentage ofLD children (Hale & Saxe, 1983;McDermott, Glutting, Jones, Watkins,& Kush, 1989) . Prifitera and Uersh(1993) found the ACIU profile morefrequently in AUHU students thanLD students, but it only identified asmall proporticm of the total of hcxh

Vol.2 tVo.21une 149

SUBTEST NAMES SS FACTOR SCORES

Name ID Score

Arithmetic 1 1 Verbal Comprehension 128 97

~oding 6 Freedom from Distractib'y 101 53Information 9J)igit Span 10 Perceptual Organisation 109 73

symbol Search 8 Processing Speed 86 18

Similariiiesf

15Vocabulary 17Comprehension 16

Name ID Score

Picture Completion 12 Verbal 124 95

Picture Arrangement 8 Performance 102 55

Block Design 14 Full Scale 1 14 82

Object Assembly 1 1

Page 6: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

groups . It is prudent to take heed ofWard et al's. (1995, p . 275)conclusion that :

at best, the ACLU and SCAUprofiles may serve as a confirmatoryvariable for a small number ofstudents. In other words, if assessmentdata indicate the presence of alearning disability and the studentdemonstrates the AC(U or SCAUprofile, then we can he moreconfident in the diagnosis. ACID andSCAU profiles should never he used asthe sole criterion for the diagnosis .

AN EDUCATIONALPRACTITIONER'S VIEW OFTHE WISC-IIIIn spite of psychometric weaknesses,the WISC-111 is a useful diagnostictool for the educational practitioner .Cognitive abilities are a majorpredictor of academic attainment . Thediagnostic value of the scores is thatthey provide working hypotheses forthe educator . A low Scaled Score onvocabulary, for instance, alerts theeducator to check for the followingissues : a possible hearing difficulty,difficulties with language skills, and/cardifficulties with concept development .Each of these hypotheses could lead tofurther assessments such as a referralto an audiological service provider, aspeech pathologist, or a speecheducator. A speech pathologist couldundertake a language assessment orthe educator could carefully analysethe student's language skills by askingthe student to follow instructions, a~

Vol.2 No. 2 June 1997

TABIE 2 : Wechsler's 1991

P R E F E R E E D P A P E R S 4

maintain conversation, or toexamining his or her written languageand comprehension responses inreading scales . Depending nn theresults of the additional assessments,interventions can be devised toaddress the particular difficulties thechild is experiencing . In summary,Scaled Scores provide information forthe selection of further assessmentinstruments . They also indicate thoseareas of skill and knowledge whichneed careful evaluation andml,nitoring in less formal or ongoingassessments.

It is the subtest Scaled Scores, notthe Verbal, Performance and FullScale I(~ scores that are most useful ingenerating diagnostic hypotheses toassist in determining the mostefficacious educational interventions .As producing a number of explanatoryhypotheses to explain the student'slearning difficulty is the main purposeof testing, it is necessary to findconfirmatory evidence from othersources to support the WISC-Iltgenerated explanation . It is prudent todiscard any hypothesis which is notconfirmed by other forms ofassessment whether they he informalor formal .

In addition to the foregoing,additional information can hegathered from the student at the timeof testing such as the child's level ofattention, type of language, mood, andstrategies in tackling the items.Sometimes, it is also possible toobserve the student': problem solving

" No factor analyses are reported in 1974 WISC-R Manual (Wechsler, 1974) .Table 2 shows the location of the WISC-R suhtests reported by Kamphaus (1993) in italics .

' Mazes omitted by Wechsler (1991) due to low loading on this factor .

AIlS 1RALIAM ~OURMAL OfIEIRNINC DISABILRI~S

strategies in discriminating keyelements, and checking thecorrectness of their answers .

In "summary, it would seem thatthe WISC-lll is not psychometricallyrobust . if practitioners are cautious inits interpretation and use it primarilyto generate hypotheses for furthertesting it is a useful educationalinstrument to use .

ADDENDUM:The Usefulness of the WISC-I11Factor StructureThis addendum is a statistically baseddiscussion regarding the WISC-Illfactor structure . This discussion iswarranted as the 13 suhtests of theWISC-I11 have been allocated toeither a Verbal or a Performance scaleon the basis of the different types ofintellectual skills they are thought tobe testing. This allocation implies thatthere is shared variance among all ofthose suhtests grouped to make theVerbal scale and shared varianceamong those suhtests grouped to makethe Performance scale . In addition,the totalling of both the Verbal andPerformance subtest scores provides anestimate of a third factor . This wascalled a general intellectual power byGalton in the 1870s and has becomeknown as g following thedevelopmental work on factor analysisby Spearman .

It will he shown below that theuse of the WISC-Ill to confirm a clearpattern of the three factors (Verbal,Performance, and R) has not beensuccessful as five factors emerge . As

four factor solution of the WISC-III with the WISC-R Three factor solution * .

Page 18

Factor 1Verbal

Comprehension(VC)

Factor 2Perceptual

Organisation(PO)

Factor 3Freedom fromDistractibility

(FDA

Factor 4Processing Speed

(PS)

Information Pichlre Complet. Arithmetic CodingInformation Picture Complet. ArithmeticSimilarities Picture Arrang't. Digit Span Symbol SearchSimilarities Picture Arrang't . Digit Span

Vocabulary Block DesignVocabulary Block Design Coding

Comprehension Object AssemblyComprehension Object Assembly

Mazes'

Page 7: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

described above, these factors havebeen labelled g , VerbalComprehension, PerceptualOrganisation, Freedom fromDistractibility, and Processing Speed .Psychometricians agree that g factorcan be found in the subtests, howeverthere is vigorous debate concerningthe value of the remaining four factors(Carrot, 1993 ; Roid et al ., 1993 ;Kamphaus, Benson, Hutchinson, &Platt, 1993 ; Kaufman, 1994) . Whilethe debate between thepsychometricians has not been settled,psychologically meaningfulinterpretations of a student's scores onthe four factors which emerged fromthe analyses of the WISC-Ill havebeen demonstrated above.

Factor analysis is a statisticalmethod that can he used to examinethe relations between tests onvariables. Accordingly, thecorrelations between the subtests inthe WISC-III can be analysed todetermine which of them belongtogether, that is, those which haveelements in common. In other words,factor analysis mathematicallyidentifies the subtests that belongtogether . If the allocation of thesubtests to either the Verbal orPerformance Scales on the basis offace validity is to he supported, afactor analysis would show the subtestsnominated as Verbal loading heavily(correlating strongly) nn a Verbalfactor while the nominatedPerformance Subtests would loadheavily on a Performance factor . Froma theoretical perspective, subtestsloading heavily on the Verbal factorwould be expected to load onlyslightly nn the Performance factor andvice versa .

The first two factors fenlnd in hexhthe WISC-R {the earlier version <~fthe test) and the WISC-Ill aredominated by the same subtests andsupport Verbal and Performanceabilities (Table 2) . The VerbalComprehension factor has the subtestsof Information, Similarities,Vocabulary, and Comprehension ; andthe Perceptual Organisation factor hasthe subtests of Picture Completion,Picture Arrangement, Block Design,and Object Assembly . Accordingly,interpretation of the two factors isrelatively straight forward .

The language dependent subtestsArithmetic and Digit Span havestrayed from their Verbal home andloaded on a factor labelled Freedom

Page 19

O R E F E R E E D P A P E R S 4

. . . factor analysismathematically

identifies the subteststhat belong together .If the allocation of thesubtests to either the

Verbal orPerformance Scaleson the basis offacevalidity is to be

sup~orted, a factoranalysis would show

the subtestsnominated as Verbal

loading heavily(correlacting strongly)on a Verbal factor

while the nominatedPerformance Subtestswould load heavily ona Performance factor.

from Distractihility . Two performancetests, Ccxling and Symbol Search,have also cohered to form anotherfactor called Processing Speed . Theremaining test of Mazes loadsminimally on all four factors and istherehxe ignored .

The Verbal Comprehension andPerceptual Organisation factors aresubstantive as together they accountfor 459f, of the variance of the subtests(Wechsler, 1991) . The two additionalfactors are far more controversial asthey only account for a smallpercentage of the variance, that is,Freedom from Distractihility, 2-3% ;Processing Speed, 4-596 . Similarly,Roid, et al ., (1993) using anotherrepresentative sample (n = 1,118)found Freedom from Distractihilityaccounted for 6%,and ProcessingSpeed for 896, compared with VerbalComprehension and PerceptualOrganisation accounting for 54% ofthe variance .

Factor analyses cf the earlierWISC-R have produced three factors ;

AUSTRALIAN ~OURNAI. OF LEARNING DISARIIJTIfS

a small third factor, called Freedomfrom Distractihility, was compcued ofArithmetic, Digit Span, and C<xling .This structure has consistentlyemerged in a variety of populations(Kamphaus, 1993 ) . The robustness ofthis third factor rather spoilt theoriginal preferred two factor solution .In addition, this third factor is difficultto interpret as there is no obviouscommon element in the three subtestsof Arithmetic, Digit Span andCoding. Kamphaus (1993) notes thatseveral names have been proposed todescribe the intellectual abilityrepresented by this factor such as"attention/distractihility, anxiety,symbolic ability, sequential prexessinQ,and memory" (p . 132) .

In an attempt to clarify thecommon element tying these threesubscales together in this third factor,a new suhtest called Symbol Searchwas added to create the WISC-111. Inthe factcx analysis with SymbolSearch included, a fourth factor calledProcessing Speed emerged . This fourthfactor was composed of Symbol Searchand Coding which previously loadedon the third factor-Freedom fromDistractihility .

The third factor (Freedom fromDistractibility) of the WISC-lll wasrepresented by the two subtests ofArithmetic and Digit Span . Wechsler(1991) suggested that the name,Freedom from Distractibility, he kept"for purposes of historical continuity"(p . 187) . This decision was not helpfulas distractibility conveys littlemeaning to parents and teachers . Theterm, distractibility, in the Australiancontext, is more likely to heinterpreted in a behavioural sense as"easily distracted" . Clearly, facility orlack of it with mental arithmeticproblems and the ability tc, recalldigits is probably not a sound predictorof a student's classrcx~m behaviour .

A number of alternative terms forthe Arithmetic and Digit Spansubtests composing this factor havebeen considered . They are numericalfacility, auditory short-term memcxy(Kamphaus, 1993) ; auditory shcut-term memory, numerical/mathematical processing andreasoning, and verbal responding(Hishinuma & Yamakawa, 1993) ;number ability and sequentialprocessing ( Kaufman, 1994 ) ; andnumerical knowledge, numerical

Vol . 2 No . 2 June 1991

Page 8: THEEDUCATIONALUSES OE THE WISC-III - Don Pugh Interests/IQ TESTING/THE... · WISC-III, gexxj usecanhe made of the individualsubtests. While the results ofthese subtestsareonly estimates

facility and quantitative reasoning(Carrol, 1993). It is unlikely that anyof these suggestions will he seriouslycilnsidered as there appears to be someconsensus between thepsychometricians that the WISC-IllFreedom from Uistractihility factor isnot robust enough to stand in its ownright. Ncx only does it account for theleast variance of the four factors, butthe Arithmetic suhtest loads almt>stequally at 0.49 on the VerhalComprehension factor compared withits 0.53 on the Freedom fnlmUistractihility factcx Roid et al .(1993)z .

In his review of the WISC-II1,Little (1992) recommends that"Freedom from Uistractihility shouldhe ignored" (p. 153) . Sattler (1992)and Thorndike (cited by C<xiper,1995) agree with Little and also argueagainst attempting an interpretationof Freedom from Distractibility . Itappears that their objections are firmlyrooted in psychometric concernsregarding the factor . Whileacknowledging these psychometricconcerns, Kaufman (1994) provides ahelpful view for the use of this factorby practitioners .

Kaufman (1994) recommendsthat the four factors should heconsidered as two pairs. The verbalpair consist of the factors of VerhalComprehension and Freedom fromUistractihility. These two factorsconsist solely of Verhal suhtests whilethe euher two factors are compcued ofjust the Performance suhtests. Heargues that Verhal 1Q is nllt a unitaryconstruct and can he well understcxxlas the "quartet of Infortnatiun-Sitttilarities-Vc,cahulary-Cotnprehension" (p . 106) andArithmetic with Digit Span . On therather hand, the fewrth factor, resultingfrom the addition of Symbol Searchmirrors the Verbal split on thePrrfe,rmance IQ . In this case, thePerfixmance pair is the facnns ofPerceptual Organisation andPrtx:essing Speed. The fmtrth factor,although small, has generalacceptance by the majority ofpsychetmetricians cited above.

REFERENCESRannatyne, A. (1974) . 1)iagncxis : A note

on the rceategorizatie~n of the WISCscaled scores . Jtxtrnal of LearningUiwbilities, 7, 272-274

Vol . 2 Nn . 2 Junc 1497

4 R E F E R E E D P A P E R S 4

Bauman, E. (1991) . Determinants ofWISC-R suhtest stability in childrenwith learning difficulties . Journal ofClinical Ps7choiv~, 47, 272-274 .

Rlaha, J ., & Wallbrown, F. H. (1996) .Hierarchical factor structure of theWcchslcr Intelligence Scale forChildren-111 . PsychvbgicalAssessment, 8(2), 214-218.

Carrol, J . R. (1993) . What abilities arcmeasured by the WISC-III'IrlurnalofPs7chrxducatirlnal Assessment,Mvnvgraph Series, Advances inPsychexducatielnal Assessment, WcchslcrIntelligence Scale fin Children : ThirdEdit&In, 135-143.

Cextper, S. (1995) . The clinical tcse andinterpretatum of the Wcchslcr IntelligenceScale fur Children : Third Edition .Illinois : Charles C. Thomas.

Cordoni, R. K., O'lkmncll, J . P.,Ramaniah, N. V., Kttrtz, J ., &Roscnshcin, K. (1981) Wcchslcr adultintelligence score patterns (or learningdisah(cd young adults . Journal ofLearningUisabilities, 14(7),404-407 .

Hale, R. L., & Saxc, J . E. (1983) . Profileanalysis of the Wcchslcr IntelligenceScale for Children-Revised . Journal ofPsychneducational Assessment, 1 155-162

Hishinuma, E. S., & Yamakawa, R.(1993) . Ccmstruct and critcrion-rclated validity of the WISC III (orexceptional students and those whoarc "at risk" . Journal ofPs7chcxducaticmal Assessment,Mvnngraph Series ; Adtnnces inPsychrxducatianal Assessment, WcchslcrIntelligence Scale fur Children : ThirdEditum, 94-104 .

Kamphatts, R. W . (1993) . Clinicalassessment of children's intelligence : Ahandlxxrk fvr professional practice .Boston : Allyn & Raum,.

Kamphatts, R. W., Rcnson, J., Htttchinson,S., & Platt, L. O. . (1994) .Identification of factor mcxlels fen theW1SC-III . Educational and PsydulLtgicalMca~urement, 5"I(1), 174- I86.

=. Kaufman, A. S. (1981) . The WISC-R andIcaming disahilitics assessment : Stateof the art . Journal of LearningDisabilities, 14(9), 520-526.

i Kaufman, A. S. (1994 ) . Intelligent testingwith the WISC 111 . Ncw York : Wilcy.

Lerncr, J . W. (1997) . Learning Disabilities :Theories, diagncx. is, and teachingstrategies . Rostun : Hcntghton MifflinCompany.

Little, S. G. (1992) . The WISC-Ill :Everything old is new again. SduxtlPsychrt4yQuarterly, 7(2), 116-142.

Mdkrmutt, P. A., Glutting, J . J ., Jones, J .N., Watkins, M. W., & Kush, J .(1989) . Jmst say no to suhtest analysis :A critique on Wcchslcr theory andpractice . Journal of P.rychoeducatirmalAssessment, R, 290-302.

AI!RRALIAfi JIII~RIVAI .IIF LF.IRNINf, DICARIIITIFS

Prifitera, A., & Ikrsh, J . (1993) . Rase ratesof WISC-Ill diagncxtic suhtestpatterns among normal, learning-disablcd, and ADHD samples. Journalof Ps7chcxducationnl Assessment,WISC-111 Monograph, 43-55.

Roid, G. H., Prifitera, A., 6t Weiss, L. G.(1993) . Replication of the WISC III(actor structure in an independentsample . Journal of Ps7choeducativnalAssessment, Monograph Series,Advances in Ps7chrxduratuma!Assessment, Wcchslcr Intelligence Scalefin Children : Third Edition, 6-21 .

Salvia, J ., Gajat,~A ., Mccnakshi, G., &Salvia, S. (1988) . A comparison ofWAIS-R profiles on nondisahledcollege freshmen and college studentswith learning disahilitics. Journal ofLearning Uiwbilities, 21(10), 6.32-636 .

Sattler, J . M . (1992) . Assessment ofchildren : WISC III and WPPS1-Rsupplement . San Diego, CA. )erome MSattler, Pub.

Spafford, C. S. (1989) . Wcchslcr DigitSpan suhtest: Diagnostic usefulnesswith dyslexic children . Perceptual andMotor SkilCs, 69, 115-125.

-~Vargo, F. E., Grosser, G. S., &Spafford, C.S. (1995) . Digit Span and otherWISC-R scores in the diagnosis n(dyslexia . Perceptual and Mvu~r Skills,80,1219-1229 .

Ward, S. R., Ward, T. J., Hats, C. V.,Young, D. L., Mollner, N . R. (1995) .The incidence and utility of theACID, AC1DS, and SCAR profiles ina referred population . Psychnlng7 in theSchrxlh, 31, 267-276.

Wcchslcr, I) . (1974) . Manual fro theWcchslcr Intelligence Scale fvr Children-Revised . Ncw York : The PsychologicalCcxpcxation.

Wcchslcr, D. (1991) . Wcchslcr lntelligenccScale for Children-Third Editum :Manual . San Diego, CA: HarccutrtRracc Jovanovich, Inc .

NOTES' The IQ score of I00 places tlu child xt

the 50th percentile . Use of the 95`X.confidence interval places the childbetween the 37th percentile (IQ score95)and the 63 percentile rank (IQscore 105) .

Carrol (1993) reported anorthogonalizcd hicrarchial factormatrix as determined from obliquesimple struxture and factcxintcrcorrelations at ages 6-7 and 14-16 .In the 6-7 age matrix Arithmeticloaded .64 on g and .41 on Frccdumfrom Distractibility (Fl)) along withSymbol Search .30 on FI). At the 14-16 age, Arithmetic loaded .79 Wn g(the highest loading with Informationnext at .70) and on FI) Arithmetic wasthe only suhtest to load with .64 .

Page 20