Wechsler Nonverbal Scale of Ability (WNV)

C H A P T E R 12

Wechsler Nonverbal Scaleof Ability (WNV)

Jack A. Naglieri, Kimberly Brunnert

INTRODUCTION

General ability (g) has been the underlyingmodel for IQ tests since the early 1900s.These tests were and continue to be comprisedof questions that are verbal (e.g., vocabularyor word analogies), quantitative (e.g., mathword problems or math calculation), and spatial(arranging blocks to match a simple design orassembling puzzles to make a common object).The spatial tests have been described as nonverbalbecause it is an easier concept to understand, notbecause of any intention to measure nonverbalability. In fact, Wechsler’s view was that ‘‘thesubtests are different measures of intelligence,not measures of different kinds of intelligence’’(1958, p. 64). The Technical and InterpretiveManual for the Wechsler Nonverbal Scale ofAbility (WNV; Wechsler & Naglieri, 2006a)also cites Boake (2002) as noting that ‘‘Wechslerviewed verbal and performance tests as equallyvalid measures of intelligence’’ (2006, p. 1).Further, Naglieri (2003b, 2008a) wrote that theterm nonverbal refers to the content of the test,not a type of ability.

There is considerable experimental supportfor the concept of general intelligence as

measured by tests such as the Wechsler andBinet (see Jensen, 1998, for a review). But thecontent of these tests sometimes presents a prob-lem for assessment of culturally and linguisticallydiverse populations. For an individual who hasnot had the chance to acquire verbal and quanti-tative skills due to limited opportunity to learn ora disability, verbal and quantitative tests designedto measure general ability may be a good pre-dictor of current academic performance but nota good reflection of their ability. For example,typical Native-American Navajo children livingon a reservation in northern Arizona earn lowscores on the Verbal scale but average scores onthe Performance scale of the Wechsler (Naglieri& Yazzie, 1983) because they speak English as asecond language and have had insufficient expo-sure to the language of a typical American child.Suzuki and Valencia (1997) argued that verbaland quantitative questions found on most tradi-tional IQ tests interfere with accurate assessmentof minority children; therefore, a nonverbal testof general ability such as the WNV (Wechsler& Naglieri, 2006b) offers a viable method forevaluating ability for these children.

The essence of a nonverbal test of generalability is that it does not contain verbal andquantitative test questions, although it may

315

316 PART III NONVERBAL INTELLIGENCE TESTS

?

1 2 3 4 5

FIGURE 12.1Simple Test Question from the WNV MatricesSubtest

involve verbal ways to solve the problem andit does require intellectual effort on the partof the examinee. For example, Figure 12.1shows a simple nonverbal test question likethose included in the Matrices subtest of theWNV. The item has shapes that vary acrossthe horizontal and vertical dimensions that theexaminee must recognize and understand toanswer the analogy expressed using geometricfigures (e.g., it is a figural analogy arrangedin a matrix). In Figure 12.1, for example, thedifferences between the top and bottom rows isthat the shape inside boxes changes (a triangleappears on the top row and a circle on the bottomrow). The difference between the first and secondcolumn included in the top row is the addition ofshading in the right column. The examinee needsto understand the interrelationships among thesevariables (shape and shading across the columnsand rows) to arrive at the correct answer (option3). The child may, or may not, use a verbaldescription (in any language) of the figurescontained within the matrix or the child maysimply look at the shapes and understand whichoption is the answer with minimal verbal analysis.

Tests that measure general ability nonverballymay have different formats, but the essentialgoal of these tests is the same: to measure abil-ity nonverbally. For example, some nonverbal

tests are comprised of one type of item, theprogressive matrix (e.g., the Naglieri Nonver-bal Ability Test—Second Edition [NNAT-2];Naglieri, 2008b) given in a group format orindividual format (Naglieri Nonverbal AbilityTest—Individual Form [NNAT-1]; Naglieri,2003b). Another method is to use several dif-ferent types of nonverbal subtests as found in theWNV (as well as the Universal Nonverbal In-telligence Test [UNIT; Bracken & McCallum,1998]; see Bracken and McCallum’s chapter inthis book). The slight variation in administrationformat and subtest composition notwithstand-ing, the goal is the same: to measure generalability nonverbally, and in so doing, provide away effectively and fairly to assess a wide varietyof individuals regardless of their educational orlinguistic backgrounds.

DESCRIPTION OF THE WNV

Subtest BackgroundThe WNV is comprised of a variety of subteststhat are intended to measure general ability indifferent ways. For example, although the non-verbal subtests on the WNV are all alike inthat they do not require language or arithmeticskills, they are diverse in their specific require-ments. For example, some of the subtests havea strong visual-spatial requirement, others de-mand recall of spatial information or recall ofthe sequence of information, and others involvepaper-and-pencil skills. This multidimensional-ity of task requirements distinguishes the WNVfrom tests that use one type of test format, suchas the NNAT-2 (Naglieri, 2008b), which usesprogressive matrices exclusively.

Most of the WNV subtests have appearedin previous editions of the Wechsler scales(Wechsler Preschool and Primary Scale ofIntelligence—Third Edition [WPPSI-III],Wechsler Intelligence Scale for Children—Third Edition [WISC-III], Wechsler Intel-ligence Scale for Children—Fourth Edition

CHAPTER 12 WECHSLER NONVERBAL SCALE 317

[WISC-IV], Wechsler Intelligence Scale forChildren—Fourth Edition—Integrated [WISC-IV Integrated], Wechsler Adult IntelligenceScale—Fourth Edition [WAIS-III], WechslerMemory Scale—Third Edition [WMS-III]), andhave an established record of reliability and va-lidity for the nonverbal measurement of generalability. Adaptation of the subtests was necessaryto accommodate the new pictorial directions for-mat, identify items that were most appropriate forthe specific ages, and provide directions in the sixlanguages. Each WNV subtest was included onlyafter careful examination of both the content andform of the items vis-a-vis the goals of this partic-ular instrument. The origins and descriptions ofthe WNV subtests are referenced in Table 12.1.

Each subtest is further described in thefollowing.

MatricesThe Matrices (MA) subtest requires the ex-aminee to discover how different geometricshapes are spatially or logically interrelated. Themultiple-choice options provide potential an-swers that vary in the degree to which each optioncompletes the relationships among the parts. Theitems are displayed using basic geometric figuressuch as squares, circles, and triangles using somecombination of the colors black, white, yellow,blue, and green. Items were constructed usingshapes and colors that would maintain interestand minimize the likelihood that impaired color

TABLE 12.1 Subtests Origin and Description

Subtest (Abbreviation) Origin and Description

Matrices (MA) This subtest was adapted from the NNAT–I. The examinee looks at anincomplete figural matrix and selects which of the four or five responseoptions is the missing piece.

Coding (CD) This subtest was adapted from the WISC-IV. The examinee follows a keythat provides symbols that correspond with shapes (Coding A) or numbers(Coding B).

Object Assembly (OA) This subtest was adapted from the WPPSI-III and the WISC–III, and hasone new item. The child is presented with puzzle pieces that are placed bythe examiner in a specified layout. The child completes the puzzle within aspecified time limit.

Recognition (RG) This is a new match-to-stimulus subtest. The child looks at a page with adesign with geometric patterns on it for three seconds. The child thenchooses which of four or five response options on the next page match theviewed stimulus.

Spatial Span (SSp) This subtest was adapted from the WMS–III. The examinee mimics theexaminer’s tapping on a series of blocks either in the same order as theexaminer (Spatial Span Forward) or in the reverse order (Spatial SpanBackward).

Picture Arrangement (PA) This subtest is adapted from the WAIS–III and a research version of theWISC–IV Integrated. The examinee uses a set of picture cards, which theexaminer has placed on the table in a specified order, to tell a logical storywithin a specified time limit.


vision would influence the scores. The WNVMatrices items are composed of a variety offormats (e.g., geometric patterns, reasoning byanalogy, and spatial visualization) previously usedin the NNAT-I. Matrices is always administered(i.e., it is given to examinees in both age bands andis included in both the 4- and 2-subtest batteries).

CodingThe Coding (CD) subtest requires the examineeto copy symbols (e.g., a dash, two vertical lines,an open parenthesis) that are paired with simplegeometric shapes or numbers according to a keyprovided at the top of the page. There are twoforms of the Coding subtest: Form A is usedin the 4-subtest battery for ages 4:0–7:11 andForm B is used in the 4-subtest battery for ages8:0–21:11. The Coding subtest is adapted for usein the WNV from the WISC-IV by eliminatingreversible shapes (e.g., left and right parentheses)for the younger age group and evenly distributingthe use of each code across each row (e.g., forCoding Form B, the stimuli range from 1 to 9).

Object AssemblyThe Object Assembly (OA) subtest is comprisedof items that require the examinee to completepieces of a puzzle to form a recognizableobject such as a ball or a car. These itemsvary in the number of pieces (from 2 to 11)and the complexity with which they have beendisassembled. Object Assembly is included in the4-subtest battery of the WNV for examineesages 4:0–7:11. The Object Assembly subtest wasadapted for use in the WNV by using items fromWPPSI-III (e.g., bear, apple, dog, star, calf, andtree), the WISC-III (e.g., ball), and one new item(i.e., glasses).

RecognitionThe Recognition (RG) subtest was created foruse in the WNV and is included in both the4- and 2-subtest batteries for examinees ages

4:0–7:11. It requires the examinee to examinea stimulus (e.g., a square with a small circle inthe center) for three seconds and then choosewhich option is identical to the stimulus that wasjust seen. The figures are colored black, white,yellow, blue, and/or green to maintain interestand minimize the likelihood that impaired colorvision will influence the scores.

Spatial SpanThe Spatial Span (SSp) subtest requires theexaminee to touch a group of blocks arrangedon an 8-by-11-inch board in a nonsystematicspatial manner in the same and reverse order ofthat demonstrated by the examiner. Spatial Spanis included in both the 4- and 2-subtest batteriesfor ages 8:0–21:11. The Spatial Span subtest wasadapted for use in the WNV from the WMS-IIIand adapted, like all the subtests, to the pictorialdirections format.

Picture ArrangementThe Picture Arrangement (PA) subtest requiresthe examinee to arrange cartoon-like illustrationsinto a sequence that is logical and makessense. Picture Arrangement is included in the4-subtest battery for examinees ages 8:0–21:11.The Picture Arrangement subtest was adaptedfor use in the WNV by merging colorizeditems from the WAIS-III (e.g., choir, speak, andshark) and items from a research edition of theWISC-IV Integrated (e.g., duck, storm, farm,shadow, and broken).

STRUCTURE OF THE TEST

The WNV is structured in four ways, combiningsubtests selected to best meet the examinee’s andexaminer’s needs. There are 4- and 2-subtest bat-teries for each age band, 4:0–7:11 and 8:0–21:11.The subtests that are in each are referenced inFigure 12.2.


Matrices* Coding

WNVFull Scale

ObjectAssembly Recognition*

Matrices*

Ages 4-7 years

Ages 8-21 years

Note: Subtests included in the 2-subtest version have an asterisk.

Coding Spatial Span* PictureArrangement

FIGURE 12.2Structure of the WNV

The examinee’s raw scores are converted to Tscores for each subtest, which have a mean of 50and a standard deviation of 10. Using the T scoresfor each subtest, analyses can be performed tocompare the examinee’s performance across thesubtests and to identify strengths and weaknesses.There is a Full Scale Score that can be calculatedfor each battery that has a mean of 100 anda standard deviation of 15. There are separateWNV norms tables based on standardizationsamples collected in the United States andCanada. For Spatial Span, there are additionalanalyses that can be performed to examine thedifference between Spatial Span Forward andSpatial Span Backward, as well as the longestspan in either direction.

ADMINISTRATION AND SCORING

Tips on AdministrationUnlike most tests of its kind, the WNV ad-ministration begins with a standardized shortintroduction that tells examinees that they needto look at the pictorial directions and that theycan ask the examiner questions if necessary. Like

the other spoken text included in the WNV,these sentences are provided in English, French,Spanish, Chinese, German, and Dutch. Ac-tual administration procedures follow carefullyscripted directions designed to ensure that thedemands of the tasks are completely understoodby all examinees. There are three steps to theadministration directions:

Step 1 uses the standardized directions, whichare always administered and should never bechanged. These directions must be adminis-tered in every case and include gestures thatcorrespond to the pictorial directions. Picto-rial directions are used at Step 1 to provide astandardized method of communicating the de-mands of the task. These pictorial directions (seeFigure 12.3) show a scene like the one the exam-inee is currently in. The frames of the directionsshow the progression of an examinee being pre-sented with the question, then thinking about theitem, and finally, choosing the correct solution.

Step 1 instructions include actions by theexaminer that must be carefully followed tomaximize the likelihood that the examineeunderstands the correspondence between thematerials and the task. Gestures are used to directthe examinee’s attention to specific portions


Note: The actual WNV Pictorial Directions appear in color.

FIGURE 12.3WNV Pictorial Direction

of the pictorial directions and to the stimulusmaterials and sometimes to demonstrate thetask itself. Sometimes simple statements are alsoincluded because they convey the importance ofboth time and accuracy to the examinee.

Step 2 directions are used only after the stan-dard directions are provided. These instructionsmust also be followed exactly and are given onlywhen an examinee is unclear about what he orshe is being asked to do. These directions includestandardized simple sentences and gestures forcommunicating the requirements of the task tothe examinee. Verbal directions provide anotherway to ensure that the examinee understands thedemands of the tasks and are provided in Englishas well as Spanish, German, French, Chinese,and Dutch. These translations are to be usedonly when the following two conditions are bothmet: (1) the WNV is being administered to anexaminee who speaks one of the languages, and(2) the examiner or a professional interpreterspeaks the language.

Step 3 gives an opportunity to provide help,but these directions are used only after the previ-ous two steps have been administered. This is theonly step of administration that gives the exam-iner flexibility. For example, the examiner maysay or sign additional instructions or questions.In general, examiners are given the opportunityto communicate in whatever manner they thinkwill best explain the demands of the subtest based

on their judgment of the examinee’s needs. Thiscould include providing further explanation ordemonstration of the task, restating or revisingthe verbal directions, or using additional wordsto describe the requirements of the task. At notime, however, is it permissible to teach the ex-aminee how to solve the items. Instead, the goalof Step 3 instructions is to provide additionalhelp to ensure that the examinee understands thedemands of the task, not to show how to teacha way to solve the items. The amount of helpprovided and the determination about when tostop is based on professional judgment.

When using an interpreter to facilitate com-munication prior to and during administration,it is important that the interpreter have guide-lines and training about what is and what is notpermitted. This person should translate a gen-eral explanation of the testing situation for theexaminee, including the introductory paragraphat the beginning of Chapter 3 in the WNVAdministration and Scoring Manual before ad-ministration begins. It is also important that theinterpreter recognize the boundaries of his or herrole in administration. For example, although itis appropriate for the interpreter to translatethe examiner’s responses to an examinee’s re-sponse to a sample item, it is not acceptablefor the interpreter to make additional statementsunless instructed to do so. Importantly, at notime should the interpreter communicate any


information that could influence the examinee’sscores. See Brunnert, Naglieri, and Hardy-Braz(2009) for more information about working withtranslators and especially when testing those whoare deaf or hard of hearing.

Administration of the WNV subtests is de-signed to be simple and easy. The Administrationand Scoring Manual includes a section prior toactual administration directions that describesthe subtest, the materials needed, start, stop,and reverse rules, scoring, as well as general is-sues unique to each subtest. The Manual alsoprovides considerable discussion of the physi-cal materials, uses and applications for the twoversions of the WNV, and general testing, ad-ministration, and scoring issues. In this chapter,we will highlight some of the most importantpoints, but the reader is advised to carefullystudy that Administration and Scoring Manual.What follows is a subtest-by-subtest discussionof specific administration issues.

Matrices

Although the Matrices subtest is a very straight-forward subtest to administer, examiners shouldalways be aware of possible responses that maysuggest concern. For example, some studentswho are particularly impulsive might select theoption that is mostly but not completely correctbecause the options were written with varying de-grees of accuracy. If an examinee is not lookingat the options closely, one of those that is almostcorrect may be selected. Similarly, if an examineetakes a long time to respond, the examiner may(after about 30 seconds) prompt a response.

Coding

The examinee is allowed to correct mistakes bycrossing out the incorrect symbol and writing hisor her response next to it, so the pencil withoutan eraser needs to be used for the Coding subtest.The examiner should ensure that the examineeworks from left to right and from top to bottomwithout skipping any items or row and by pro-viding the appropriate instruction when needed.

The examiner also provides instruction that in-forms the examinee to work as quickly as possible.For that reason, examinees should not be allowedto spend too much time making corrections.

Object Assembly

The examiner should always set up the puzzlepieces on the same side of the Stimulus Book asthe examinee’s dominant hand. Then remove theStimulus Book before administering the SampleItem. The examiner should also ensure that theexaminee works as quickly as possible. If theexaminee is still completing a puzzle when thetime limit expires, the examiner should placehis or her hand over the puzzle to stop theexaminee’s progress, and record the examinee’sanswer. If the examinee seems upset at thatpoint, the examiner should allow the examineeto finish but not consider any additional workfor scoring purposes. It is also important toremember to begin timing after the last word ofthe instruction is provided. Assembling the piecesfor the examinee requires a specific methodfully articulated in the WNV Administration andScoring Manual (Wechsler & Naglieri, 2006c).Essentially, the method requires that the puzzlepieces are put before the child in a specific formatface down. Once all the pieces are before the child,then they are turned over in the order indicatedby the number on the back of each piece.

Recognition

Examiners must be sure to expose each stimuluspage for exactly 3 seconds. To do so will likelyrequire that the page be exposed at a precisetime when the stopwatch strikes an exact secondand then being prepared (e.g., having yourfinger under the page so that you are ready)to turn the page exactly when the 3 secondshave elapsed. Do not allow examinees to turn thepages.

Spatial Span

The Spatial Span Board must be placed so thatthe examinee can easily reach all cubes on the


board and only the examiner can see the numberson the back of each blue block. Also, always setthe Spatial Span Board on the same side of theStimulus Book as the examinee’s dominant hand.Tap the blocks at a rate of one per second andraise the hand approximately one foot abovethe Spatial Span Board between each tap. If theexaminee does not respond after the examinertaps a sequence, the examiner can say, ‘‘It’syour turn.’’ Always administer both Spatial SpanForward and Spatial Span Backward, regardlessof the examinee’s performance on Spatial SpanForward and always administer both trials of anitem regardless of the examinee’s performanceon the first trial.

Picture Arrangement

The examiner should always place the PictureArrangement Cards on the same side of theStimulus Book as the examinee’s dominant handand remove the Stimulus Book with PictorialDirections before administering the SampleItem. Having the cards in the box in the order inwhich they are to be exposed to the examineeis an excellent way to efficiently deliver theitem. When the examinee completes the item,record his or her sequence, then resequencethe cards in the presentation order for thenext administration. If the examinee is goingvery slowly, it is permissible to ensure that theexaminee realizes that he or she should workas quickly as possible. If the examinee ordersthe cards from right to left instead of left toright, the examiner should ask, ‘‘Where does itstart?’’

If the examinee is in the midst of completinga story when the time limit expires, the examinershould place his or her hand over the story tostop the examinee’s progress, and record theexaminee’s answer. If the examinee seems upsetthat he or she was stopped while completing thestory, the examiner should allow the examineeto finish. However, the examiner should notconsider any additional work by the examineefor scoring purposes.

How to Score the TestFive of the six subtests (i.e., Matrices, Coding,Recognition, Spatial Span, and Picture Arrange-ment) are scored by summing the number ofpoints earned during administration. The sixthsubtest (i.e., Object Assembly) has time bonusesfor some items that might be part of the rawscore.

The raw scores are converted to T scores.The sum of T scores is converted to a Full ScaleScore, which has a corresponding percentile rankand confidence interval.

Information can also be assessed at the subtestlevel for the 4-subtest battery by comparing theT score an examinee earned on a subtest to themean T score for all four subtests administered.If the difference between the subtest T score andthe mean T score is significant, then that subtestis considered a strength or weakness. Base ratesare provided for each difference. Additionally,optional analyses can be performed on the SpatialSpan results as described later in this chapter.

Use of Scoring and Report WritingSoftware

The WNV Scoring Assistant provides an auto-mated way to obtain all standard and derivedscores based on the U.S. as well as the Canadianversions of the test. The report writing feature ofthe software provides reports that are appropri-ate for clinicians as well as parents. The parentreport is available in English, French, and Span-ish. The software also provides links betweenthe WNV and the WIAT-II and all the abilitycomparisons to achievement.

STANDARDIZATION, NORMS,AND PSYCHOMETRICS

Characteristics of theStandardization Sample

There were two samples collected for thecreation of the WNV norms: one in the United


States and the other in Canada. There werealso samples collected of special groups and ofother tests for validity research. The U.S. sampleconsisted of 1,323 examinees stratified across5 demographic variables: age (4:0–21:11), sex,race/ethnicity (Black, White, Hispanic, Asian,and Other), education level (8 years or lessof school, 9–11 years of school, 12 years ofschool [high school degree or equivalent], 13–15years of school [some college or associate’sdegree], and 16 or more years of school [collegeor graduate degree]), and geographic region(Northeast, North Central, South, and West).Note that education level was determined by theparent education for examinees ages 4:0–17:11and by the examinee’s own education for ages18:0–21:11.

The Canadian sample consisted of 875 exami-nees stratified across five demographic variables:age (4:0–21:11), sex, race/ethnicity (Caucasians,Asians, First Nations, and Other), educationlevel (less than a high school diploma; highschool diploma or equivalent; college/vocationaldiploma or some university, but no degree ob-tained; and a university degree), and geographicregion (West, Central, and East). Additionally,the Canadian sample consisted of 70% Englishspeakers, 18% French speakers, and 12% speak-ers of other languages.

Reliability of the ScalesThe reliability estimates for the WNV wereprovided by subtest and Full Scale Scores byage and over all ages. There are reliability esti-mates provided for the U.S. normative sample,for the Canadian normative sample, and forall the special groups that are reported in theWNV Technical and Interpretive Manual. Thereliability estimates for the U.S. normative sam-ple ranged from .74 to .91 for the subtestsand were .91 for both Full Scale Scores acrossages. The reliability estimates for the Cana-dian normative sample ranged from .73 to .90for the subtests, were .90 for the Full ScaleScore: 4-Subtest Battery, and .91 for the Full

TABLE 12.2 Reliability Estimate Ranges bySpecial Study

ReliabilitySpecial Group Estimate Range

Gifted .77–.97MR Mild .80–.93MR Moderate .87–.93Reading and WrittenExpression LearningDisorders

.72–.88

Language Disorders .74–.97English Language Learners .70–.96Deaf .77–.98Hard of Hearing .75–.97

Scale Score: 2-Subtest Battery. The reliabilityestimates for the studies with examinees thatwere diagnosed with or classified as being gifted,having mild mental retardation, moderate men-tal retardation, reading and written expressionlearning disorders, language disorders, Englishlanguage learners, Deaf, and Hard of Hearingare shown in Table 12.2.

Other reliability information is provided inthe WNV Technical and Interpretive Manual.This other information includes the standarderror of measurements (SEM), confidence in-tervals, and test–retest stability estimates. SEMsare provided for both the U.S. and the Cana-dian normative samples by subtest and FullScale Scores by age in years, for ages 4:0–15:11,and by age bands, from ages 16:0–21:11. Theconfidence intervals, which are calculated withthe standard error of estimate (SEE), are pro-vided alongside the T Score to Full ScaleScore conversion tables at both the 90%and 95% levels (see Tables A.2 and A.3 inthe WNV Administration and Scoring Manual).Test–retest stability estimates were also pro-vided in the WNV Technical and InterpretiveManual.


USE OF THE WNV

Interpretation MethodsLike any test, WNV test results should alwaysbe interpreted with consideration of the manyfactors that can influence obtained scores. Per-haps the most important are issues such as thebehaviors observed during testing, and relevanteducational and environmental backgrounds, andphysical and emotional status, all within thecontext of the reason for referral. In order toobtain the greatest amount of information fromthe WNV, there are some important methodsof interpretation that warrant discussion. Someof these methods are the same for the 4- and2-subtest batteries, and others are unique to eachversion. In this chapter, the issues that apply toboth batteries will be covered first and then thefiner points of interpretation relevant to eachversion will be examined separately.

Interpretation of the Two WNVVersions

Both versions of the WNV are comprised ofsubtests (set at a mean of 50 and SD of 10) thatare combined to yield a Full Scale score (setat a mean of 100 and SD of 15). The WNVsubtest scores are set on the T score metric of 50and 10 (as opposed to a traditional scaled scorewith a mean of 10 and SD of 3). This formatwas selected because the individual subtests hadsufficient range of raw scores, allowing for theuse of T scores, which have a greater rangeand precision than scaled scores. For example,the WNV subtest T scores range from 10 to90, yielding 81 possible different scores, whereasa scaled score typically ranges from 3 to 20,yielding only 18 different scores. The use of theT score also provides greater precision on eachsubtest, allowing for higher reliability coefficientsof the Full Scale score.

The WNV Full Scale scores are standardscores with a mean of 100 and SD of 15 basedon either the 4- or 2-subtest batteries. This score

provides a nonverbal estimate of general abilitythat has excellent reliability and validity. Boththe Full Scale and subtest T scores are based onthe U.S. or Canadian standardization samplesand can be used to measure general abilitynonverbally. It is important to recognize that thenonverbal label refers to the fact that the test itemsdo not contain verbal and quantitative contentand does not suggest that a specific type of abilityis being assessed. Additionally, even though theWNV subtests have different demands—thatis, some are spatial (e.g., Matrices or ObjectAssembly), others involve sequencing (PictureArrangement and Spatial Span), require memory(e.g., Recognition and Spatial Span), or usesymbol associations (Coding)—they all measuregeneral ability. The WNV measures generalability nonverbally. General ability allows us, forexample, to understand spatial as well as verbaland mathematical concepts, remember visualrelationships as well as quantitative or verbalfacts, and work with sequences of informationof all kinds. The content of the questionsmay be visual or verbal, and require memoryor recognition, but general ability (sometimesreferred to as g) underlies performance on allthese kinds of tasks and the WNV Full Scale isan excellent measure of g.

WNV 4-Subtest BatteryInterpretation

Step 1: The first step in interpretationof the 4- and 2-subtest versions of theWNV is to examine the Full Scale score.This score is the most reliable and validrepresentation of general ability on thescale. The Full Scale score should bereported with its associated percentile score,categorical description (Average, AboveAverage, etc.), and confidence interval. Astatement such as the following illustrateshow these concepts might be included in awritten document:


Sam obtained a WNV Full Scale score of 91,which is ranked at the 27th percentile and fallswithin the Average classification. This means thathe performed as well as or better than 27% ofexaminees his age in the normative sample. Thereis a 90% chance that his true Full Scale score fallswithin the range 85–99.

Step 2: The second step in interpretation ofthe 4-subtest version of the WNV is to ex-amine the T scores the examinee earned onthe subtests. Analysis of this type must takeinto consideration the lower reliability ofthese scores and the increased probabilitythat variability will reflect measurement er-ror. Examination of the four WNV subtestsshould also be conducted with consider-ation that even though the subtests areall nonverbal measures of general abil-ity they do have unique attributes (i.e.,some involve remembering information,others spatial demands, etc.). Additionally,variability across the subtests should beexpected and, therefore, statistical guide-lines should be followed to ensure that anydifferences interpreted are beyond thosethat could be expected by chance. In fact,Wechsler and Naglieri (2006a) reported thecumulative percentages, mean, and SD ofsubtest scatter (e.g., highest–lowest sub-test score for each individual in the U.S.standardization group) in Table B.5 in theWNV Administration and Scoring Manual.The mean score was 16.5 (SD = 7.5),indicating that practitioners can expect dif-ferences among the WNV subtests. In fact,approximately 50% of that sample had arange of subtest scores that was between 0and 16. If unusual is defined at 10% of theU.S. standardization sample, then a rangeof 27 or more would meet that criterion.Scatter is one way to determine whetherthe WNV subtests vary, but when the goalis to determine whether there is significantvariability and to relate the unique contri-butions of each subtest to other findings,

a different method is recommended. Thatmethod requires that the examinee’s sub-test scores are compared to that examinee’smean.

The values needed for significance whencomparing a WNV subtest for an examinee tothat examinee’s mean T score are provided in theWNV Administration and Scoring Manual (TableB.1) and in more detail by Brunnert, Naglieri,and Hardy-Braz (2009) and summarized herein Table 12.3. These values were computedutilizing Davis’s (1959) formula (and Silverstein’s1982, modification of this procedure) for thedifference between the average of several scoresobtained by one examinee and each of his orher scores included in the average. For example,when four WNV subtest T scores are comparedwith the mean of the four T scores, a z value thattakes into account the number of comparisonsneeds to be used. The standard errors ofmeasurement for the WNV subtest T scoresof the U.S. and Canadian samples were obtainedfrom the WNV Technical and Interpretive Manual(Wechsler & Naglieri, 2006a, Tables 4.4 and4.5) and were used for these calculations. (SeeSilverstein, 1982, for more information andNaglieri and Paolitto, 2005, for values for theWISC-IV.)

The following steps should be used to com-pare each of the four WNV subtest T scores tothe child’s mean subtest T score:

1. Calculate the mean of the four subtest Tscores.

2. Calculate the difference between each subtestT score and the mean.

3. Subtract the mean from each of the subtest Tscores (retain the sign).

4. Find the value needed for significance usingthe examinee’s age group and the desiredsignificance level in Table 12.3.

5. If the absolute value of the difference is equalto or greater than the value in the table, theresult is statistically significant.


TABLE 12.3 Differences Required for Significance When Comparing Each WNVSubtest for the 4-Subtest Battery T Scores to the Examinee’s Average Subtest T Score forthe U.S. and Canadian Standardization Samples by Age Group

Country Age p-Value MA CD OA RG

United States 4:0–7:11 .10 7.4 7.9 8.7 8.6.05 8.3 8.8 9.7 9.6

MA CD SSp PA8:0–21:11 .10 6.8 8.7 7.5 9.3

.05 7.6 9.7 8.4 10.4

MA CD OA RGCanada 4:0–7:11 .10 8.1 8.0 9.1 8.6

.05 9.0 8.9 10.1 9.6

MA CD SSp PA8:0–21:11 .10 6.8 8.7 7.4 9.6

.05 7.6 9.7 8.3 10.7

NOTE : MA = Matrices, CD = Coding, OA = Object Assembly, RG = Recognition, SSp = Spatial Span,and PA = Picture Arrangement.

6. If the subtest difference from the mean islower than the mean, then the difference isa weakness; if the subtest difference fromthe mean is greater than the mean, then thedifference is a strength.

For example, if an 18-year-old from theUnited States obtained T scores of 65 onMatrices, 42 on Coding, 39 on Spatial Span, and61 on Picture Arrangement, the mean T scorewould be 51.8. Using the values from Brunnertet al. (2009) and the .05 level of significance,Matrices would be considered a relative strengthfor this examinee (65 – 51.8 = 13.3, whichexceeds the critical value of 7.6 for the .05level of significance). The T score for SpatialSpan is significantly lower than the examinee’smean T score (39 – 51.8 = –12.8, which exceedsthe critical value of 8.4 for the .05 level ofconfidence), and would be considered a relativeweakness. Similarly, the T score for Coding issignificantly lower than the mean T score (42 –

51.8 = –9.8, which exceeds the critical value of9.7 for the .05 level of confidence), and is also arelative weakness.

When there is significant variability in theWNV subtests, it is also important to determinewhether a weakness relative to the examinee’soverall mean is also sufficiently below theaverage range. Determining whether a child hassignificant variability relative to his or her ownaverage score is a valuable way to determinestrengths and weaknesses relative to the child’smean score, but Naglieri (1999) cautioned thata relative weakness could also be significantlybelow the normative mean. He recommendedthat any subtest score that is low relative to thechild’s means should also fall below the averagerange to be considered a noteworthy weakness(e.g., < 1 SD below the mean). In the exampleabove, the Spatial Span T score of 39 wouldbe considered a weakness from the ipsative andnormative perspectives, but the Coding scoreof 42 would not. This would strengthen the


TABLE 12.4 Differences Required forSignificance When Comparing Recognition orSpatial Span to Matrices on the WNV 2-SubtestBattery for U.S. and Canadian StandardizationSamples by Age Group

Age Group in Years4:0–7:11 8:0–21:11

UnitedStates

p = .10 MA vs. RG 9.9 —MA vs. SSp — 5.3

p = .05 MA vs. RG 11.8 —MA vs. SSp — 6.3

Canadap = .10 MA vs. RG 9.2 —

MA vs. SSp — 8.1p = .05 MA vs. RG 11.0 —

MA vs. SSp — 9.7

level of concern about this finding and morestrongly suggest that additional examinationusing a multidimensional measure of abilitysuch as the Cognitive Assessment System (seeNaglieri, 1999) could be appropriate. Subtestdifferences that are significant (in addition to theFull Scale score) should be described in a mannersimilar to the following:

His scores on the individual WNV subtest scores variedsignificantly, suggesting a relative strength on Matrices,a subtest that requires understanding the relationshipsamong spatial designs. A relative weakness was foundon Coding and his Spatial Span subtest T score wassignificantly below his overall subtest mean and the nor-mative mean of 50. Both of these subtests require recallof information, and this finding suggests that furtherexamination of immediate memory may be indicated.

WNV 2-Subtest BatteryInterpretation

The differences required for significance wheneach pair of WNV subtests included in the

2-Subtest battery are compared are provided inTable 12.4. These values are used, for example,to determine whether a T score difference of11 points between Matrices and Recognition issignificant for an examinee who is 7:3 years old.To use this table, simply subtract one subtestT score from the other (ignore the sign) andcompare the result to the value in the table thatcorresponds to the desired level of significance(.10 or .05). If the obtained value is equal toor greater than the value in the table, then theresult is significant. Determining how often adifference of a specific magnitude occurred in theU.S. or Canadian samples would further clarifythe importance of the finding. For example,Table B.4 of the WNV Administration and ScoringManual provides the base rates of subtest T scoredifferences by the direction of the difference.Using that table, we find that about 11% ofexaminees aged 8:0–21:11 obtained T scores forMatrices that were 14 or more points higher thantheir T scores for Spatial Span. This informationcan be used to augment the interpretation ofthe significance of the difference between thescales.

WNV Full Scale Score InterpretationThe WNV Full Scale score is a nonverbalmeasure of general ability that should be reportedwith the corresponding confidence interval,percentile rank, and classification. The Full Scaleand subtest T scores could be described in amanner similar to the following:

Gene obtained a WNV Full Scale score of 98, whichfalls in the Average range and is ranked at the 45thpercentile. This means he did as well as or better than45% of examinees his age in the normative sampleon this nonverbal measure of general ability. Thereis a 90% chance that his true Full Scale score fallswithin the range of 91 to 105. There was significantvariability between the two subtests that comprise theWNV. Gene performed significantly better on a subtestrequiring reasoning with geometric designs (Matrices)than he did on a subtest requiring recall of a sequence ofmovements arranged in a spatial array (Spatial Span).


This difference is also unusual, with only 4.3% of thepopulation obtaining differences this large or larger,suggesting that further investigation is appropriate.

In-depth Interpretationof Spatial Span

The WNV Spatial Span subtest Forward andBackward scores can be interpreted separately,particularly when this test is embedded within thegreater context of a comprehensive assessment.The sizes of the differences required for statisticalsignificance by age and for the U.S. and Canadiansamples are 11 and 13 for the .10 and .05levels for the United States, and 10 and 13for the Canadian standardization samples forthe combined ages 8:0–21:11. This comparisoncan be accomplished using Table C.1 of theWNV Administration and Scoring Manual, whichprovides a way to convert the raw scores to Tscore equivalents for Spatial Span Forward andSpatial Span Backward. According to Table C.2of the Manual, a difference of 9 T score pointsis needed at the .15 level (13 at the .05 level)to have a significant difference between thesetwo scores. The frequency with which Forwardand Backward score differences occurred in thenormative sample are provided in Table C.3 ofthe test Manual, which presents the bases forthe T score differences for the U.S. sample. Thetables provide the base-rate data by the directionof the difference. For example, about 24% ofexaminees obtained Spatial Span Forward scores7 or more points higher than their Spatial SpanBackward scores, whereas about 25% obtainedSpatial Span Backward scores 7 or more pointshigher than their Spatial Span Forward scores.

Comparison of the differences between Spa-tial Span Forward and Backward T scores mayprovide useful information, but it should beintegrated within the greater context of a com-prehensive assessment. For example, if a SpatialSpan Forward T score is 14 points higher than theSpatial Span Backward T score for a 13-year-oldU.S. examinee, the difference would be signifi-cant (only 12 points are needed at the .05 level),

and the difference occurs in only 9% of thenormative sample (see WNV Administration andScoring Manual, Table C.3). This informationwould be expected to be similar to other similartest scores, such as WISC-IV Digit Span For-ward vs. Digit Span Backward. Both of these tests’Backward scores should be related to the Plan-ning Scale of the Cognitive Assessment System(see Naglieri, 1999) and may suggest that theexaminee has difficulty with development andutilization of strategies for reversing the order ofserial information.

IDENTIFICATION OF SPECIALPOPULATIONS

Specific Learning DisabilitiesThe Individuals with Disabilities Education Im-provement Act (IDEIA) of 2004 has broughtabout widespread discussion of policies regard-ing the identification of children with SpecificLearning Disabilities (SLD). The law no longerrequires, but does not disallow, the use of anability–achievement discrepancy for eligibilitydetermination, but the law clearly states thata comprehensive evaluation is required and thatevaluation must include tests that are not cultur-ally or linguistically discriminatory. This needfor a comprehensive evaluation has also been ad-vocated by the National Association of SchoolPsychologists’ ‘‘Position Statement on Identifi-cation of Children with Specific Learning Dis-abilities’’ (2007). Some researchers have arguedthat SLD is best identified by examining a patternof strengths and weaknesses in cognitive pro-cessing scores (e.g., Hale, Flanagan, & Naglieri,2008; Naglieri & Kaufman, 2008). How does theWNV fit into the process of evaluating childrenwith SLD?

The WNV provides a nonverbal measure ofgeneral ability that can be compared to cur-rent achievement test scores to help determinewhether a child is demonstrating academic per-formance that is commensurate with ability.


Unlike other measures of general ability thatcontain verbal and quantitative tests, the WNVprovides an evaluation of ability that is not influ-enced by achievement-like content (see Naglieri& Bornstein, 2003, or Naglieri, 2008a, for morediscussion of the similarity of test questions onability and achievement tests). The WNV alsoprovides a measure of ability that can be viewedas nondiscriminatory on the basis of race, eth-nicity, language, and disability (see the followingsections). The WNV is not, however, designedto be a test of basic psychological processes, andother tools should be used for that purpose (e.g.,the Cognitive Assessment System; Naglieri &Das, 1997).

Practitioners who wish to compareWNV scores with the Wechsler IndividualAchievement Test—Second Edition (WIAT–II; Pearson, 2005) can do so using thepredicted-difference and simple-differencemethods. The predicted-difference methodtakes into account the reliabilities and thecorrelations between the two measures. In thismethod, the ability score is used to predict anachievement score, and the differences betweenpredicted and observed achievement scoresare compared. Tables B.1–B.7 in the WNVTechnical and Interpretive Manual provide thevalues needed for significance when conductingthis analysis for children in the United Statesand Tables B.8–B.14 are used for comparingscores using the WNV Canadian normativesample with the WIAT–II (see Wechsler &Naglieri, 2006a for more details).

Simple differences between the WNV4- and 2-subtest Full Scale standard scorescompared with other achievement tests areprovided by Brunnert et al. (2009). Theachievement tests include the Kaufman Testof Educational Achievement—Second Edition(K–TEA II; Kaufman & Kaufman, 2004),Woodcock-Johnson Tests of Achievement—Third Edition (Woodcock, McGrew, & Mather,2001), Diagnostic Achievement Battery—ThirdEdition (Newcomer, 2001), and the Wide RangeAchievement Test—Fourth Edition (Wilkinson

& Robertson, 2006). Regardless of whichmethod is used, the examiner can augment thosefindings with the base rates provided in theWNV Technical and Interpretive Manual’s TablesB.1–B.14 for the WIAT–II. Using both thereliability of the difference and the rate at whichability and achievement test score differencesoccur in the normal population will provide aneffective comparison of these two measures.

Attention Deficit HyperactivityDisorders

The attention deficit hyperactivity disorders(ADHD) are evaluated on the basis of the corre-spondence of behavioral characteristics describedby parents with the DSM-IV-TR criteria. Theessential feature of ADHD is currently describedas a consistent pattern of inattention and/orhyperactivity-impulsivity found more frequentlyand more severely than is typical in individualsat a comparable level of development. Althoughthe DSM-IV-TR states that those with ADHDmay demonstrate variability in IQ, the diagnosticcriteria are based on behavioral rather than cog-nitive factors even though some have argued thatcognitive processing scores should play a role inidentification (Goldstein & Naglieri, 2006). TheWNV, therefore, like any test of general ability,plays a minimal role in the diagnostic process; itsonly role may be in helping evaluate the child’slevel of general ability.

GiftedThe underrepresentation of minority children inclasses for the gifted has been and continues to bean important educational problem (Ford, 1998;Naglieri & Ford, 2005). In fact, Naglieri andFord (2003) stressed that Black, Hispanic, andNative-American students are underrepresentedby 50% to 70% in gifted education programs.In recent years, addressing this problem has be-come more focused on the types of tests usedwhen evaluating the ability of children poten-tially eligible for gifted programming. Some have


argued that the verbal and quantitative contentof some of the ability tests used and proceduresfollowed are inconsistent with the characteristicsof culturally, ethnically, and linguistically diversepopulations (Naglieri & Ford, 2005; Naglieri,Brulles, & Lansdowne, 2009). That is, becauseIQ has traditionally been defined within a ver-bal/quantitative/nonverbal framework, studentswith limited English-language and math skillsearn lower scores on the Verbal and Quantita-tive scales these tests include because they donot have sufficient knowledge of the languageor training in math, not because of low ability(Bracken & Naglieri, 2003; Naglieri, 2008a). If astudent has not had the chance to acquire verbaland quantitative skills due to limited opportunityto learn, or a disability, verbal and quantitativetests designed to measure general ability maybe a good predictor of current academic per-formance but an inaccurate reflection of theirability to learn especially after instruction is pro-vided. One way to address this issue is to includetests that measure general ability nonverbally.Naglieri and Ford (2003) demonstrated the ef-fectiveness of using a group nonverbal measureof general ability (the Naglieri Nonverbal AbilityTest—Multilevel Form [NNAT-ML; Naglieri,1997]) for increasing the identification of His-panic and Black students. Similarly, the WNV

provides an individually administered way toassess general ability nonverbally and increasethe participation of minorities in gifted classes.There is evidence that children in gifted educa-tion programs earn high scores on the WNV.

The WNV was administered to gifted chil-dren who were carefully matched to controlsubjects included in the standardization sampleon the basis of age, race/ethnicity, and educa-tion level. The differences between the meanswere calculated using Cohen’s (1996) formula(i.e., the difference between the means of thetwo groups divided by the square root of thepooled variance). The study included 41 exam-inees, all of whom had already been identifiedas gifted using a standardized ability measurewhere they performed at 2 standard deviationsabove the mean or more. The students in thegifted programs performed significantly betterthan their matched counterparts from the nor-mative sample with effect sizes that were large forthe Full Scale Score: 4-Subtest Battery and FullScale Score: 2-Subtest Battery. See Table 12.5for more details.

Mentally RetardedNaglieri and Rojahn (2001) suggested that as-sessment of mental retardation should take into

TABLE 12.5 WNV Means, SDs, and Effect Sizes for Special Populations and Matched Control Groups

Special Population Matched Sample

Mean SD Mean SD n Effect Size

Moderate MentalRetardation

45.9 8.9 99.5 14.1 36 –4.5

Mild Mental Retardation 67.3 12.9 97.4 15.3 51 –2.1Gifted 123.7 13.4 104.2 12.3 41 1.5English LanguageLearners

101.7 13.4 102.1 13.4 55 0.0

Hard of Hearing 96.7 15.9 100.5 14.2 48 –0.3Profoundly Deaf 102.5 9.0 100.8 14.3 37 0.1


consideration the appropriateness of the con-tent of the IQ test used for the individual beingevaluated. They found that minority childrenwho were previously identified as having mentalretardation earned lower Verbal than Perfor-mance IQ scores on the WISC-III, presumablybecause of the academic content of the verbaltests (Naglieri & Bornstein, 2003). When thesesame subjects were evaluated using a measure ofability that does not contain verbal and quanti-tative test questions (the Cognitive AssessmentSystem; Naglieri & Das, 1997), the number ofchildren who would still qualify as having men-tal retardation was reduced. They concluded,therefore, that measures of ability that do notcontain verbal and quantitative questions aremore desirable when assessing children for men-tal retardation, particularly when the childrenhave had limited educational opportunities andhave culturally and/or linguistically diverse back-grounds. The WNV is a tool that can evaluategeneral ability nonverbally, and thereby providea way to assess ability without undue influenceof language and knowledge for individuals sus-pected of having mental retardation.

There were two special studies of examineeswith mental retardation performed with the stan-dardization of the WNV, one for examinees whowere diagnosed with Mild Mental Retardationand the other for examinees who were diagnosedwith Moderate Mental Retardation. Examineeswho were diagnosed with Mild Mental Retarda-tion were included in the study if their cognitiveability was between 55 and 70 on a standardizedability test, if they were not institutionalized, andif they met the general criteria for inclusion ina special study (see Appendix C of the WNVTechnical and Interpretive Manual). There were51 examinees included in this study; 43 of theseexaminees were 8–21 and performed significantlyworse than their matched counterparts from thenormative sample with effect sizes (Cohen’s d)large for the Full Scale Score: 4-Subtest Bat-tery and Full Scale Score: 2-Subtest Battery. SeeTable 12.5 for more details.

Examinees who were diagnosed with Mod-erate Mental Retardation were accepted in thestudy if their cognitive ability was between 40and 55 on a standardized ability test, if they werenot institutionalized, and if they met the generalcriteria for inclusion in a special study (see Ap-pendix C of the WNV Technical and InterpretiveManual). There were 31 examinees included inthis study; 28 of these examinees were 8–21 andperformed significantly worse than their matchedcounterparts from the normative sample with ef-fect sizes (Cohen’s d) large for the Full ScaleScore: 2-Subtest Battery and Full Scale Score:4-Subtest Battery. See Table 12.5 for moredetails.

Deaf and Hard of HearingThere are numerous issues surrounding the eval-uation of ability for individuals who are deafor hard of hearing. A thorough discussion ofthese issues and those particularly relevant to theWNV can be found in Brunnert et al. (2009).In general, however, the assessment issues centeron (1) content of the test and (2) communi-cating test requirements to the examinee. Theformer issue has been covered in the previ-ous sections regarding evaluation of mentallyretarded and gifted. The issue of communicat-ing test requirements has also been discussed inthe administration portion of this chapter. Inessence, because the directions are given pic-torially, and can be augmented with additionalstatements and/or communication using sign lan-guage, the WNV offers considerable advantagesfor appropriate evaluation of individuals who aredeaf or hard of hearing as the research studiesthat follow illustrate.

Profoundly Deaf

Wechsler and Naglieri (2006a) reported a studyof profoundly deaf examinees who were com-pared with cases from the standardization of theWNV who were matched on a number of im-portant demographic variables. This included 37examinees who ‘‘must not have been able to hear


tones to interpret spoken language after the ageof 18 months, must not lip read, must not betrained in the oral or auditory-verbal approach,and must not use cued speech (i.e., they must haveroutine discourse by some means of communicat-ing other than spoken language). They must havesevere to profound deafness (hearing loss mea-sured with dB, Pure Tone Average greater thanor equal to 55)’’ (Wechsler & Naglieri, 2006a,p. 65). Thirty-one of these examinees were 8–21and performed minimally differently than theirmatched counterparts from the normative samplewith effect sizes (Cohen’s d) negligible for FullScale Score: 4-Subtest Battery and Full ScaleScore: 2-Subtest Battery. There is additional in-formation about this sample of examinees from asurvey collected with the WNV standardization(Wechsler & Naglieri, 2006a, Appendix D). SeeTable 12.5 for more details.

Hard of Hearing

The WNV Manual also describes a study of indi-viduals who were hard of hearing and comparedtheir WNV scores to a group from the standard-ization sample who were matched on a numberof important demographic variables. This studyincluded 48 examinees who ‘‘could have a uni-lateral or bilateral hearing loss or deafness, andthe age of onset of their inability to hear couldbe any age and [they] could have cochlear im-plants’’ (WNV Technical and Interpretive Manual,pp. 65–66). Sixteen of these examinees were 4–7and performed minimally differently than theirmatched counterparts from the normative samplewith effect sizes (Cohen’s d) that were negligi-ble for Full Scale Score: 4-Subtest Battery, andFull Scale Score: 2-Subtest Battery. Thirty-twoof these examinees were 8–20 and performedminimally differently than their matched coun-terparts from the normative sample with effectsizes (Cohen’s d) that were negligible for FullScale Score: 4-Subtest Battery, and Full ScaleScore: 2-Subtest Battery. See Table 12.5 formore details; for more details about the sample,see the WNV Technical and Interpretive Manual(Appendix D).

English as a Second LanguageIt is clear that as the United States continues tobecome more diverse, the number of individualswhose primary language is not English willcontinue to increase. The largest of thesegroups is the Hispanic population, which isapproximately 37 million or about 13% ofthe U.S. population, making it the largestminority group (Ramirez & de la Cruz, 2002).This population of Hispanics is dominated byindividuals of Mexican origin (66.9%) who residein the Western (44.2%) and Southern (34.8%)regions of the country. Hispanics aged 25 andolder are less likely to have a high school diplomathan non-Hispanic whites (57.0% and 88.7%,respectively). Importantly, 27.0% of Hispanicshave less than a ninth-grade education comparedwith only 4.0% of non-Hispanic whites (Ramirez& de la Cruz, 2002). The large number ofimmigrants in this country makes clear theneed for psychological tests that are appropriatefor those who come from these working-classhomes with parents who have limited academicand English-language skills. Nonverbal tests ofgeneral ability such as the WNV are, therefore, aparticularly useful way to assess minority childrenbecause they yield smaller race and ethnicdifferences (which is attributed to the differencein content) while these instruments retain goodcorrelations with achievement, and can helpidentify minority children for gifted programs(Bracken & McCallum, 1998; Naglieri & Ford,2003; Naglieri & Ronning, 2000a, 2000b).

Wechsler and Naglieri (2006a) provide evi-dence of the utility of the WNV for individualswho are learning English. The study involvedexaminees who speak English as a second lan-guage who were compared to a matched samplefrom the WNV standardization sample. Thisincluded 55 examinees aged 8–21 years whose‘‘native language was not English, they spoke alanguage other than English at home, and theexaminee’s parents had resided in the UnitedStates less than 6 years’’ (Wechsler & Naglieri,2006a, p. 63). There were 27 Hispanics and 28


examinees who specified their primary languagewas Cantonese, Chinese (unspecified), Korean,Russian, Spanish, or Urdu. Additional informa-tion about this sample is available in the WNVTechnical and Interpretive Manual (Appendix D).These examinees performed very similarly totheir matched counterparts from the normativesample with negligible effect sizes (Cohen’s d)for the Full Scale Score: 4-Subtest Battery andFull Scale Score: 2-Subtest Battery, as shown inTable 12.5.

Interventions Based on Test ResultsThe WNV can be used in at least two wayswhen a plan for instruction is being developed.At the global level, the Full Scale score providesan indication of general ability that can suggestspecific instructional needs based on the charac-teristics of the child. For example, a child whohas limited English-language skills who earns ahigh WNV Full Scale score as part of an as-sessment for possible placement in Gifted shouldreceive these special services in an environmentthat recognizes his or her strengths and needs.Naglieri, Brulles, and Lansdowne (2008) pro-vide explicit suggestions about the manner inwhich children who are low in academic skills buthigh on a nonverbal measure of general abilityshould be taught. Their book includes strategiesthat are particularly useful for diverse popula-tions of gifted students. Instructional topics forbilingual students include critical and creativethinking, appropriate gifted educational objec-tives, and student-directed learning in areas suchas language arts, social studies, mathematics, andscience.

The specific subtest scores of the WNV,like other tests of general ability, could beused to develop hypotheses about characteristicsof the student that may have implications forinstruction. These hypotheses should be furtherevaluated by other evidence and, when possible,tested to see whether a desirable response toinstruction occurs. Subtest scores on the WNVcould suggest hypotheses that could be further

investigated or evaluated on the basis of thechild’s response to intervention. For instance,a Spatial Span Forward T score of 54 and aSpatial Span Backward T score of 38 couldsuggest a weakness with Planning as definedby Naglieri (2008c). These two scores are bothsignificantly different and the difference occursin only 6.2% of the U.S. standardization sample.If this finding is corroborated by a weakness onthe Planning scale of the Cognitive AssessmentSystem (Naglieri & Das, 1997) and academicproblems are found, the application of methodsdescribed by Naglieri and Pickering (2003)for teaching strategy use should be attempted.There is research that supports the value ofteaching children to be strategic and the positiveinfluence this instruction has on math andreading comprehension (see Naglieri, 2008c, fora summary).

Subtest performance on the WNV may alsosuggest a weakness on tests of general ability thatrequire recall of information. The Recognitionsubtest requires recall of information, and if thescore on that subtest is significantly lower thanthe child’s mean and less than one SD from themean (< 40), the hypothesis should be consid-ered. Once corroborated with other test dataand if accompanied by academic problems intasks that require recall of information, teachingstrategies for remembering should be imple-mented. This would include teaching chunkingor other mnemonic methods of recall of infor-mation, for which there is considerable evidenceof effectiveness (Mastropieri & Scruggs, 2006;Minskoff & Allsopp, 2003). Ways in which thesemethods can be communicated to teachers andparents in instructional handouts are providedby Naglieri and Pickering (2003). The WNVcould, therefore, help develop possible explana-tions for problems associated with difficulty inthe classroom.

SummaryThe research studies summarized here suggestthat the WNV offers an effective measure of


general ability that yields expected results. Indi-viduals identified as having mental retardationearned low scores with those with the mostsevere retardation earning lower scores thanthose with mild retardation. In contrast, chil-dren identified as being gifted earned very highscores on the nonverbal measure of general abil-ity. Importantly, individuals who were acquiringEnglish-language skills performed very similarlyto the normative mean of 100; in fact their scorewas 101.7, which was very similar to the matchedcontrol group, which had a score of 102.1. Addi-tionally, the results for individuals who are hardof hearing, as well as those for the deaf, showthat these two groups also earned scores on theWNV that were very similar to the matchedcontrol groups. In summary, these data providestrong support for the use of the WNV withdiverse populations.

VALIDITY

Relationships between WNV andOther Ability and Achievement TestsThe WNV is strongly related to other nonverbalmeasures of ability and other measures ofability that contain verbal and nonverbal scales.For example, Wechsler and Naglieri (2006a)reported that the WNV 4-subtest battery FullScale score correlated .79 with the UNIT and.73 with the NNAT-1. Similarly, the WNV FullScale score correlated .82 with the WPPSI-III,.82 with the WISC-IV, .84 with the WAIS-III,and .83 with the WISC-IV Spanish. Thesefindings provide evidence that the WNV is agood test of general ability even though it doesnot contain verbal and quantitative subtests.

The WNV is strongly correlated withachievement (.66) as measured by the WIAT-II.The correlations between the WNV and theWIAT-II are consistent with other studies of therelationship between ability and achievement,which have yielded a correlation of about .6(Naglieri, 1999). Previous research (e.g., Naglieri

& Ronning, 2000b) with the NNAT-ML hasshown a correlation of .5 to .6 for large samplesof children in grades K–12. The results illus-trate that the 4-subtest battery Full Scale score iseffective as a predictor of academic achievement.

Demographic DifferencesEnglish Language Learners

There is good evidence that the WNV is an ap-propriate measure of general ability for thosewho have limitations in either knowledge oruse of the English language. Wechsler andNaglieri (2006a) provide evidence of the util-ity of the WNV with examinees who haveEnglish-language limitations and hearing prob-lems that limit their ability to acquire informa-tion. The first study included examinees whosenative language was not English, where the pri-mary language they spoke was not English, alanguage other than English was spoken at home,and their parents resided in the United States lessthan six years. When compared to a group fromthe normative sample matched on basic demo-graphics, they found that the examinees learningEnglish earned essentially the same score as thematched control of English-speaking examineesin the normative group (effect sizes for the 4-and 2-subtest batteries were .03 and .04, respec-tively). While these results suggested that theWNV measures general ability effectively andfairly for those with limited English-languageskills when combined with the studies involvingstudents with hearing limitations, the strength ofthis instrument is more clearly understood.

The findings summarized above for Englishlanguage learners suggest that the WNV may beuseful for addressing the underrepresentation ofminority children in classes for the gifted. Thishas been described as one of the most importantproblems facing educators of gifted students(Ford, 1998; Naglieri & Ford, 2005). Onesolution has been to use nonverbal tests of generalability as a part of the identification procedure,particularly for children whose primary languageis not English. Support for the use of a nonverbal


test in this context is amply documented byNaglieri (2008a). The logic is based on the factthat traditional measures of ability include teststhat require knowledge of English words and theuse of language even when questions involvingmathematics are used, which poses a barrier forEnglish language learners. Suzuki and Valencia(1997) argued that these verbal and quantitativequestions interfere with accurate assessmentof minority children. Naglieri & Ford (2005)maintained that tests like the WNV provide aneffective way to assess these individuals. Becausethe WNV does not penalize English languagelearners, it is an effective tool for assessinggeneral ability and, therefore, should be usedas part of the process to identify gifted examineeswhose primary language is not English.

Deaf and Hard of Hearing

Wechsler and Naglieri (2006a) reported that theWNV Full Scale scores are also very similarfor two matched samples of deaf and hearingexaminees and a study with hearing-impairedexaminees matched with the same demographiccharacteristics to hearing examinees from thenormative sample. Those in the deaf samplehad not ever heard spoken language; they hadnever heard tones after the age of 18 months,could not lip read or use cued speech, and wereclassified as having severe-to-profound deafness.The hard-of-hearing sample had exposure tospoken language, either through hearing or lipreading, and could have a unilateral or bilateralhearing loss or deafness. Their inability to hearcould have occurred at any age and they couldhave cochlear implants. The results for thesetwo groups, like the sample of examinees withEnglish-language limitations, earned WNV FullScale scores that were close to average and werevery similar to the matched control group. Theeffect sizes are considered negligible and small.Taken as a whole, these studies suggest thatlanguage has a negligible effect on the WNVFull Scale score and illustrate the strength of thisinstrument for assessment of individuals withhearing as well as language limitations.

Gender Differences

Gender differences in ability has been a topicof considerable interest for some time, resultingin a substantial body of literature on the topic(e.g., Fennema & Sherman, 1977; Geary, 1996;Halpern, 1997). Hyde and Linn (1988) found asmall mean effect size (favoring females) of .11 inverbal skills for students aged 5 through 18 years,but the differences between genders were notuniform across tasks. Halpern (1997) concludedthat females outperform males on tests of verbalfluency, foreign language, fine motor skills,speech articulation, and reading and writing,but males do better on tasks that involve mentalrotation, mechanical reasoning, math and scienceknowledge, and verbal analogies. Lynn andIrwing (2004) argued that sex differences must beviewed developmentally and with considerationof the role played by biology. Based on hisresearch using Raven’s Progressive Matrices,Lynn (2002) argued that females are slightlybetter than males between the ages of 10 and13 and that after 14 males catch up and overtakefemales, ending up with an advantage that reachesabout 2 IQ points among adults. The differencesbetween WNV scores by gender were recentlystudied by Brown (2008).

Table 12.6 provides the effect sizes forWNV Full Scale and subtest scores for 1,300girls and boys aged 4–21 years (broken intofour age groups: 4–7, 8–10, 11–14, and 15–21)who participated in the standardization sample.The results indicate that both the 4- and2-subtest Full Scale scores showed minimalgender differences. The 4-subtest effect sizesranged from 0.0–.16 and the 2-subtest FullScale effect sizes ranged from .01–.10. Theseare negligible and inconsequential differences.Interestingly, the subtest differences were alsosmall with the exception of the Coding subtest,which showed a female superiority of .31, .33, .17,and .48 across the four age groups. These findingsindicate that the WNV Full Scale scores yieldvalues that are very similar by gender and thedifferences found for Coding have little influenceon the total test score.


TABLE 12.6 Male/Female Effect Sizes for the 4-and 2-Subtest Batteries Full Scales and IndividualSubtests by Age in Years in the StandardizationSample

Age (in Years)4–7 8–10 11–14 15–21

Full Scale—4 −.16 −.06 .00 −.05Full Scale—2 −.09 .05 .01 .10Matrices −.04 .08 −.06 .16Coding −.31 −.33 −.17 −.48Object Assembly −.02 — — —Recognition −.12 — — —Spatial Span — .01 .07 .01Picture Arrangement — .07 .10 .20

NOTES :N = 1,300.Positive effect sizes indicate male scores were higher thanfemale scores.Effect Size = (X1 − X2) / SQRT [(n1 * SD1

2+ n2 * SD22)/(n1

+ n2)].

Empirical Support for the TestStructure

The WNV offers subtest-level data as well as aFull Scale score. Empirical support for this struc-ture was examined in three ways and reported inthe Technical and Interpretive Manual. Thesethree ways are intercorrelations, confirmatoryfactor analysis: communality, specificity, and er-ror variance, and confirmatory factor analysis:model fit.

The intercorrelations for all comparisons(subtest to subtest, subtest to full-scale score,and full-scale score to full-scale score) matchedthe patterns reported for other Wechsler scales(e.g., WISC-IV, WPPSI-III, and WAIS-III).Additionally, the intercorrelations for the FullScale scores were moderate to high for both agebands (4:0–7:11 and 8:0–21:11) for both the U.S.and the Canadian standardization samples.

The next measure of the strength of the teststructure is comparison of the specificity anderror variance as well as an examination on theloadings of each subtest on g. ‘‘For a subtestto provide a unique contribution to the latentvariable (g), it is expected that the specificitywill exceed the error variance’’ (Wechsler &Naglieri, 2006a, p. 46). In each way the specificityand error variance were reported (by age band:4:0–7:11 and 8:0–21:11; and by smaller groupingsfor the older age band: 8:0–10:11, 11:0–14:11,and 15:0–21:11) the specificity exceeded theerror variance, allowing the conclusion thateach subtest measures something unique. Anadditional conclusion from the g loadings is thatall of the subtests load on g.

Finally, the model fit for a single factor (g)model was examined. The findings indicate goodfit for this model for each of the two agebands (i.e., 4:0–7:11 and 8:0–21:11). However,when the older age band was broken down, the11:0–14:11 students did not fit as well as all otherage bands, but still showed adequate fit.

Overall, the conclusion about the test struc-ture based on empirical support is that the WNVprovides useful information both at the subtestlevel and at a single, general factor level. Sub-test scores as well as Full Scale scores are usefuland important when reporting the results of theWNV.

SUMMARY

The WNV was explicitly designed to provide anonverbal measure of general ability that wouldbe appropriate for a wide variety of culturally andlinguistically diverse populations. The selectionof tests that are described as nonverbal inconjunction with Pictorial Directions and oraldirections in five languages provides a uniqueapproach to measuring general ability. Theevidence provided in this chapter supportsthe utility of the test for fair assessment ofcognitive ability of those from culturally diversebackgrounds as well as those with language


differences or deficiencies as well as those who aredeaf or hard of hearing. The research providedin the test manual provides a base to supportthe use of the instrument but additional researchis needed, especially regarding the utility of theinstrument within diverse clinical environments.

REFERENCES

Boake, C. (2002). From the Binet-Simon to theWechsler-Bellevue: Tracing the history of intelli-gence testing. Journal of Clinical and ExperimentalNeuropsychology, 24, 383–405.

Bracken, B. A., & McCallum, R. S. (1998). UniversalNonverbal Intelligence Test. Itasca, IL: Riverside.

Bracken, B. A., & Naglieri, J. A. (2003). Assessingdiverse populations with nonverbal tests of gen-eral intelligence. In C. R. Reynolds & R. W.Kamphaus (Eds.), Handbook of psychological andeducational assessment of children (pp. 243–273).New York: Guilford.

Brown, C. (2008). Gender differences on the Wech-sler Nonverbal Scale of Ability. Manuscript inpreparation.

Brunnert, K., Naglieri, J. A., & Hardy-Braz, S. (2009).Essentials of WNV Assessment. New York: Wiley.

Cohen, B. H. (1996). Explaining psychological statistics.Pacific Grove, CA: Brooks & Cole.

Davis, F. B. (1959). Interpretation of differencesamong averages and individual test scores. Journalof Educational Psychology, 50, 162–170.

Fennema, E., & Sherman, J. (1977). Sex-relateddifferences in mathematics achievement, spatialvisualization, and affective factors. American Edu-cational Research Journal, 14, 51–71.

Ford, D. Y. (1998). The underrepresentation ofminority students in gifted education: Problemsand promises in recruitment and retention.Journal of Special Education, 32, 4–14.

Geary, D. C. (1996). Sexual selection and sex differ-ences in mathematical abilities. Behavioral andBrain Sciences, 19, 229–247.

Goldstein, S., & Naglieri, J. A. (2006). The role ofintellectual processes in the DSM-V diagnosis ofADHD. Journal of Attention Disorders, 10, 3–8.

Hale, J. N., Flanagan, D. P., & Naglieri, J. A. (2008,June). Alternative research-based methods for

IDEA 2004 identification of children with specificlearning disabilities. Communique, 36(1), 14–15.

Halpern, D. F. (1997). Sex differences in intelligence.American Psychologist, 52, 1091–1102.

Hyde, J. S., & Linn, M. C. (1988). Gender differencesin verbal ability: A meta-analysis. PsychologicalBulletin, 104, 53–69.

Jensen, A. R. (1998). The ‘‘g’’ factor: The science ofmental ability. Westport, CT: Praeger.

Kaufman, A. S., & Kaufman, N. L. (2004). KaufmanTest of Educational Achievement—Second Edition.San Antonio: Pearson.

Lynn, R. (2002). Sex differences on the progressivematrices among 15–16 year olds: Some data fromSouth Africa. Personality and Individual Differences,33, 669–673.

Lynn, R., & Irwing, P. (2004). Sex difference onprogressive matrices: A meta-analysis. Intelligence,32, 481–498.

Mastropieri, M. A., & Scruggs, T. E. (2006). Inclusiveclassroom: Strategies for effective instruction (4th ed.).Upper Saddle River, NJ: Prentice Hall.

Minskoff, E., & Allsopp, D. (2003). Academic successstrategies for adolescents with learning disabilities andADHD. Baltimore: Brookes.

Naglieri, J. A. (1997). Naglieri Nonverbal Ability Test—Multilevel Form. San Antonio: Psychological Cor-poration.

Naglieri, J. A. (1999). Essentials of CAS assessment. NewYork: Wiley.

Naglieri, J. A. (2003a). Naglieri nonverbal ability tests:NNAT and MAT-EF. In R. S. McCallum (Ed.),Handbook of nonverbal assessment (pp. 175–190).New York: Kluwer.

Naglieri, J. A. (2003b). Naglieri Nonverbal AbilityTest—Individual Form. San Antonio: Psychologi-cal Corporation.

Naglieri, J. A. (2008a). Traditional IQ: 100 yearsof misconception and its relationship to mi-nority representation in gifted programs. In J.VanTassel-Baska (Ed.), Critical issues in equityand excellence in gifted education series alternative as-sessment of gifted learners (pp. 67–88). Waco, TX:Prufrock Press.

Naglieri, J. A. (2008b). Naglieri Nonverbal AbilityTest—Second Edition. San Antonio: Pearson.

Naglieri, J. A. (2008c). Best practices in linkingcognitive assessment of students with learningdisabilities to interventions. In A. Thomas &


J. Grimes (Eds.), Best practices in school psychology(5th ed., pp. 679–696). Bethesda: NASP.

Naglieri, J. A., & Bornstein, B. T. (2003). Intelli-gence and achievement: Just how correlated arethey? Journal of Psychoeducational Assessment, 21,244–260.

Naglieri, J. A., Brulles, D., & Lansdowne, K. (2009).Helping gifted children learn. San Antonio, TX:Pearson.

Naglieri, J. A., & Das, J. P. (1997). Cognitive AssessmentSystem. Austin, TX: ProEd.

Naglieri, J. A., & Ford, D. Y. (2003). Addressingunder-representation of gifted minority chil-dren using the Naglieri Nonverbal Ability Test(NNAT). Gifted Child Quarterly, 47, 15–160.

Naglieri, J. A., & Ford, D. Y. (2005). Increasingminority children’s participation in gifted classesusing the NNAT: A response to Lohman. GiftedChild Quarterly, 49, 29–36.

Naglieri, J. A., & Kaufman, A. S. (2008). IDEIA2004 and specific learning disabilities: What roledoes intelligence play? In E. Grigorenko (Ed.),Educating individuals with disabilities: IDEIA 2004and beyond (pp. 165–195). New York: Springer.

Naglieri, J. A., & Paolitto, A. W. (2005). Ipsativecomparisons of WISC-IV Index scores. AppliedNeuropsychology, 12, 208–211.

Naglieri, J. A., & Pickering, E. B. (2003). Helpingchildren learn: Intervention handouts for use at schooland home. Baltimore: Brookes.

Naglieri, J. A., & Rojahn, J. (2001). Evaluation ofAfrican-American and White children in specialeducation programs for children with mentalretardation using the WISC-III and CognitiveAssessment System. American Journal of MentalRetardation, 106, 359–367.

Naglieri, J. A., & Ronning, M. E. (2000a). Comparisonof White, African-American, Hispanic, and Asianchildren on the Naglieri Nonverbal Ability Test.Psychological Assessment, 12, 328–334.

Naglieri, J. A., & Ronning, M. E. (2000b). Therelationships between general ability using theNNAT and SAT Reading Achievement. Journalof Psychoeducational Assessment, 18, 230–239.

Naglieri, J. A., & Yazzie, C. (1983). Comparison ofthe WISC-R and PPVT -R with Navajo children.Journal of Clinical Psychology, 39, 598–600.

Newcomer, P. (2001). Diagnostic AchievementBattery—Third Edition. Austin, TX: ProEd.

Pearson. (2005). Wechsler Individual AchievementTest—Second Edition. San Antonio, TX: Pearson.

Ramirez, R. R., & de la Cruz, G. (2002). The His-panic population in the United States: March 2002.Current Population Reports, 20-545, U.S. Cen-sus Bureau, Washington, DC.

Silverstein, A. B. (1982). Two- and four-subtestshort forms of the Wechsler Adult IntelligenceScale—Revised. Journal of Consulting and ClinicalPsychology, 50, 415–418.

Suzuki, L.A., & Valencia, R. R. (1997). Race-ethnicityand measured intelligence. American Psychologist,52, 1103–1114.

Wechsler, D. (1958). The measurement and appraisal ofadult intelligence (4th ed.). Baltimore: Williams &Wilkins.

Wechsler, D. (1991). The Wechsler Intelligence Scalefor Children—Third Edition. San Antonio, TX:Pearson.

Wechsler, D. (1997a). Wechsler Adult IntelligenceScale—Fourth Edition, San Antonio, TX: Pearson.

Wechsler, D. (1997b). Wechsler Memory Scale—ThirdEdition. San Antonio, TX: Pearson

Wechsler, D. (2002). Wechsler preschool and primaryscale of intelligence—Third Edition. San Antonio,TX: Pearson.

Wechsler, D. (2003). Wechsler Intelligence Scalefor Children—Fourth Edition, San Antonio, TX:Pearson.

Wechsler, D., Kaplan, E., Fein, D., Kramer, J.,Morris, R., Delis, D.C., & Maerlender, A. (2004).Wechsler Intelligence Scale for Children—FourthEdition—Integrated. San Antonio, TX: Pearson.

Wechsler, D. & Naglieri, J. A. (2006a). WechslerNonverbal Scale of Ability Technical and InterpretiveManual. San Antonio, TX: Pearson.

Wechsler, D. & Naglieri, J. A. (2006b). WechslerNonverbal Scale of Ability. San Antonio, TX:Pearson.

Wechsler, D. & Naglieri, J. A. (2006c). WechslerNonverbal Scale of Ability Administration andScoring Manual. San Antonio, TX: Pearson.

Wilkinson, G. S., & Robertson, G. J. (2006). WRAT4Wide Range Achievement Test professional manual(4th ed.). Lutz, FL: Psychological AssessmentResources.

Woodcock, R. W., McGrew, K.S., & Mather, N.(2001). Woodcock-Johnson III Test of Achievement.Itasca, IL: Riverside Publishing Company.

Wechsler Nonverbal Scale of Ability (WNV)

Documents

Transcript of Wechsler Nonverbal Scale of Ability (WNV)