development and application of the reynolds intellectual ... · naglieri c05.tex v1 - 04/30/2009...

Naglieri c05.tex V1 - 04/30/2009 1:15 P.M. Page 95

C H A P T E R 5

Development and Applicationof the Reynolds Intellectual

Assessment Scales (RIAS)Cecil R. Reynolds, R. W. Kamphaus∗

This chapter provides the reader with anextensive introduction to the RIAS (Reynolds& Kamphaus, 2003), a recently published mea-sure of intelligence for children and adults. Abrief overview of the tests is provided, followedby a review of the theory and structure of theRIAS, framed primarily around its goals for de-velopment. A more extensive description is thenprovided of the subtests and their administrationand scoring. Psychometric characteristics of theRIAS are next presented along with guidelinesfor interpretation. Clinical applications of theRIAS are discussed, followed by a brief reviewof the characteristics and use of the ReynoldsIntellectual Screening Test (RIST). The chaptercloses with a case study using the RIAS as thefeatured measure of intelligence.

The RIAS is an individually administered testof intelligence appropriate for ages 3–94 yearswith a co-normed, supplemental measure ofmemory. The RIAS includes a two-subtest Ver-bal Intelligence Index (VIX) and a two-subtest

Nonverbal Intelligence Index (NIX). The scaledsums of T scores for the four subtests are com-bined to form the Composite Intelligence Index(CIX), which is a summary estimate of globalintelligence. Administration of the four intelli-gence scale subtests by a trained, experiencedexaminer requires approximately 20 to 25 min-utes. A Composite Memory Index (CMX) isderived from the two supplementary memorysubtests, which require approximately 10 to 15minutes of additional testing time. The Com-posite Intelligence Index and the CompositeMemory Index represent the combination ofboth verbal and nonverbal subtests. Table 5.1provides an overview of the indexes and subtestsof the RIAS.

For those who consider memory to be animportant element in the determination of IQ,the RIAS Professional Manual as well as the RIASscoring and interpretive software includes all nec-essary psychometric and normative informationfor creating intelligence indexes that combine

∗Portions of this chapter are adapted with permission of PAR, Inc. from Reynolds and Kamphaus (2003). We also wish to disclosethat we are the authors of the RIAS.

95


96 PART II INTELLIGENCE TESTS MEASURING DIVERSE ABILITIES

the VIX, NIX, and CMX subtests into TotalBattery Indexes.

THEORY AND STRUCTURE

The RIAS was designed to meld practicaland theoretical aspects of the assessment ofintelligence. While the models of Cattell andHorn (Horn & Cattell, 1966; Kamphaus, 2001)and Carroll (1993) were the primary theoreticalguides, the RIAS also followed closely thedivision of intelligence into verbal and nonverbaldomains, due to the practical benefits of assessingverbal and nonverbal intelligence. The latter areclosely related to concepts of fluid and crystallizedintelligence (aka the Cattell-Horn-Carroll orCHC approach) but are clearly not entirelycognate concepts. Memory was included as aseparate scale on the RIAS due to the growingimportance of working memory in models ofintelligence and the practical aspects of memoryto everyday diagnostic questions faced by thepractitioner (e. g., see Bigler & Clement, 1997;Goldstein & Reynolds, 1999; Reynolds & Bigler,1994; Reynolds & Fletcher-Janzen, 1997). Tounderstand the theoretical underpinnings of theRIAS as well as its practical aspects and structure,a review of the goals for development of the testprovides a strong heuristic.

Development GoalsReynolds and Kamphaus (2003) describe a setof eight primary goals for development of theRIAS, derived based on their experiences over theyears in teaching, administering and interpreting,and researching intelligence tests (for moreextensive review and discussion, see Reynolds &Kamphaus, 2003, especially Chapters 1 and 6).

Goal 1: Provide reliable and valid measurementof g and its two primary components, verbaland nonverbal intelligence, with close corre-spondence to crystallized and fluid intelligence.The general intelligence factor, g, is the

most reliable component present in anymultifactorial view of intelligence (Jensen,1998). In the Cattell-Horn model (Horn& Cattell, 1966; Kamphaus, 2001) of in-telligence, g is the dominant factor in thehierarchy of multiple abilities, with the nexttwo dominant facets being crystallized andfluid intelligence. The RIAS includes sub-tests that match both approaches closely inorder to share in the theoretical supportof the Cattell-Horn model of crystallizedand fluid intelligence, while taking ad-vantage of the very practical division ofintelligence into verbal and nonverbal com-ponents. Verbal and nonverbal componentsof intelligence also have strong supportfrom factor-analytic work (e.g., Kaufman,1994) and the brain sciences (e.g., Reynolds,Castillo, & Horton, 2008; Riccio & Hynd,2000).

Goal 2: Provide a practical measurement devicein terms of efficacies of time, direct costs, andinformation needed from a measure of in-telligence. Time, cost, and efficiency havealways been necessary considerations in thedelivery of effective psychological and psy-choeducational services, but the advent ofmanaged care in the 1990s (for private prac-tice and clinical settings) and the explosionof services for children with disabilities inschools has made time a crucial considera-tion for practitioners. A useful intelligencetest needs to provide an objective, reliable,and valid assessment of the major constructsthat underlie psychometric intelligence. In-tellectual assessment can be done efficientlyand at a significantly lower cost than is of-ten the case with other tests. One goal forthe RIAS was for it to provide an efficientmeasure of intelligence that, at the sametime, meets regulatory guidelines for theIndividuals with Disabilities EducationalImprovement Act and other rules.

Goal 3: Allow continuity of measurement acrossall developmental levels from ages 3 through


CHAPTER 5 DEVELOPMENT AND APPLICATION OF THE (RIAS) 97

94 years for both clinical and research purposes.

AQ1

Individuals frequently require reevaluationof intellectual function over a period ofyears. As they age, different versions of in-telligence tests may be required, and thesevarious tests have different subtests, mea-sure different aspects of intelligence, werenormed in different years, and may havesample stratifications that do not match.Thus, scores obtained over time may notbe comparable due to measurement artifactand not to any real changes in the indi-vidual’s cognitive level or structure (Sattler,2001). There is clinical utility in havinga common set of subtests and a commonreference group for such comparisons.

Goal 4: Substantially reduce or eliminate depen-dence on motor coordination and visual-motorspeed in the measurement of intelligence. Themajority of current individually admin-istered intelligence tests rely heavily onvisual-motor coordination and motor speedfor accurate assessment of intelligence.However, many children referred to specialeducation services have visual-motor diffi-culties or frank motor impairment or theymay have other neurodevelopmental disor-ders that produce motor-related problems(Goldstein & Reynolds, 1999). Individualswith traumatic brain injury or central ner-vous system disease commonly have motorproblems of various forms, including finemotor, gross motor, and strength problems(Reynolds & Fletcher-Janzen, 1997). Inolder populations, the incidence of tremorand related motor problems is quite high.To attempt to measure the intelligence ofsuch individuals (who form a significantportion of referrals involving intellectualassessments) with tasks that require rapidmanipulation of blocks, small cardboardpieces, or even pencil markings where speedand accuracy of performance are substan-tial contributors to the resulting IQ orcognitive index seems inappropriate. It is

our view that intelligence tests should em-phasize thinking, reasoning, and problemsolving.

Goal 5: Eliminate dependence on reading inthe measurement of intelligence. Tasks wherethe ability to read the English languagefacilitates individual item performance con-found the measurement of intelligence withschool instruction. Certainly intelligencecannot be assessed completely independentof prior knowledge despite many failed at-tempts to do so (e.g., culture-free tests;see Anastasi & Urbina, 1997; Kamphaus,2001; Reynolds, Lowe, & Saenz, 1999).However, to confound intellectual assess-ment with clues obtained from the abilityto read a vocabulary card or to fill in blankswithin printed words (which also adds aconfound with the ability to spell) makessuch tests inappropriate for nonreaders orthose with limited English-reading skills.Reading tasks also penalize individuals withvisual impairments whose intellectual func-tioning is assessed traditionally via verbaltasks.

Goal 6: Provide for accurate prediction of basicacademic achievement at levels that are at leastcomparable to that of intelligence tests twice itslength. Prediction of academic achievementand acquired knowledge in such areas asreading, language, and math is an importantfunction for intelligence tests. Prediction ofachievement should remain a function ofany new intelligence test.

Goal 7: Apply familiar, common concepts that areclear and easy to interpret, coupled with simpleadministration and scoring. Formal intelli-gence testing via Binet-type tasks is morethan a century old. During this time, innu-merable tasks have been devised to measureintelligence and related abilities. Many ofthese tasks are quite good at the measure-ment of intellectual function and possesslong histories in psychology and educa-tion. The use of familiar, well-researched



tasks has many advantages over the use ofnovel, less-well-established tasks. Many ofthese tasks are simple and easy to admin-ister despite the complex mental functionsrequired for deriving a correct solution. Ob-jective scoring can be facilitated by avoidingtasks that require lengthy verbal responsesor split-second timing for awarding bonuspoints. Tasks that are simple to adminis-ter and objective to score nearly eliminateadministration and scoring errors.

Goal 8: Eliminate items that show differentialitem functioning (DIF), associated with genderor ethnicity. The problem of cultural biashas long produced debate in psychology, ineducation, and in the lay press (e.g., Brown,Reynolds, & Whitaker, 1999; Reynolds,Lowe, & Saenz, 1999). Following years ofdebate, a host of methods for detectingtest items that function differentially acrossnominally defined groups have been devised(Reynolds, 2000). Despite the importanceof this issue, it is seldom discussed in sig-nificant detail in test manuals. However,in view of the availability of sound statis-tical approaches for identifying such testitems, all intelligence test items should bescrutinized during development and stan-dardization to determine whether any arein fact biased.

The RIAS provides a more reliable assessmentof memory, both verbal and nonverbal, thanother intelligence tests, and also treats memoryfunction as a separate scale. The RIAS includesassessment of memory function because it iscrucial to the diagnostic process for numerousdisorders of childhood (Goldstein & Reynolds,1999; Reynolds & Fletcher-Janzen, 1997) andadulthood (Goldstein & Reynolds, 2005), par-ticularly in later adulthood (Bigler & Clement,1997). In fact, assessment of memory functionprovides a better window into the integrity ofbrain function than does the assessment of in-telligence (e.g., see Adams & Reynolds, 2009).Memory functions are more easily disrupted than

general intellectual skill in the face of central ner-vous system (CNS) compromise due to traumaor disease. Although intelligence will often sufferunder such conditions (Joseph, 1996), memorywill suffer sooner and is typically more affected.The RIAS CMX does not provide a comprehen-sive memory assessment, but it does cover thetwo areas of memory historically assessed by in-telligence tests, which are considered by many tobe the two most important memory functions toassess (e.g., Bigler & Clement, 1997; Reynolds& Bigler, 1994): memory for meaningful verbalmaterial and visual memory.

There is also utility in having memory testsco-normed with a measure of intelligence ona fully overlapping sample. This co-normingpresents the best possible scenario for contrastingtest scores (Reynolds, 1984–1985), allowing theclinician directly to compare the examinee’s IQwith these key memory functions.

TheoryThe RIAS was designed to measure four impor-tant aspects of intelligence—general intelligence(of which the major component is fluid or rea-soning abilities), verbal intelligence (sometimesreferred to as crystallized abilities, which is aclosely related though not identical concept),nonverbal intelligence (referred to in some the-ories as visualization or spatial abilities and closelyallied with fluid intelligence), and memory (sub-tests comprising this ability have been labeledvariously as working memory, short-term memory,or learning). These four constructs are measuredby combinations of the six RIAS subtests (seeTable 5.1).

The RIAS subtests were selected and designedto measure intelligence constructs that have asubstantial history of scientific support. In ad-dition, Carroll’s (1993) seminal and often-citedthree-stratum theory of intelligence informedthe creation of the RIAS by demonstratingthat many of the latent traits tapped by in-telligence tests were test-battery independent.



TABLE 5.1 Structure and Components of the RIAS

Composite Intelligence Index (CIX)

Subtests of the Verbal Intelligence Index (VIX)Guess What. Examinees are given a set of 2 or 3 clues and asked to deduce the object or concept being

described. This subtest measures verbal reasoning in combination with vocabulary, language development,and overall fund of available information.

Verbal Reasoning. Examinees listen to a propositional statement that essentially forms a verbal analogy and areasked to respond with one or two words that complete the idea or proposition. This subtest measuresverbal-analytical reasoning ability but with fewer vocabulary and general knowledge demands than GuessWhat.

Subtests of the Nonverbal Intelligence Index (NIX)Odd Item Out. Examinees are presented with a picture card containing five to seven pictures or drawings and

asked to designate which one does not belong or go with the others. This subtest measures nonverbalreasoning skills but will also require the use of spatial ability, visual imagery, and other nonverbal skills onvarious items. It is a form of reverse nonverbal analogy.

What’s Missing. A redesign of a classic task present on various ability measures, examinees are shown a picturewith some key element or logically consistent component missing and are asked to identify the missingessential element. This subtest assesses nonverbal reasoning wherein the examinee must conceptualize thepicture, analyze its Gestalt, and deduce what essential element is missing.

Composite Memory Index (CMX)Verbal Memory Index (VMX). This scale consists of a single verbal memory subtest. Depending on the

examinee’s age, a series of sentences or brief stories are read aloud by the examiner and then recalled by theexaminee. This task assesses the ability to encode, store briefly, and recall verbal material in a meaningfulcontext where associations are clear and evident.

Nonverbal Memory Index (NMX). This scale consists of a single visual memory subtest. It contains a series ofitems in which a stimulus picture is presented for five seconds, following which an array of pictures ispresented. The examinee must identify the target picture from the new array of six pictures. It assesses theability to encode, store, and recognize pictorial stimuli that are both concrete and abstract or withoutmeaningful referents.

He clearly demonstrated, for example, that nu-merous tests measured the same crystallized,visual-perceptual, and memory abilities. How-ever, Kamphaus (2001) concluded that thesesame test batteries did not measure fluid abilitiesto a great extent.

The RIAS focuses on the assessment ofstratum-three and stratum-two abilities fromCarroll’s (1993) three-stratum theory. Stratumthree is composed of one construct only, g.Psychometric g accounts for the major portionof variance assessed by intelligence test batteries.More important, however, is the consistent

finding that the correlations of intelligence testswith important outcomes, such as academicachievement and occupational attainment, arerelated to the amount of g measured by the testbattery. In other words, so-called g-saturatedtests are better predictors of important outcomesthan are tests with low g saturation. Althoughthe nature of g is yet to be fully understood, thescores from g-saturated tests have known utility,especially in terms of prediction.

The second stratum in Carroll’s (1993) hi-erarchy consists of traits that are assessed by



combinations of subtests, or stratum-one mea-sures. A stratum-one measure is typically a singlesubtest that measures the trait of interest. Com-binations of stratum-one subtests, such as thoseused to form VIX and NIX, are consideredstratum-two measures and should result in en-hanced measurement of complex traits such asverbal and nonverbal intelligence. Combiningstratum-two index measures into an overarchingcomposite measure, such as CIX, allows for themeasurement of a complex stratum-three trait,such as general intelligence.

There are, however, several stratum-two traitsto choose from. These second-stratum traitsinclude fluid intelligence, crystallized intelli-gence, general memory and learning, broadvisual perception, broad auditory perception,broad retrieval ability, broad cognitive speed,and processing speed (i.e., reaction time or de-cision speed). Of importance, however, is thefinding, from hundreds of investigations, sug-gesting that these abilities are ordered by theirassessment of g (Kamphaus, 2001). Specifically,subtests that tap reasoning abilities are excellentmeasures of g, making the first few stratum-twofactors the best for inclusion in an intelligencetest like the RIAS, especially one that seeks to bea time-efficient test.

Any test of g must measure so-called higher-order cognitive abilities, those associated with fluidabilities, such as general sequential reasoning,induction, deduction, syllogisms, series tasks,matrix reasoning, analogies, quantitative reason-ing, and so on (Carroll, 1993). Kamphaus (2001)advocated the following definition of reasoning:‘‘that which follows as a reasonable inference ornatural consequence; deducible or defensible onthe grounds of consistency; reasonably believedor done.’’ (Oxford Press, 1999). This definitionemphasizes a central cognitive requirementto draw inferences from knowledge. Thischaracteristic of general intelligence is measuredbest by two RIAS subtests, Verbal Reasoningand Odd Item Out, although all of the subtestshave substantial g saturation (see Reynolds &Kamphaus, 2003, especially Chapter 6).

Kamphaus (2001) has suggested that the termcrystallized for this second-order factor does notfully capture the centrality of language processesinvolved in successful performance on subteststypically associated with this ability. He proposedthe term verbal to describe the latent constructtapped by subtests like those selected for theRIAS.

Nonverbal tests have come to be recog-nized as measures of important spatial andvisual-perceptual abilities—abilities that mayneed to be assessed for a variety of examinees,including those with brain injuries. In the 1963landmark ETS Kit of Factor-Referenced CognitiveTests, spatial ability was defined as ‘‘the abilityto manipulate or transform the image of spa-tial patterns into other visual arrangements’’ (ascited in Carroll, 1993, p. 316). The RIAS What’sMissing and Odd Item Out subtests follow inthis long tradition of tasks designed to measurevisuospatial abilities.

Digit recall, sentence recall, geometric de-sign recall, bead recall, and similar measuresloaded consistently on a ‘‘general memory andlearning’’ stratum-two factor identified by Car-roll (1993) in his numerous analyses. The RIASVerbal Memory and Nonverbal Memory sub-tests are of this same variety, although morecomplex than simple confrontational memorytasks such as pure digit recall. Carroll’s find-ings suggest that the RIAS Verbal Memoryand Nonverbal Memory subtests should begood measures of the memory construct thathas been identified previously in so many in-vestigations of a diverse array of tests. Car-roll described memory span as ‘‘attention toa temporally ordered stimulus, registration ofthe stimulus in immediate memory, and out-put of its repetition’’ (p. 259). This operationaldefinition is an accurate description of theRIAS memory subtests and composite. Mem-ory is typically considered a complex trait withmany permutations, including visual, verbal,long term, and short term. Carroll’s analysisof hundreds of data sets supports the organi-zation of the RIAS, in that he found ample



evidence of a general memory trait that maybe further subdivided for particular clinicalpurposes.

Description of SubtestsSubtests with a familiar look and feel and withessentially long histories in the field of intellec-tual assessment were chosen for inclusion onthe RIAS. There are a total of four intelli-gence subtests and two memory subtests. Theintelligence subtests were chosen also due totheir complex nature—each assesses many intel-lectual functions and requires their integrationfor successful performance (also see Reynolds &Kamphaus, 2003, Chapter 6, under the head-ing of ‘‘evidence based on response processes’’).The memory subtests were chosen not onlyfor complexity but also due to their repre-sentation of the primary content domains ofmemory.

Guess What

This subtest measures vocabulary knowledgein combination with reasoning skills that arepredicated on language development and fundof information. For each item, the examinee isasked to listen to a question that contains cluespresented orally by the examiner and then to givea verbal response (typically one or two words)that is consistent with the clues.

Verbal Reasoning

The second verbal subtest, Verbal Reasoning,measures analytical reasoning abilities. More dif-ficult items of necessity also require advancedvocabulary knowledge. For each item, the exam-inee is asked to listen to an incomplete sentence,presented orally by the examiner, and then to givea verbal response, typically one or two words,that completes the sentence, most commonlycompleting a complex analogy. Completion ofthe sentences requires the examinee to evaluatethe various conceptual relationships that exist

between the physical objects or abstract ideascontained in the sentences.

Odd Item Out

This subtest measures general reasoning skillsemphasizing nonverbal ability in the form of areverse analogy. For each item, the examinee ispresented with a picture card containing fromfive to seven figures or drawings. One of thefigures or drawings on the picture card has adistinguishing characteristic, making it differentfrom the others.

What’s Missing

This subtest measures nonverbal reasoning skillsthrough the presentation of pictures in whichsome important component of the picturedobject is missing. Examinees must understandor conceptualize the pictured object, assess itsgestalt, and distinguish essential from nonessen-tial components. For each item, the examineeis shown a picture card and asked to examinethe picture and then to indicate, in words or bypointing, what is missing from the picture. Nam-ing the missing part correctly is not required solong as the examinee can indicate the location ofthe missing component correctly.

Verbal Memory

This subtest measures the ability to encode,briefly store, and recall verbal material in ameaningful context. Young children (ages 3–4)are asked to listen to sentences of progressivelygreater length as each is read aloud by the exam-iner and then asked to repeat each sentence backto the examiner, word for word, immediately af-ter it is read aloud. Older children and adults areasked to listen to two stories read aloud by theexaminer and then to repeat each story back tothe examiner, word for word, immediately afterit is read aloud. The sentences and stories werewritten to provide developmentally appropriatecontent and material of interest to the targetedage group. Specific stories are designated forvarious age groups.



Nonverbal Memory

This subtest measures the ability to encode,briefly store, and recall visually presented ma-terial, whether the stimuli represent concreteobjects or abstract concepts. For each item, theexaminee is presented with a target picture forfive seconds and then a picture card containingthe target picture and an array of similar pictures.The examinee is asked to identify the target pic-ture among the array of pictures presented onthe picture card. For each item, the examinee isgiven two chances to identify the target picture.The pictures are, at the upper levels, primarilyabstract, and at the lower age levels, commonobjects. The use of naming and related languagestrategies, however, is not helpful due to the de-sign of the distractors. For example, one earlyitem presents as a target stimulus a picture of acat. On the recall page, six cats are presented,each different (save one) in some characteristicfrom the target stimulus.

Scales and StructureThe six RIAS subtests are divided into fourcomposite indexes as depicted in Tables 5.1and 5.2: the Verbal Intelligence Index (VIX),the Nonverbal Intelligence Index (NIX), theComposite Intelligence Index (CIX, composedof the VIX + NIX), and the Composite memoryIndex (CMX). Interpretations of these indexesare discussed in a later section.

Administration and ScoringThe RIAS was specifically designed to be easyto administer and objective to score. For allsubtests except Verbal Memory, there are clearobjective lists of correct responses for each testitem and seldom are any judgment calls required.Studies of the interscorer reliability of these fivesubtests produced interscorer reliability coeffi-cients of 1.00 by trained examiners (Reynolds& Kamphaus, 2003). On Verbal Memory, somejudgment is required when examinees do notgive verbatim responses; however, the scoring

TABLE 5.2 RIAS Subtest Composition of IQ andIndex Scores

IndexesMean = 100 SD = 15

SubtestsMean = 50SD = 10 VIX NIX CIX CMX

GWH % %VRZ � %OIO % %WHM % %VRM %NVM %

NOTE: VRZ = Verbal Reasoning; GWH = Guess What;OIO = Odd Item Out; WHM = What’s Missing; VRM =Verbal Memory; NVM = Nonverbal Memory; VIX = VerbalComposite Index; NIX = Nonverbal Composite Index; CIX= RIAS Composite Index; CMX = Composite MemoryIndex.

criteria provide clear examples and guidelines forsuch circumstances, making the Verbal Mem-ory subtest only slightly more difficult to score.The interscorer reliability study of this subtestproduced a coefficient of .95.

The time required to administer the entireRIAS (including the intelligence and the mem-ory subtests) averages 30 to 35 minutes once theexaminer has practiced giving the RIAS and hasbecome fluent in its administration. Basal andceiling rules along with age-designated startingpoints were employed to control the adminis-tration time and each was derived empiricallyfrom the responses of the standardization sam-ple. A detailed description of the methods usedfor setting these administration parameters isgiven in Chapter 2 of Reynolds and Kamphaus(2003). Also, to facilitate ease of administrationand to make it more efficient, the RIAS recordform contains all of the instructions and ex-aminer guides necessary to administer the test.



Experienced examiners as well as graduate stu-dents have consistently reported that the RIAS issurprisingly easy to administer and score.

SCORING AND INTERPRETIVESOFTWARE

There are two separately available computer-ized scoring and interpretive programs relatedto the RIAS. One is the RIAS-IR (Reynolds &Kamphaus, 2007a), a scoring and interpretiveprogram devoted just to the RIAS (and the RI-AST), and the RIAS/WRAT-4-DIR (Reynolds& Kamphaus, 2007b), which analyzes discrepan-cies between the RIAS and the WRAT4 scoresof examinees based on a linking sample betweenthe two measures.

The RIAS-IR (Scoring andInterpretive Report)

The unlimited-use RIAS-IR is designed to assistclinicians with scoring, profiling, and interpret-ing the performance of individuals (ages 3–94) onthe RIAS and the Reynolds Intellectual Screen-ing Test (RIST; only the RIAS features are dis-cussed here). After manual entry of an examinee’sraw scores into the software, the RIAS-IR cangenerate up to three Score Reports, two Feed-back Reports (for parents/guardians of minors,teachers, or adult examinees after a discussion ofthe test results), and two Interpretive Reports.Program functionality includes report editing.Report options for the RIAS and the RISTinclude:

RIAS Score Report

After input of all four subtest scores, the RIASScore Report includes the examinee’s demo-graphic information, the RIAS Score SummaryTable, the RIAS Profile, brief interpretive text,and an Extended Score Summary Table.

RIAS Total Battery Score Report

After input of all six subtest scores, the RIASTotal Battery Score Report includes the exami-nee’s demographic information, the RIAS TotalBattery Score Summary Table, a Total BatteryProfile, and brief interpretive text.

Feedback ReportsRIAS Feedback Report

After input of all four subtest scores, the RIASFeedback Report provides easy-to-understandinformation about the examinee’s performanceon the RIAS written in lay terms for par-ents/guardians, teachers, or an adult examinee.

RIAS Total Battery Feedback Report

After input of all six subtest scores, theRIAS Total Battery Feedback Report provideseasy-to-understand information about the exam-inee’s performance on the RIAS Total Batterywritten in lay terms for parents/guardians, teach-ers, or an adult examinee.

Interpretive ReportsRIAS Interpretive Report

After input of either four or all six subtest scores,the RIAS Interpretive Report includes the ex-aminee’s demographic information, the RIASScore Summary Table, the RIAS Profile, anExtended Score Summary Table, and extensiveinterpretive text with intervention and additionaltesting recommendations. This report also pro-vides clinicians with the option to include theTotal Battery scores (i.e., Total Test Battery,Total Verbal Battery, Total Nonverbal Battery).

The RIAS-IR provides clinicians with greaterflexibility to obtain the report they desire byenabling the selection of specific report com-ponents for inclusion and by providing built-inreport-editing features that also allow cliniciansto cut-and-drop features into their own wordprocessing programs. It also provides flexibil-ity within the Interpretive Reports to generate

Naglieri c05.tex V1 - 04/30/2009 1:15 P.M. 4


setting-specific feedback and recommendations(i.e., school, employment, long-term care [as-sisted living, nursing home]) in addition to thesoftware’s default general feedback and recom-mendations.

Aptitude–Achievement DiscrepancySoftware: RIAS/WRAT4-DIR

The RIAS/WRAT4-DIR addresses the evalua-tion of discrepancies between intelligence testperformance and achievement test performanceby providing and evaluating discrepancies be-tween scores on the Reynolds Intellectual As-sessment Scales (RIAS)/Reynolds IntellectualScreening Test (RIST) and on the Wide RangeAchievement Test 4 (WRAT4). By comparinggeneral ability levels as assessed by the RIAS toachievement levels as obtained by the WRAT4,the RIAS/WRAT4-DIR provides the informa-tion necessary for assisting in the determinationof special education eligibility and the presenceof specific learning disabilities in individuals incases where discrepancy analyses are useful ornecessary. Although evaluation of such discrep-ancies is now optional in school settings, manyother settings continue to adhere to a require-ment of aptitude–achievement discrepancies forlearning disability determination.

Two types of scoring methodologies areused to derive the discrepancy scores andto evaluate the statistical significance of thescore and its prevalence within the popu-lation. The simple difference method exam-ines the difference between an obtained RIASscore and an obtained WRAT4 score. Thepredicted-achievement method uses the indi-vidual’s obtained RIAS score to predict his orher performance on the WRAT4. It then ex-amines the difference between the predictedWRAT4 score and the individual’s obtainedWRAT4 score. Using these methods, three typesof discrepancy reports that contain a total of 40discrepancy scores are generated by the software:

1. RIST/WRAT4 Discrepancy Interpretive Report:provides discrepancy scores and analysis

between the RIST index score and eachWRAT4 subtest/composite score (i.e., WordReading, Sentence Comprehension, Spelling,Math Computation, Reading Composite)

2. RIAS/WRAT4 Discrepancy Interpretive Report:provides discrepancy scores and analysis be-tween each RIAS Index score (i.e., VIX,NIX, CIX, CMX) and each WRAT4 sub-test/composite score

3. RIAS/WRAT4 Total Battery Discrepancy In-terpretive Report: provides discrepancy scoresand analysis between the RIAS Total BatteryIndex scores (i.e., TTB, TVB, TNB) andeach WRAT4 subtest/composite score

The RIAS/WRAT4-DIR Professional ManualSupplement provides normative data about theseparate linking sample developed specificallyfor derivation of this interpretive program. Itconsisted of 410 participants ages 5–24 and wasmatched to U.S. Bureau of the Census populationdata for gender, ethnicity, and educational attain-ment. The Professional Manual Supplement alsoprovides overviews of the RIAS and the WRAT4,including original standardization information,descriptions of the two different scoring meth-ods, and suggestions for the interpretation of thediscrepancy scores. It is unlimited-use softwarethat generates discrepancy scores and interpre-tive reports based on a clinician’s entry of anindividual’s raw scores and provides an efficientfile-handling system that enables the user to cre-ate examinee files where all protocols and reportfiles for each examinee are managed.

PSYCHOMETRIC PROPERTIES

The psychometric characteristics of any mea-surement device and its scores are certainlycrucial in determining its utility. In this section,a review of the standardization procedures andthe scaling of the RIAS is presented, followedby a summary of the reliability of the scores de-rived from the instrument and evidence related



to the validity of score interpretations. Due tothe length restrictions in a single book chapter,a discussion of the developmental process of thetest simply cannot be provided. However, theRIAS underwent years of development, includ-ing tryout and review of the items on multipleoccasions by school, clinical, and other psychol-ogists, including neuropsychologists. Items werewritten to conform to clear specifications consis-tent with the goals for development of the testas given previously in this chapter. Items werereviewed by panels of expert psychologists forcontent and construct consistency and by expertminority psychologists, as well, to ascertain thecultural saliency of the items and any potentialproblems of ambiguity or offensiveness in vari-ous settings. The developmental process speaksdirectly to the psychometric characteristics ofthe tests and is described in far more detail inReynolds and Kamphaus (2003) and should beconsidered carefully in any full evaluation of theinstrument.

StandardizationThe RIAS was normed on a sample of 2,438participants residing in 41 states between theyears 1999 and 2002. United States Bureau ofthe Census projected characteristics of the U.S.population, initially to the year 2000 and thenupdated through 2001, were used to select apopulation proportionate sample. Age, gender,ethnicity, educational level (parent educationallevel was used for ages 3–16 and the participants’educational level was used at all other ages), andregion of residence were used as stratificationvariables. To facilitate some of the analyses ofcultural bias in the item pool, minorities wereoversampled in some cells, particularly at theearly ages. The resulting norms for the RIASwere calculated on a weighted sampling thatprovided a virtually perfect match to the censusdata. The overall sample was a close match to thepopulation statistics in any regard (see Reynolds& Kamphaus, 2003, especially Tables 4.2–4.5).

Norm Derivation and ScalingStarting Points and Basal/Ceiling Rules

During standardization, starting points andbasal/ceiling rules were set to ensure that partic-ipants received the maximum number of itemsthey would be expected to receive on the fi-nal version. Once the final RIAS items weredetermined (after standardization), they were re-ordered by ascending item difficulty index. Aniterative process of item selection and variousbasal and ceiling rules was then applied to lo-cate the points that captured the performance ofmost (over 90% across all ages) of the examineeshad they been administered all items on the test.The reliability of scores obtained under the basaland ceiling rules that best fit these criteria werethen examined to be certain accuracy of scoreestimation was not adversely impacted across ex-aminees by applying a particular set of rules.Finally (after scaling was completed using thechosen rules), the scores of all examinees werecompared under the application of the chosenbasal/ceiling rules and under the condition oftaking all items as a final check to ascertain theimpact on the scores of individual examinees. Itwas rare under the final rules for any examinee’ssubtest scaled score to be affected by more thanone point and most were in fact unaffected. Oncethe basal and ceiling rules were established, theraw score distributions were determined and thevarious scaled scores derived.

Scaling Methods

All standard scores for the RIAS were derived viaa method known as continuous norming. Continu-ous norming is a regression-based methodologyused to mitigate the effects of any sampling ir-regularities across age groupings and to stabilizeparameter estimation. An important feature ofcontinuous norming is that it uses informationfrom all age groups, rather than relying solelyon the estimates of central tendency, dispersion,and the shape of the distributions of a single agegrouping for producing the norms at each chosen



age interval in the normative tables for a partic-ular test (Zachary & Gorsuch, 1985). As such,the continuous norming procedure maximizesthe accuracy of the derived normative scores andhas become widespread in its application to thederivation of test norms over the last 20 years.

T-scores were chosen for the RIAS subtestsover the more traditional mean = 10 scaled scorespopularized by Wechsler (e.g., Wechsler, 1949)due to the higher reliability coefficients obtainedfor the RIAS subtest scores and expandeditem pools. With high degrees of reliability oftest scores, the use of scales that make finerdiscriminations among individuals is possible,producing a more desirable range of possiblescores. For the convenience of researchers andexaminers who wish to use other types of scoresfor comparative, research, or other purposes, theRIAS Manual (Reynolds & Kamphaus, 2003,Appendix B) provides several other commontypes of scores for the RIAS indexes, includingpercentiles, T-scores, z-scores, normal curveequivalents (NCEs), and stanines, along witha detailed explanation of each score type.

Score ReliabilitySince the RIAS is a power test, the internalconsistency reliability of the items on theRIAS subtests was investigated using Cronbach’scoefficient alpha. Alpha reliability coefficients forthe RIAS subtest scores (Cronbach, 1951) andthe Nunnally reliability estimates (see Nunnally,1978, p. 249, formula 7-15) for the index scoresare presented in Reynolds and Kamphaus (2003)for 16 age groups from the total standardizationsample. The reliability estimates are rounded totwo decimal places and represent the lower limitsof the internal consistency reliability of the RIASscores.

According to the tables in Chapter 5 ofReynolds and Kamphaus (2003), 100% of thealpha coefficients for the RIAS subtest scoresreach .84, or higher, for every age group. Themedian alpha reliability estimate for each RIASsubtest across age equals or exceeds .90. This

point is important because many measurementexperts recommend that reliability estimatesabove .80 are necessary and those above .90 arehighly desirable for tests used to make decisionsabout individuals. All RIAS subtests meet theserecommended levels. As shown in Table 5.2of Reynolds and Kamphaus (2003, p. 78), thereliability estimates for all RIAS indexes havemedian values across age that equal or exceed .94.These reliability estimates are viewed as excellentand often exceed the reliability values presentedfor the composite indexes or IQs of tests twoor three times the length of the RIAS. Thus,one can have strong confidence in the relativereliability and accuracy of both subtest and indexscores derived from standardized administrationof the RIAS.

One cannot always assume that because atest is reliable for a general population, it willbe equally reliable for every subgroup withinthat population. It is thus instructive to viewthe various reliability estimates for the RIAS(or any test) for smaller, meaningful subgroupsof a population (Reynolds, 2003). When calcu-lated separately for male and female examinees,the reliability coefficients are high and relativelyuniform (see Table 5.3 of Reynolds & Kam-phaus, 2003, for the full table of values) withno significant differences in test score reliabil-ity at any age level as a function of gender.

TABLE 5.3 Uncorrected (and Corrected) PearsonCorrelations between Reynolds IntellectualAssessment Scales (RIAS) Indexes and WJ-III Scoresfor a Referral Sample

WISC-III WAIS-III

RIASIndex VIQ PIQ FSIQ VIQ PIQ FSIQ

VIX .86 .44 .78 .71 .61 .70NIX .60 .33 .60 .67 .71 .71CIX .81 .42 .76 .74 .70 .75CMX .67 .40 .66 .76 .76 .79



Reliability estimates were also calculated sepa-rately for whites and African Americans. Thesevalues are reported in Chapter 5 of the RIASManual. Although the reliability estimates acrossethnic groups display more variability than acrossgender, the values are consistently high exceptfor several instances. Of the 320 values cal-culated, only five drop below .80 (i.e., rangefrom .66 to .79). When the values are testedfor the presence of statistically significant dif-ferences (again using the Feldt technique, seeReynolds, 2003) and the Bonferroni correctionis applied (as recommended in the APA taskforce on statistical significance testing officialreport; Wilkinson & Task Force on Statisti-cal Inference, 1999), no significant differencesare found. Thus, the RIAS shows uniformlyhigh internal consistency reliability estimatesacross age (3–94) and across gender and acrossethnicity.

Test score stability (sometimes referred to aserror due to time sampling) refers to the extentto which an individual’s test performance isconstant over time and is usually estimated by thetest–retest method. The stability of RIAS scoresover time was investigated using the test–retestmethod with 86 individuals ages 3–82. Theintervals between the two test administrationsranged from 9 to 39 days, with a mediantest–retest interval of 21 days. The correlationsfor the two testings, along with mean scores andstandard deviations, are reported in detail in theRIAS Manual (Reynolds & Kamphaus, 2003) inTables 5.7–5.11 for the total test–retest sampleand for four age groups: 3–4 years, 5–8 years,9–12 years, and 13–82 years.

The obtained coefficients are of sufficientmagnitude to allow confidence in the stabilityof RIAS test scores over time. In fact, the valuesare quite good for all of the subtests, but espe-cially for the index scores. Both the uncorrectedcoefficients and the corrected, or disattenuated,coefficients are reported (the corrected coeffi-cients being corrected for the alpha for eachsubtest). The uncorrected coefficients are allhigher than .70, and 6 of the 10 values are in

the .80s. The corrected values are even more im-pressive, with all but two values ranging from .83to .91. When viewed across age groups, the valuesare generally consistent with the values obtainedfor the total test–retest sample. The test–reteststability coefficients for scores on the RIAS sub-tests and indexes are quite strong and provideevidence of more than sufficient short-term tem-poral stability of the scores to allow examiners tobe confident in the obtained results.

Validity of RIAS Test Scores asMeasures of intelligence

According to the Standards, validity, in this con-text, refers to ‘‘the degree to which evidenceand theory support the interpretations of test scoresentailed by proposed users of tests’’ (emphasisadded, p. 9). Reynolds (1998) defined validitysimilarly, arguing that validity refers to the ap-propriateness and accuracy of the interpretationof performance on a test, with such performanceusually expressed as a test score. Validation ofthe meaning of test scores is also a process, onethat involves an ongoing, dynamic effort to ac-cumulate evidence for a sound scientific basisfor proposed test score interpretations (AERA,APA, & NCME, 1999; Reynolds, 1998). Validityas such will always be a relative concept becausethe validity of an interpretation will vary accord-ing to the purpose for which test scores are beingused, the types of individuals or populations be-ing examined, and the specific interpretationsbeing made.

The Standards (AERA et al., 1999, pp. 11–17)suggests a five-category scheme for organiz-ing sources of evidence to evaluate proposedinterpretations of test scores, although clearlyrecognizing that other organizational systemsmay be appropriate. What follows is a sum-mary of the currently available validity evidenceassociated with the RIAS/RIST scores as mea-sures of intelligence organized according to therecommendations just noted.



Evidence Based on Test Content

The Standards states that ‘‘test content refers tothe themes, wording, and format of the items,tasks, or questions . . . as well as guidelines . . . foradministration and scoring’’ (p. 11). In discussingthe types of evidence that may be appropriatein this context, the Standards concludes thatevidence related to test content may be logicalor empirical, and that expert judgments of therelationship between the test and the intendedconstructs to be assessed are appropriate.

Subtests with familiar formats were chosenfor the various intelligence and memory subtestsof the RIAS. Although the individual items ofthe RIAS are unique, each of the formats forthe four intelligence and two memory taskshas a long history in the field of psychologicalassessment. Tasks similar in format and cognitiveprocessing requirements to What’s Missing andVerbal Reasoning, for example, can be traced tothe early efforts of Alfred Binet at the turn ofthe twentieth century (Binet & Simon, 1905)and also to Yerkes (1917) and his team ofpsychologists, who built aptitude measures forthe U.S. military early in the last century. Verbalanalogies such as those represented on VerbalReasoning are present on many tests of verbalability, including early and modern versions ofthe SAT and GRE examinations, the MillerAnalogies Test (Miller, 1926), as well as manyindividually administered intelligence tests (e.g.,McCarthy, 1972). Item formats corresponding tothose of Guess What and Odd Item Out appearshortly after these early types and are commonformats on various tests of general informationand matrix reasoning. Validity evidence for theseitem formats as measures of intellectual skill isvoluminous and reviewed in a variety of sources,the most readily accessible being Eliot and Smith(1983), Kamphaus (2001), Kaufman (1990, 1994),and Sattler (2001).

Likewise, the two memory tasks chosen forinclusion in the RIAS have long, rich historiesin the field of psychological testing, dating backto the 1920s and 1930s (e.g., Kamphaus, 2001;Reynolds & Bigler, 1994). Over many decades,

similar tasks have been included on a varietyof tests designed to measure memory function(see Benton, 1974; Reynolds & Bigler, 1994;Wechsler, 1997b).

Additionally, during the first item tryout, apanel of psychologists representing a variety ofethnic, cultural, and linguistic groups, all with ex-perience in assessment, reviewed all RIAS itemsfor appropriateness as measures of their respec-tive constructs and for applicability across variousU.S. cultures. Items questioned or faulted by thispanel were eliminated from the RIAS item poolor were revised. Another panel of five psycholo-gists with doctoral degrees in school psychology,clinical psychology, clinical neuropsychology,and measurement also reviewed all items in theitem pool for appropriateness. Items questionedor found faulty by this group were either elimi-nated outright or modified.

Evidence Based on Response Processes

Evidence based on the response processes of thetasks is concerned with the fit between the na-ture of the performance or actions in which theexaminee is actually engaged and the constructsbeing assessed. The four RIAS intelligence sub-tests are designed to measure general intelligencein the verbal and nonverbal domains. As such, thetasks are complex and require the integration ofmultiple cognitive skills, thereby avoiding con-tamination by irrelevant, noncognitive responseprocesses.

Because of their relationship to crystallized in-telligence, the two verbal subtests invoke vocab-ulary and language comprehension. However,clearly academic or purely acquired skills suchas reading are avoided. The response process re-quires integration of language and some generalknowledge to deduce relationships; only min-imal expressive language is required. One- ortwo-word responses are acceptable for virtuallyall items. The response process also is not con-taminated by nonintellectual processes such asmotor acuity, speed, and coordination. Rather,problem solving through the processes of deduc-tive and inductive reasoning is emphasized.



Likewise, the two nonverbal tasks avoid con-tamination by extraneous variables such as motoracuity, speed, and coordination. Examinees havethe option of responding by pointing or with aone- or two-word verbal indication of the cor-rect answer. As with any nonverbal task, someexaminees might attempt to use verbal encodingto solve these tasks, and some will do so success-fully. However, the tasks themselves are largelyspatial and are known to be more affected byright—rather than by left—hemisphere impair-ment (e.g., Joseph, 1996; Reynolds, Castillo, &Horton, 2008), a finding that supports the lack ofverbal domination in strategies for solving suchproblems. When examiners suspect that verbalstrategies are being applied, they should questionthe examinee regarding his or her strategies afterall the subtests have been administered in order togain relevant information for the interpretationof the examinee’s test performance.

Response processes of the two RIAS memorysubtests also avoid contamination from readingand various aspects of motor skills. Althoughgood language skills undoubtedly facilitate ver-bal memory, it is not the dominant skill involved.The RIAS memory tasks are very straightfor-ward, with response processes that coincide withtheir content domain—verbal in verbal memoryand nonverbal in nonverbal memory. Even so, inthe latter case, examinees who have severe mo-tor problems may give a verbal indication of theanswer they have selected.

Evidence Based on Internal Structure

Analyses of the internal structure of a test pro-vide information about the interrelationships ofthe items or subtests of a larger test and canreveal how this structure might conform to thehypothesized construct(s) being assessed (AERA,APA, & NCME, 1999). Such an examinationof evidence requires both empirical and logicalanalyses. Here, evidence from studies of internalconsistency and studies of the factor structure arereviewed in the context of proposed interpreta-tions of RIAS scores.

Evidence based on the internal structure ofthe RIAS is provided from two sources—itemcoherence (or internal consistency) and factoranalyses of the intercorrelations of the subtests.The internal consistency evidence has beenreviewed in the section on the reliability of testscores derived from the RIAS, and the evidencefor coherence is certainly strong for each ofthe subtests as well as the composite scores.Factor analysis is another method of examiningthe internal structure of a scale that lends itselfto assessing the validity of recommended scoreinterpretations.

Two methods of factor analysis have beenapplied to the intercorrelation matrix of the RIASsubtests, first with only the four intelligencesubtests examined and then with all six subtestsexamined under both techniques of analysis.Exploratory analyses were undertaken first andwere followed by a set of confirmatory analyses toassess the relative goodness-of-fit of the chosenexploratory results to mathematically optimalmodels.

For purposes of factor analyses of the RIASsubtest’s intercorrelations, the sample was di-vided into five age groups (rather than one-yearinterval groups) to enhance the stability and thegeneralizability of the factor analyses of the RIASscores. These age groupings reflect common de-velopmental stages. The five age groupings wereearly childhood, ages 3–5; childhood, ages 6–11;adolescence, ages 12–18; adulthood, ages 19–54;and senior adulthood, ages 55–94.

When the two-factor and three-factor solu-tions were subsequently obtained, the two-factorvarimax solution made the most psychologicaland psychometric sense both for the set of fourintelligence subtests and for all six RIAS sub-tests. In the four-subtest three-factor solution,no variables consistently defined the third factoracross the four age groupings. In the six-subtestthree-factor solutions, singlet factors (i.e., factorswith a single salient loading) appeared, com-monly representing a memory subtest; What’sMissing tended to behave in an unstable man-ner as well. Summaries of the two-factor varimax



solutions are presented in Reynolds and Kam-phaus, 2003 for the four-subtest and six-subtestanalyses. In each case, the first, unrotated factoris a representation of g as measured by the RIAS.

The g factor of the RIAS is quite strong.Only the intelligence subtests have loadings thatreach into the .70s and .80s. All four intelligencesubtests are good measures of g; however, ofthe four, the verbal subtests are the strongest.Odd Item Out and What’s Missing follow, thelatter being the weakest measure of g among thefour intelligence subtests. The strength of thefirst unrotated factor is, however, indisputableand indicates that first and foremost the RIASintelligence subtests are measures of g and thatthe strongest interpretive support is given, inthese analyses, to the CIX.

At the same time, the varimax rotation ofthe two-factor solution clearly delineates twocomponents of the construct of g among theRIAS intelligence subtests. For every age group,the verbal and nonverbal subtests clearly breakinto two distinct factors that coincide withtheir respective indexes, VIX and NIX. Thesix-subtest solution also breaks along contentdimensions, with Verbal Memory joining thetwo verbal intelligence subtests on the firstrotated factor and Nonverbal Memory joiningthe two nonverbal intelligence subtests on thesecond rotated factor. However, in view of theanalysis of the content and response processesas well as other evidence presented throughoutthis manual, we continue to believe that theseparation of the Verbal Memory and NonverbalMemory subtests into a separate memory index(i.e., CMX) is more than justified. Memory isclearly a component of intelligence. The twomemory tasks that were chosen for the RIASare relatively complex, and both are strongpredictors of broader composites of verbal andnonverbal memory (Reynolds & Bigler, 1994),characteristics that are at once an asset and aliability. Although these two memory tasks aregood measures of overall, or general, memoryskill, they tend to correlate more highly withintelligence test scores than do very simple,

confrontational measures of working memory,such as forward digit repetition. Given thepurpose of providing a highly reliable assessmentof overall memory skill, such a compromise iswarranted.

The stability of the two-factor solution acrossother relevant nominal groupings and the po-tential for cultural bias in the internal structureof the RIAS were also assessed. For this pur-pose, the factor analyses were also calculatedseparately for males and females and for whitesand African Americans, according to recom-mendations and procedures outlined in detailby Reynolds (2000). Tables 6.3 through 6.6 inReynolds and Kamphaus (2003) present these re-sults for each comparison. The similarity of thefactor-analytic results across gender and acrossethnicity was also assessed. Two indexes of fac-torial similarity were calculated for the visuallymatched rotated factors and for the first unro-tated factor, the coefficient of congruence (rc)and Cattell’s (1978) salient variable similarity in-dex (s), as recommended in several sources (e.g.,Reynolds, 2000). In all cases, the factor structureof the RIAS was found to be highly consistentacross gender and ethnicity.

Subsequent to the exploratory factor analy-ses, a series of confirmatory factor analyses wereconducted to examine the fit of our choice ofexploratory analyses to a more purely mathemat-ical model. Based on the theoretical views of thestructure of the RIAS discussed earlier in thischapter, three theoretical models were tested.The models were defined as follows: Model 1,The RIAS is a measure of general intellectualabilities; Model 2, The RIAS is a measure ofverbal and nonverbal abilities; and Model 3, TheRIAS is a measure of verbal, nonverbal, andmemory abilities.

The resulting chi square (χ2), residuals, rootmean square error of approximation (RMSEA),and other model-fit statistics were then comparedusing the LISREL-VI program (Joreskog &Sorbom, 1987) to test the relative fit of thethree models. Model 1, general intelligence,clearly fit better when only the four intelligence



subtests were included (χ2 = 8.17–20.57 andRMSEA ranging from .10 to .14 depending onthe age range studied) than when six subtestswere included. Although these models suggest,much in the same way as the exploratory factoranalyses showed, that the RIAS is dominated bya large first factor, the RMSEAs were still highenough to suggest that Models 2 and 3 should beexplored.

Model 2 was a very good fit to the dataparticularly when four subtests were included inthe model versus six subtests. For the model thatincluded four subtests, the chi-square values werebetween .22 and 1.49. Similarly, the RMSEAswere less than .01 for the first four age groups(i.e., 3–54 years) and .04 for ages 55 and older,values suggesting that two factors explainedvirtually all of the variance between the foursubtests. These findings indicated that the fit ofa three-factor model was not likely to be as good,and in fact Model 3 with six subtests included (χ2

= 14.14–37.48 and RMSEA ranging from .01 to.09) did not fit nearly as well as Model 2.

In summary, the results of the confirmatoryfactor analyses suggest that the CIX, VIX, andNIX possess evidence of factorial validity. TheCMX in particular requires further research witha variety of clinical and nonclinical samples.

Evidence Based on Relations with Other(External) Variables

Another important area in the validation processis the evaluation of the relationship of scores onthe instrument of interest to variables that areexternal to the test itself. A variety of externalvariables were chosen for investigation with theRIAS, including developmental variables (e.g.,age), demographic variables, relations with othertests, and clinical status.

Developmental Trends

As a developmental construct, intellectual abilitygrows rapidly in the early years, begins to plateauin the teens but with some continued growth(particularly in verbal domains), and eventually

declines in the older years. This decline generallybegins sooner and is more dramatic for nonver-bal, or fluid, intelligence (Kaufman, McLean,Kaufman-Packer, & Reynolds, 1991; Kaufman,Reynolds, & McLean, 1989; Reynolds, Chastain,Kaufman, & McLean, 1987). If raw scores on thetasks of the RIAS reflect such a developmentalprocess or attribute, then relationships with ageshould be evident. The relationship between age(a variable external to the RIAS) and performanceon the RIAS was investigated in two ways.

First, the correlation between age and rawscore for each subtest was calculated for theprimary developmental stage, ages 3–18, forthe entire sample and for a variety of ethnicgroups and by gender. The correlations for allgroups are uniformly large, typically exceeding.80 and demonstrating that raw scores on theRIAS increase with age and in a relatively con-stant manner across subtests. When the valuesfor each subtest are compared across nominalgroupings (gender and ethnicity), there are verysmall differences observed, none being statisti-cally significant or clinically meaningful, indicat-ing common developmental patterns across allgroups to the extent correlational data can re-veal such trends. Second, to examine the issue inmore detail, lifespan developmental curves weregenerated for each subtest from ages 3 through94. These curves are presented in the RIAS Man-ual (Reynolds & Kamphaus, 2003) and show aconsistent pattern of score increases and declines(with aging) across all groups.

Correlations with the Wechsler Scales

Measures of intelligence generally should corre-late well with one another if they are measuringg and related constructs. Thus, to understand anew measure and its appropriate interpretations,it is instructive to assess the relationship of thenew measure to other measures of intelligence.For children (ages 6–16), the best known andmost widely researched scale over the years andone that has maintained a reasonably consistentstructure is the Wechsler Intelligence Scale for



Children, in its third edition (WISC-III; Wech-sler, 1991) when the RIAS was published, andnow in its fourth.

WISC-IV Edwards (2006) reported on corre-lations between the RIAS and the WISC-IVfactor indexes for two referral samples. In thefirst group of 83 children ages 7–9, the correla-tions reported were: RIAS VIX-WISC-IV VCIr = .83, NIX-PRI r = .42, CIX-WMI r = .58,CIX-PSI r = .36, and CIX-FSIQ r = .75. Ina sample of 121 children and adolescents ages6–16, Edwards reported the same set of corre-lations as: RIAS VIX-WISC-IV VCI r = .83,NIX-PRI r = .54, CIX-WMI r = .62, CIX-PSIr = .45, and CIX-FSIQ r= .79.

In another study, Edwards and Paulin (2007)reported on another referral sample, of 26 boysand 24 girls ranging in age from 6–12 years,administered both the RIAS and the WISC-IV.In this study, some additional correlations werereported and the correlations were generallylarger as well. Edwards and Paulin reportedcorrelations between related constructs of: RIASVIX-WISC-IV VCI r = .90, NIX-PRI r = .72,and CIX-FSIQ r = .90. They also reporteda correlation between the RIAS CIX and theWISC-IV GAI of .90. The RIAS compositescores are thus seen to be very highly correlatedwith the cognate composite scores on theWISC-III and WISC-IV, except for the PRI,which on the Wechsler Scales is confoundedby motor and speed of performance—factorseliminated as confounds on the RIAS and factorsless confounding in the assessment of adults.

WISC-III Table 5.4 summarizes RIAS cor-relations with the WISC-III and WAIS-III.The RIAS indexes all correlate highly with theWISC-III Full Scale IQ (FSIQ), with correla-tions ranging from a low of .60 (NIX to FSIQ)to a high of .78 (VIX to FSIQ). The relativelylower correlations between the RIAS NIX andthe WISC-III IQs are most likely attributable tothe increased emphasis on motor and languageskills on the WISC-III Performance IQ (PIQ)

relative to that in the RIAS. The WISC-III PIQsubtests also rely on sequencing (e.g., Picture Ar-rangement and Coding), whereas the RIAS hasno subtest in which sequencing is crucial to thetask. Speed of performance and motor skills arealso less important on the RIAS, taking awayconfounds in the measurement of nonverbal in-telligence present on the Wechsler Scales at thechild level. For adults, these are far less salientissues. With this one exception, the pattern ofcorrelations is much as was predicted, namely,the highest correlations were between those as-pects of the tests most closely associated with g(from their respective factor analyses).

WAIS-III A group of 31 adults were ad-ministered the RIAS and the Wechsler AdultIntelligence Scale—Third Edition (WAIS-III;Wechsler, 1997a) in a counterbalanced design.All but two of the correlations exceed .70; theVIX-PIQ correlation was the lowest at .61. All ofthe RIAS indexes correlate at or above .70 withthe WAIS-III FSIQ (see Table 5.3).

Correlations with Measures of AcademicAchievement

One of the major reasons for the development ofthe early, individually administered intelligencetests was to predict academic achievement levels.Intelligence tests have done well as predictorsof school learning, with typical correlationsin the mid .50s and .60s (see Kamphaus,2001, and Sattler, 2001, for summaries). Toevaluate the relationship between the RIASand academic achievement, 78 children andadolescents were administered the RIAS and theWechsler Individual Achievement Test (WIAT;Wechsler, 1992).

School learning is fundamentally alanguage-related task, and this fact is clearlyevident in the data presented in Reynolds andKamphaus (2003). Although all of the RIASindexes correlate well with all of the WIATcomposite scores, the highest correlations areconsistently between the VIX and CIX andthe WIAT composites. These correlations are



TABLE 5.4 Correlations between RIAS Indexes and WISC-III and WAIS-III IQs

RIAS Index

WJ-III Score VIX NIX CIX

Reading Composite .88 (.85) .74 (.75) .88 (.86)Basic Reading Composite .79 (.74) .59 (.60) .76 (.73)Comprehension .88 (.86) .74 (.75) .87 (.85)Low Identification .85 (.81) .62 (.64) .81 (.78)Reading Fluency .37 (.33) .06 (.06) .26 (.24)Passage Comprehension .85 (.81) .71 (.72) .85 (.82)Math Composite .80 (.75) .70 (.71) .82 (.79)Calculation Composite .72 (.67) .57 (.58) .72 (.68)Math Reasoning .86 (.83) .85 (.86) .90 (.88)Calculation .76 (.72) .63 (.64) .76 (.72)Math Fluency .68 (.63) .55 (.56) .66 (.63)Applied Problems .79 (.75) .76 (.77) .84 (.81)Written Language Composite .56 (.51) .33 (.34) .54 (.50)Written Expression Composite .58 (.53) .55 (.57) .64 (.60)Writing Fluency .50 (.45) .34 (.35) .49 (.45)RIAS Mean 94 99 96RIAS SD 17 14 16

NOTE: N varies by subtests. Not all were administered to all children. Total N = 121, but most correlations in this table arebased on Ns from 50 to 81. Corrections are for the population estimated of the SDs of the scores.SOURCE: Data courtesy of Charles Szasz of West Virginia.

predominantly in the .60s and .70s, indicatingthat the RIAS has strong predictive value foreducational achievement as measured by theWIAT.

Reynolds and Kamphaus (2007b) report on acorrelational study between the RIAS and theWRAT4, based on a sample of 410 nonreferredindividuals ranging in age from 5 to 24 years. Themajority of the 16 correlations reported werein the .50s and .60s, with the VIX correlatinghighest overall with the WRAT4 achievementvariables of Sentence Comprehension, WordReading, and the Reading Composite. Corre-lations with the Math Computation scores weremostly in the .40s. The RIAS CMX correlated

in the .40s and .50s with all of the WRAT4 vari-ables. When the RIAS Total Battery scores areused, the correlations increase, some reachinginto the .70s. In another study, Reynolds andKamphaus (2007c), including over 2,000 chil-dren and adults, reported the RIAS to correlatein the .40s and .50s with a simple measure ofrapid word calling.

Beaujean et al. (2006) reported the RIASto be a significant predictor of SAT scores aswell. Correlations of .58 were determined in avery-high-achieving sample (mean SAT > 1.5SDs above the population mean) between theCIX and the SAT using two methods of centeringSAT scores as described therein.



In an unpublished study conducted for aschool district, Szasz (2006, personal communi-cation) reported correlations between the RIASand the Woodcock-Johnson-III for a sampleof 121 students referred for potential specialeducation placement. Table 5.4 provides thecorrelation table from this study. In general,the correlations for the RIAS and the WJ-IIIscores trend higher than with other achievementtests, some reaching into the high .80s. Indeed,most of the correlations with reading achieve-ment, especially complex reading tasks such ascomprehension, are quite high. These correla-tions indicate the RIAS has strong predictivevalue for educational achievement as measuredby the WJ-III as well as a wide range of otheracademic achievement tests and across all agelevels, predicting achievement well for children,adolescents, and adults.

Performance of Clinical Groups

Examination of performance on the RIASby groups of clinically identified individuals(wherein the RIAS was not a component ofthe process of clinical diagnosis) can also beinstructive. For example, individuals with mentalretardation (MR), dementia, and related cog-nitive problems associated with intellectual im-pairment should earn lower scores on the RIASthan the normally functioning examinees in theRIAS standardization sample. In interpretingsuch scores of preselected samples, however, es-pecially when those samples are selected on thebasis of extreme scores on a cognitive measure,one must always consider the problem of re-gression to the mean on a second testing. Thus,scores obtained from a sample with MR will typ-ically be higher on such a testing, but the scoresshould still be well below the population mean.

During the standardization of the RIAS, 15different clinical groups were identified, and theirscores on the RIAS analyzed to supplement thevalidation of the interpretation of RIAS scores asreflecting intelligence and memory processes. Ineach instance, the primary diagnosis given by theagency serving the examinee was accepted, and

no independent review or diagnosis was under-taken. The samples included those with traumaticbrain injury (TBI), dementia, and stroke/CVA,mental retardation, deaf and hearing impaired,learning disabilities, and a variety of others.The various impairments represented a varietyof organic deficits within each category, alongwith diffuse brain lesions. There were samplesof children with learning disabilities (LD) andattention-deficit/hyperactivity disorder (ADHD)as well as the cognate adult samples, which camefrom more than one source. In reviewing theoutcomes of testing with the RIAS with these 15clinical samples, Reynolds and Kamphaus con-cluded that all of the various clinical groups inthese studies demonstrated some levels of devi-ation from the population mean on the RIAS.Although most deviations are small decrements,as is commonly found in the literature, the sam-ples with more severe disorders showed greaterdecrements in performance. Again, however, thepurpose for presenting these data is not to makedefinitive statements about these clinical groupsbut to describe how the RIAS scores function foreach group. Moreover, these data are preliminaryand not definitive as to any score patterns on theRIAS that may emerge for clinical groups. Repli-cation with larger and more carefully definedsamples will be necessary before firm conclusionscan be drawn. The evidence thus far is quite sup-portive because the score patterns do conformwell to known patterns in the related literatureand can be reviewed in detail in Reynolds andKamphaus (2003), Chapter 6.

Evidence Based on the Consequences ofTesting

This area of the validation process is the mostcontroversial of all the aspects of the process aspresented in the standards. It is most applica-ble to tests designed for selection and may dealwith issues of bias or loss of opportunity. Howthese applications should be evaluated for clini-cal diagnostic tests is largely unclear. However,accurate diagnosis might be one anticipated con-sequence of testing and should be the key to



evaluating the ‘‘consequential’’ validity of a clin-ical instrument. The evidence reviewed in thepreceding sections demonstrates the ability ofthe RIAS to provide an accurate estimate of in-tellectual ability and certain memory skills and todo so accurately across such nominal groupingsas gender and ethnicity. Cultural biases in theformat and content of tests, when apparent, havealso been found to produce undue consequencesof testing. Evidence pointing toward a lack ofcultural bias in the RIAS is extensive for maleand female examinees, whites, African Ameri-cans, and Hispanic Americans. This evidence hasbeen reviewed previously in this chapter and isdiscussed in detail in Chapters 4–6 of the RIASManual. Studies of potential cultural bias in theRIAS items were extensive in both objective andsubjective formats and resulted in the removal ofmany items and modification of others. Evalua-tion of the mean scores of different ethnic groupsfrom the RIAS standardization sample indicatedthat mean score differences across groups forwhites, African Americans, and Hispanic Ameri-cans were about half the size typically reported inthe research literature for traditional intelligencetests such as the Wechsler and the Binet series.

In evaluating evidence for test score interpre-tations, examiners must always consider theirpurposes for using objective tests. Evidenceclearly supports the use of the constructs repre-sented on the RIAS. The potential consequencesof knowing how an individual’s performancecompares to that of others are many and com-plex and not always anticipated. The RIAS wasdesigned to eliminate or minimize any culturalbiases in the assessment of intelligence and mem-ory for individuals reared and educated in theUnited States (who are fluent in the Englishlanguage). The data available to date indicatethe RIAS precludes undue consequences towardminority groups and women who fit the tar-get population. Examiners must nevertheless actwisely, consider the need for objective testingof intelligence and memory, and work to mini-mize or eliminate unsupported interpretations ofscores on such tests.

APPLICATIONS OF THE RIAS

As a measure of intelligence, the RIAS is appro-priate for a wide array of purposes and should beuseful when assessment of an examinee’s intel-lectual level is needed. The RIAS will be usefulwith preschool and school-aged children for pur-poses of diagnosis and educational placementand for diagnosis of various forms of childhoodpsychopathology (especially developmental dis-orders) where intellectual functioning is an issue.Diagnosis of specific disorders, such as mentalretardation, learning disabilities, the various de-mentias, and the effects of central nervous systeminjury or compromise, most often calls for theuse of an intelligence test as a component of pa-tient evaluation, and the RIAS is appropriate forsuch applications. Clinicians who perform gen-eral clinical and neuropsychological evaluationswill find the RIAS very useful when a measureof intelligence is needed. Practitioners will alsofind the RIAS useful in disability determinationsunder various state and federal programs, suchas the Social Security Administration’s disabilityprogram and Section 504 regulations.

Intelligence tests, including the RIAS, havemany additional uses and perform certain func-tions particularly well. Performance on an in-telligence test predicts a host of outcomes. Forexample, the relationship between intelligenceand persistence and success in school and re-lated academic environments is one of the mostwidely, soundly documented relationships in allof psychology (Jensen, 1998). Intellectual levelpredicts a plethora of other quite varied outcomesas well, such as job performance in numerousoccupations and professions and recovery fol-lowing traumatic brain injury (Golden, Zillmer,& Spiers, 1992).

Although the RIAS is rapid to administer rel-ative to the majority of other comprehensivemeasures of intelligence, it is not an abbreviatedmeasure or a short form of intellectual assess-ment. The RIAS is a comprehensive measure ofverbal and nonverbal intelligence and of general



intelligence, providing the same level of useful in-formation often gleaned from other intelligencetests much greater in length. When the mem-ory subtests are also administered, the RIAS canprovide even more useful information than typ-ical intelligence tests currently used. The majorclinical uses of intelligence tests are generallyclassification (most commonly, diagnostic) andselection. The RIAS has broad applicability ineach of these areas. Some of the more commonuses in these areas are discussed here.

Learning DisabilityFor the evaluation of a learning disability,assessment of intelligence is a common activity.However, when children and adults are evaluatedfor the possible presence of a learning disability,both verbal and nonverbal intelligence shouldbe assessed. Individuals with a learning disabilitymay have spuriously deflated IQ estimates in oneor the other domain due to the learning disabilityitself. Lower verbal ability is the most commonin the school population and among adjudicateddelinquents (Kaufman, 1994). However, theconcept of the nonverbal learning disabilityis gaining momentum. For individuals withnonverbal learning disability, verbal ability willoften exceed nonverbal ability. The assessmentof functioning in both areas is important andthe RIAS provides a reliable assessment of thesedomains as well as a composite intelligence index.The three RIAS intelligence index scores (i.e.,VIX, NIX, and CIX) allow for an accurateestimation of intelligence simultaneously withthe diagnosis of learning disability.

Mental RetardationMost definitions, including those of the Amer-ican Association on Mental Retardation (2002)and the Diagnostic and Statistical Manual of Men-tal Disorders—Fourth Edition, Text Revision(DSM-IV-TR; American Psychiatric Associa-tion, 2000), require the administration of anindividually administered test of intelligence for

diagnosis of mental retardation. The RIAS is ap-plicable to the diagnosis of mental retardationfor which the evaluation of verbal and nonver-bal intelligence as well as adaptive functioningis necessary. Mental retardation is a pervasiveintellectual problem and not limited to seriousproblems in only the verbal or nonverbal do-main. The range of scores available on the RIASwill also make it useful in distinguishing levelsof severity of mental retardation. Lower levels offunctioning such as profound mental retardationare difficult to assess accurately on nearly all testsof intelligence, and this is likewise true of theRIAS. Although normed on children as youngas 3 years of age, the RIAS also has limited dis-criminative ability below mild levels of mentalretardation in the 3-year-old age group.

Intellectual GiftednessMany definitions of giftedness include referenceto superior levels of performance on measuresof intelligence. Here again, measures of boththe verbal and nonverbal domains are useful dueto the influences of schooling and educationalopportunity on verbal intelligence. The RIASalso has eliminated issues of motor skill andtiming to enhance the assessment of intelligencefree of irrelevant confounds, which makes itespecially useful for identifying high-IQ children(see Bruegeman, Reynolds, & Kamphaus, 2006,for a detailed discussion of these issues). Therange of index scores available on the RIAS (upto index scores of 160, four SDs above the mean)is adequate at all ages for identifying personswith significantly above-average levels of overallintellectual function as well as in the verbal andnonverbal domain.

Visual ImpairmentIndividuals with significant visual impairmentsthat are uncorrected should not be administeredthe RIAS nonverbal subtests. The verbal subtests,however, should be useful with such individuals.



Hearing DeficiencyIndividuals with significant hearing impairmentmay require examination by specially trainedexaminers. In such instances, the examinermust judge whether the particular individual’simpairment makes any portion of the RIAS moreor less useful. Because of the extreme variabilityin the levels and types of hearing impairmentsand the communication skills (both form andlevel) of individuals with significant hearingimpairments, no general rule for applicabilityis offered. Judgments of applicability shouldbe made on a case-by-case basis by examinerswith special training and experience with thesepopulations.

Physical/Orthopedic ImpairmentThe RIAS will be particularly useful in theevaluation of intellectual functioning among in-dividuals with any significant degree of physicalor motor impairment. The RIAS has no realdemands for speed or accuracy of fine motormovements. If necessary, the pointing responsesby the examinee on the RIAS nonverbal taskscan all be replaced with simple verbal responses,designating the location of the chosen response.It is, however, very important for examiners tohave knowledge of the physical impairments ofany examinee and to make any necessary modifi-cations in the testing environment, doing so in amanner consistent with appropriate professionalstandards (e.g., Standards for Educational andPsychological Testing; American EducationalResearch Association [AERA], American Psycho-logical Association [APA], & National Councilon Measurement in Education [NCME], 1999).

Neuropsychological ImpairmentThe RIAS can form one central component of theneuropsychological evaluation of children andadults. Assessment of general intelligence levelsis important in establishing comparison levelsfor the interpretation of the highly specific tasks

that are commonly used by neuropsychologists.The RIAS has been shown to be sensitive tointellectual decline in the various dementiasand in cerebrovascular accident and traumaticbrain injury (Reynolds & Kamphaus, 2003). Thebrevity of the RIAS also makes it attractive inthe context of neuropsychological assessment, inview of the extensive length of such evaluations.The RIAS can provide strong and efficientmeasurement of general intelligence and createmore time for the neuropsychologist to assessthe many specific cognitive and psychomotorfunctions that are desirable components of thecomprehensive neuropsychological examination.The RIAS memory subtests also can provide apreliminary indication of the need for a moreextensive memory evaluation.

Memory ImpairmentThe information gleaned from evaluating mem-ory functions can provide valuable clinical infor-mation above and beyond what is traditionallyassessed using IQ measures. Memory is gener-ally recognized as a focal or discrete subset ofcognitive functions and as such is often quite vul-nerable to central nervous system (CNS) traumaand various other CNS events. Disturbances ofmemory and attention are the two most frequentcomplaints of children and adults following trau-matic brain injury at all levels of severity as wellas other forms of CNS compromise (e.g., viralmeningitis, AIDS-dementia complex, and othersystemic insults). Therefore, it is not unusual formemory functioning to be affected even whenthere is little or no impact on general intellectualability. The memory measures on the RIAS offerclinicians valuable assessment tools with whichto evaluate recent or more immediate memoryfunctioning in both the auditory (i.e., VerbalMemory) and visual (i.e., Nonverbal Memory)modalities.

For both children and adults, memory deficitscan occur in a broad range of cognitive andpsychiatric disorders. For example, memory



dysfunction is often seen in individuals diag-nosed with learning disability (LD), attentiondeficit/hyperactivity disorder (ADHD), trau-matic brain injury, and stroke/cerebrovascularaccident (CVA). Many of the common neurode-velopmental and genetic disorders of childrenhave memory and/or attention problems as anelement of the symptom complex (Goldstein &Reynolds, 1999; Reynolds & Fletcher-Janzen,1997). One significant value of including mem-ory measures in an assessment is that it allows theexaminer to expand the scope of available infor-mation to include clinically meaningful measuresthat pertain to the examinee’s current func-tioning in daily activities such as learning andmemory.

The RIAS memory measures add importantinformation in terms of the ecological validity ofthe evaluation. For example, the ability to focusattention and learn new information is criticalfor academic success. The verbal and nonverbalmemory measures can give the examiner a verygood preliminary indication of a student’s abilityto learn and remember material. This can lead tomore meaningful and better informed evaluationrecommendations and remediation strategies.

It is important to note that the RIAS doesnot provide memory measures to assess delayedmemory. Delayed memory functioning (e.g.,recall and retrieval) can be affected adverselyeven though an examinee performs at expectedlevels on immediate or short-term memorymeasures. As such, the RIAS does not provide acomprehensive measure of all memory domains.

Emotional DisturbanceIndividuals with various forms of emotionaland/or psychotic disturbance (e.g., depression,

schizophrenia) may exhibit cognitive impair-ments to varying degrees. Often clinicians do notassess the intelligence of such individuals due tothe time required. The RIAS offers the cliniciana more efficient means of gathering informationon the psychometric intelligence of individualswith emotional problems. Often, especially in thecase of children, other cognitive disorders, suchas LD, may contribute to the emotional prob-lems observed. An intellectual assessment can behelpful in identifying cognitive and intellectualdifficulties in such cases. Level of intellectualfunction may also influence the choice of ap-propriate treatments. For example, cognitive be-havioral therapies and analytic/insight-orientedapproaches to treatment would not be appro-priate for individuals with borderline levels ofintellectual function and, in some cases, evenlow average intelligence. Knowledge of the in-tellectual level of a referred child or adult withemotional problems as the primary complaint canprovide a broader context for treatment, and theRIAS seems especially well suited for use in suchreferrals.

Job PerformanceIn personnel settings, IQ tests are sometimesused to predict success in job training pro-grams and, in other instances, lower limits areset on IQ levels for specific jobs. The RIASand the RIST are strong predictors of aca-demic performance, and the tasks involved andconstructs assessed on these instruments matchup well with known predictors of job perfor-mance in the form of other IQ tests. Whenintelligence level is a question in such situa-tions, the RIAS and the RIST are appropriatechoices.

SUMMARY OF A CASE STUDY

This case summary is an illustration of the use of RIAS results. While a much more extensiveevaluation was conducted, this section highlights only the findings most relevant to the use ofthe RIAS and does not present the entire case due to space limitations.



Brad is a 9-year-old boy referred for difficulties in learning to read. Brad was born followinga full-term, uneventful pregnancy. Currently, Brad is in good physical health. Overall, Bradpresented himself as well behaved and cooperative during testing. He seemed engaged in mosttesting procedures and appeared to put forth his best effort. Results of the evaluation are viewedas valid estimates of Brad’s intellectual abilities, academic achievement, and social-emotionaladjustment. Achievement testing as well as consultation with the school confirmed grade-levelperformance in math but very poor reading skills.

On testing with the RIAS, Brad earned a Composite Intelligence Index (CIX) of 88. On theRIAS, this level of performance falls within the range of scores designated as below average andexceeds the performance of 21% of individuals at Brad’s age. Brad earned a Verbal IntelligenceIndex (VIX) of 78, which falls within the moderately below-average range of verbal intelligenceskills and exceeds the performance of 7% of individuals Brad’s age. Brad earned a NonverbalIntelligence Index (NIX) of 103, which falls within the average range of nonverbal intelligenceskills and exceeds the performance of 58% of individuals Brad’s age. Brad earned a CompositeMemory Index (CMX) of 75, which falls within the moderately below-average range of workingmemory skills. This exceeds the performance of 5% of individuals Brad’s age. On testing withthe RIAS, Brad earned a Total Test Battery (TTB) score of 80. This level of performance on theRIAS falls within the range of scores designated as below average and exceeds the performanceof 9% of individuals at Brad’s age. Brad’s Total Verbal Battery (TVB) score of 75 falls withinthe range of scores designated as moderately below average and exceeds the performance of 5%of individuals his age. The chances are 90 out of 100 that Brad’s true TVB falls within the rangeof scores from 71 to 81.

Brad’s Total Nonverbal Battery (TNB) score of 91 falls within the range of scores designatedas average and exceeds the performance of 27% of individuals his age.

RIAS Discrepancy Score Summary Table

Discrepancy Score Score DifferenceStatisticallySignificant?

Prevalence inStandardizationSample

VIX < NIX 25 yes 9.20%CIX > CMX 13 yes 40.00%VRM > NVM 1 no 94.60%TVB < TNB 16 yes 28.30%

VIX is the Verbal Intelligence Index, NIX is the Nonverbal Intelligence Index, CIX is the CompositeIntelligence Index, CMX is the Composite Memory Index, VRM is the Verbal Memory Subtest, NVM isthe Nonverbal Memory Subtest, TVB is the Total Verbal Battery Index, and TNB is the Total NonverbalBattery Index.

Discrepancy Norm-Referenced Interpretations

Although the CIX is a good estimate of Brad’s general intelligence, a statistically significantdiscrepancy exists between his NIX of 103 and his VIX of 78, demonstrating better-developednonverbal intelligence or spatial abilities. The magnitude of the difference observed betweenthese two scores is potentially important and should be considered when drawing conclusions



about Brad’s current status. A difference of this size is relatively uncommon, occurring in only9% of cases in the general population. In such cases, interpretation of the CIX or generalintelligence score may be of less value than viewing Brad’s verbal and nonverbal abilitiesseparately.

When compared to Brad’s measured level of general intelligence as reflected in Brad’s CIX, itcan be seen that his CMX falls significantly below his CIX. This result indicates that Brad is ableto engage in intellectual problem solving and general reasoning tasks at a level that significantlyexceeds his ability to use immediate recall and working memory functions. Although the sizeof the observed difference is reliable and indicates a real difference in these two cognitivedomains, the magnitude of the difference observed is relatively common, occurring in 40% ofthe population. Within the subtests making up the CMX, Brad’s performance was substantiallyequivalent on verbal and nonverbal memory tasks. This result indicates that Brad functionsabout equally well when called on to engage in recall or use working memory functions in eitherthe verbal or nonverbal domain.

School Feedback and Recommendations

Brad’s CIX score of 88 and TTB score of 80 indicate mild deficits in overall development ofgeneral intelligence relative to others at Brad’s age. Individuals earning general intelligencescores in this range frequently experience at least some difficulty acquiring information throughtraditional educational methods provided in the classroom setting.

The TTB measures the same general construct as the CIX with the exception that six testsare included rather than four. Evidence in the RIAS/RIST Professional Manual (Reynolds &Kamphaus, 2003) documents the equivalence of these two scores based on evidence that afirst-factor solution is defensible at all age levels of the RIAS whether four or six subtests areused. There also is evidence from a variety of intelligence tests to suggest the ‘‘indifference of theindicator.’’ In other words, general intelligence may be assessed using a variety of cognitive tests,providing further evidence that for most individuals the TTB and CIX will be interchangeable.There will be exceptions to this well-documented scientific finding, in the case of severe braininjury, for example, where significant memory impairment may be present, but these cases willbe exceptions rather than the rule.

Since most instructional programs presume at least average intellectual ability and involvelecture, note taking, and other typical instructional approaches, with the exception ofdemonstrative and repetitive methods commonly used with young children, difficulties inacquiring information when these methods are used is anticipated. Given Brad’s deficits, specialteaching methods might be considered, including special class placement for severe deficitsin general intellectual development. Teachers should prepare an individualized curriculumdesigned for students who learn at a slower rate than others of the same age and gradelevel. Alternative methods of instruction should be considered that involve the use of repeatedpractice, spaced practice, concrete examples, guided practice, and demonstrative techniques.Individuals with general intelligence scores in this range often benefit from repeated practiceapproaches to training because of problems with acquisition and long-term retrieval, as well as anindividualized instructional method that differs significantly from that of their age-mates. It alsowill be important to assist Brad in developing strategies for learning and studying. Although it isimportant for all students to know how to learn and not just what to learn, low scores on generalintelligence indices make the development of learning and study strategies through directinstruction even more important. If confirmed through further testing, co-occurring deficitsin adaptive behavior and behavioral problems should be added to the school interventionprogram.



Brad’s VIX score of 78 and TVB score of 75 indicate moderate deficits in the developmentof verbal intellect relative to others at Brad’s age. Verbal memory problems of varying severitylevels are commonly evident in this group of individuals, and special attention to Brad’s VRMscore is necessary.

Verbal ability is important for virtually every aspect of activity because language iskey to nearly all areas of human endeavor. A multitude of research investigations havedocumented the importance of verbal ability for predicting important life outcomes. Verbalability should be considered equivalent to the term crystallized intelligence. As assessed by theRIAS, verbal ability (like crystallized intelligence) is highly related to general intelligence,and as such its relationship to important life outcomes is easily correlated. Verbal abilityalso is the foundation for linguistic knowledge, which is necessary for many types oflearning.

With the exception of the early grades, along with Kindergarten and pre-K settings, schoolis principally a language-oriented task. Given Brad’s relative verbal deficits, special teachingmethods might be considered, including special class placement in the case of severe deficitsin verbal intellectual development. The examiner should also consider either conducting, ormaking a referral for, an evaluation for the presence of a language disorder. Alternative methodsof instruction that emphasize ‘‘show me’’ rather than ‘‘tell me’’ techniques, or as a minimumpair these two general approaches, are preferred.

Although linguistic stimulation likely cannot counteract the effects of verbal ability deficitsthat began in infancy or preschool years, verbal stimulation is still warranted to improveadaptation or at least prevent an individual from falling further behind peers. Verbal conceptand knowledge acquisition should continue to be emphasized. A simple word-for-the-dayprogram may be beneficial for some students. Verbal knowledge builders of all varieties maybe helpful, including defining words, writing book reports, a book reading program, and socialstudies and science courses that include writing and oral expression components. Alternatively,assistive technology (e.g., personal digital assistance devices, tape recorders, MP3 players, oriPods) may be used to enhance functioning in the face of the extensive verbal demands requiredfor making adequate academic progress.

In addition, teachers should rely more heavily on placing learning into the student’sexperiential context, giving it meaning and enabling Brad to visualize incorporating each newlylearned task or skill into his life experience. The use of visual aids should be encouraged andmade available to Brad whenever possible. Academic difficulties are most likely to occur inlanguage-related areas (e.g., the acquisition of reading), especially early phonics training. Theacquisition of comprehension skills also is aided when the verbal ability falls into this level by theuse of language experience approaches to reading, in particular. Frequent formal and informalassessment of Brad’s reading skills, as well as learning and study strategies (the latter with aninstrument, e.g., the School Motivation and Learning Strategies Inventory—SMALSI; Stroud& Reynolds, 2006) is recommended. This should be followed by careful direct instruction inareas of specific skill weaknesses and the use of high-interest, relevant materials. It also willbe important to assist Brad in developing strategies for learning and studying. Although it isimportant for all students to know how to learn and not just what to learn, low scores withinthe verbal intelligence domains make the development of learning and study strategies throughdirect instruction even more important.

Brad’s CMX of 75 falls within the moderately below-average range and indicates moderatedifficulties with recall of verbal and visual/spatial information relative to others Brad’s age.Individuals who score in this range frequently experience consternation in the acquisition ofnew information and may need some assistance in developing strategies for the day-to-day recallof verbal and visual/spatial information.



Students with deficits in memory ability relative to others their age can benefit fromspecial accommodations in the classroom and from instruction on developing better memorytechniques. The use of multiple modalities is typically recommended to increase retention, suchas routinely pairing visual/spatial stimuli with verbal stimuli in order to enhance recall. The useof lists, visual, oral language, and written language directions, and verbal and visual remindersmay be especially helpful. Repetition of information during learning is often helpful along withreview of important material at regular intervals after initial instruction to improve retention.

Memory is a complex function that heavily involves attention. Brad’s low performanceon the CMX most likely reflects complex difficulties with memory as well as difficultieswith attention to at least some degree. Although specific instructional strategies must await acomprehensive assessment of memory functions as noted in subsequent sections, some generalrecommendations for classroom management of Brad’s memory problems can be provided. Forexample, students referred for learning problems who score 75 on the CMX should be seatednear the front of the classroom to ensure that the teacher can obtain the maximum degreeof attention to classroom instruction and directions for completing work. The instructions,particularly for younger students, should be specific, simple, and given one at a time. Theteacher may wish to stand close to Brad when giving instruction and directions in the classroom.Brad’s teacher should ensure that homework assignments are provided in writing or are writtendown by Brad prior to leaving the classroom for the day. It would be helpful to have theteacher check to ensure that any homework instructions or instructions for any other projectsto be completed are written accurately and completely. Maintenance of a daily calendar andinstruction in the use of general types of memory strategies also will be helpful to most studentswith this level of performance on the CMX. Instruction in the use of such memory strategies asrehearsal, chunking, visualization, and related mnemonic aides also can be useful in improvingmemory function.

Brad’s VRM score of 35 suggests that he will experience moderate difficulties in learningand recall of verbal material, even when presented in a meaningful or experiential context,which should be performed because it aids recall. When presenting verbal information to Brad,cues for future recall should be provided and a discussion of strategies for the recall of verbalinformation would be helpful. Moderate attentional difficulties also are common with VRMscores at this level.

Students with verbal memory deficits relative to others their age may benefit from instructionon developing better memory techniques as related to verbal material. Students such as Bradshould be seated near the classroom instructor, who should be sure to ascertain that he orshe has Brad’s attention during any form of lecture or verbal presentation. Direct instructionin reading and study strategies also may be quite helpful. Memory strategies can be taught,including strategies such as rehearsal, chunking, and skill in making verbal associations. CallingBrad’s name prior to giving instructions or other verbal information that will need to berecalled may be helpful along with the use of other techniques to assure his specific attentionto the material being presented. Pairing verbal information with visual presentations includingpictures or demonstrations of what is expected or what is to be recalled also may be quite useful.A determination of existing strategies employed for learning verbal material can be obtainedfrom administration of the School Motivation and Learning Strategies Inventory for studentsbeginning at age 8 years (Stroud & Reynolds, 2006).

Brad’s NVM score of 34 suggests that he will experience moderate difficulties in the learningand recall of visual/spatial material, including maps, figures, graphs, drawings, locations, the ar-rangement of items in a written work, signs, faces, and directions that require visualization. Thesedifficulties will likely occur even when Brad is presented material in a meaningful or experientialcontext, which should be performed because the use of concrete objects often aids recall.



Discrepancy Feedback and Recommendations

The magnitude of discrepancy between Brad’s VIX score of 78 and NIX score of 103 isrelatively unusual within the normal population. Although this is the most common patternwithin referral populations, the magnitude of the discrepancy occurring for Brad makesthe difference noteworthy. In general, this pattern represents substantially disparate skillsin the general domains of verbal and nonverbal reasoning with clear superiority evidentin the nonverbal domain. Relative to their verbal reasoning and general language skills,individuals who display this pattern will experience greater success in tasks involving spatialreasoning, visualization skills, the use of mental rotation, reading of nonverbal cues, andrelated aspects of nonverbal reasoning and communication that usually includes nonverbaland visual memory skills. Nonverbal ability is less influential in others’ appraisal of generalintellectual functioning. Because NIX is greater than VIX, Brad’s general intellectual functioningmay appear lower than is reflected by his CIX score. Whenever possible, one should takeadvantage of Brad’s relatively higher levels of performance in the nonverbal domain byalways providing visual cues and an explanation of tasks, expectations, or a demonstration ofwhat is expected to be learned (where possible). Experiential learning is typically superior totraditional lecture and related pedagogical methods for individuals with this score pattern.The synthesis of information as opposed to the analysis of information is often a relativestrength.

Teaching should emphasize the use of visual images, spatial representations of relationships,experiential learning, and the synthesis of information as opposed to methods of deduction inlearning. Difficulties are likely to occur with traditional pedagogical styles such as lecturing andthe completion of reading and written assignments. An emphasis on the spatial relationshipsof numbers and the construction of problems is likely to be the most effective meansfor teaching math versus the memorization and the learning of step-by-step rules forcalculation. A heavy emphasis on learning by example and by demonstration is likely tobe most effective with students with this intellectual pattern. Also common are problemswith sequencing, including sequential memory and, in the early grades, mastery of phonicswhen synthesizing word sounds into correct words. Emphases on holistic methods of learningare likely to be more successful in addition to experiential approaches. The practical side oflearning and the application of knowledge can be emphasized to enhance motivation in thesestudents.

Often, these students do not have good study, learning, and test-taking strategies. It isoften useful to assess the presence of strategies with a scale such as the School Motivation andLearning Strategies Inventory and then to target deficient areas of learning strategies for directinstruction (Stroud & Reynolds, 2006).

Recommendations for Additional Testing

In cases where the CMX falls below 90, additional follow-up assessment with a morecomprehensive memory battery often provides additional clues and insights into appropriateinstructional practices as well as rehabilitative exercises that may be most useful. Follow-upevaluation with a comprehensive memory battery should be given even stronger considerationwhen the CMX is below 90 and also falls at a level that is significantly below the CIX. Evaluationwith a comprehensive memory battery such as the Test of Memory and Learning—SecondEdition (TOMAL-2; Reynolds & Voress, 2007) is recommended for such assessments because ofthe high degree of variability in performance produced by even small changes in memory tasks.

Brad’s NIX score of 103 is significantly higher than his VIX score of 78. As such, follow-upevaluation may be warranted. Although this is the most common pattern in referral populations,



additional information is almost always helpful in making a diagnosis, in treatment planning,and/or in making vocational recommendations. Evaluations that consider disturbances inlanguage and verbal functions in general (including receptive and expressive language) andother left-hemisphere-related tasks may prove helpful. Although empirical research at thispoint is lacking, clinical experience with the RIAS indicates that when the VIX score issignificantly below the NIX score and the absolute value of the VIX is less than 90, thereis a high probability of the presence of a language disorder that may have an adverseimpact on academic attainment or success in any academically related vocational trainingprogram. When this pattern occurs, as in the case of Brad, screening for a language disorderis recommended at a minimum and a more comprehensive language assessment should beconsidered.

SUMMARY

The RIAS was devised from theories of in-telligence that have practical value and em-pirical support. Therefore, the RIAS indexesare firmly rooted in modern intelligence testtheory of the 1990s, thus making the results in-terpretable. In addition, usage of modern testtheories means that the most important and pre-dictive intellectual abilities are assessed. Morethan this, however, the RIAS is designed topromote empirically supported practice in in-tellectual assessment. The practices of the pastthat lack empirical support, most prominentlysubtest level profile analysis of IQ-tests, havebeen the principal reason for giving lengthy sub-test batteries of intelligence. The RIAS is steepedin empirical research that not only fails to sup-port such practices, but also refutes their utility.The RIAS takes advantage of well-known andwell-researched tasks to provide the examinerwith measurement of the constructs of primaryutility when questions of intelligence need to beanswered.

The RIAS was developed with special atten-tion to well-reasoned modern standards for psy-chological and educational assessment (AERA,APA, NCME, 1999). Consistent with these stan-dards, the RIAS provides many opportunities forvalidation research to uncover new or unfore-seen and contraindicated inferences based ontest scores.

REFERENCES

Adams, W., & Reynolds, C. R. (2009). Essentials ofmemory assessment with the WRAML2 and theTOMAL-2. New York: John Wiley & Sons.

American Association on Mental Retardation. (2002).Mental retardation: Definition, classification, andsystems of supports (10th ed.). Washington, DC:Author.

American Educational Research Association, Amer-ican Psychological Association, and NationalCouncil on Measurement in Education. (1999).Standards for educational and psychological test-ing. Washington, DC: American PsychologicalAssociation.

American Psychiatric Association. (2000). Diagnosticand statistical manual of mental disorders (4th ed.,text revision). Washington, DC: Author.

Anastasi, A., & Urbina, S. (1997). Psychological testing(7th ed.). Upper Saddle River, NJ: PrenticeHall.

Beaujean, A., Firmin, M., Knoop, A., Michonski, J.,Berry, T., & Lowrie, R. (2006). Validation ofthe Frey and Detterman (2004) IQ predictionequations using the Reynolds Intellectual Assess-ment Scales. Personality and Individual Differences,41, 353–357.

Benton, A. L. (1974). Benton visual retention test (rev.ed.). New York: The Psychological Corporation.

Bigler, E. D., & Clement, P. F. (1997). Diagnosticclinical neuropsychology (3rd ed.). Austin, TX:University of Texas Press.

Binet, A., & Simon, T. (1905). New methods for thediagnosis of the intellectual level of subnormals.L’Annee Psychologique, 11, 191–244.



Brown, R. T., Reynolds, C. R., & Whitaker,J. S. (1999). Bias in mental testing since bias inmental testing. School Psychology Quarterly, 14(3),208–238.

Bruegemann, A., Reynolds, C. R., & Kamphaus,R. W. (2006). The Reynolds Intellectual Assess-ment scales (RIAS) and assessment of intellec-tual giftedness. Gifted Education International, 21,122–136.

Carroll, J. B. (1993). Human cognitive abilities: A surveyof factor analytic studies. New York: CambridgeUniversity Press.

Cattell, R. B. (1978). Matched determiners vs. fac-tor invariance: A reply to Korth. MultivariateBehavioral Research, 13(4), 431–448.

Cronbach, L. J. (1951). Coefficient alpha and theinternal structure of tests. Psychometrika, 16,297–334.

Edwards, O. (2006). WISC-IV VS. RIAS: Will thereal IQ please stand up!! Paper presented at theannual meeting of the Florida Association of SchoolPsychologists, Orlando, FL.

Edwards, O. & Paulin, R. V. (2007). Referred stu-dents’ performance on the Reynolds IntellectualAssessment Scales and the Wechsler IntelligenceScale for Children— Fourth Edition. Journal ofPsychoeducational Assessment, 25(4), 334–340.

Eliot, J., & Smith, I. M. (1983). An international dic-tionary of spatial tests. Windsor, Berks, England:Nefer-Nelson. (Distributed in the USA by Hu-manities Press, Atlantic Highlands, NJ, 07716).

Golden, C. J., Zillmer, E., & Spiers, M. (1992). Neu-ropsychological assessment and intervention. Spring-field, IL, England: Charles C. Thomas.

Goldstein, S., & Reynolds, C. R. (1999). Handbook ofneurodevelopmental and genetic disorders in children.New York: Guilford Press.

Goldstein, S., & Reynolds, C. R. (2005). Handbook ofneurodevelopmental and genetic disorders in adults.New York: Guilford Press.

Horn, J. L., & Cattell, R. B. (1966). Refinement andtest of the theory of fluid and crystallized generalintelligences. Journal of Educational Psychology,57(5), 253–270.

Jensen, A. (1998). The suppressed relationship be-tween IQ and the reaction time slope parameterof the Hick function. Intelligence, 26(1), 43–52.

Joreskog, K., & Sorbom, D. (1987). LISREL6.13: User’s reference guide. Chicago: ScientificSoftware.

Joseph, R. (1996). Neuropsychiatry, neuropsychology, andclinical neuroscience: Emotion, evolution, cognition,language, memory, brain damage, and abnormal be-havior (2nd ed.). Baltimore: Williams & Wilkins.

Kamphaus, R. W. (2001). Clinical assessment of child andadolescent intelligence (2nd ed.). Boston: Allyn andBacon.

Kamphaus, R. W., & Reynolds, C. R. (2003). Reynoldsintellectual screening test. Odessa, FL: PAR, Inc.

Kaufman, A. S. (1990). Assessing adolescent and adultintelligence. Needham Heights, MA: Allyn &Bacon.

Kaufman, A. S. (1994). Intelligent testing with theWISC-III. New York: Wiley.

Kaufman, A. S., McLean, J. E., Kaufman-Packer,J., & Reynolds, C. R. (1991). Is the pattern ofintellectual growth and decline across the adultlifespan different for men or women? Journal ofClinical Psychology, 47(6), 801–812.

Kaufman, A. S., Reynolds, C. R., & McLean,J. E. (1989). Age and WAIS-R intelligence ina national sample of adults in the 20 to 74 yearrange: A cross-sectional analysis with educationlevel controlled. Intelligence, 13(3), 235–253.

McCarthy, D. (1972). McCarthy scales of children’s abil-ities. New York: The Psychological Corporation.

Miller, W. S. (1926). Miller Analogies Test. New York:The Psychological Corporation.

Nunnally, J. C. (1978). Psychometric theory (2nd ed.).New York: McGraw-Hill.

Oxford Press. (1999). New shorter Oxford Englishdictionary. Oxford: author.

Reynolds, C. R. (1984–1985). Critical measurementissues in learning disabilities. Journal of SpecialEducation, 18, 451–476.

Reynolds, C. R. (1997). Measurement and statisticalproblems in neuropsychological assessment ofchildren. In C. R. Reynolds & E. Fletcher-Janzen(Eds.), Handbook of clinical child neuropsychology(2nd ed., pp. 180–203). New York: Plenum Press.

Reynolds, C. R. (1998). Fundamentals of measurementand assessment in psychology. In A. Bellack & M.Hersen (Series Eds.) & C. R. Reynolds (Vol. Ed.),Comprehensive clinical psychology: Vol. 4. Assessment(pp. 33–56). Oxford, England: Elsevier Science.

Reynolds, C. R. (2000). Methods for detecting andevaluating cultural bias in neuropsychologicaltests. In E. Fletcher-Janzen, T. Strickland, &C. R. Reynolds (Eds.), Handbook of cross-culturalneuropsychology (pp. 249–286). New York: Plenum.



Reynolds, C. R., & Bigler, E. D. (1994). Test of memoryand learning. Austin, TX: PRO-ED.

Reynolds, C. R., Castillo, C. L. & Horton, A. M.,Jr. (2008). Neuropsychology and intelligence: Anoverview. In A. M. Horton, Jr. and D. Wedding(Eds.), The neuropsychology handbook (3rd ed., pp.69–86). New York: Springer.

Reynolds, C. R., Chastain, R. L., Kaufman, A. S., &McLean, J. E. (1987). Demographic influences onadult intelligence at ages 16 to 74 years. Journalof School Psychology, 25(4), 323–342.

Reynolds, C. R., & Fletcher-Janzen, E. (Eds.). (1997).Handbook of clinical child neuropsychology (2nd ed.).New York: Plenum Press.

Reynolds, C. R., & Kamphaus, R. W. (2003). Reynoldsintellectual assessment scales. Odessa, FL: PAR, Inc.

Reynolds, C. R., Kamphaus, R. W., & PAR Staff(2007a). Reynolds Intellectual Assessment Scales In-terpretive Report. Lutz, FL: Psychological Assess-ment Resources.

Reynolds, C. R., & Kamphaus, R. W. (2007b). ReynoldsIntellectual Assessment Scales/Wide Range Achieve-ment Test 4 Discrepancy Interpretive Report profes-sional manual supplement. Lutz, FL: PsychologicalAssessment Resources.

Reynolds, C. R., & Kamphaus, R. W. (2007c). Testof irregular word reading efficiency. Lutz, FL:Psychological Assessment Resources.

Reynolds, C. R., Lowe, P. A., & Saenz, A. (1999). Theproblem of bias in psychological assessment. In C.R. Reynolds and T. B. Gutkin (Eds.), Handbook ofschool psychology (3rd ed., pp. 549–595). New York:Wiley.

Reynolds, C. R., & Voress, J. (2007). Test of memoryand learning (2nd ed.). Austin, TX: Pro-Ed.

Riccio, C. A., & Hynd, G. W. (2000). Measurablebiological substrates to verbal-performance dif-ferences in Wechsler scores. School PsychologyQuarterly, 15(4), 386–399.

Sattler, J. M. (2001). Assessment of children: Cognitiveapplications (4th ed.). La Mesa, CA: Jerome M.Sattler.

Stroud, K., & Reynolds, C. R. (2006). School motivationand learning strategies inventory (SMALSI). LosAngeles: Western Psychological Services.

Wechsler, D. (1949). Wechsler Intelligence Scale forChildren. San Antonio, TX: The PsychologicalCorporation.

Wechsler, D. (1991). Wechsler Intelligence Scale forChildren— Third Edition. San Antonio, TX: ThePsychological Corporation.

Wechsler, D. (1992). Wechsler Individual AchievementTest. San Antonio, TX: The Psychological Cor-poration.

Wechsler, D. (1997). Wechsler Adult Intelligence Scale—Third Edition. San Antonio, TX: The Psycholog-ical Corporation.

Weinberg, R. A. (1989). Intelligence and IQ: Land-mark issues and great debates. American Psycholo-gist, 44(2), 98–104.

Wilkinson, L., & Task Force on Statistical Infer-ence. (1999). Statistical methods in psychologyjournals: Guidelines and explanations. AmericanPsychologist, 54(8), 594–604.

Yerkes, R. M. (1917). The Binet versus the Pointscale method of measuring intelligence. Journalof Applied Psychology, 1, 11–122.

Zachary, R., & Gorsuch, R. (1985). Continuousnorming: Implications for the WAIS-R. Journalof Clinical Psychology, 41, 86–94.


Queries in Chapter 5AQ1. We have shortened the recto running

head since it exceeds the page width.Please confirm if this is ok.

development and application of the reynolds intellectual ... · naglieri c05.tex v1 - 04/30/2009...

Documents