UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Pantomime administration of the WISC-III and SB:FE to hearing and otitis prone Native Indian students Handy, Lynda Anne 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1995-983817.pdf [ 2.82MB ]
JSON: 831-1.0088999.json
JSON-LD: 831-1.0088999-ld.json
RDF/XML (Pretty): 831-1.0088999-rdf.xml
RDF/JSON: 831-1.0088999-rdf.json
Turtle: 831-1.0088999-turtle.txt
N-Triples: 831-1.0088999-rdf-ntriples.txt
Original Record: 831-1.0088999-source.json
Full Text

Full Text

PANTOMIME ADMINISTRATION OFTHE WISC-III AND SB:FE TO HEARING AND OTITIS PRONENATIVE INDIAN STUDENTSbyLYNDA ANNE HANDYB.Sc. University of Toronto, 1978B.Ed. University of Toronto, 1979MA. University of British Columbia, 1985A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF EDUCATIONInTHE FACULTY OF GRADUATE STUDIESEducational Psychology & Special EducationWe accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIA10 April 1993Lynda Anne Handy, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of‘ipThe University of British olumbiaVancouver, CanadaDate____________DE-6 (2/88)ABSTRACTAdministration of nonverbal tests of intelligence to deaf or hard of hearingand Native Indian children has been problematic. Communication difficultiesbetween examiner and examinee have resulted in modifications to theadministration of the test. A standard method of test administration has notbeen reported in the literature.The current study investigated the use of pantomime instruction to administerperformance scales of the Wechsler Intelligence Scale for Children- ThirdEdition (WISC-IlI) and the Stanford Binet: Fourth Edition (SB:FE). Subjectsincluded a group of otitis prone Native Indian children (n=87) and hearing peers(n=80). Subjects in the hearing peer group were randomly assigned to verbal(n=40) or pantomime instruction (n=40) to compare verbal and pantomimeadministration of the tests.T-test and ANOVAs were used to investigate comparisons between verbaladministration method (hearing peer) and between groups receiving pantomimeinstruction (otitis prone and hearing peer). Multiple regression analyses wereused to determine whether nonverbal tests of intelligence were predictors ofacademic achievement.Results demonstrated pantomime instruction gave subjects information requiredIIto complete test and subtest tasks of the WISC-lll and SB:FE. There was nota significant difference between verbal and pantomime instruction for thehearing peer group. When the otitis prone group and the hearing peer groupwere compared there was no significant differences between groups for theWISC-lll. Item analysis provided additional support to pantomime instructiongiving adequate information to understand tasks. Multiple regression analysesdemonstrated that nonverbal measures of intelligence should not be used topredict math or reading achievement for this population.Language measures administered yielded very low scores across all threegroups. Although this was expected for the otitis prone group, low scores werenot expected for the hearing peer group. The short term memory subtests forthe SB:FE were also very low indicating another area of weakness for all threegroups which was not expected.Results of the present study provides evidence to indicate pantomime instructionis a viable method of administration when used with otitis prone and hearingNative children. Further research is necessary to determine if standardpantomime administration can be utilized to administer nonverbal measures ofintelligence to other groups of children with delayed language skills orcommunication difficulties.iiiTABLE OF CONTENTSAbstract iiTable of ContentsList of Tables viiAcknowledgments ixI. Introduction 1A. Background 1B. Use of Intelligence Scores 2C. Assessment of Native Indian Children 4D. Difficulties Associated with Assessment 6E. Assessment of the Hearing Impaired 8F. Difficulties Assessing the Hearing Impaired 11G. Summary 13II. Review of the Literature 16A. Assessment of Native Indian Children 161. Background 162. Intellectual Assessment of Native Indian Children 173. Assessment of Academic Skills 204. Intelligence and Achievement 215. Difficulties Associated with Assessment 246. Summary of Assessment Research of NativeIndian Children 26B. The Impact of Chronic Otitis Media 261. Background 262. Otitis Media and Cognitive Development 283. Speech/Language Difficulties 304. Retrospective and Prospective Studies 32a. Retrospective Studies 32b. Prospective Studies 345. Otitis Media and Native Indian Children 366. Summary of Research on the Impact of Otitis Media 39ivC. Assessment of Hearing Impaired Children 401. Background 402. Intellectual Assessment of the Hearing Impaired 413. Assessment of Academic Skills 444. Academic Achievement and Nonverbal Intelligence 475. Test Modifications for Hearing Impaired 496. Difficulties Associated with Assessment 547. Summary of Research on Assessment of HearingImpaired Children 57D. SummaryIII. Statement of the Problem 60A. Background 60B. Purpose of the Study 61C. Research Questions & Hypotheses 62Hypothesis #1 63Hypotheses #2a & 2b 64Hypothesis #3a & 3b 65Hypothesis #4a & 4b 66IV. Methodology 68A. Sampling 68B. Procedure 72C. Instruments 731. The Weschler Intelligence Scale for Children-Ill 732. The Stanford Binet Test of Intelligence 773. Stanford Achievement Test of the Hearing Impaired 804. Peabody Picture Vocabulary-Revised 82D. Data Analyses 83V. Results 86A. Hypothesis #1 87B. Hypothesis #2a 96C. Hypothesis #2b 98D. Hypothesis #3a & 3b 100E. Hypothesis #4a 106F. Hypothesis #4b 111G. Item Analysis 113H. Summary of Results 121VVI. Discussion 123A. Review of the Purposes of the Study 123B. Hypothesis #1 123C. Hypothesis #2a 130D. Hypothesis #2b 131E. Hypothesis #3a 133Hypothesis #3b 136F. Hypothesis #4a 137G. Hypothesis #4b 139H. Influence of Otitis Media 145I. Pantomime Instruction 147J. Limitations to the Study 150K. Implications of the Study and Future Research 154VII. References 158VIII. AppendicesAppendix A 172Appendix B 176Appendix C 178viLIST OF TABLESTable 1: Subjects by group and method of instruction 71Table 2: WISC-Ill subtest for performance I.Q. andIndex scores Group 77Table 3: SB:FE Composite score, area socres andsubtest scores 79Table 4: Descriptive Statistics for WISC-Ill by Group 89Table 5: Descriptive Statistics for SB:FE andPPVT-R by Group 90Table 6: T-test for the WISC-lll Group when comparedto standardization sample 93Table 7: T-test for the SB:FE and PPVT-R by Group whencompared to standardization sample 94Table 8: Confidence Intervals and Means for theWSPIQ, WSPO, WSPS, SB:FE Composite scoreand Area score 97Table 9: Regression Equations, R Squared Values andMultiple Correlation Coefficients for ReadingComprehension and Math Computation forGroups I, II and III 103Table 10: Standardized Regression Coefficients, R SquaredValues and Multiple Correlation Coefficients forReading Comprehension and Math Computation forGroups I, II, and Ill 102Table 11 : Correlations Between Measures of Intelligenceand Index/Area scores for Group I 107Table 12: Correlations Between Measures of Intelligenceand Index/Area scores for Group II 108Table 13: Correlations Between Measures of Intelligenceand Index/Area scores for Group Ill 10911ITable 14: Correlations Between Subtests of the WISC-llland Partial Composite SB:FE, Index andArea scores for Group I 111Table 15: Correlations Between Subtests of the WISC-llland Measures of Intelligence, Index andArea scores for Group II 112Table 16: Correlations Between Subtests of the WISC-Illand Measures of Intelligence, Index andArea scores for Group III 113Table 17: Correlations Between Nonverbal Measures ofIntelligence and SAT-HI (Groups I, II, Ill) 114Table 18 : Group II (Pantomime Instruction) ParameterEstimates Responses for the Picture Completionsubtest (WISC-Ill) 118Table 19 : Group Ill (Verbal Instruction) ParameterEstimates Responses for the Picture Completionsubtest (WISC-lll) 119.vlIIACKNOWLEDGEMENTSMany individuals influence decisions and choices you make throughout your life.I have been very fortunate to have a very supportive and loving family, caringfriends and colleagues who have supported me through decisions.Thanks are extended to committee members Dr. Janet Jamesion, Dr. NandKishor, and Dr. Art More for their time and encouragement through theconceptualization of the study to the final drafts and revisions. Thanks are alsoextended to Dr.Emily Goetz and Laureen Niki for their assistance in collectionof the data.Dr. Perry T. Leslie served as both research supervisor and committee chair formy doctoral work. He has been a mentor, friend and colleague, alwayssupportive and encouraging. His influence over my professional life has beenfar reaching. I only hope I can emulate the support, integrity andencouragement he has shown to me to others.John provides love and laughter to my life. I thank him for his support,encouragement and helping me to stand back and look at the “big picture”.His encouragement required me to expect the best from myself whichsometimes was difficult through the dissertation process.ix/1I. INTRODUCTIONA. BackgroundDefinitions of intelligence emphasize the ability to perform, adjust or adapt to theenvironment, the ability to learn, to perform abstract thinking tasks (Zachary, 1990)and to assess learning that has occurred in a variety of life experiences. Thedivergence in definitions and measurements of intelligence can be attributed to thenotion that intelligence is an attribute, not an entity, and that it reflects thesummation of learning experiences of the individual (Weinburg, 1991).Cognition and intelligence are not synonymous terms (Horowitz & Leake, 1979).Behaviour which is labelled as cognition involves actions which are influenced byassumptions the child makes regarding how the world operates. Intelligence is amore global term and reflects the child’s relative standing in ability in relation tochildren the same age.As early as 1939, Wechsler defined intelligence as “the aggregate or globalcapacity of the individual to act purposefully, to think rationally, and to deal with hisenvironment” (Wechsler, 1939, p.3). Wechsler’s definition states that intelligenceis composed of qualitatively different abilities, not the sum of abilities. He took apragmatic view of intelligence stating intelligence is known by what it enables usINTRODUCTION /2to do (Sattler, 1988).It is important to provide some clarification as to how nonverbal intelligence isdefined. For the purposes of this study nonverbal intelligence will be defined asthe measure of intelligence or score obtained by administration of the PerformanceScales of the WISC-llI and the nonverbal subtests of the Stanford Binet: FourthEdition.B. Use of Intelligence ScoresTraditionally IQ scores have been used in the diagnosis of learning problems,labelling and classification of children. As with hearing children, lQ scores havebeen used as predictors for academic achievement for hearing impaired children(Zachary, 1990). Significant differences between IQ scores (predictedperformance) and achievement scores (actual performance) have been used ascriteria for class placement with both hearing and hearing impaired children.Educators have also used IQ scores as a reference point for underachievement,to determine whether a child is making adequate progress (Watson, Goldgar,Kroese, & Lotz, 1986).The WISC-R has been the instrument of choice when assessing intelligence orINTRODUCTION /3ability of Native Indian children(Hynd & Garcia, 1979). The WISC-R Performancesubtests are thought to be the most accurate indicator of potential. However, inseveral studies (Hynd & Garcia, 1979; Chrisjohn & Lanigan, 1986), verbally loadedsubtests (WISC-R, Verbal Scales) have accurately predicted academic levels withinan English speaking academic setting (Hynd & Garcia, 1979).Kaufman (1982) disputed this statement somewhat by reporting that the WISC-RPerformance IQ scores possessed a reasonable degree of construct and predictivevalidity with groups of minority children. He felt the WISC-R Performance IQ couldbe used to predict academic performance of Native Indian children in a traditionalEnglish speaking class. Studies present conflicting results, indicating a need forfurther investigation.The relationship between lQ and achievement has been investigated with hearingimpaired children (Brill, 1962; Watson et al. 1986; & Phelps & Branyan, 1990).After reviewing the literature examining the relationship between nonverbal 10 andachievement with hearing impaired samples, there is some doubt as to whethernonverbal IQ scores can be used as a reference point for educational planning(Watson, et al., 1986)INTRODUCTION /4However, Phelps & Branyan (1990) demonstrated that the relationship betweennonverbal IQ scores and achievement had similar significance for hearing impairedchildren as for their hearing peers. Subjects in this study were hearing impairedstudents in a public school setting. Results indicated a recommendation to useappropriate and up to date nonverbal intellectual measures that can facilitate theidentification of underachievement in populations of hearing impaired children.C. Assessment of Native Indian ChildrenIntelligence and achievement measures have been administered to Native Indianchildren for the purpose of identification and placement (Seyfort, Spreen, &Lahmer, 1980). Studies of intellectual ability of Native Indian children indicate thathigher scores are obtained on performance or nonverbal tests than verbal tests(McShane & Plas, 1984). This pattern may be reflective of language difficulties,cultural and/or experiential factors. Consequently, only nonverbal measures areconsidered to be valid estimates of ability with this population (McShane & Plas,1984).Additional explanations could include weak English language skills and hearingloss due to chronic otitis media. The incidence of otitis media is extremely high inINTRODUCTION / 5many Canadian Native communities (28-33 %) (Bachor & Crealock, 1988;Scaldwell, 1989). This rate has been noted as being significantly higher than inother minority populations (Beery, Doyle, Cantekin, Bluestone & Wilt, 1980). Otitismedia is known to be difficult to diagnose and is recurrent, even when properlytreated (Roberts & Schuele, 1990).Children with the “condition” otitis media are faced with continual changes inauditory input. Otitis media is extremely difficult to detect due to its fluctuatingnature. Holme and Kunze (1969) published one of the first reports linking a historyof chronic otitis media with impaired performance on linguistic and academic tasks.It is apparent after a review of current research that even a mild fluctuating lossdoes impede speech and language development, as well as the acquisition ofacademic skills (Roberts & Schuele, 1990). All of these factors make learning ina predominately aural/oral environment very difficult and frustrating.Horowitz and Leake (1979) investigated the relationship between cognitivedevelopment, language development and tested intelligence. Their study revealedthat consistent and repeated auditory input is important to language and cognitivedevelopment. If this input is missing, as when a child experiences recurrent otitisINTRODUCTION I6media, the child is likely to be missing important language experiences. Recurrentotitis media and associated periodic hearing loss typically indicate an inconsistentlearning environment, which makes learning through oral instruction extremelydifficult and frustrating.After reviewing the literature on assessment of Native Indian children, it is apparentthat the child’s hearing status has not been considered as a variable which mayeffect the outcome of test results. It is hypothesized that current hearing statusand history of otitis media are variables which are of the utmost importance whenassessing intelligence and achievement skills. Although nonverbal measures ofintelligence are routinely administered, a method of standard administration whichdoes not rely on verbal instructions requires investigation and study.D. Difficulties Associated with AssessmentSternberg (1984) has had an influence in adding to the understanding ofassessment of Native Indian children. His contextualist view of intelligence statessimply that intelligence is considered in respect to the circumstances under whichan individual develops and operates. Therefore, tests developed and formed onINTRODUCTION / 7the white population will reflect those skills and capabilities which are adaptive andrelevant to the white culture. They will not necessarily be those which are adaptiveand relevant to the Native way of life (Common & Frost, 1988).Irvine and Saunders (1972) also warn that one cannot assume that a testmeasures the same “thing” in disparate groups. They suggested a number ofmodifications for increasing the comparability of tests across cultures. First, itemcontent was changed to reflect Canadian rather than American life. These changesare common in the use of the WISC-R by Canadian psychologists, but refer mainlyto the Verbal Scales. A second suggestion was to involve community membersfor their input concerning possible problems in tests and procedures. Third,members of the community were trained to administer the test under thesupervision of qualified psychologists.Nonverbal measures of intelligence administered to Native Indian children oftenrequire verbal instructions. This presents two major difficulties. The first is thelevel of linguistic competence necessary to comprehend instructions. If English isnot the child’s first language, instructions are often translated. The second asChrisjohn, Towson, Pace and Peters (1989) suggest is to modify time limits andINTRODUCTION /8alter the ceiling criterion or cut off levels as possible alternatives whenadministering standardized tests.Native Indian children may be placed at a further disadvantage during theassessment process if a hearing loss, as a result of chronic otitis media, isundetected. In order to address and investigate the difficulties of testadministration with hearing impaired Native Indian children, literature pertaining tothe assessment of the hearing impaired should be reviewed.Children with a sensorineural loss are, typically, subjects in the majority of researchpertaining to the assessment of the hearing impaired. It is important to review thisliterature to assess whether results can be generalized to a native population withconductive loss due to chronic otitis media.E. Assessment of Hearing Impaired ChildrenThe attempt to determine a nondiscriminatory method of evaluating hearingimpaired individuals has a long history (Bragman, 1982). Pintner and Patterson(1915) found that deaf children scored within the mentally handicapped rangewhen given verbal IQ tests. They hypothesized this was due more to languagedeprivation than deafness, and developed the Pintner Non-Language Test (Pintner,1924).INTRODUCTION /9Using this test, deaf children continued to score significantly lower than hearingsamples (Vernon, 1966). Contrary to these results, Graham and Shapiro (1953)found that when language was not a factor, deaf and hearing children wereapproximately equal in mental ability. The controversy surrounding intellectualequality has been longstanding and will be discussed further in Chapter II.During the next thirty years investigators continued to explore alternative methodsfor measuring the intelligence of hearing impaired children (Bragman, 1982). Themajor implication of this body of research is that verbal intelligence measures aremore valid for measuring hearing impaired children’s language difficulties than theiractual mental capacity or intelligence (Sullivan, 1982).The WISC-R Performance Scales have been found to be both reliable and validinstruments for assessment of hearing impaired children (Hirshoren, Hurley &Kavale, 1977; Hurley, Hirshoren, Kavale & Hunt, 1979). Although the subtests are“nonverbal” measures of intelligence (Sullivan, 1982), they are administered orally,relying on the child’s receptive language and auditory skills. Verbal administrationof test instructions has been recognized as inappropriate for severely andprofoundly hearing impaired children (Graham & Shapiro, 1953; Murphy, 1957;Sullivan, 1982).INTRODUCTION / 10Modifications have been suggested by various investigators in an attempt toalleviate this problem. Reed (1970) administered practice items for the majority ofsubtests on the WISC in an attempt to increase task comprehension. The use ofvisual aids to increase understanding of the task were investigated by Murphy(1957) and Neuhaus (1967). Graham and Shapiro (1953) suggested pantomimeinstruction as an alternate to standard verbal directions. Overall research resultswere inconclusive.What remains consistent is that verbal administration or instructions for nonverbaltests are inappropriate for severely and profoundly deaf children (Sullivan, 1982).The effect of modifications on the reliability and validity of test results with hearingimpaired subjects has been questioned (Sattler, 1988). Substitution of instructionsin the administration process violates standard procedures. Prior research has notadequately demonstrated whether substitutions adversely affect the quality of thehearing impaired child’s performance (Sullivan, 1982).Achievement levels of hearing impaired children have been found to be much lowerthan one would expect from performance on intelligence tests (Levine, 1971).Reading achievement is consistently below expected or predicted scores. Minimalexposure to language stimulation is the factor thought to account for the limitedINTRODUCTiON /11acquisition of reading skills(BrilI, 1962; Levine, 1971; Phelps & Branyan, 1990).Low or minimal language skills provides one explanation for low achievement levelsresulting in low or nonsignificant correlations between achievement scores andpredicted intelligence or ability (Brooks & Riggs, 1980). Average correlationsreported for hearing subjects between IQ and academic achievement is .50(Jensen, 1980).Empirical research is lacking on a similar relationship with hearing impairedchildren (Phelps & Branyan, 1990). Pollard and Oakland (1985) concluded that theWISC-R Performance lQ was the only variable to correlate with SAT-HI reading andmathematics scores. Watson et al. (1986) found the measure of intelligence criticalwhen investigating the correlations between intelligence and achievement. Theyconcluded that the Learning Quotient (LQ) obtained from the Hiskey Nebraskaaccounted for only one third of the variance in achievement scores. Therefore,intelligence tests administered and IQ scores obtained are only one factor whenexamining individual differences.F. Difficulties Assessing the Hearing ImpairedNonverbal performance tests have been recommended for use with hearingINTRODUCTION I 12impaired children for many years. Pintner and Patterson (1915) found that deafchildren scored within the mentally handicapped range when given a verbal testof intelligence. They hypothesized that this was due to Language deprivation morethan to the child’s hearing loss. A valid assessment instrument that measuresintellectual capacity yet requires a nonverbal response continues to be a need inthe field (Phelps & Ensor, 1986).Administration modifications suggested include pantomime (Graham & Shapiro,1953; Murphy, 1957) and visual aids (Murphy, 1957; Neuhaus, 1967; Reed, 1970)to assist the hearing impaired child’s understanding of test directions.Unfortunately, the method(s) of administration have not been reported accuratelyin research studies. One cannot assume uniformity of presentation of instructionswith a heterogeneous population of hearing impaired children (Ray, 1982). Lackof standard procedures leads to problems in replication studies, may limit thefurther use of the test, and leads to questions about the tests’ validity due to thesemodifications (Gerweck & Yssledyke, 1975; Sattler, 1988).Goetzinger and Houchins (1969) investigated effects of pantomime and verbalinstruction on student performance. The Raven’s Progressive Matrices wasadministered to 40 deaf and 40 hearing children. Pantomime instructions wereINTRODUCTION /13given to all deaf subjects, whereas half the hearing subjects received pantomimeinstructions and half received verbal instructions. When the two groups of hearingsubjects were compared, no differences were found. Likewise, when hearingimpaired and hearing subjects who received pantomime instructions werecompared, no differences in test scores were found.The question remains as to whether information relayed via pantomime is adequateto ensure comprehension of the task. Standard modifications of instruction for usewith standardized measures would allow examiners consistency in theiradministration of a test, regardless of the hearing impaired child’s linguistic abilitiesor communication mode. It is important to determine the effects of pantomimeinstructions on test performance. The need to determine procedures for testadministration that do not discriminate for or against a specific population requiresfurther research (Bragman, 1982).G. SummaryDespite identification of problems associated with assessment of Native Indianchildren (Goldstein, 1988), presence of a hearing loss has not been considered animportant variable in the assessment process. The hearing status of the NativeINTRODUCTION / 14Indian child is of utmost importance if test results are to be valid and reliable foreducational decisions. Therefore, the literature pertaining to the assessment ofhearing impaired was reviewed.The body of literature revealed many difficulties associated with assessment ofintelligence and academic skills. Lack of standardized tests and standard testinstructions due to the heterogeneity of the population was identified as asignificant difficulty or problem to overcome (Bragman, 1982). It appears a methodof administering tests of intelligence and achievement utilizing a standardizedmethod should be investigated. The most viable method of administration appearsto be pantomime instruction (Courtney, Hayes, Couch & Frick, 1984), as testscould be administered to Native Indian children regardless of their hearing statusor language level.In the present study, nonverbal subtests of the WISC-lll and the SB:FE wereadministered to otitis prone Native Indian students using standard pantomimeinstructions (Appendices A and B). Results were then compared to the Nativechild’s immediate peer group (hearing students) and to the normative sample. Inaddition, it was important to assess whether measures of intelligence administeredwith standard pantomime instruction are predictive of academic achievement.INTRODUCTION /15The purposes of this study are to investigate: 1) the use of standardizedpantomime instructions for the nonverbal scales of two recently revised measuresof intelligence, namely the WISC-lll and the SB:FE and 2) to examine whethernonverbal intelligence measures administered with pantomime instruction arepredictive of academic performance for hearing and otitis prone Native Indianchildren.Research pertinent to assessment of Native Indian children, the relationshipbetween otitis media, intelligence and academic skills and assessment of otitisprone children will be reviewed in Chapter II./ 16II. REVIEW OF THE LITERATUREThe literature will be reviewed in three areas critical to the study: assessment ofNative Indian children, effects of otitis media on cognitive and academic skills, andthe assessment of hearing impaired children.A. Assessment of Native Indian Children1. BackgroundPsychologists and educators have been interested in understanding thedevelopment of intellectual abilities of the Native Indian child (Hynd & Garcia, 1979;Seyfort, Spreen, & Lahmer, 1980). Although these children comprise a significantproportion of the school population in the U.S. and Canada, very little is knownabout them (Goldstein, 1988).McShane and Plas (1984) reviewed 100 studies invoMng the WISC and WISC-Rwith Native Indian children in the U.S. and Canada. These studies are typicallycomparative in nature (Goldstein, 1988), and leave many questions regarding validassessment of Native children unanswered. Hynd and Garcia (1979) state that thelack of information regarding the assessment of Native Indian children suggestsan area of critical need.REVIEW OF THE LITERATURE 117If psychologists are unfamiliar with typical patterns of behavior within the Nativeculture, then an over reliance on standardized assessment devices may give invalidresults which are used when making educational decisions (Common & Frost,1988).2. Intellectual Assessment of the Native Indian ChildThe Performance Scales of the WISC-R, the Visual reasoning subtest of theStanford Binet : Fourth Edition (SB:FE), the nonverbal scales of the KaufmanAssessment Battery for Children (K-ABC), and the Raven’s Progressive Matricesare most frequently administered to Native Indian children as nonverbal measuresof intelligence (Sattler, 1988).The majority of studies investigating Native Indian intelligence have relied heavilyon the WISC-R and its forerunner, the WISC (Hynd & Garcia, 1979). Results ofthese studies provide a consistent pattern of findings; Native individuals do notperform in the same manner as the Non-Native population on standardized testsof intelligence (Common & Frost, 1988). There have been several Canadianstudies which have examined Native Indian children’s intelligence by administeringREVIEW OF THE LITERATURE I 18the WISC and WISC-R. Seyfort, Spreen, and Lahmer (1980) found Native childrendisplayed average perceptual motor skills and poor verbal skills. Results of theirstudy indicated that those using WISC-R scores should interpret their findings withextreme caution. -When interpreting results of research, it is important to note that the majority ofsubjects live on reserves where indigenous cultural values would most likelyprevail. Orientation towards speed, accuracy, and verbal fluency play a minor rolein Native Indian culture (Goldstein, 1988). Familiarity with the spoken and writtenword was influenced by the degree of bilingualism as elders and peers frequentlyconversed in their Native language (Seyfort, Spreen & Lahmer, 1980).Wilgosh, Mulcahy, and Watters (1986) reported the assessment results for 366Inuit children from two school districts in the Northwest Territories. Verbal Scoreson the WISC-R were below 70 for 77.4% of the children in the study, whereas thepercentage of children scoring below 70 on the Performance Scale was 5.74%.The resulting Full Scale lQ scores indicated that 32.24% of the children in thesample scored below 70 and, according to test norms, would have been classifiedas mentally retarded.St. John and Krichev (1976) studied Cree and Ojibwan children. Their results wereREVIEW OF THE LITERATURE / 19similar to those from the Wilgosh et al.(1 986) study. Verbal IQ scores obtained bySt. John and Krichev ranged from 69.70 to 91.10 for the Cree and Ojibwan childrenin the study. All of their Verbal lQ scores were significantly lower than thePerformance lQ scores, which were average according to the test norm (about100).Seyfort, Spreen, and Lahmer (1980) examined WISC-R results of 177 Nativechildren in southwestern B.C. The children all lived on reserves, but attended localcommunity schools. Results supported previous research where subjectsdisplayed average Performance lQ scores in contrast to poor Verbal lQ scores.The three verbal subtests which seemed to contribute to differences in the studyand the test forming population were Information, Vocabulary, andComprehension (Common & Frost, 1986).The pattern of at least average Performance lQ scores and of significantly lowerVerbal IQ scores has been obtained by researchers working in different parts ofCanada, studying Natives from different linguistic and indigenous peoples’ groups.The groups have varied in the degree to which they have had contact with thenon-Native or non-Indian culture. Some children lived and attended school inisolated settlements. Others attended integrated schools and lived on theirREVIEW OF THE LITERATURE /20reserves, and still others attended schools and lived off the reserve in large urbanareas. Goldstein (1988) has labelled this verbal-performance discrepancy the“Native Profile”.3. Assessment of Academic SkillsReported achievement levels of American and Canadian Indian children have beenconsistently below the performance of the majority group. Native children performconsistently lower on measures of achievement when compared to the majoritygroup, using either grade or age norms (McShane, 1982).MacArthur (1968) found that by the end of the third grade, the achievement lag is1/2 to 1 1/2 years below grade level, and by the end of grade twelve, Indianstudent achievement was 2 to 3 1/2 years below grade level. McShane (1982)stated it is generally assumed that there is a sharp drop in achievement at the endof grade three followed by a period of substantial gain from the fourth to seventhgrades. Bryde (1968) coined the term “crossover phenomenon” to indicate amarked depression of achievement scores at the end of primary, gains to theseventh grade, and a marked decline of skill acquisition after grade seven.Tomusiak (1983) and Scaldwell and Frame (1 985a) have noted that Indian childrenREVIEW OF THE LITERATURE I 21score consistently lower than the normative group on Spelling, Mathematics, andReading tests across all age levels.Scaldwell (1989) believes otitis media to be one of the more important factors toinfluence functional language skills among Native Indian school children. If thereis a connection between acquisition of language skills and academic skills such asreading and spelling, then hearing status is an important variable.4. Intelligence and AchievementInvestigators who have been concerned with low Native Indian achievement levelshave often focused attention on Indian scoring patterns on the WISC-R scales.Several hypotheses have been advanced in an attempt to partially explain Indianachievement and intelligence test performance (McShane, 1982). Despite uniqueachievement and intelligence test patterns of Native students, little systematicattention has been given to the relationship between academic achievement andspecific skills measured by intellectual instruments. If factors contributing todepressed achievement are to be understood, skills that are related to the Indianchild’s growth in reading and mathematics achievement need to be identified(McShane & Plas, 1984). This can be accomplished if valid instruments areREVIEW OF THE LITERATURE /22administered to Native Indian children, and if important research findings movefrom the theoretical to the practical domain.Underachievement patterns of American Indian children were investigated byMcShane and Plas (1984). They believed overall intelligence scores on standardinstruments masked the unique patterns of abilities characteristic of Indian children.The authors used the Illinois Test of Psycholinguistic Abilities (ITPA) and theWISC-R as measures of cognitive skills. Results indicated selected subtests of theITPA were significantly associated with reading, language and mathematics skills.The ITPA proved to be superior to the WISC-R as a predictor of academicachievement.Several hypotheses have been advanced as a partial explanation of lower NativeIndian achievement and intelligence test performance relative to the non-Nativepopulation. These include socio—economic factors, cultural conflict, racialintegration difficulties (Miller, 1968), and the excessive prevalence of chronic middleear disease (otitis media) among young children from all tribes (McShane & Plas,1984; Wallace, 1973).It has been suggested by Scaldwell (1989) and others that the Performance lQfrom the WISC-R is the least biased measure currently available to assess theREVIEW OF THE LITERATURE /23intellectual skills of Native Indian children. Although Performance subtests of theWISC-R may be the best indicator of potential to perform, it would appear thatverbally loaded tests can continue to serve as indices of predicted levels offunctioning within an English speaking setting, or more specifically within academicsettings (Hynd & Garcia, 1979).Scaldwell (1989) concluded that the WISC-R may be the preferred instrument foruse, with the Full Scale being a good estimator of general intelligence, and theVerbal Scale a good predictor of academic success in school. Common and Frost(1988) do not agree with this position, especially the calculation of a Full Scale lQscore. A Full Scale lQ score can mask the significant differences between theVerbal lQ and Performance IQ scores. In addition, Common and Frost (1988)argued that Scaldwell (1989) did not have any basis for stating that the VerbalScale is a good predictor of academic achievement, as the study did not examinethe relationship of children’s 10 scores and their achievement in school. St. Johnand Krichev (1976) found neither the Verbal or Performance Scale of the WISC tobe predictive of academic achievement for groups of Canadian Native Indianchildren.Another important finding from the St. John and Krichev study was that WISCscores were unrelated to school performance as measured by the grades achievedREVIEW OF THE LITERATURE /24by students. If one purpose of administering measures of intelligence is to identifystudents as exceptional and to make predictions regarding academic potential(Watson et al., 1986), it is important to study this relationship. St. John andKrivchev (1976) concluded that the use of the W)SC with Native students couldlead to misclassification of large numbers of children as mentally handicapped.5. Difficulties Associated with AssessmentSchool psychologists faced with assessment of ethnic or minority children arefaced with the problem of administering an instrument which has beenstandardized on a population representative of the general population (Seyfort etal., 1980). Although the particular ethnic group may be part of the standardizationsample, the pattern or profile unique to that group is lost in the overall populationvariance. This is one of the difficulties suggested by Chrisjohn and Lanignan(1986) which may limit the use of research findings.Additional difficulties include: 1) small sample sizes, 2) use of improperinstruments, 3) lack of fundamental psychometric research, and 4) lack of theory.Small sample size has been identified as a difficulty with research (Goldstein,1988). Limited sample size has made any results obtained difficult to generalizeacross age range, and between Native groups across Canada. Often results haveREVIEW OF THE LITERATURE /25been generalized to the larger population without an attempt to examine testresults and their impact on Native Indian children (Scaldwell, 1989).Native children’s communication style differs from the non-Native culture (Everett,Proctor, & Cartwell, 1983). For example, Native Indian children may have difficultyanswering questions, as they may perceive them as rude. They may also be slowin accepting help or answering questions and may maintain long silences duringinterviews. These results become an important consideration when administrationof intelligence measures is taken into account. If nonverbal intelligence measuresare administered using verbal directions, students may not comprehend theinstructions, nor may they understand the importance of answering (Sattler, 1988).If English is not the child’s first language, instructions are often translated, yetanother violation of standard administration procedures (Goldstein, 1988).McShane and Plas (1984) state the prevalence of middle ear infection in youngNative Indian children may provide a partial explanation of poor test performance.The authors do not state how test performance may be affected. The predictionthat otitis media has a significant effect on test performance and results shouldinfluence how tests are administered to children with this condition.REVIEW OF THE LITERATURE / 266. Summary of Assessment Research of Native Indian ChildrenThe WISC-R is the preferred instrument to assess Native Indian children. Researchhas demonstrated that Verbal IQ scores are significantly lower than PerformanceIQ (PIQ) scores when the WISC-R is administered. A pattern of average PIQ forNative Indian children has emerged.Difficulties in assessment have been identified as including small sample size,improper instruments, lack of theory, and cultural differences. In addition,Scaldwell (1989) reports that the high incidence of otitis media in Native Indianchildren may influence functional language and acquisition of speech and languageskills and academic skills. Otitis media is a variable which requires examinerattention in the assessment of Native Indian children.B. The Impact of Chronic Otitis Media1. BackgroundOtitis media, or inflammation of the middle ear, is a common childhood disease(Friel-Patti, 1990). In its chronic form, it has been shown to result in hearing lossin many of the children affected (Bluestone & Klein, 1983). Due to the deleteriouseffects of hearing loss on language and learning (Menyuk, 1980), it has beenREVIEW OF THE LITERATURE / 27hypothesized that children with chronic otitis media may be at risk for languagedisorders which often result in diminished academic performance (TalIal, 1991).Estimates from outpatient clinics and from private practices (Klein,1980) indicatethat between 75% and 95% of all preschool children are affected by otitis mediafor some period of time. Children who are treated for chronic otitis media, wheresymptoms persist for an eight week period, were found to be at risk for speechand language problems which lead to learning and academic problems in school(Menyuk & Rosener, 1988; Zinkus & Gottlieb, 1980).Several studies which suggest a causal relationship between a history of chronicotitis media and language impairment have been criticized on methodologicalgrounds. Leviton (1980), in reviewing the literature, stated there was lack ofcontrol over other factors known to effect language development, such as IQ,socioeconomic status, and language stimulation in the home.Children used as subjects in various studies were often assessed during anepisode of otitis media, making it difficult to distinguish whether lower scores werea result of a history of chronic otitis media or a general malaise at the time oftesting.REVIEW OF THE LITERATURE /28Otitis media, although prevalent, often escapes detection since the symptoms areoften absent or unapparent, or otoscopic examination is difficult to accomplish.Tympanic membrane abnormalities are often subtle (Paradise et al., 1976) makingdetection very difficult if appropriate evaluation has not occurred. Pure toneaudiometry has traditionally been used as a screening test, but many have foundthe pure tone measure does not adequately detect middle ear abnormalitiesassociated with otitis media (Northern, & Downs 1978). Impedance screening willnot only identify those individuals who have otitis media, but Paradise et al. (1976)state it is “superior” for detection. A full audiological evaluation is always preferred,and necessary if otitis media is to be detected.2. Otitis Media and Cognitive DevelopmentCognitive development is a broad term that subsumes the growth of problemsoMng, thinking, judgement, and reasoning or the higher mental processes (Gray,1983). Gray (1983) discusses the index of cognitive development associated withmeasurement, the indMdual intelligence test. She states that although tests ofintelligence are useful tools, they cover only certain manifestations of cognitivedevelopment and tell nothing of the underlying mechanisms or structures. Mosttests are heavily dependent on receptive and expressive language.REVIEW OF THE LITERATURE /29Consistent and discriminated auditory input is obviously necessary for languagedevelopment to take place (Horowitz & Leake, 1982). Frequent bouts of otitismedia mean frequent periods of fluctuating hearing loss. Otitis media might beexpected to affect cognitive development, if the connection between language andcognition is assumed (Gray, 1983). This may be especially true in the first fewyears of life when the prevalence of otitis media tends to be the highest, and whenauditory experience is essential for skill development.Downs (1977) found children who suffered frequent otitis media in the first year oflife to have lower IQ scores. She postulated that poor hearing that accompaniesotitis media and the subsequent hearing loss leads to an inability on the part ofthe child to learn all the strategies necessary to understand sounds.HoIm and Kunze (1969) administered the ITPA to 16 children, ages 5 to 9, whohad chronic otitis media. They were matched with a control group on age, sexand socioeconomic status. Performances on the ITPA were compared with resultsindicating depressed scores for the otitis media group on subtests involvingauditory discrimination and attention. Subtests stressing visual and motor aspectswere not as low as auditory subtests, but below age expectations.Zinkus, Gottlieb and Shapiro (1978) studied 40 children with otitis media matchedREVIEW OF THE LITERATURE /30with control peers. When examining differences between the two groups, theyfound no difference on developmental milestones or on subtests invoMng visualcompetence. However, when examining individual results, children with severeotitis media histories had difficulty in reading and spelling but not arithmetic.Poorer performance was also noted on visual tasks of the WISC when an auditorycomponent was involved.3. Speech/Language and Academic DifficultiesMenyuk (1986) identified a controversy in the literature concerning the relationshipbetween persistent otitis media and academic and language problems. Oneargument presented (Roberts, Sanyal, Burchinal, Collier, Ramey, & Henderson,1986) is that otitis media has no effect on speech and language because childrenexperience only mild to moderate hearing losses (i.e., from 10 to 40 dB). Thisargument assumes that after the otitis media episode, the child’s hearingthresholds return to normal, and that chronic otitis media decreases dramaticallyafter the age of three (Friel-Patti, 1990).There are others who argue that chronic otitis media with effusion (OME) in earlychildhood places children at risk for speech and language difficulties, which mayREVIEW OF THE LITERATURE /31lead to academic difficulties in school (Klein, Chase, Teele, Menyuk & Rosner,1988). This position is based on an assumption that fluctuating hearing loss duringthe early years of life presents the child with a degraded and inconsistent speechsignal on which to base language learning. Evidence from several studies wouldsupport that this unstable auditory signal actually creates more problems forspeech and language learning than the stable auditory input which results from amild to moderate sensorineural hearing loss (Downs & Bergman, 1982; Menyuk,1980).Additional support for the position that a mild conductive hearing loss can affectspeech and language development comes from research findings indicating thatthe 10 to 40 Db loss associated with otitis media can impair speech discrimination(Dobie & Berlin, 1979; Mustain, 1979). The results also indicated that an averagethreshold of 20 to 25 dB loss for older children with otitis media may not be thesame as for infants and young children.Nozza (1988) argued that infants and younger children as a group have a greaterreduction in auditory capability than is commonly thought during episodes of OMEand that they do not respond to sound until it is presented at higher intensitylevels than are required by older children with OME.REVIEW OF THE LITERATURE /324. Retrospective and Prospective Studiesa. Retrospective StudiesA number of studies investigating otitis media are retrospective in design. Thesestudies, according to Friel-Patti (1990), are suspect because of confounds posingthreats to both internal and external validity. Typically, subjects for these studiesare chosen on the basis of an already identified speech, language, or academicproblem. Investigators then attempt to document several factors which mayprovide some connection between a history of otitis media and deficit in the child’sperformance. Typically, the history of OME is taken from a physician’s records orparental report of middle ear infections.Zinkus, Gottlieb, & Shapiro (1978) state that these methods are unreliable, andperhaps that the most important data, the child’s current hearing status, is omitted.Friel-Patti (1990) argues that hearing measures should be taken when the child isotitis free as well as at times when there is OME present.Paradise (1981) has criticized retrospective studies making the following points:there is often a lack of measures of hearing levels in early life, matching of subjectsis difficult resulting in heterogeneous groups, subjects are often selected fromgroups that vary widely, and there is usually a small sample size.REVIEW OF THE LITERATURE 133Research findings from retrospective studies on children with histories of otitismedia have reported general delays in speech and language development (Zinkuset aL,1978), lower performance on auditory perceptual and auditory processingtests (Zinkus & Gottlieb, 1980), and a higher incidence of academic problems,especially reading speech and language development (Zinkus et aL, 1978), lowerperformance on auditory perceptual and auditory processing tests (Zinkus &Gottlieb, 1980), and a higher incidence of academic problems, especially readingand spelling disorders (Needleman, 1977).Paradise (1981) also states that in retrospective studies comparing an “otitis media”group with a “control” group, there may be misclassification, so that some childrenin the control group may in fact have had otitis media or vice versa. Differencesbetween groups need not indicate a causal relationship and may, rather, reflect theinfluence of factors associated with otitis media, such as home background orpresence of other diseases.Another important point made by Paradise (1981) is that studies of OME fall intotwo categories: those where the experimental group has conductive hearing lossat the time of the study, and those where the “otitis media” group consists ofchildren with a history of recurrent otitis media but normal hearing at the time ofREVIEW OF THE LITERATURE /34the study. This concern is a valid one, but may be difficult to control, as OME isa fluctuating condition (Feagan, Sanyal, Henderson, Collier, & Applebaum, 1987).b. Prospective StudiesKlein et al. (1988) state that a prospective research plan is the preferredexperimental design for addressing the question of otitis media and academic orspeech/language delay. Basically, this design is a longitudinal or cohort designwhere a group of children are identified at a certain point, and then followed fora period of time. Children in a cohort study where language is the outcomemeasure should be followed for several years. If language is extended to includereading skills, it is necessary to study subjects for a number of years of school(Friel-Patti, 1990).Roberts and his colleagues (Roberts et al., 1986) reported results of a prospectivestudy of 61 economically disadvantaged children enroled in a university baseddaycare centre. The extent of OME was quantified by accumulating the numberof days with effusion. There were no significant correlations between days witheffusion and the outcome measures of standardized test scores for both verbaland nonverbal intelligence and academic functioning, It is interesting to note thatalthough the subjects had hearing assessed on an ongoing basis, degree ofREVIEW OF THE LITERATURE /35hearing loss was not mentioned.Using the same sample as the Roberts et al. study, Feagan et al. (1987) found arelationship between days with effusion and narrative ability but no relationship forOME and mean length of utterance. Friel-Patti (1990) gave some clear directionat the end of her review of the literature on otitis media and language development.She feels research is needed to identify factors which increase the risk forlanguage differences. Also longitudinal research relating early experiences withotitis media, conductive hearing loss, and speech/ language development areneeded to explore possible long term effects on reading and academicachievement.In addition, it is necessary to design small studies that examine how children withearly experience of otitis media and conductive hearing loss learn in noisyenvironments, Investigating the effect of unilateral as compared to bilateral OMEand hearing loss (Friel-Patti, 1990) may prove to be important, as otitis mediabecomes a recognized variable in the assessment process.REVIEW OF THE LITERATURE /365. Otitis Media and Native Indian ChildrenAs stated previously, otitis media is a relatively common ear disease of childhood.However, most professionals working with children are relatively unaware of thecondition’s incidence, symptomatology, and developmental consequences(McShane, 1982). Even less awareness exists that otitis media constitutes one ofthe most serious health problems and disabling conditions in the American Indianand Alaska Native populations (McShane, 1982).The incidence of otitis media among various Indian groups has been reported asconsistently high. The rate has been noted as much higher than in otherpopulations (Scaldwell, 1989), although the reasons remain unclear. McShane(1982) presents a summary of possible relationships including genetic,socio-cultural, environmental and economic factors. Whatever the reasons for thehigh incidence of otitis media in Indian children, the problem is a severe one notonly from a health viewpoint, but also from an educational one (Scaldwell & Frame,1985b).Ling, McCoy, and Levinson (1969) found that children having academic difficultyREVIEW OF THE LITERATURE 137experienced more chronic ear disease than children who were successfulacademically. McShane (1982) also stated that because Indian groups experienceboth otitis media and academic difficulties in proportions which greatly exceedthose found within the larger non-Native population, it seems a comprehensiveunderstanding of these relationships, in Native Indian groups, is of vital importance.An important longitudinal study by Kaplan (1975) reported the pathogenesis ofotitis media within a birth cohort of Alaskan Eskimos. This study gave significantattention to the academic correlates of middle ear disease. The sample of 380children were given the WISC and achievement measures. The mean WISC VerbalScore reported was 77, while the mean Performance Score was 98. Testadministration procedures were not reported for this study.Children who experienced four or more episodes of otitis media during the first twoyears of life scored significantly lower than those who had experienced one tothree episodes. Academically, over a third of the children were behind theirexpected grade placement, with 63% of them having experienced an initial otitisepisode before age two. Children with otitis related hearing loss were furtherbehind in reading, math and language achievement than their peers who had noREVIEW OF THE LITERATURE /38history of an ear disease. The achievement gap between children affected by otitismedia and peers with no known history tended to widen with age (McShane,1982).Scaldwell and Frame (1985b) noted a high incidence (about 38%) of severe andearly ear infections, upper respiratory infections, and allergies among childrenreferred as demonstrating school learning problems in language areas, particularlyword recognition and spelling. In a project in a federal school for Indian children,Scaidwell & Frame (1985b) found a similar pattern where 52% of the childrenreferred for assistance in word recognition and spelling had a history of chronicmiddle ear infection and respiratory ailments.Indian children have been noted as scoring lower on standardized reading teststhan children of the majority population, especially if they have lived on reserves(Tomusiak, 1983). Naturally, there is no one single reason for this difference.Teachers and researchers have identified lower reading and spelling performanceas the result of lMng in isolated areas along with a first language other thanEnglish, and having a different cultural and experiential background.Scaldwell (1989) linked otitis media to later academic performance. Results fromREVIEW OF THE LITERATURE /39this study demonstrated that Native children, grades one to eight, who had ahistory of otitis media, were consistently behind their hearing Native peers.Although the study was conducted in a reserve school, Scaidwell felt the effect ofotitis media on later reading competence may be applicable to all children.6. Summary of Research on the Impact of Otitis MediaOtitis media may place children at risk for speech and language difficulties as wellas academic difficulties in school (Zinkus & Gottlieb, 1980). The fluctuating mildconductive loss associated with the condition of otitis media also has an effect onthe acquisition on how children develop cognitive skills, which has a direct impacton the acquisition of language and academic skills.Difficulties associated with research have been linked to small sample size, difficultyin matching subjects to control and experimental groups, as otitis media is afluctuating condition, and the heterogeneity of the population. Otitis media has notbeen considered as an important variable in the assessment process despite thehigh incidence of the condition among Native Indian children (Scaldwell, 1989).REVIEW OF THE LITERATURE /40C. Assessment of Hearing Impaired Children1. BackgroundIntellectual assessment of deaf and hard of hearing children has been a source ofcontroversy for many years. Few psychologists have either the communicationskills or the knowledge needed for adequate administration of intelligence tests todeaf and hard of hearing persons (Braden, 1992). The use of intelligence tests toexplain or predict academic achievement levels is a common practice amongschool psychologists (Phelps & Branyan,1988). Regardless of whether expectancyformulae or differences in standard scores are used, the ability achievementdiscrepancy is an accepted method of determining academic underachievement.With the hearing population, this relationship has been established; however, lessis known about predicting academic achievement from measures of ability for thehearing impaired population.When examining this relationship within the deaf and hard of hearing group, a fullunderstanding of the effects of hearing loss on language development andcommunication competence is necessary. Therefore, age at onset and age atREVIEW OF THE LITERATURE 141identification are paramount (Mindel & Vernon, 1971). As the loss becomes moresevere, communication competence of the child is an important factor. Whenchildren do not have a consistent auditory signal prior to the usual onset oflanguage, impairment results often causing a significant language disability (Vernon& Alles, 1986).2. Intellectual Assessment of the Hearing ImpairedNonverbal performance tests have been recommended for use with deaf childrensince the turn of the century (e.g., Pintner & Patterson, 1915). Braden (1992)reviewed all published and unpublished literature on assessment of deaf and hardof hearing people to determine assessment practices, outcomes of assessmentand the relationship between the two. The instruments of choice for assessmentof intellectual abilities of deaf and hard of hearing children has been the WISC-RPerformance Scales and the earlier version, the WISC PIQ (Braden, 1992; Levine,1971; McQuaid & Alovietti, 1981).Braden (1992) also concluded that before a test of intelligence is selected, thepurpose of the assessment should be clarified, If the examiner wishes to assessREVIEW OF THE LITERATURE /42intelligence, nonverbal tests are recommended as verbal measures yield lower IQscores.In reviewing the research, results and findings are inconsistent. Likely, this is aresult of the methodological difficulties associated with assessment of a specialpopulation. These include: small sample size, residential populations as subjects,and the various methods investigators have chosen to administer test instruments.Kelly and Braden (1990) point out that as experimental control has increased,observed differences between the hearing and hearing impaired populations havedecreased. Evidence in a study by Hirshoren, Hurley, and Kavale (1979)demonstrated the internal reliability of the WISC-R PIQ. They found the internalreliability for the WISC-R to be consistent with results from the normative group.In addition, Braden (1985b) reported the construct validity of the WISC-R to besimilar for hearing and hearing impaired children.Subjects in the majority of research are typically from residential schools and usea form of manual communication as their primary means of communication.Subjects using an aural/oral approach or those with a mild to moderate loss aretypically not included as subjects in the studies reviewed (Braden, 1992).REVIEW OF THE LITERATURE /43There are two major theoretical positions to consider in undertaking the intellectuaiassessment of hearing impaired children. The first position considers the deaf asintellectually normal. Though early research indicated the deaf were intellectuallyinferior (Pintner, 1924), a more contemporary formulation views apparent deficitsdeaf people display as a function of the type of test used (verbal tests) and/or theinability of the examiners to communicate with deaf and hard of hearing subjects.The second position views deaf people as cognitively different, and asserts thatdifferences encountered in the intellectual assessment of deaf children do indeedreflect a difference in abilities (Levine, 1971; Myklebust, 1960). This positionpredicts that hearing impaired children should score differently on intelligence,although overall lQ scores should be within the average or normal range (whencompared to the hearing population).Concurrent validity studies have been performed on several of the most widelyused instruments. Most of the research has been completed on institutionalized orstate-school deaf populations (Phelps & Branyan, 1988). In addition, a singlestudy has not evaluated all of the instruments used with the same set of subjects.There have been several studies that compared the Hiskey Nebraska Test ofLearning Aptitude (H-NTLA) or the K-ABC to the WISC-R PIQ when the norms forREVIEW OF THE LITERATURE /44deaf children were utilized. Of interest are two recent studies, both of which foundnonsignificant differences between IQ scores on these instruments when norms fordeaf students were used (Braden, 1985a).Phelps and Branyan (1988) compared four measures: WISC-R PIQ, LeiterInternational Performance Scales, Hiskey-Nebraska, and the K-ABC NonverbalScale. Results demonstrated that the K-ABC Nonverbal Scale scores weresignificantly lower than those from the Leiter and the WISC-R PIQ. The conclusionof the study was that the K-ABC underestimated intellectual skills of the hearingimpaired population, relative to other measures. When choosing an appropriateinstrument, the results from this study have far reaching effects. It appears allinstruments administered to populations of hearing impaired children do not yieldsimilar results.3. Assessment of Academic SkillsStandardized achievement measures are frequently administered to hearing andhearing impaired students to make important decisions at all levels of education(Allen, White, & Karchmer, 1983). Assessment within the special educationframework is very important. Therefore, it is critical that instruments used byexaminers be valid indicators of academic achievement for the population ofstudents to be assessed.REVIEW OF THE LITERATURE 1 45In an attempt to accommodate the special needs of the hearing impaired studentand to provide a standardized test, the Office of Demographic Studies (ODS)modified the Stanford Achievement Test. Modifications of this instrument tookresearch results into consideration. Four major actMties surrounding themodifications included: a) modifying test instructions to make them adaptable tothe variety of communication situations used in various educational programs, b)developing a screening procedure assigning students to the appropriate level testbattery, C) reassembling the subtests within a battery level so that the entirebattery, as it progresses from level to level, represents the growth patterns ofhearing impaired students, d) forming of test results with a representativesample of 6,873 hearing impaired students. The norms make it possible tocompare the hearing impaired student’s performance to both hearing impaired andhearing distributions (Allen, White, & Karchmer, 1983).The Stanford Achievement Test-B (SAT-8) was published by the PsychologicalCorporation in 1988. Special procedures for using the SAT-8 with hearing impairedstudents were developed in 1990, with a national (U.S.) sample of hearing impairedchildren. Practice items as well as screening tests are given prior to each subtest.The SAT-B test manual states hearing impaired students do poorly on achievementtests, not because they lack the skills, but because they do not understand theREVIEW OF THE LITERATURE /46tasks they are required to perform. In addition, the manual states it is importantto communicate the intent of tasks required for tests. Therefore, the method ofcommunication to be used in the administration of the test is the method normallyemployed in the instructional context.The authors state that lack of standardization allows flexibility, but actually itillustrates one of the difficulties in assessment. It would seem that standardizedmeasure requires a standardized method of administration if results are to be validand reliable.In addition to appropriate instrumentation, academic levels of functioning havebeen a concern for educators of hearing impaired students. Readingcomprehension is thought to be one of the most critical areas of schoolachievement (Trybus & Karchmer, 1977). A study completed by ODS in 1977found the average reading level for students aged 20 to be at less than a midgrade four level. At best only 10% of hearing impaired 18 year-olds can read ator above an eighth grade level.Discouraging as these scores may seem, they are consistent with similar findingsobtained by ODS in 1969 with 12,000 children and in 1971 with 19,000 children.Whatever the limitations of an achievement test, it is clear that a majority of hearingREVIEW OF THE LITERATURE /47impaired children are at a substantial educational disadvantage when comparedto their hearing age peers (Trybus & Karchmer, 1977).4. Academic Achievement and Nonverbal IntelligenceTests of achievement are frequently administered in addition to tests of intelligence.The SAT-HI is a group administered measure of achievement. Many studies haveutilized indMdual achievement measures such as the Peabody Achievement Test(PIAT), the Woodcock Reading Mastery Test, and the Woodcock Johnson Battery- Achievement Part II (WJ). Watson et al. (1986) used the WJ and the PIAT toassess achievement, and then correlated achievement scores with intelligencescores. The achievement measures were administered by an interpreter usingtotal communication, violating standard administration procedures. Also, theinstruments used were developed and formed for hearing populations.Several studies have examined criterion-related correlations with achievement(Hirshoren et al., 1979; Kelly & Braden, 1990). These studies report variedcorrelations between the Stanford Achievement Test (SAT-HI) scores and theWISC-R PIQ. The Hirshoren et al. (1979) study found that the WISC-R PIQ did nothave criterion-related validity with hearing impaired children’s achievement scores.REVIEW OF THE LITERATURE /48In contrast to these findings, Kelly & Braden (1990), found the WISC-R PIQ to haveadequate criterion-related validity with the SAT-HI subtest scores when appropriatecriterion metrics are adopted. The selection of the criterion scale, scaled scoresvs. percentiles, affects the relationship between the IQ score and achievement.SAT-HI percentile scores were suggested as a better metric than scaled scores orgrade equivalents for testing the relationship between aptitude and achievementamong hearing impaired children.Phelps and Branyan (1990) demonstrated that the relationship between cognitiveability and achievement is as significant for the public school hearing impairedpopulation as it is for the normal population. They suggested that the use ofcurrent and appropriate measures of nonverbal intelligence could facilitate theidentification of underachievement in hearing impaired children.Brooks and Riggs (1980) found the WISC-R to have validity for predicting readingachievement of hearing impaired students. They found the WISC-R to be a validmeasure of nonverbal intelligence among hearing impaired school populations.Scores for the hearing impaired sample approximated those for the hearingpopulation on the Performance Scales. Contradictory findings from the Watson etal. (1986) study demonstrated that the role of the WISC-R PIQ in interpretingachievement levels was not supported across all age levels. An interesting findingREVIEW OF THE LITERATURE /49in this study was that significant correlations were found for younger children butnot for children in the middle grades (five through nine), when grade equivalentscores were used for achievement measures.5. Test Modifications for Hearing Impaired ChildrenModification of test instructions are useful for research and may provide additionalinformation about the child’s ability (Bragman, 1982). However, it is important tonote these practices can change the meaning of the scores, and little is knownabout whether modifications facilitate or interfere with understanding of the taskand the response of the individual (Ray, 1982). Additionally, norms developed forthe test are no longer efficient once standard test instructions have been modified(Anastasi, 1988; Sattler, 1988).Bragman (1982) states the importance of assessing the effects of modifications totest instructions on performance. The most frequent modifications have includedalternate ways or methods of presenting task demands to subjects, i.e., testinstructions. Modifications have included: pantomime, demonstration, andlanguage communication invoMng speech, lipreading, reading written words,fingerspelling, and translation of speech into a manual communication systemREVIEW OF THE LITERATURE /50(Bragman, 1982).Documentation of modification to test instructions are rarely included in researchstudies (Sattler, 1988). When studies are reviewed in this section, it will beimportant to recall the heterogeneity of the population, and how viablemodifications are to the deaf and hard of hearing population.When Braden (1992) examined all literature pertinent to assessment practices ofthe deaf and hard of hearing subjects (N=285), examination of administrationmethods revealed that the most popular method of administration was signaccompanied with speech (N=69). Oral or verbal administration followed (N=36),gestural (N=35), and written communication (N=5). It is important to note thatmore than half of the studies (N =175) failed to describe and document testadministration procedures.Administration of the WISC-R PIQ to subjects in the Anderson and Sisco (1977)study provided the data for special norms for hearing impaired children. Anunfortunate aspect of the Anderson & Sisco study was that there was no attemptto adhere to standardized administration procedures. Although special normsinitially appeared promising (Anderson & Sisco, 1977), empirically, there is noevidence to indicate that special norms result in a more accurate diagnosisREVIEW OF THE LITERATURE /51(Braden, 1992).Research related to appropriate language/communication of instruction withhearing impaired children has been limited (Levine, 1971; Sattler, 1988). There isa confounding variable involved, namely that research indicates changes instandardized testing procedures may cause the test to become invalid (Sattler,1988; Salvia & Yssledyke, 1991), or results may not be reliable. In addition, Smith(1974) found that translating a test into another language changes the difficultylevel and the validity of the test.The majority of research tends to focus on methods to convey instructions thatapproximate the language of the child. Therefore, studies have investigatedspeech reading with and without voice and manual-based approaches includingthe Rochester Method, Total Communication, Siglish, and Ameslan (White &Stevenson, 1975).Ray (1982) published an adaptation of the WISC-R Performance Scaleadministration procedures. These involve the use of supplemental items fordemonstration as well as alternative instructions for signing. The use ofdemonstration on extra practice items to modify test administration may haveserious effects on the validity of the test, because additional items may provideREVIEW OF THE LITERATURE /52opportunities for rehearsal (Perley-McField, 1990). Sattler (1988) believesdemonstration as a type of test modification is a more “drastic alteration” thanpantomime instructions and less likely to produce valid results.Graham and Shapiro (1953) investigated the effects of pantomime and verbalinstruction of the WISC. A group of 20 hearing impaired children were given theWISC using pantomime instructions. One group of hearing subjects was given theWISC using pantomime; verbal directions were used with the second group ofhearing subjects. No differences in total performance scores were found betweendeaf children given pantomime directions and hearing children receMngpantomime or verbal instructions.Goetzinger and Houchins (1969) also investigated the effects of pantomime andverbal instructions on student performance. They studied 40 hearing and hearingimpaired children to determine the effects of pantomime and verbal directions. Aset of pantomime instructions for the Raven was developed and administered toall hearing impaired subjects. Half of the hearing subjects received pantomimeinstructions and half received verbal directions. When the two groups of hearingsubjects were compared, no differences in performance were found. There wasalso no difference between the hearing and hearing impaired groups whenpantomime instructions were administered (Bragman, 1982).REVIEW OF THE LITERATURE /53Courtney, Hayes, Couch, and Frick (1984) concluded that pantomime instructionis an acceptable method for administering the WISC-R PIQ. They stated that theresults provided “minimum estimates of the intellectual capabilities of the hearingimpaired”(p.6). They continued to state that the IQ measure should always beconsidered one component of a total assessment of a child and that pantomimeinstruction may be an alternative if it is the only common ground between theexaminer and child.Braden (1 985d) contested the findings of the Courtney et al. (1984) study, referringto the Graham and Shapiro (1953) and Sullivan (1982) studies which indicatehearing impaired children obtained lower lQs under pantomimed conditionsthan when they are tested by a psychologist who signs directions, or when theyare assessed via a sign language interpreter (Ray, 1982). Braden (1985d)concludes his response to the Courtney et al. (1984) study with a “warning” statingthat pantomime administration of the WISC-R PIQ to hearing impaired childrenshould be discouraged.Hayes and Courtney (1985) responded to Braden (1 985d) by clarifying conclusionsfrom their original work and discussing once again the need for alterative testadministration procedures. The authors advocate and support the use of the bestoption for a child in any given situation. Sullivan (1982) held a similar view,REVIEW OF THE LITERATURE /54stating that although total communication is preferable when possible,modifications such as pantomime should be used cautiously with hearing impairedchildren. Whatever modifications a psychologist chooses, they should be identifiedas administration modes in reports.One cannot assume psychologists are using consistent procedures whenassessing deaf or hard of hearing children. This issue will be addressed in thepresent study.6. Difficulties Associated with AssessmentLevine (1971) stated a key problem in assessment of the deaf and hard of hearingpopulation is the use of standardized tests with an “unstandardized” population.The heterogeneity of the population is perhaps the most overwhelming variable,making common or standard practices of test administration for this group difficult.Few instruments are standardized on hearing impaired subjects, to the manyneeds and situations that call for valid test information (Levine, 1971). As a result,instruments that are standardized on hearing subjects are typically administeredto hearing impaired subjects (Ulissi, Brice, & Gibbins, 1989).REVIEW OF THE LITERATURE /55Two intelligence tests have norms reported for hearing impaired indMduals, namelythe Performance Scales of the WISC-R, and the Hiskey-Nebraska Test of LearningAptitude (H-NTLA) (Hiskey, 1966). More recently, the Nonverbal Scale of theK-ABC has been used to assess both cognitive and academic skills. Thedevelopment of this scale, although not standardized on a hearing impairedpopulation, included hearing impaired subjects in the standardization sample. Thesubtests have pantomime instructions, and can be responded to motorically (Ulissi,et al., 1989).Ray (1982) suggested that there were three problems contributing to the lack ofstandardization of instruction: 1) limitations of the instruments used in assessment,2) limitations of the psychologist’s training and 3) heterogeneity of the hearingimpaired population. Ray (1982) attempted to solve some of these difficulties byusing supplemental and alternate instructions for each of the Performance subtestsof the WISC-R.Supplemental instructions and items were designed to communicate the task ofeach subtest by means of unscored examples similar in nature to WISC-R items.After the subjects had completed the supplemental items, the alternate instructions,which were linguistically less complex than the original Wechsler items, wereadministered. Limitations to the study were discussed by the author, even thoughREVIEW OF ThE LITERATURE /56the findings supported the idea that the hearing impaired child did not differsignificantly from hearing children on the WISC-R PIQ. There may have beenincreased or improved comprehension scores resulting from practice effects fromthe supplemental items.Difficulties with translating verbal instructions into sign systems have been identified(Levine, 1971; Sullivan & Vernon, 1979). Often English grammar does not lenditself to a sign translation. Younger children who may be linguistically immatureoften have difficulty comprehending instructions when they are translated usingSigned English (Vernon & Alles, 1986).Phelps and Branyan (1988) discussed the importance of utilizing a communicationapproach that would remain consistent regardless of the child’s hearing loss orcommunication system. Pantomime instruction was used in the administration ofthe WISC-R and the K-ABC to a sample of hearing impaired children. Resultsindicated pantomime and total communication administration yielded similar results.There were no statistical differences in lQ scores between the two groups.One of the purposes of the Phelps and Branyan study was to investigate aconsistent method of test administration. Pantomime may be an alternative to totalcommunication for this population of hearing impaired children, as differencesREVIEW OF THE LITERATURE / 57between the scores were not apparent.These findings dispute those from Sullivan’s (1982) study, which demonstratedadministration via total communication yielded significantly higher scores on theWISC-R than pantomime or a verbal/gesture set of instructions. It should benoted that in this study, the sample size was extremely small, and it appears fromthe methodology that children were given combinations of communicationmethods. Conclusions Sullivan reached are questionable, as methodology cannot be replicated as standard instructions were not used nor reported.7. Summary of Research on Assessment of Hearing Impaired ChildrenThe WISC and WISC-R have been the instruments of choice when assessingintelligence of hard of hearing and deaf children. The nonverbal subtests (PIQ)have been administered to children, as verbal subtests are a measure of linguisticcompetence rather than a measure of intelligence.Deaf and hard of hearing children are not a homogeneous group. Language skillsand mode of communication have made administration of any assessmentmeasure difficult. Instructions for many deaf subjects must be interpreted into aform of manual communication. Other modifications to administration proceduresREVIEW OF THE LITERATURE /58have been to give subjects additional sample items, to demonstrate the taskrequired in greater detail, and to use an interpreter to aid the examiner who maynot be fluent in the subject’s first language. These modifications typically are notreported in the research, which makes it very difficult to determine how the testwas administered.Alternate forms of test administration have been examined. In addition to manualcommunication, gestures, lipreading, and pantomime instruction have been utilized.Pantomime instruction may provide examiners a method by which tests ofintelligence such as the WISC-Ill and SB:FE could be administered to allexaminees, regardless of their mode of communication or hearing loss.D. SummaryThe body of literature reviewed in Chapter II reveals many difficulties in theassessment of intelligence and academic skills in Native Indian children. Lack ofstandardized tests and test instructions, use of verbal instructions, small samplesize, and language difficulties are several of the problems associated with theassessment process.In addition to difficulties associated with the assessment of Native Indian children,REVIEW OF THE LITERATURE /59the incidence of otitis media is a variable of the utmost importance. The incidenceof otitis media has been reported as higher among the Native population thanwithin the general population. The importance of the research on otitis media isstressed for two reasons. First, if hearing loss due to otitis media is notconsidered, test results would not be valid or reliable. The child may notcomprehend verbal instructions as a result of delayed language, or the testadministration may have been altered to ensure comprehension of the task. Bothfactors may cause the examiner to violate standard administration procedures.Second, the research also indicates that children with recurring otitis media are atrisk for, and often demonstrate, speech and language difficulties which mayinfluence test results.Hearing status of the child is important to the assessment process. It is logicalthat research pertaining to the assessment of hearing impaired children should becritically reviewed. Research strongly supports the use of nonverbal tests.Difficulties arise when verbal/oral instructions are used to administer nonverbalintelligence measures. Test scores must be questioned and standardadministration of assessment instruments must be investigated./ 60III. STATEMENT OF THE PROBLEMA. BackgroundNative Indian children are assessed to determine levels of cognitive, linguistic, andacademic functioning. Typically, nonverbal measures of intelligence areadministered to compensate for the children’s limited linguistic skills. A difficultywith this rationale is that many of the instructions for nonverbal intelligencemeasures are administered with verbal instructions.Examiners have attempted to compensate for the child’s linguistic difficulties byusing interpreters, altering instructions by making them more simplistic, oradministering demonstration items. All of these modifications violate standardizedadministration procedures. There is a need to study the effects of utilizing astandard method of test administration.A high incidence of otitis media has been reported in Native communities inCanada. Mild to moderate hearing loss is typically associated with conductivehearing loss resulting from chronic otitis media. However, hearing status of theNative Indian child has not been considered an important variable in theassessment process. If test scores are to be valid and reliable estimates of abilitySTATEMENT OF THE PROBLEM /61and academic achievement levels, it is essential that hearing loss of the Nativechild be considered.These problems will be addressed by examining the use of a standardizedpantomime procedure to administer nonverbal tests of intelligence. Empiricalevidence pertaining to standardized test modifications for hearing impaired childrenwill be utilized with groups of hearing and otitis prone Native Indian children.B. Purpose of the StudyThere are two purposes of the study. The first is to study consistent orstandardized administration procedures using pantomime instructions with tworecently revised measures of intelligence, the Wechsler Intelligence Scales forChildren-Ill, Performance Scales (WISC-lll PS) and selected subtests of theStanford Binet:Fourth Edition (SB:FE). These nonverbal measures of intelligencewill be administered to otitis prone Native Indian children and their hearing peers.The second is to determine if nonverbal measures of intelligence (WISC-Ill andSB:FE), administered with standard pantomime instructions, can be used to predictacademic achievement for hearing and otitis prone Native Indian children.STATEMENT OF THE PROBLEM /62Definitions of intelligence emphasize the ability to perform or adapt to theenvironment, the ability to learn, and to perform abstract thinking tasks (Zachary,1990). Sternberg (1984) states intelligence should be considered in respect to thecircumstances under which an indMdual develops and operates.The nonverbal measures, according to Wechsler’s (1939) definition serve toexamine a construct of intelligence which is composed of quantitatively differentabilities. Results of these measures have been used as predictors of achievementfor hearing impaired children (Zachary, 1990), and are thought to be the mostaccurate indicator of academic potential for Native Indian children (Hynd & Garcia,1979). These assumptions were addressed by the two purposes of the study andwere specifically addressed by the hypotheses.C. Research Questions & HypothesesResearch Questions1. Will nonverbal measures of intelligence, administered with pantomimeinstructions to hearing Native children, yield scores which are similar to those fromthe standardization sample?STATEMENT OF ThE PROBLEM /65These questions generated Hypothesis #3a3a. The WISC-lII Performance IQ score and the nonverbal Partial Composite scorefrom the SB:FE will be predictors of achievement for the otitis prone and hearingpeer groups.The literature in the area is conflicting. Kelly and Braden (1990) state that theWISC-R performance scales have adequate criterion validity when appropriatecriterion measures were utilized. If predicted IQ scores, which have utilizedpantomime instructions, are similar to peer groups or the normative data,correlations similar to those found within the normative group will be predicted forboth of the groups.Research Questions7. What is the best combination of subtests of measures of intelligence to predictachievement in hearing Native Indian children?8. What is the best combination of subtests of measures of intelligence to predictachievement in the otitis prone Native Indian group?STATEMENT OF THE PROBLEM 166These questions generated Hypothesis #3b3b. Specific subtests from the WISC-Ill Performance Scales and subtests from thenonverbal Partial Composite SB:FE will be predictors of achievement for the otitisprone and hearing peer group.Sisco and Anderson (1978) and Phelps and Branyan (1988) have found specificsubtests to be more accurate predictors of academic skills for hearing impairedsubjects. Similar results were found by Chrisjohn and Lanigan (1988) for NativeIndian children. Therefore, it is expected that specific subtests, on both measures,will be accurate predictors of academic skills for the otitis prone group.Research Questions9. When pantomime instruction is utilized, what is the strength of associationbetween the measures of intelligence for the hearing Native Indian group?10. When pantomime instruction is utilized, what is the strength of associationbetween measures of intelligence for the otitis prone Native Indian group?These questions generated Hypothesis #4a4a. It is expected there will be a significant correlation between the WISCIIISTATEMENT OF THE PROBLEM /67Performance IQ scores and the SB:FE Partial Composite score for the hearing andotitis prone groups.Within the “typical” population there is a high correlation between the WISC-R FullScale score and the Stanford-Binet Test Composite score (Zachary, 1990). Similarresults are expected for both the hearing and hearing impaired Native Indiangroups.Research Questions11. What is the strength of association between measures of nonverbalintelligence and measures of achievement for the hearing peer group?12. What is the strength of association between measures of nonverbalintelligence and measures of achievement for the otitis prone group?These questions generated Hypothesis #4b4b. It is expected there will be a statistically signficant correlation betweennonverbal measures of intelligence and achievement for the hearing peer and otitisprone groups./ 68IV. METHODOLOGYA. SamplingThe population for this study included all school age children of 13 Indian bandsaffiliated with the Carrier-Sekani Tribal Council in British Columbia. Members ofthis Tribal Council live in both rural and urban areas. Several of the rural areas arevery remote and isolated with little contact or access to medical facilities. Theurban areas in small towns and villages are less isolated with medical facilitiesavailable. The language and culture of the Carrier-Sekani was evident in all of theschools included in the study. Elders are involved in school life, and Carrierlanguage classes are taught within the school setting.Approximately 900 children enroled in five Band or Federal Reserve Schools, threeCatholic separate schools, and twenty one schools in three school districts werescreened. Each student was screened by a nurse or trained research assistantusing a pure tone air conduction audiometric sweep test using recommended B.C.Ministry of Health Guidelines. Subjects were screened utilizing the followingfrequencies and intensities: 500Hz-25dB, 1000 Hz-2OdB, 2000 Hz-2OdB,4000Hz-25d8. Children who passed the screen, and had no known histories ofhearing loss (as confirmed by school nurse, Community Health Representative,and school personnel) became possible subjects for the hearing control group inthe study.METHODOLOGY /69Those who failed one frequency, in either ear, were screened again. A failure onthe second screen resulted in a referral to an audiologist for a full audiologicalassessment, including a pure tone test, impedance measure, and a speechreception test. Students who passed screening tests but had a known history ofhearing problems or recurrent colds and upper respiratory problems were alsoreferred to the audiologist.It was expected that 20-30% of the Native chidren screened would not pass thehearing screen. In fact 210 children, or approximately 24% of the original 900children screened, were assessed by the audiologist, and had a full audiologicalassessment. The incidence of children failing the hearing screening and/or havinga history of upper respiratory infection was within expected limits (Scaldwell &Frame, 1 985a).Paradise (1981) expressed some concern that there may be a misclassificationwhen assigning children in an experimental study to either an otitis media groupor a control group, as hearing loss due to otitis media is a fluctuating condition.It was decided for the present study, therefore, that factors other than pure toneloss should be considered for inclusion in the otitis prone group. Additional factors,such as home background and socioeconomic status, which were also a concernexpressed by Paradise (1981), were controlled by the investigators in the presentMETHODOLOGY /70study.Audiologic data were then examined for each subject to determine whether theymet the criteria for inclusion in the experimental or otitis prone group.The criteria were as follows: 1) conductive hearing loss, either unilateral or bilateral,2) significant negative ear pressure, .3) history of recurrent ear infection/upperrespiratory infection, 4) scarring or inflammation of the tympanic membrane fromcolds, etc. Native Indian children were included as “otitis prone” if they meet twoof the four criteria.Children ranged in age from 6-0 to 16-11, which is the age range used in theWISC-Ill standardization sample. After examining the audiological data, 137 NativeIndian children met the criteria for the otitis prone group. lndMdual test data werecollected at the beginning of the school year. Subjects were “lost” for severalreasons: closure of one the Independent schools, transfer of schools as familiesmoved over the summer, lack of return parental permission forms, or children’sabsence during the data collection.After examining the audiological data, 87 otitis prone Native Indian childrencomprised the experimental group. They were matched with hearing Native IndianMETHODOLOGY /71controls by the following criteria: 1) they were the same age, 2) they were from thesame school, 3) they were from the same class, and 4) they were the same sexSubjects in the hearing peer group were randomly assigned to one of twoadministration methodologies; as shown in Table 1: a) those who receivedstandard verbal instructions for the measures of intelligence as stated in theWISC-lll and. SB:FE manuals, b) those who received standard pantomimeinstructions (refer to Appendices A and B) for administration of the measures ofintelligence.Table 1. Subjects by group and method of instructionOtitis Prone Hearing Control GroupGroupN=87 N=40 N=40Pantomime Verbal PantomimeInstruction Instruction InstructionMETHODOLOGY /72B. ProcedureThe WISC-llI and the SB:FE were administered to the otitis prone group withstandard pantomime instruction (Appendices A and B). Examiners administeringthe tests of intelligence were trained school psychologists. Examiners receivedtraining in the administration of the standard pantomime instructions. In an attemptto control for reliability and examiner error, two examiners administered theWISC-tll and SB:FE to subjects in the otitis prone and hearing groups whoreceived pantomime instruction.The Vocabulary subtest from the Stanford Binet:Fourth Edition and the PeabodyPicture Vocabulary Test-Revised (PPVT-R) were administered indMdually tosubjects in both the otitis prone and control group using standard verbalinstructions.Stanford Achievement Test-Hearing Impaired (SAT-HI) screening tests wereadministered to determine the appropriate level for Reading Comprehension andMath Computation. The SAT-HI (Reading Comprehension and Math Computationsubtests) were administered to subjects in small groups by trained examiners.Norms for the hearing impaired were used to calculate standard scores for theReading Comprehension and Math Computation subtests for the SAT-HI.METHODOLOGY /73Standard scores and percentiles were derived using norms for hearing children forboth the nonverbal intelligence measures. Scores obtained included IQ scoes,index scores and subtest scaled scores. Norms for hearing children were aisoused to calculate standard scores for the SB:FE Vocabulary subtest, and thePPVT-R.Extensive demographic data collected with the assistance of the child’s parents,school personnel, the regional health nurse, CHR, and band members employedby the research team are valuable when investigating the effects of variables suchas school history, length of time at the present school, and provision of medicaland educational services on test performance.C. Instruments1. The Wechsler Intelligence Scales for Children-IllThe WISC-R has undergone revision and was published in the Fall of 1991(Psychological Corporation) as the Weschler Intelligence Scale for Children-ThirdEdition (WISC-III). The WISC-R Performance Scales have been the most populartest for assessing the intelligence of hearing impaired children (Braden, 1 985a) andMETHODOLOGY /74Native Indian children (Goldstein, 1988). Psychologists administering the WISC-Rto deaf children have had the option of utilizing norms developed from a largesample of deaf children (Anderson & Sisco, 1977).It has been assumed that deaf norms were superior to the use of norms for thestandardization sample. Braden (1985a, 1990) stated that there is no reason toassume hearing norms are improper or inaccurate for use with hearing impairedchildren. Empirically, there is no evidence that special norms result in a moreaccurate diagnoses of general intelligence (Braden, 1992).Recent studies investigating the intelligence of Native Indian children have reliedon the WISC-R. Research with Native children from various areas of Canada hasdemonstrated that even though children have varied linguistic, socioeconomic, andcultural backgrounds, they demonstrate average Performance lQ scores, or haveaverage nonverbal intelligence.Verbal lQ scores have been found to be significantly lower than Performance IQscores in the Native group. Therefore, administration of the Performance Scalesas a nonverbal measure has become recommended practice (Goldstein, 1988).METHODOLOGY /75The WISC-lIl was recently revised and published in September, 1991. Thestandardization sample included 2,200 children throughout the U.S. The samplewas matched by the variables of age, sex, race/ethnicity and parent educationlevel (PEL). PEL is defined by the test developers as “average number of yearsof education completed by parents.” U.S. census data from March, 1988,determined the proportion of the sample needed in each category.A validation study of the WISC-lll is being conducted in Canada concurrently withthe U.S. development of the WISC-lll. Approximately 1,000 Canadian children willbe tested, according to a stratified sampling plan. The same demographicvariables will be used as in the U.S. sample, with modifications to reflect theCanadian population.If all subtests of the WISC-lll are administered, Verbal, Performance and Full ScalelQ scores can be derived. In addition, factor based index scores can also becalculated (Wechsler, 1991). Results of previous research indicated two majorfactors, namely Verbal Comprehension and Perceptual Organization, and a weakerthird factor, Freedom from Distractibility, emerged (Kaufman, 1982).METHODOLOGY /76One of the goals for developing the WISC-lIl was to further investigate factor orindex scores. The Symbol Search subtest was added to the WISC-lll for thispurpose. For each item of the Symbol Search (SS) the child visually scans twogroups of symbols, a target group and a search group. The child indicates, bymarking the appropriate box, whether or not the target symbol appears in thesearch group.With the inclusion of Symbol Search, a fourth factor emerged, in addition to thethree existing factors, Processing Speed (PS). PS includes the Coding and theSymbol Search subtests, and was calculated for subject responses in the study.In the current study, the Performance subtests were administered to the otitisprone and control group, utilizing standard verbal instructions and standardpantomime instructions (Table 1). Symbol Search is not included in the calculationof the Performance IQ It is, however, included in the calculation of the ProcessingSpeed Index score.METHODOLOGY /77Table 2. WISC-lll subtests for Performance IQ and Index scoresPerformance IQ Perceptual Processing(PIQ) Organization SpeedIndex (P0) Index (PS)Picture Completion Picture Completion CodingPicture Arrangement Picture Arrangement Symbol SearchBlock Design Block DesignObject Assembly Object AssemblyCoding(Symbol Search)2. The Stanford-Binet:Fourth Edition (SB:FE)The SB:FE was published in 1986, and represents a marked departure in itsrevised form when compared to earlier editions. The test is based on a three-levelhierarchical model. The model postulates g (a general intelligence factor) is at thehighest level of interpretation; crystallized, fluid, and short term memory factors areat the second level; and more specific factors such as verbal reasoning,quantitative reasoning, and abstract visual reasoning are at the third level. Thespecific factors at the third level, plus short term memory at the second level, formthe four area scores in the SB:FE.METhODOLOGY /78Area scores are a result of subtest scores, and the composite score is an adjustedsum of area scores. In the current study, the composite score was comprised ofthe Short Term Memory area score and the Abstract/Visual reasoning area score.The Vocabulary subtest was not included as the composite score was to benonverbal.The SB:FE was standardized on a sample consisting of 5,013 children in 17 agegroups. The sample selected was representative of the U.S. population accordingto the 1980 census data. Stratification variables included geographic region,community size, ethnic group, age, gender, and socioeconomic status.The current study utilized SB:FE subtests for examinees with limited Englishproficiency, or those who are non-language proficient (Delaney & Hopkins,1987).Delaney and Hopkins (1987) assume that those administering the SB:FE or anyother intelligence test will not use English directions if students do not understandor speak English. Another suggestion made is that tests should be administeredin the child’s native language by a bilingual examiner. Difficulties associated withaltering or changing standard administration of a test were discussed in ChapterII.When evaluating subjects who have limited proficiency in English, and are notMETHODOLOGY /79fluent in their native language, certain subtests of the SB:FE, lend themselves tononverbal administration, namely Bead Memory, Pattern Analysis, Memory forObjects, and Matrices. Entry levels for the nonverbal tests are determined fromresults of the Bead Memory subtest. It is thought the Bead Memory subtest is themost representative of the level of cognitive development and is appropriate as anentry level (Delaney & Hopkins, 1987).Table 3. SB:FE composite score, area scores and subtest scoresComposite Score(g)Area ScoresShort Term Memory Visual/AbstractArea Score Reasoning Area ScoreSubtest scoresBead Memory Pattern AnalysisMemory for Objects MatricesStandard pantomime instructions for the SB:FE subtests used for the purposes ofthe current study are presented in Appendix B. These instructions were usedMETHODOLOGY /80when the SB:FE was administered to otitis prone subjects and hearing controlsubjects receiving pantomime instruction.The Vocabulary subtest from the SB:FE was also administered to hearing and otitisprone subjects to give an estimate of expressive language skills. It was not theintent of the study to give extensive information on language function from theadministration of this subtest. Rather, results from the Vocabulary subtest will lendsupport to previous research indicating delayed expressive language skills.Previous research indicates a correlation between PPVT-R and the Verbal Areascore of .74 (Hunter, 1992). Correlations between the two language measures willalso be examined in the current study.3. Stanford Achievement Test for the Hearing ImpairedThe Office of Demographic Studies (ODS), now known as the Center forAssessment and Demographic Studies (CADS) of Gallaudet University, has studiedthe educational achievement of hearing impaired students for over two decades.Trybus and Karchmer (1977) were the first to conduct a study investigating theeducational achievement of hearing impaired children, using a large sample. TheMETHODOLOGY /81only other study to report achievement results of any magnitude was theWrightstone, Aronow, and Moskowitz (1963) study.In 1990, a Special Edition of the SAT for the Hearing Impaired (SAT-HI) wasdeveloped by CADS. The test was standardized nationally in the U.S. on astratified random sample of 6,871 hearing impaired children. The stratifyingvariables were program type and enrolment size, so that a full range of specialeducational settings, from large residential schools for the deaf to small integratedclassrooms with support services, were represented in the standardizationaccording to the proportion in the population.Prior to the administration of the SAT-HI, a screening instrument is administeredto determine the child’s entry level. The SAT-HI (8th Edition) assesses ReadingComprehension, Spelling, Concept of Number, Math Applications, MathComputation Language, Social Studies, and Science.For the purposes of the study, selected subtests of the SAT-HI were administeredto both the hearing peers and the otitis prone Native Indian groups. The subtestsadministered were Reading Comprehension and Math Computation.METHODOLOGY /82The SAT-HI reports grade equivalents, scaled scores and percentile ranks. Scaledscores, based on data from hearing students, represent approximately equal unitson a continuous, criterion-referenced scale. Percentile ranks are based onnormative data from hearing impaired students, whose referent groups aredetermined by age.It has been recommended (Kelly & Braden, 1990) that percentile rank scores beused to describe a hearing impaired child’s achievement levels, as they are a morevalid criterion measure than scaled scores or grade equivalents. Kelly andBraden(1990) also state percentile ranks are a more accurate metric whencomparisons are made between aptitude (intefligence) and achievement.4. Peabody Picture Vocabulary Test-Revised (PPVT-R)The PPVT-R was administered as a measure of receptive vocabulary or “hearingvocabulary” (Hunter, 1992). The purpose of this study was not to measurelanguage levels of the subjects. From review of the literature, it could be assumedthat the language levels of the subjects were delayed as the majority of thesubjects were English as a Second Language children and had a history of otitismedia. The administration of this measure was to confirm this assumption.METHODOLOGY /83D. Data AnalysesJackson (1984) states that it is of utmost importance to compare test results ofchildren to their immediate peer group. Therefore a methodology similar to thatused by Goetzinger and Houchins (1969) will be applied in the present study toinvestigate the use of pantomime instructions with the WISC-lll and SB:FE forhearing and otitis prone Native Indian children.Goetzinger and Houchins (1969) investigated the effects of pantomime and verbalinstructions on the performance of hearing and hearing impaired children when theColoured Raven’s Progressive Matrices were administered. To accomplish this,hearing subjects were dMded into two groups, those who received standard verbalinstructions and those who received standard pantomime instructions.Comparisons were then made between the two methods of administration for thehearing group and between the hearing and hearing impaired group. Resultsindicated there was no difference between the hearing and hearing impairedgroups when the Raven’s was administered with pantomime instructions.METHODOLOGY /84In the current study, the hearing and hearing impaired Native Indian groups will bematched by age, sex, and geographic region. Members of the hearing group willbe randomly assigned to one of two groups: standard verbal administrationor standard pantomime administration. Random assignment to each of the twoadministration groups ensured a representative sample of children in the twogroups.St. John and Krichev (1976) and Wilgosh, Mulcahy, and Watters (1986) have foundWISC-R Performance IQ scores of Native Indian children to be similar to normativedata (test norms). WISC-lll and SB:FE scores from the hearing Native groupreceiving standard verbal directions will be compared to the normative data. It ishypothesized that they will approximate the normative data, having similar meansand standard deviations.Administering both pantomime and verbal directions to the hearing Native groupwill enable valid and reliable comparisons of administration procedures within thehearing group and administration procedures between the hearing and hearingimpaired Native Indian groups.In addition, to answer the research questions, multiple regression analyses wereused to investigate the relationship between IQ scores (predictors) andMETHODOLOGY /85achievement scores (criterion). ANOVAs were used to investigate the within andbetween group differences/similarities. Correlations between the achievement andintelligence subtests, and between the measures of intelligence, were completedto answer the remaining research questions.Comparison of group scores for measures of intelligence and achievement wereused for the majority of the analyses. Item analyses were attempted to investigateindividual differences at the item level. It was important to investigate results ofglobal scores, subtest scores and item responses to have a complete picture ofthe effect of pantomime instruction.In addition to investigation of pantomime and verbal administration at the test andsubtest level, item analyses were completed for the Picture Completion (PC)subtest of the WISC-lll to examine subjects’ response to each item. Responsesfrom Group II and Group Ill were investigated for each item to determine whetherthere was a difference between the subjects’ response to items administered usingeither pantomime or verbal instruction./86V. RESULTSResults throughout this chapter are presented for the three groups: Group I- otitisprone, pantomime instruction (N=87), Group II- hearing peer, pantomimeinstruction (N=40), Group Ill- hearing peer, verbal instruction (N=40). Sample sizeis reported for the analyses. In several cases there are small differences from theoriginal sample size. Missing data were a result of absentee subjects, subjectswho could not complete the test as they did not have the prerequisite skills, orsubjects who were outside of the age range for the test.WISC-lll results include means and standard deviations for Performance IQ score,two Index Scores; namely Perceptual Organization (PC) and Processing Speed(PS), and scaled scores for the subtests. SB:FE results include a compositescore, area scores and subtest scores. Results are also presented for languagemeasures administered, the PPVT-R and the Vocabulary subtest of the SB:FE.Conservative levels of significance (p < .01) were used for the ANOVA analyses.For t-tests p < .05 was used; however, when correlations were examined p < .005and p < .001 were used. The Bonferroni t procedure was used to correct for alphaslippage over repeated comparisons to guard against Type I error. Item analysiswas completed as a post hoc measure to provide complete information to answerhypothesis #2a and #2b.RESULTS / 87A. Hypothesis #1It is expected nonverbal measures of intelligence administered with pantomimeinstructions will yield results that approximate the standardization sample for boththe otitis prone and hearing peer groups.T-tests were used to test hypothesis #1. Results are presented in Tables 6 and7. In addition to t-tests, descriptive data were examined to answer the researchquestions. When examining hypotheses it is important to note that desired resultswere that there would be no difference between administration methods, or thatthe null hypothesis would not be rejected.1. Descriptive StatisticsMeans and standard deviations were calculated for all WISC-lll and the SB:FEsubtest scores and lQ scores. Subtest means were tested to determine whetherthey differed from the expected population means (i.e. X=1O, SD =3, for the WISCIll subtests).Several subtests were more than one standard deviation (SD) below the mean, orexpected score for the standardization group (Tables 4 and 5). In Group I, thePicture Arrangement subtest (ss 6.46) of the WISC-Ill, and the Matrices (SAS41.47), Bead Memory (SAS 39.71), Abstract/Visual Reasoning Area (SAS 83.67),RESULTS /88Short Term Memory Area (SAS 79.17) and the Composite Score (SAS 79.41)yielded scores which were more than 1 SD below the mean. This also was thecase for the Vocabulary SB:FE (SAS 83.06) and the PPVT-R (SS 75.85).RESULTS / 89Table 4. Descriptive statistics for WISC-Ill by groupOtitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Group I Group II Group IllSubtest N=87 N=40 N=40mean SD mean SD mean SDPicture 7.96 3.52 8.83 3.67 9.82 3.40CompletionCoding 6.98 3.15 7.88 2.79 7.34 3.50Picture 6.46 3.28 7.35 3.01 7.55 3.82ArrangementObject 8.88 3.71 9.58 3.55 9.42 2.46AssemblyBlock 9.11 2.30 9.35 3.03 9.68 2.89DesignSymbol 9.23 3.74 9.43 3.27 10.03 3.77SearchPerformancelQ Score(PIQ) 87.08 13.86 92.25 14.44 92.21 16.35PerceptualOrganization(P0) 89.55 14.30 93.75 14.52 94.92 15.44ProcessingSpeed(PS) 91.48 15.01 94.03 12.99 94.32 16.45RESULTS /90Table 5. Descriptive statistics for SB:FE and PPVT-R by groupOtitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Subtest Group I Group II Group EliN87 N=40 N=40mean SD mean SD mean SDPattern 43.90 5.96 45.77 6.49 47.23 6.99AnalysisMatrices 41.37 5.28 43.73 6.01 43.94 5.17BeadMemory 39.71 5.97 42.03 6.54 41.85 7.57Memory forObjects 42.28 5.79 42.97 5.73 44.03 6.04Abstract/VisualReasoning 83.67 11.18 89.03 11.97 89.82 12.40Short TermMemory 79.17 11.92 83.46 12.16 83.18 12.68Partial TestComposite 79.41 10.90 84.77 12.03 84.64 12.43VocabularySB:FE 83.06 11.56 87.55 12.79 86.13 10.66PPVT-R 75.85 15.42 77.60 17.12 76.30 17.92RESULTS 191Group II had several subtests which were also low, Bead Memory (SAS 42.03) andthe Short Term Memory Area (SAS 83.46) from the SB:FE. The PPVT-R score (SS77.60) was also more than one SD below the mean or average.When hearing peers (Group Ill) were given verbal directions for subtests, the ShortTerm Memory Area SB:FE (SAS 83.18) and the PPVT-R (SS 76.30) were the onlymeasures one standard deviation below the mean.2. T-testsT-tests were used to compare all measures administered to subjects in Group Iwith the standardization sample. Results are presented in Tables 6 and 7. Whenresults were examined, it was evident that there were statistically significantdifferences between Group I and the standardization sample. The null hypothesiswas rejected in each case with the exception of Picture Arrangement (t (83) =1.80,p> .05, and Symbol Search (t (86)=1 .88), p> .05.Group II data were then compared to the standardization sample. The nullhypothesis, that Group II scores would approximate the standardization sample,was not rejected for Object Assembly (t (37)= .76, p> .05), Block Design(t (37)=1 .36, p> .05) and Symbol Search (t (37)=1 .11, p> .05).RESULTS /92The null hypothesis was rejected for the remaining subtests of the WISC-lll andSB:FE (Tables 6 and 7).The null hypothesis was not rejected for Group Ill, for Picture Completion (t(37)=.33, p> .05), Object Assembly (t (37)=1 .45, p> .05), Block Design (t(37)=.68,p> .05), and Symbol Search (t (37)=.04, p> .05). The remaining test and subtestscores from the WISC-lll and SB:FE caused hypothesis #1 to be rejected.RESULTS / 93Test/subtestPictureCompletionCodingPictureArrangementObjectAssemblyBlock DesignSymbol SearchPIQP0PS*p<05Group I, n=86, Group II,Table 6. T-Test for the WISC-lll by group when compared to the standardizationsampleOtitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Group I Group II Group Ill527* 203* 033202* 481* 470*1.80 557* 395*2.75* 0.76 1.452.71* 1.36 0.681.88 1.11 0.04818* 340* 301*896* 272* 208*533* 2.92* 2.19*n=38, Group Ill, n=38RESULTS /94Table 7. T-Test for the SB:FE and PPVT-R by group when compared to thestandardization sampleOtitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Test/subtest Group I Group II Group IllPatternAnalysis 9.32* 4j3* 2.47*Matrices 14.91* 6.56* 8.14*Bead Memory 15.70* 7.71* 6.73*Memory forObjects 12.13* 775* 8.11*Abstract/VisualReasoning 13.31* 5.80* 5.20*Short TermMemory 15.00* 8.60* 8.40*SB:FEPartial Composite 17.21* 8.01* 7.82*SB:FE Vocabulary 13.17* 6.16* 8.23*PPVT-R 14.28* 8.20* 8.25**p<05RESULTS/954. Confidence IntervalsConfidence intervals provide another means of expressing the precision of testscores (Weschler, 1991). They also assist in the test interpretation by providingthe examiner with a range of scores in which the child is likely to fall. It is commonpractice to establish confidence intervals based on SEM values and centre theinterval around the child’s obtained score. The reporting of confidence intervalsserves as a reminder that the observed score contains some amount of error(Weschler, 1991).Intelligence scores are not a single number or score; therefore, confidenceintervals were calculated for the PIQ scores, index scores on the WISC-lll, and thecomposite and area scores on the SB:FE (Table 8).When these figures are examined, it is evident each of the groups has a lowermean or average than the standardization sample; however, there is overlapbetween the confidence intervals for the three groups and the standardizationsample.Confidence intervals for the t-tests were calculated for each of the non significantsubtests from the three groups. For Groups I, Picture Arrangement 5.76 to 7.16and Symbol Search 8.52 to 9.93. Group II, Object Assembly 9.02 to 10.14, BlockRESULTS /96Table 8. Confidence intervals and means for the WSPIQ, WSPO, WSPS,SB:FE Composite Score, and Area ScoresOtitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Group I Group II Group IllWSPIQ 84.06 to 90.11 87.63 to 96.87 86.84 to 97.59(87.08) (92.25) (92.21)WSPO 86.43 to 92.68 89.11 to 98.39 89.85 to 100.00(89.55) (93.75) (94.92)WSPS 88.20 to 94.76 89.87 to 98.18 88.91 to 99.72(91.48) (94.03) (94.32)SBTOT 77.09 to 81.74 80.87 to 88.67 80.61 to 88.67(79.41) (84.77) (84.64)SBSTM 76.79 to 81.56 79.52 to 87.40 79.07 to 87.29(79.17) (83.46) (83.18)SBAVR 81.28 to 86.05 85.15 to 92.91 85.80 to 93.84(83.67) (89.03) (83.18)Design 9.02 to 10.14 and Symbol Search 8.87 to 10.05. Group Ill, ObjectAssembly 8.63 to 10.20, Block Design 8.75 to 10.61 and Symbol Search 8.82 to11.24.The confidence intervals for Groups II and Ill are within average limits on the WISCRESULTS! 97Ill. Results for Group I indicate the confidence interval for the Symbol Searchsubtest is within the average range, with the Picture Arrangement interval beingmore than one SD below the mean.B. Hypothesis #2aIt is expected scores from nonverbal measures of intelligence administered withpantomime instructions will approximate scores when measures are administeredwith verbal instructions for the hearing peer group.1. Comparison of Administration MethodsOne way ANOVA’s were used to calculate differences between groups to test thehypothesis. A conservative level of significance (.01) was used for the analyses.Data from Groups II and Ill were examined initially. After examining results,statistically significant differences between the two administration methods forGroups II and Ill were not found. Results from the WISC-lII (Group II, n=40and Group Ill, n=38) were as follows: WSPIQ, F(1,77)=.0001, p>.01, WSPS,F(1 ,77)= .12, p> .01, WSPS, F(1 ,77)= .008, p> .01. ANOVA results for the SBTOTwas F(1 ,77) = .002, p> .01. These results caused hypothesis # 2a not to berejected.RESULTS /98C. Hypothesis #2bIt is expected that nonverbal measures of intelligence administered with pantomimeinstructions to the otitis prone group will yield results that approximate those of thehearing peer group receMng pantomime instruction.1. Group Comparisons with Pantomime AdministrationData were examined by comparing Group I (n=84) and Group II (n=38) usingANOVA. Subjects in these groups had received pantomime instruction. For eachof the ANOVAs the following assumptions were examined: normality, heterogeneityof variance, and independence of residuals. Normality was assessed for each ofthe ANOVAs by examining the distribution of the error component. Heterogeneityof variance was examined by Bartlett’s F test (Appendix C). In addition, asmentioned previously a more stringent level of significance was applied in theanalysis (p> .01). Independence of residuals was accepted as observations withinthe groups were assumed to be independent of each other (Glass & Hopkins,1984). Assumptions for each ANOVA were tested and not violated.Groups for the analysis were unequal. Collyer and Enns (1987) suggestunweighted means or proportional means as possible solutions. This was dealtRESULTS /99with when Groups II and Ill were collasped into one group for the purposes of theanalysis, as described later in this chapter.Statistically significant differences between Group I (n=83) and Group II (n=40)were not evident when the WISC-lll data were examined (p < .01). ANOVA resultsfor the WISC-Ill were as follows: WSPIQ, F(1 ,122)=3.65, p>.01, WSPO,F(1,122)=2.30, P>.01, WSPS, F(1,122)=.84, p>.01. Significant differencesbetween Group I (n=87) and Group II (n=39) for the SB:FE were not found.Results of the SB:FE were: SBTOT, F(1,125)=6.09, p>.01, SBSTM,F(1,125)=3.75, p>.01, SBAVR, F(1,125)=5.93, p>.01. These results causedhypothesis #2b not rejected for IQ scores, index scores, and subtests of theWISC-lll and SB:FE.As stated previously, hearing peers were randomly assigned to either verbal orpantomime administration. Analyses completed to test hypothesis #2ademonstrated there were no differences between administration methods forGroups II and Ill. Groups II and III were collapsed into one group (n=76) forpurposes of analysis. Although groups remained unequal, subjects in each groupwere approximate (otitis prone, n=84, hearing peer, n=76).ANOVA’s were then completed for Group I (otitis prone) and Groups Il/Ill (hearingRESULTS /100peer. There were no significant differences between groups (p < .01) for the WISCIll and SB:FE scores with the exception of the SB:FE Abstract/Visual Reasoningscore (Appendix C). Results of the WSPIQ, F(1 ,160)=5.01, p< .01 (Group I, n=83and Group Il/Ill, n=78) and SBTOT, F(1,164)=8.7, p <.01 (Group I, n=87 andGroupll/Ill, n=78) causing hypothesis #2b not to be rejected. A statisticallysignificant difference was found between groups for the SB:FE Abstract/VisualReasoning area score (F(1 ,164) = 10.08, p> .01). The hypothesis #2b wasrejected for this area score.The power of the F-tests (effect size .05) was examined to determine how likelycorrect conclusions had been made regarding the null hypotheses. For analysisof administration methods (Group II and Group Ill) at .01, the power wasapproximately .75. For group comparisons for pantomime administration (GroupI, n=84, Group II, n=76) at .01, the power was approximately .80 (Glass &Hopkins, 1982, p.542).D. Hypothesis #3aThe WISC-lll Performance IQ score and the SB:FE Partial Composite score will bepredictors of achievement for the otitis prone and hearing peer groups.RESULTS /101E. Hypothesis #3bSpecific subtests of the WISC-Ill Performance Scales and subtests from the SB:FEPartial Composite will be predictors of achievement for the otitis prone and hearingpeer groups.To test for Hypothesis .#3a and #3b regression equations were used to constructthe best possible linear equation for predicting the score on one variable when thescore on another variable is known. The Biomedical Computer Program (BMDP)was used to test the research questions which queried which subtests or groupof subtests predicted math and reading scores.BMDP, 9R Program, “AN Possible Subtests Regression,” examines regressionequations for the “best” subtests of predictor variables and provides detailedresidual analysis. Best subtests are those which satisfy the specified selectioncriterion, better than other subtests, or the best subset overall, which satisfies thecriteria.Reading and math were the dependent variables. The best possible combinationof independent variables as predictors were generated. Regression equations andthe multiple correlation coefficients are reported in Table 9. Standardizedregression coefficients were examined in Table 10.RESULTS/ 102Standard scores from the SAT-HI manual were used for the regression analysis.Kelly and Braden (1990) had recommended the use of percentiles with a group ofdeaf children who had severe to profound hearing loss. The otitis prone group inthe present study had subjects had a mild to moderate loss. Hearing peers andsome members of the otitis prone group did not have a pure tone loss, makingthis sample very different from the subjects used in the Kelly and Braden (1990)study. In addition, although subjects’ ages ranged from 6-0 to 16-11, the majorityused Primary I, II, or lit Reading Comprehension tests making age-basedpercentiles inappropriate.Huebner (1989) warns against the use of percentiles as intervals are unequal andoften misleading. Small differences in the centre of the distribution areexaggerated and differences near the extremes of the distribution are minimized.Use of ordinal data (percentiles) in general is ambiguous and often open tomisinterpretation because of the problem of unequal intervals (Reynolds, 1981).Subjects’ chronological age (CA) ranged from 6-0 to 16-11 for the study. Thisrange was consistent with the age range used for the WISC-IlI. CA was oneindependent variables entered in the regression equation to answer hypothesis#3a. CA was then partialled out to determine whether the R squared values wouldchange significantly.RESULTS /103Table 9. Regression equations, R squared values, multiple correlation coefficients,and R squared values with CA partialled out for reading comprehension and mathcomputation for groups I, II and IllMuftiple R withGroup Regression Equation Correlation CAZ. Partlalled.A)IIIcIenLI Reading=n=56 3.13 WSOA + .91 SBVOC + .85 CA + 359.52 .63 .40 .19n=62 Math=-22.36 WSCOD + 19.62 WSPIQ + -16.06 WSPO+ 1.35 CA + 282.89 .73 .53 .13II Reading=n=27 1.41 SBTOT + 2.09 PPVTR + 1.64 CA + 68.44 .90 .81 .75Math=6.87 WSGOD + 4.85 SBMOB + 1.87 PPVTR +2.2 CA + -24.29 .89 .79 .47III Reading=n=23 4.9 WSPC + 8.2 WSGOD + -9.82 SBBM+ -7.46 SBMOB + 6.32 SBSTM + 1.15 PPVTR+ 0.86 CA + 465.48 .90 .81 .69n=32 Math=1.811 PPVTR + 1.54 CA + 229.07 .75 .56 .44WSOA WISC-lll, Object Assembly, SBBM = SB:FE, Bead MemoryWSCOD = WISC-lll, Coding subtest, CA = Chronological AgeWSPIQ = WISC-lll, Performance IQ, WSPO = WISC= Ill, PerceptualOrganization,PPVTR = Peabody Picture Vocabulary - Revised, SBMOB = SB:FE, Memory forObjects,SBSTM SBFE, Short Term Memory Area, SBBM = SB:FE Bead Memory.RESULTS I 104Table 10. Standardized regression coefficients, R squared values andmultiple correlation coefficients for reading comprehension and mathcomputation for groups I, II and IlLGroup Standardized RegressIon Coefficients MultipleCorrelationCoefficient RReading =n=56 0.23 WSOA + 0.25 SBVOC + 0.46 CA + 7.4 40n=62 Math=-1.1 WSCOD + 4.3 WSPIQ + -3.6 WSPO÷0.64CA+4.49 73 53II Reading=n=27 0.27 SBTOT + 0.5 PPVTR + 0.78 CA + 1.1.81n=31 Math=0.22 WSCOD + 0.31 SBMOB + 0.39 PPVTR +0.86 CA + -029 89 79Ill Reading=n=23 0.24 WSPC + 0.46 WSGOD + -1.3 SBBM+ -0.72 SBMOB + 1.35 SBSTM + 0.28 PPVTR+ 0.41 CA +. .81n=32 Math=0.40 PPVTR + 0.57 CA + 2.71 56WSOA WISC-lIl, Object Assembly, SBBM = SB:FE, Bead MemoryWSCOD WISC-lll, Coding subtest, CA = Chronological AgeWSPIQ WISC-lll, Performance IQ, WSPO = WISC= III, PerceptualOrganization,PPVTR = Peabody Picture Vocabulary - Revised, SBMOB SB:FE, Memory forObjects,SBSTM = SBFE, Short Term Memory Area, SBBM = SB:FE Bead Memory.RESULTS /105Results of the regression analysis indicated that for Group I, Reading, the Rsquared value was 40%, indicating the independent variables contributed toapproximately half of the variance. In the regression equation, CA contributed tothe R squared value, with WSOA and SBVOC also making contributions. Theregression equation for math, indicated that the R squared value was 53% CAcontributed to the variance with WSPO, WSPIQ, WSCOD and WSPIQ makingcontributions to the regression equation.Group II results for reading indicated 81% of the variance could be explained viathe independent variables. CA contributed to the variance with PPVT-R andSBTOT also contributing. In math the R squared value was 79%. CA made thelargest contribution to the variance, with additional variables PPVTR and SBMQBmaking contributions.Group III results for math indicated the R squared value was 56%. Alsocontributing to the variance were CA, and the PPVTR score. In reading, the Rsquared value was 81%, with CA, WSPC, WSCOD, SBBM, SBMOB, SBSTM, andthe PPVTR all contributing to the variance.When CA was partialled out of the regression equations, the R squared values didnot change for Group Ill math and reading, or for Group II reading.RESULTS /106Differences between the R squared values and the R squared values with CApartialled out, were found for Group I reading and math and for Group II math.The R squared values when CA was partialled out were lower, indicating theindependent variables contributed less to the variance.Results of the regression analysis caused hypothesis #3a and #3b to be rejected.The WSPIQ or the SB:FE as measures of nonverbal intelligence did not predictachievement for the otitis prone or hearing peer groups.E. Hypothesis #4aIt is expected there will be a statistically significant correlation between the WISC-IllPerformance IQ score and the SB:FE Partial Composite score for the otitis proneand hearing groups.Pearson-product moment correlations were completed using variables for each ofthe three groups. The Bonferroni t procedure (Glass & Hopkins, 1984) was usedto correct for alpha slippage over repeated comparisons between test and subtestscaled scores. The research question related specifically to the measures ofintelligence, and whether they were correlated. In addition, correlations wereexamined between the subtests of the measures of intelligence (Table 11).RESULTS /107Table 11. Correlations between measures of intelligence and Index/Area Scoresfor Group IWSPIQ WSPO WSPS SBAVR SBSTM SBTOTWSPIQ 1.00WSPO 0.96* 1.00WSPS 0.51* 0.32 1.00SBAVR 0.50* 0.54* 0.13 1.00SBSTM 052* 056* 021 056* 100SBTOT 0.58* 0.63* 0.20 0.88* 0.87* 1.00*p(001 n=83WSPIQ = WISCIII, Performance IQ, WSPO = WISC-llI, Perceptual Organization,WSPS = WISC-IIl, Processing Speed, SBAVR = SB:FE, Abstract/Visual ReasoningArea, SBSTM SB:FE, Short Term Memory Area, SBTOT SB:FE , PartialComposite ScoreCorrelations for Group I indicate the majority of tests of intelligence ,Index Scores,Area Scores correlate with each other (p < .001). The exception is the correlationbetween the Processing Speed Index Score (WISC-lll), and SBAVR, r=.13,SBSTM, r=.21 and SBTOT, r=.20.RESULTS /108Table 12. Correlations between measures of intelligence and Index/Area Scoresfor Group IIWSPIQ WSPO WSPS SBAVR SBSTM SBTOTWSPIQ 1.00WSPO 0.98* 1.00WSPS 0.58* 0.49* 1.00SBAVR 0.60* 0.60* 0.41 1.00SBSTM 0.39 0.39 0.40 0.58* 1.00SBTQT 0.57* 0.57* 0.46 0.90* 0.87* 1.00*p<001 n=38Results indicate the majority of correlations between the measures of intelligence,area scores and Index Scores are significant at the .001 level. Correlationsbetween SBSTM and WSPIQ, r=.39, WSPO, r=.39 and WSPS, r=.40 were notsignficant when p < .001. In addition, correlations between the WSPS Index Scoreand SBAVR, r=.41, SBSTM, r=.40, and SBTOT, r=.46 were not significant whenp.cz.001.RESULTS /109Table 13. Correlations between measures of intelligence and Index/Area Scoresfor Group IllWSPIQ WSPO WSPS SBAVR SBSTM SBTOTWSPIQ 1.00WSPO 0.98*WSPS 079* 068* 100SBAVR 065* 066* 054* 100SBSTM 058* 055* 052* 056* 100SBTOT 0.65* 0.62* 0.61 * 0.87* 0.88* 1.00*p<00l n=38Results indicate all measures of intelligence, Area Scores and subtest scores aresignificantly correlated for Group Ill.In addition to determining significant correlations between measures of intelligencefor Groups I, II, and Ill, it was also important to determine whether there was adifference between correlation coefficients. The Fisher z transformation (Glass &Hopkins, 1984) was used to place confidence intervals about the correlations toRESULTS /110determine whether there were significant differences between correlations forGroups I, II, and Ill.The correlations between WSPS/WSPIQ for Groups I and II, r=0.51, r=0.58(confidence interval .36 to .77) and WSPS/WSPO for Groups I and II, r=0.32,r=0.49 (confidence interval .13 to .53) were significantly lower than the correlationscalculated for Group Ill. The correlations for Group Ill, WSPS/WSPIQ, r=.79, andWSPS/WSPQ, r=.68 were outside of the confidence interval established by theFisher z.In Group I, correlations for SBAVR/WSPS, r=O.13 (confidence interval .01 to .33),SBSTM/WSPS, r=0.21 (confidence interval .01 to .39) and SBTOT/WSPS, r=O.20(confidence interval .01 to .41) were lower than correlations for Groups II and III.For Group II SBSTM/WSPIQ, r=O.39 (confidence interval .24 to .56), andSBSTM/WSPS, r=O.40 (confidence interval .25 to .60) were outside of theconfidence interval established by the Fisher z, indicating a lower correlations thanthose reported for Groups I and Ill.Although correlations are significant between test and subtest scores for all threegroups (Tables 11, 12, and 13), when using the Fisher z transformation, there aredifferences between the significant correlations reported.RESULTS /111F. Hypothesis #4bIt is expected there will be a correlation between nonverbal measures ofintelligence and achievement for the hearing peer and otitis prone groups.To test this hypothesis, correlations were completed between nonverbal measuresof intelligence and achievement for all three groups.Table 14. Correlations between subtests of the WISC-lll and Partial CompositeSB:FE, Index and Area scores for Group IWSPC WSCOD WSPA WSOA WSBD WSSSWSPIQ 073* 055* 073* 066* 066* 03&WSPO 0.75* 0.31 0.71* 0.70* 0.69* 0.27*WSPS 0.23 0.81* 0.22 028 0.29 0.87*SBAVR 0.34* 0.05 0.34* 0.40* 0.49* 0.12SBSTM 0.36* 0.07 0.40* 0.35* 0.47* 0.22SBTOT 0.45* 0.06 0.41* 0.41* 0.56* 0.22p<.0O1, n=83RESULTS /112Table 15. Correlations between subtests of the WISC-lll and measures of intelligence,Index and Area scores for Group IIWSPC WSCOD WSPA WSOA WSBD WSSSWSPIQ 084* 062* 051* 076* 075* 038WSPO 0.86* 0.50* 0.48* 0.75* 0.81* 0.34WSPS 0.36 0.80* 0.34 0.43 0.34 0.87*SBAVR 0.47* 0.40 0.47* 0.38 0.56* 0.29SBSTM 0.31 0.36 0.28 0.31 0.30 0.32SBTOT 0.45 0.43 0.45 0.39 0.49* 0.34*p<J1 n=39Results indicate correlations are significant between the WISC-llI index scores andWISC-fll subtest scores. Correlations were not evident when the SB:FE area scoreswere correlated with WISC-lll subtest scores.RESULTS /113Table 16. Correlations between subtests of the WISC-Ill and measures of intelligence,Index and Area scores for Group IllWSPC WSCOD WSPA WSOA WSBDWSPIQ 081* 079* 083* 073* 057* 057*WSPO 083* 064* 083* 080* 078* 054*WSPS 059* 086* 064* 057* 036 089*SBAVR 0.54* 0.45 0.57* 0.52* 0.50* 0.51*SBSTM 0.48 0.56* 0.37 0.39 0.45 0.36SBTOT 0.52* 0.57* 0.49* 0.50* 0.48* 0.50**p<J1 n=38Subtest scores were also examined. It appears SBSTM area score does notcorrelate with subtests of the WISC-lll. This could be evidence to indicate that shortterm memory skills on the SB:FE do not correlate with skills necessary to completesubtests of the WISC-lII.RESULTS /114Table 17. Correlations between nonverbal measures of intelligence andSAT-HI (Groups I, II, Ill)Otitis Prone Hearing Peer Hearing Peer(pantomime) (pantomime) (verbal)Group I Group II Group IllMath Reading Math Reading Math ReadingWSPIQ .31 43* .10 .34 .40 55*WSPO .26 •39* .10 .31 .37 .52*WSPS .15 .17 .11 .09 .41 .51*SBAVR .05 .13 .07 .29 .19 .29SBSTM .03 .15 .01 .10 .21 .26SBTOT .04 .14 .02 .25 .25 .30*p<.001Correlations between the WSPIQ and Group I reading,(r=.43) and Group Ill reading,(r= .55) were significant. The SB:TOT score did not correlate significantly withmeasures of achievement. Scores from the WISC-Ill appear to correlate withreading measures for Groups I and Ill, whereas the SBTOT as a measure ofintelligence does not correlate with measures of achievement.RESULTS /115Therefore hypotheses #4a and #4b were not rejected for Groups I, II, and Ill.There were significant correlations between measures of nonverbal intelligence, andcorrelations were found between measures of nonverbal intelligence andachievement.G. Item AnalysisScores from the three groups have been compared using t-tests, ANOVAs,correlations, and by examination of the descriptive statistics. All of the abovecomparisons examine either test or subtest results for each group. It was decidedto examine beyond global results of test and subtest standard scores, and toexamine individual differences on item response. It was important to determine theeffect of pantomime and verbal instructions for each of the items.This was difficult for the majority of subtests administered. WISC-lll subtests relyon time and accuracy for a higher “point” response. A response can receive partialcredit or differing scores depending on the speed of the response. PictureCompletion was the only subtest administered which had responses scored ascorrect or incorrect. Therefore, responses from Group II and Group Ill wereinvestigated for each item, with the purpose of determining whether there was anydifference between the items administered using either pantomime or verbalRESULTS /116instruction.Item response theory (IRT) is a statistical theory which consists of a family ofmodels that express the probability of observing a particular response to an itemas a function of certain characteristics of the item and the ability level of theexaminee (Crocker & Algina, 1986). IRT typically requires a large sample size (N= 500). Although the current sample is significantly smaller, it was decided theanalysis could provide information as to how subjects using pantomime or verbalinstruction responded to each item of the Picture Completion subtest.Results of the item analysis should be viewed as exploratory and should beinterpreted with caution as the sample size is considerably smaller thanrecommended and only one subtest was utilized for the analysis. Results that aregenerated are likely to be “liberal” estimates, and may be questionable as result ofthe small sample.The item parameter estimation program (ASCAL) was used from the MicroCATprogram. The three parameter model in this program is applicable to questionsscored in a dichotomous (for example correct-incorrect) manner. The (a) parameterin this model represents the item’s capability of discriminating between levels ofability, It corresponds to the rate at which the probability of a correct response toRESULTS/I 17an item changes as a function of the ability level of each examinee. The (b)parameter is an index of the item’s difficulty. The (c) parameter is equal to theprobability that an examinee of extremely low ability can answer the item correctly,or in other words, that the item can be answered correctly by guessing. If an emcan not be guessed correctly, the parameter (c) = 0.00.RESULTS /118Table 18. Group II (Pantomime Instruction) parameter estimates responsesfor the Picture Completion subtest (WISC-Ill)a b c N Chi dfSquareItem1 1.486 -2.696 0.00 39 0.167 182 1.887 -2.142 0.00 39 0.704 183 1.460 -1.729 0.00 39 8.728* 184 1.503 -1.675 0.00 39 9.956* 185 1.283 -1.971 0.00 39 11.013* 186 1.171 -1.180 0.00 39 34.366* 187. 0.500 -2.428 0.00 39 28.619* 188 1.894 -0.593 0.00 39 6.542 189 1.364 -0.398 0.00 39 19.761* 1810 0.500 -0.697 0.00 39 23.364* 1811 1.920 -0.674 0.00 39 6.513 1812 1.207 -0.510 0.00 39 24.913* 1813 1.409 -0.441 0.00 37 30.453* 1814 1.328 0.081 0.00 37 24.332* 1815 1.727 -0.182 0.00 37 15.090* 1816 0.865 -0.668 0.00 36 17.490* 1817 1.783 -0.143 0.00 36 9.660* 1818 1.661 0.059 0.00 34 9.677* 1719 1.182 0.014 0.00 34 16.413* 1720 1.515 0.483 0.00 33 17.725* 1621 1.443 0.279 0.00 32 13.320* 1622 1.114 0.635 0.00 30 12.737* 1423 0.905 1.644 0.00 26 12.620* 1424 1.244 0.568 0.00 26 12.122* 1425 0.900 1.009 0.00 23 8.500 1226 1.060 1.367 0.00 21 17.159* 1227 1.219 1.652 0.00 18 14.883* 1028 1.482 1.567 0.00 17 1.425 1029 1.329 1.822 0.00 16 2.944 930 0.978 2.869 0.00 14 33.969* g* p<.05 (n=38)RESULTS /119Table 19. Group Ill (Verbal Instruction) parameter estimates responses for thePicture Completion subtest (WISC- Ill)a b c N Chi dfSquareItem1 1.494 -2.636 0.00 38 0.211 182 1.600 -2.210 0.00 38 3.284 183 1.600 -2.210 0.00 38 3.284 184 1.600 -2.210 0.00 38 3.284 185 1.275 -1.810 0.00 38 22.179* 186 1.898 -1.738 0.00 38 0.870 187 1.898 -1.738 0.00 38 0.870 188 1.430 -1.756 0.00 38 19.227* 189 1.124 -0.660 0.00 38 8.002 1810 1.320 -0.298 0.00 37 16.876* 1811 1.312 -1.085 0.00 37 43.638* 1812 1.373 -0.866 0.00 37 12.182* 1813 1.600 -0.923 0.00 37 6.177 1814 1.502 -0.109 0.00 37 19.318* 1815 1.855 -0.738 0.00 37 7.579 1816 1.405 -0.983 0.00 37 8.091 1817 1.540 -0.224 0.00 36 9.468* 1818 2.075 -0.708 0.00 36 6.293 1819 1.101 -0.430 0.00 35 20.081* 1820 1.583 -0.380 0.00 35 8.661 1821 0.525 -0.357 0.00 35 18.925* 1822 1.556 0.257 0.00 35 4.682 1823 0.500 1.345 0.00 35 29.005* 1824 1.052 1.058 0.00 35 41.035* 1825 0.500 1.941 0.00 31 11.196* 1726 1.130 0.380 0.00 29 25.782* 1527 1.957 1.083 0.00 26 3.751 1428 1.749 1.571 0.00 24 2.214 1429 1.083 1.111 0.00 22 8.934* 1330 1.300 2.212 0.00 19 2.517 12*p<05 (n=38)RESULTS /120The chi-square value is called the “goodness of fit” test because it tests how closelyobserved frequencies from a sample fit theoretically expected frequencies based onthe null hypothesis (Shavelson, 1988).Results of the item analyses can be examined in several ways. If chi-squared valuesare examined, the “goodness of fit” for the items on the Picture Completion subtestare not consistent for Groups II and Ill. Results for Group II indicated 19 of the 30items have values which are significant (Table 17). When Group III data wereexamined, 14 of the 30 items were significant (Table 18).Chi-square values were then examined for both Group II and Ill. Only 10 of theitems have values which are significant for both groups. Examining items in thismanner did indicate that subjects responded differently to items that wereadministered via pantomime or verbal instruction. Although there was not asignificant difference between the two administration method when ANOVAs werecalculated, at an item level there was a difference as to how the subjects respondedto instruction.Therefore, the 10 items that had a “good fit” for both groups were examined. Whenconsidering the 10 items only, for Group lithe mean was 5.95 with a standarddeviation of 2.64. In Group Iii, the mean was 5.76 with a standardRESULTS /121deviation of 2.37. Subject responses on those items were used to complete a ttest. Results of the t-test (t (9) = .031, p < .05) indicated there was not a statisticallysignificant difference between the verbal and pantomime administration.Items were also examined under the parameter which examines item difficulty (b).When item difficulty was examined there was a difference as to how subjects inGroups II and Ill responded to each item (Tables 18 and 19, Item Difficulty (b)). Adifference in response could be dependent on method of test administration, eitherpantomime or verbal directions. This will be discussed in Chapter VI.H. Summary of ResultsResults demonstrate pantomime instruction gave subjects information required tocomplete tasks of the WISC-llI and SB:FE despite the diversity of the groups. Therewas not a significant difference between verbal and pantomime instruction for thehearing peer group. When data from the otitis prone group and the hearing peergroup were compared, statistically significant differences between the groups for theWISC-lll were not evident.Multiple regression analyses demonstrated that nonverbal measures of intelligencewere not predictors math or reading achievement as measured by the SAT-HI. ForRESULTS /122all three groups, chronological age (CA) was a variable which was consistentlycontributing to the variance in the regression equations.When CA was partialled out from the regression equations the R squared values didchange for Group I reading and math , and Group II math. Lower R squared va!uesindicated less of the variance could be explained by the independent variables in theregression equation.Language measures administered yielded very low scores across all three groups.Although this was expected for the otitis prone group, low scores were notexpected for the hearing peer group. The short term memory subtests from theSB:FE were also another area of weakness for all three groups. This finding wasnot expected and will be discussed further in Chapter VI./123VI. DISCUSSIONA. Review of the Purposes of the StudyThere were two purposes of the study. The first was to study the use of consistentor standard administration procedures using pantomime instruction, with tworecently revised measures of intelligence. The second was to determine ifnonverbal measures of intelligence administered with standard pantomimeinstructions could be used to predict academic achievement for otitis prone andhearing Native Indian children.B. Hypothesis #1It is expected nonverbal measures of intelligence administered with pantomimeinstructions will yield results that approximate the standardization sample for boththe otitis prone and hearing peer groups.1) Mean ScoresWhether Performance IQ and the Stanford Binet Partial Composite score are lowerthan the standardization sample has not been reported in studies reviewed.Previous studies (Braden, 1985b; Brooks & Riggs, 1980; Watson etal., 1986) usingthe WISC and WISC-R to assess Native Indian and hearing impaired children foundthe Verbal IQ means to be lower than the standardization group (WISC-R,DISCUSSION/124mean = 100 SD =15, and SB:FE, mean = 100 SD = 16). Performance IQ scoresin the present study were within the average or expected range (Tables 4 and 5).SuNivan and Schulte (1992) found the mean of the WISC-R PIQ to be 107.70 witha SD of 17.80 for hard of hearing subjects with a loss of 60 dB or less.Mailer and Braden (in press) report the mean for the WISC-Ill PIQ to be 105.83 fora sample of deaf children.In the present study WISC-lll PIQ scores, for all three groups, were lower than thisscore reported by Mailer and Braden (105.83), but within one standard deviationof the mean (Group I, 87.1, Group II, 92.3, Group III, 92.2). Mean scores for theSB:FE found Group I (79.4) to be more than one standard deviation below themean. Group 11(84.8) and Group III (84.6) were within average or expected limits,or within one standard deviation of the mean.Means were also calculated for index scores, area scores and subtest scores forthe otitis prone and hearing groups. These means were different from thosereported by Sullivan and Schuite (1992). They completed a factor analysis of theWISC-R with subjects who were hard of hearing as well as profoundly deaf.Subtest scores reported were Picture Arrangement (mean 11.57) and Coding(mean 10.25) for hard of hearing subjects. These mean scores were higher thanDISCUSSION / 125those reported in the present study for all three groups.The present study reported Group I WISC-lll subtest means for Coding (mean6.98) and Picture Arrangement (mean 6.46). The coding subtest requires the childto complete a visual motor task requiring accurate visual motor skill, attention totask, and short term memory skills, whereas Picture Arrangement requires thesubject to sequence a series of pictures to tell a story.Zinkus (1986) found that visual processing skills were comparable betweenchildren with otitis media and hearing counterparts. The observation was madethat visual tasks which had a language component or were language basedshowed delay or impairment in the group with early ear disease. The Codingsubtest may require internal verbal mediation for successful completion. Subjectswho can complete the task quickly and accurately also use short term memoryskills to associate the “number” and the “correct shape.”Picture Arrangement is a performance or nonverbal test; however, internallanguage skills are necessary if events are to be sequenced correctly. Tosequence the pictures into a logical series of events or to “tell a story,” the childneeds an internal language or system to identify the sequence of events. Sattler(1988) thought the Picture Arrangement subtest to be verbally loaded. The childDISCUSSION /126does not give a verbal response, yet must have a language system intact tocomplete the task successfully.When t-test results were examined for Group I, the Picture Arrangement subtestwas not significantly different from the standardization sample. This could beexplained by examining individual differences, and the range which scores ofGroup IThe low mean score obtained by the otitis prone group could be another indicationof the language difficulties experienced by this group. This finding is consistentwith previous research which found the Picture Arrangement subtest of the WISC-Rto be difficult for hearing impaired children (Sullivan, 1982).SB:FE results for Group I found the Short Term Memory Area score (79.17), BeadMemory subtest (39.71), and Matrices (41.37) were more than one standarddeviation below the mean. Results for Group II and Group Ill did not indicateweaknesses on these subtests.Zinkus (1986), in a study with children with otitis media and their siblings, foundvisual sequential memory skills were a strength for the group with otitis media.Investigators suggested this visual strength was evidence that children with otitisDISCUSSION / 127media compensated for their auditory processing deficits by developing strengthsin the visual area. Results of the present study did not support Zinkus’ (1986)theory. The Short Term Memory Area, which assesses visual memory on theSB:FE, is a weakness when means are examined.Subjects in both groups had very low language skills when compared to thestandardization sample. It is possible that short term visual memory also relies oninternal language to some degree. To recall items on Bead Memory, for example,the assumption is that the subject will attach labels to the bead shapes to aidrecall. Although various strategies may be utilized to assist in recall, languagemust play a role, thus in part accounting for this finding.2) Standard DeviationsStandard deviations for the WISC-lIl and the SB:FE are similar to those reportedfor the standardization sample (WISC-lll, SD=15, SB:FE, SD=16). It appearsthe dispersion of scores or standard deviations reported for the otitis prone andhearing Native Indian children are similar to the standardization sample.Despite the lower means, scores were normally distributed and had standarddeviations that were similar to the standardization sample. When confidenceDISCUSSION / 128intervals for the WSPIQ and SB:FE were calculated it was evident that the intervalsreported overlapped those from the standardization sample in that they were withinone SD of the mean.Seyfort et al. (1980) discussed the lower means and confidence intervals inrelationship to placement and identification. Mean scores lower than the average,but with normal distributions, would indicate fewer children would be achieving atthe higher end of the scale (above average). A larger percentage of the populationwould then fall within the low average and borderline range.This could be explained by the Native Indian child’s lack of English language or thefact that many of the subjects had English as their second language, and hearingloss as a result of chronic otitis media.3) Language MeasuresAs stated in Chapter II there has been controversy in the literature as to whetherotitis media has an effect on speech and language development of children(Menyuk, 1980). One position is that children who experience episodes of otitismedia also experience mild to moderate hearing loss which presumably returns tonormal after the episodes (Menyuk, 1986).DISCUSSION / 129The other side of the theoretical argument is that persistent otitis media presentsan auditory signal that is difficult to process. An unstable signal creates difficultiesin establishing speech and language skills as a result of the mild to moderate loss(Nober & Nober, 1977). Research reviewed in Chapter II suggested best practicewas not to administer Verbal lQ measures to deaf and Native Indian children.Therefore, intelligence measures that rely on language skills were not administeredto subjects in the present study.It was important to document the language skills of the subjects in Groups I, II,and Ill. The Vocabulary subtest from the SB:FE was administered to determineexpressive language and the PPVT-R was administered to determine receptivelanguage skill level.Means of the PPVT-R were more than one standard deviation below the mean forall three groups (Group I, x=75.85, Group II, x=77.6, Group Ill, x=76.3). Whent-tests were completed, there was not a significant difference between groups.Administration of the Vocabulary subtest SB:FE placed the mean scores for GroupII (87.55), and Group Ill (86.13) within one standard deviation of the mean. GroupI (83.06) mean was lower than the means for Groups II and Ill. T-tests did notindicate significant differences between groups. Results confirmed previousresearch (Goldstein, 1988; Leviton, 1980) which found expressive and receptiveDISCUSSION / 130language skills of Native children were well below average and age expectations.Low language skills of all subjects give support to the use of nonverbal measuresof intelligence. If verbal measures had been administered they would be ameasure of incidental language learning, not a measure of intelligence (Braden,1992). Low receptive vocabulary scores across groups gives added support topantomime instruction as many children may not comprehend the verbalinstructions.C. Hypothesis #2aIt is expected scores from nonverbal measures of intelligence administered withpantomime instructions to the hearing peer group will approximate scores whenmeasures are administered with verbal instructions for the hearing peer group.1) Comparison of Administration MethodsThere was no significant difference between the two instructional methods forsubjects in the hearing peer group (Group II and Ill). Results of the ANOVAindicated there were no significant differences between the two groups formeasures of nonverbal intelligence and subtests of the measures.DISCUSSION I 131Pantomime instruction gave subjects information necessary to complete the tasksrequired on subtests of the WISC-lll and the SB:FE. This provides evidence tosupport the use of pantomime instruction as an alternate form of testadministration.2) Group Comparisons with Pantomime AdministrationHypothesis #2bIt is expected nonverbal measures of intelligence administered with pantomimeinstructions to the otitis prone group will yield results that approximate those of thehearing peer group receMng pantomime instruction.Results substantiate subjects had sufficient information via pantomime instructionto complete subtest tasks on the WISC-IIl. The statistically significant differencebetween groups for the Abstract/Visual Reasoning Area score may be a result ofinsufficient information for subtest completion utilizing pantomime instruction.Additionally, only two subtests have been used to calculate the Abstract/VisualReasoning Area Score: Matrices and Pattern Analysis. Although these subtestsare non verbal or do not require a verbal response, they do require reasoning andproblem soMng skills, which are dependent on language (Gray, 1983).DISCUSSION I 132The limited number of subtests and assessment of different” cognitive skills couldprovide an explanation which could account for the differences between groups.Subtests of the SB:FE appear to assess skills which differ from those assessed onthe WISC-lll. For example, the WISC-lll does not have any one subtest that tapsvisual memory skills (Bead Memory), or a subtest that examines problem soMngskills as required for the Matrices subtest. Sattler (1988) states that the SB:FEand the WISC-lll are comparable but not equal.In addition, insufficient information and the different skills required to completesubtests, there may be other factors other that affect test outcomes. These mayinclude the subjects’ receptive and expressive language skills, family and schoolbackground, and that English was not the first language of many of the subjects.D. Hypothesis #3aThe WISC-Ill Performance IQ score and the SB:FE Partial Composite score will bepredictors of achievement for the otitis prone and hearing peer groups.There have been conflicting results presented in the research reviewed in ChapterII as to whether tests of intelligence can predict academic achievement for hearingimpaired and Native Indian children.DISCUSSION I 133The use of intelligence tests to explain or predict achievement levels is commonpractice among school psychologists (Phelps & Branyan, 1990), and the ability-achievement discrepancy is an accepted method of determining academicunderachievement. Watson et a). (1986) state intelligence scores can be used asa reference point when planning and identifying underachievement for hearingimpaired children. Others (Brooks & Riggs, 1980) have found the WISC-R to havevalidity for predicting reading achievement of hearing impaired children.The present study used regression analyses to indicate the “best” predictor ofachievement. Chronological age was found to be the best predictor of math andreading achievement. Achievement typically increases as a function of age or aschildren become older. Therefore, age as a predictor can be explained. However,it was not expected that chronological age would account for a large percentageof the variance in the majority of the regression equations.Besides chronological age as a predictor, regression analyses indicated tests orsubtests which made a contribution to the regression equation, or measures thatpredict achievement. The PPVT-R, a measure of receptive language, appears tobe one of the predictors for reading comprehension in Groups II and Ill (hearingpeer). The SBVOC, an expressive language measure, is one of the predictors ofreading for Group I (otitis prone). Expressive language and receptive language,DISCUSSION / 134both verbally loaded tests, were also predictors of reading ability.Language measures, predicting reading ability, could also be explained as orallanguage and understanding of language have been related to reading ability.Results could be explained in that nonverbal measures of intelligence are notlanguage based. There may not be a direct relationship between nonverbalintelligence and academic skills, specifically reading skills. Verbal intelligence islikely to be linked or related to reading skills, more than nonverbal skills, asmeasured by nonverbal intelligence.CA was then partialled out of the regression equation. The R squared values didnot change for Group Ill reading and math or for Group II reading. R squaredvalues did change for Group I reading and math and for Group II math. Thechanges indicate CA as an independent variable contributed to the explainedvariance. Entering CA as an independent variable did contribute to the R squaredvalues for Group I reading and math and for Group II math.Braden (1 985c) discussed the criterion-related validity of the WISC-R PIQ. Resultsof his investigation found the WISC-R PIQ did not predict academic achievement.DISCUSSION / 135Academic achievement may be an inappropriate criterion, as it is severelydepressed. The subjects in the current sample had extremely low achievementscores, thus a situation similar to that described by Braden (1 985c) existed in thepresent study.When predicting math computation skills, again chronological age was found to bea predictor for all three groups. The PPVT-R contributed to the regressionequation for Groups II and Ill. The WSCOD contributed to regression equationsfor Groups I and II. Results of the current study did not pro4de support to indicatenonverbal measures of intelligence should not be used to predict achievement forsubjects from the otitis prone or hearing peer groups.Measures of nonverbal intelligence may be indicators of potential, or a referencepoint; however, they were not predictive of achievement as measured by the SATHI. Rather than examining a global score, subtests from measures of intelligence,which could determine cognitive strengths and weaknesses, problem solving skillsand reasoning skills may provide information that is more concrete and usefulwhen planning and programming for Native Indian children.Nonverbal lQ scores may reflect different psychological processes than those usedin either reading or math abilities. Mild hearing loss and the subjects who haveDISCUSSION I 136English as a second language may have reduced incidental learning. Anotherpossibility is that the processes tapped by nonverba] lQ measures can not beapplied to academic tasks by children. It is important for those administeringnonverbal tests of intelligence to question how results can best be utilized in aschool setting, if tests results are not predictive of academic achievement.Fraser (1969) stated studies should explore areas of assessment beyond generalintelligence of Native Canadian children. Measures of intelligence can giveessential information, although other assessment options should be investigated.Alternate methods of assessment could give further information for programmingand learning styles of the Native child.E. Hypothesis #3bSpecific subtests of the WISC-lll Performance scales and subtests of the SB:FEPartial Composite score will be predictors of achievement for the otitis prone andhearing peer groups.Subtests of measures of nonverbal intelligence were not predictors of achievementfor the otitis prone or hearing peer group. Tasks required on various subtests ofmeasures of intelligence are not predictive of performance on measures of eithermath computation or reading comprehension.DISCUSSION! 137Subtests may not be predictive of math and reading, yet results provide valuableinformation regarding the students’s learning style, and strengths and weaknesses.When planning programs or designing strategies to specifically meet the child’sneeds, assessment results, in conjunction with additional assessment, could beutilized. This will be discussed in greater detail in this chapter.F. Hypothesis #4aIt is expected there will be a correlation between the WISC-lIl Performance lQscore and the SB:FE Partial Composite score for the hearing and otitis pronegroups.1. Nonverbal Measures of IntelligencePrevious research used correlations to examine the relationship between theWISC-R Full Scale lQ score and the Stanford Binet Composite score. Correlationsranging from .66 to .83 have been reported. Previous research results suggestscales are similar, but they are not interchangeable (SaWer, 1988).Measures of nonverbal intelligence may be assessing different skills and abilities,but they may measure elements that are similar (Salvia & Ysseldyke, 1991).DISCUSSION / 138Scores have not been investigated for the revised measures, the WISC-lll and theSB:FE or for index scores on the WISC-lll and area scores from the SB:FE.Research has not investigated the relationship or correlation between nonverbalmeasures of intelligence, or “portions of the instrument.NStatistically significant correlations were found between measures of nonverbalintelligence for all three groups in the present study. When the WISC-lll PIQ andthe Partial Composite score SB:FE were compared, Group I, r = 0.58 , Group II,r = 0.57, Group Ill, r = 0.65 correlations were not high but they were significantat the .001 level.Correlations were also examined for the PIQ with PC and PS for each of the threegroups. Logically, the PIQ, and P0/PS index scores would have high correlations.The P0 has identical subtest configuration as the PIQ, with the exception of theCoding subtest. This assumption held true for the PIQ/P0 correlations rangingfrom Group I, r = 0.96, Group II, r = 0.98 and Group Ill, r = 0.98. Correlationsbetween PIQ/PS were slightly lower with Group I, r = .51, Group II, r = .58 andGroup Ill, r =.79, (p<.001).Similar correlations were then investigated for the SB:FE composite score and thearea scores. Correlations between the composite score and the area scoresDISCUSSION / 139should be high as both use similar subtests. For Group I correlations betweenSBTOT and SBAVR/SBSTM were: r = 0.88, r = 0.87, Group II r = 0.90, r = 0.87,and Group Ill r = 0.87, r = 0.88, (p<.001). These correlations approximate thosein the WISC-lll manual (Weschler, 1991) and in the SB:FE Technical Manual(Thorndike, Hagen, & Sattler, 1986). Correlations reported for the standardizationsample in these manuals are similar to the results in the present study.G. Hypothesis #4bIt is expected there will be a correlation between nonverbal measures ofintelligence and achievement for the hearing peer and otitis prone groups.2. Non Verbal Intelligence and Achievement MeasuresCorrelations between intelligence and academic achievement for hearing sampleshas been reported as .50 (Jensen, 1980). Few studies have examined thecorrelation between academic achievement and nonverbal intelligence for hearingimpaired or Native children, however Watson et al. (1986) found correlationsbetween the WISC-R PIQ and the SAT-HI Reading Comprehension to be .36 andWISC-R PIQ and Math Computation .53.DISCUSSION / 140Brooks and Riggs (1980) report a correlation of .19 between the WISC-R PIQ andthe SAT-HI reading comprehension subtest for hearing impaired children with amoderate to severe hearing loss. Porter and Kirby (1986) report a correlation of.54 for reading and .68 for math on the Metropolitan Achievement Test, when theK-ABC Nonverbal scales were administered with pantomime instruction for hearingimpaired children with a severe to profound loss.Braden (1989) found the correlations between WISC-R and the SAT-HI low withnon significant values. This led to the assumption that the WISC-R PS lackedcriterion validity. Kelly and Braden (1990) used percentile scores from the SAT-HIand standard scores from the WISC-R PS for correlations. Correlations werehigher and statistically significant when percentile ranks were used which lead tothe recommendation of the use of percentiles for analysis.Percentile ranks were not used for the present analyses for several reasonspresented earlier. Use of percentiles may result in ambiguous results (Huebner,1989). Those scoring within the lower end of the scale, where intervals are spacedfurther apart would minimize results with differences near the centre of distributionexaggerating results (Huebner, 1989). Again the argument is presented that thissample is different from the subjects used in Kelly and Braden’s (1990) research.DISCUSSION / 141Mailer and Braden (in press) studied the relationship between the WISC-IlI PIQ andreading and math subtests from the SAT-HI. Subjects in this study wereprofoundly deaf, and a small sample size was noted (n=30). Correlations reportedwere .46 between reading and PIQ, and .63 between math and PIQ. Wechsler(1991) reports correlations for the WISC-lII PIQ and the Wide Range AchievementTest-Revised to be .43 for reading and .58 for math. It appears correlationsbetween intelligence and achievement are comparable for deaf and hearingsubjects.Mailer and Braden (in press) also administered the Verbal subtests from the WISCIll. They found the Verbal IQ Score was a better predictor of achievement thanthe Performance IQ for deaf children. This finding was consistent for reading andlanguage measures but not for math tests. PIQ and VIQ were not different as bothwere predictive of math achievement. This was explained by Mailer and Braden(in press) as indicating less language and incidental learning are required for theacquisition of math skills. This does not really provide an explanation as to whyVIQ would predict math achievement.Research with Native children indicates low correlations between achievement testsand the WISC-R. The strongest relationship found in the McShane and Plas (1984)study was between reading and language based measures, (i.e. between readingDISCUSSION / 142and Verbal IQ scores). Correlations were not found between the WISC-R PIQ andachievement measures were not evident in the McShane and Plas (1984) study.The present study examined correlations between achievement scores and thenonverbal measures of intelligence or subtests of the WISC-IlI and SB:FE.Correlations for Group I indicated a significant relationship between reading andthe PIQ (r=.43). The Picture Arrangement subtests from the WISC-lll correlatedwith reading, (r= .42) and was significant at the .001 level. Sequencing of picturesto indicate a series of events are related to reading achievement and wereexpected results.Correlations between math computation and measures of nonverbal intelligencewere examined. A significant correlation was not found between math and PIQ.However, there was a significant correlation between the Coding subtest and Math,(r= .54). Research with hearing subjects indicates a relationship between math andthe Coding subtest from the WISC-R (Sattler, 1988).Achievement scores for the otitis prone group (Group I) and hearing peers(Groups II and lii) are low overall. A correlation was not expected between readingcomprehension and math computation (r= .77) for the otitis prone group.DISCUSSION I 143The correlation between math and reading was not expected after the examinationof previous research with Native Indian and hearing impaired children. Previousresearch indicated math skills were higher than reading skills for both thesegroups.St. John & Krichev (1976) found Native children and hearing impaired children toperform consistently higher on math achievement than on measures of readingcomprehension. Low language levels as a result of an ESL background and otitismedia could effect all areas of achievement. The mean scores for Group I (theSBVOC x=83.1 and PPVTR x=75.85) are well below average scores. Low scoreson the expressive language measure (SBVOC) correlates with both math andreading scores.Results of Group II are essentially the same as reported for Group I. PictureCompletion (r= .39) and Picture Arrangement (r= .41) demonstrated significantcorrelations with the SAT-HI reading comprehension subtest. The SBVOC (r= .42)and PPVTR (r= .60) also demonstrated significant correlations. These results wereexpected. Significant correlations between the SAT-HI math computation and anysubtest of the WISC-lll or SB:FE were not found.The PPVTR was administered as a measure of receptive language skills and wasDISCUSSION I 144the only measure that demonstrated a significant correlation with the SAT-HI mathcomputation subtest (r= .42, p < .001). This was not expected from the researchreviewed, but can be explained if low language levels are taken into account.Receptive language skills, or the understanding of language concepts is requiredfor completion of the PPVTR. These skills, especially the understanding ofconcepts may also be necessary for math skill development.Group III results demonstrate significant correlations between the WISC-III PIQ andreading comprehension (r=.55) and math computation (r=.40). The PictureCompletion subtest (r=.60), Coding (r=.51), and Picture Arrangement (r.45),when p < .001, were correlated with reading comprehension. These results aresimilar to the results reported for Groups I and II. Language measures alsodemonstrated correlations with reading comprehension, PPVTR (r=.56), SBVOC(r=.61), and with math computation PPVTR (r=.43). Reading comprehension andmath computation were correlated for Group Ill (r= .67) as they had been forGroups I (r=.77) and II (r=.83).In addition to language skills scoring as well below average, short term visualmemory skills were very low when compared to the standardization sample. Asstated previously language skills are necessary in order to complete tasks whichDISCUSSION / 145are nonverbal or visual. Use of internal language (Conrad, 1979) to label thosevisual images to aid recall, may then in fact be linked to a subjects’ level oflanguage ability. Low language skills may directly impact performance of verbaltasks (Zinkus and Gottlieb, 1980). Children who are able to complete visual shortterm memory tasks with success typically rely on a strategy which utilizes theirlanguage skills (Ellis & Deshler, 1991).H. Influence of Otitis MediaThe presence or absence of otitis media has not been considered an importantvariable in the assessment of Native Indian children. Otitis media is a fluctuatingcondition making it very difficult to ascertain how it affects acquisition of language,speech, academic, and cognitive skills. Evidence has been presented to supportotitis media as a variable which warrants attention in the assessment process(Scaldwell, 1989). Expressive and receptive language levels are typicallydepressed, and there is evidence to suggest certain cognitive skills are delayed(Jensen, 1991).McShane (1982) concluded that because Indian groups experience both otitismedia and academic difficulties in proportions which greatly exceed those foundwithin the non Indian population, it is important to investigate the relationshipDISCUSSION /146between otitis media and academic achievement. Results from this study wouldindicate that the otitis prone and the hearing peer groups had achievement scoreswhich were well below the scores reported for hearing children. When norms wereused for children with mild to moderate loss (Holt, Traxier, & Allen, 1992), subjectsin the present study scored below the reported standard scores. The controlgroup also scored well below the mean.Nonverbal tests of intelligence are administered to Native Indian children and itappears receptive language should also be considered. In the present study, theotitis prone group, as well as the hearing peers scored well below the mean, forboth measures of expressive and receptive language ( PPVT-R/SBVOC, Group I75.8/83.1, Group Il 77.6/87.6, Group III 76.3/86.1).Results of the ANOVA indicated a significant difference between Groups I and Il/Illfor expressive language skills as measured by the Vocabulary subtest of theSB:FE. Receptive language or the understanding of language or oral directionswas difficult for many children in the present study. ANOVA results indicated therewas not a significant difference between groups for the PPVT-R, although meanscores were well below average, or approximately two standard deviations belowthe mean.DISCUSSION I 147In addition to low expressive and receptive language levels, the majority of thesubjects in the otitis prone and hearing groups were English as a Secondlanguage (ESL) children. The combination of low language levels and ESLprovides a rationale as to why English oral directions may be an ineffective methodof test administration to this population. If oral directions to nonverbal measuresare not an option for this population, then a standard method of test administrationrequires investigation. This was one of the objectives of the present study.There was not a difference between groups when scores were examined from theWISC-lll and SB:FE. Although differences were not evident in the descriptivestatistics, t-tests or ANOVA’s, it was felt that examining these “group” scores andindicating otitis media was not a factor may be inaccurate. lndMdual differencesof the subjects should be taken into account, not forgetting the characteristics ofthe children in the sample, the nature of otitis media as a fluctuating condition.I. Pantomime InstructionThe need for valid assessment of intelligence or intellectual capacity of the hearingimpaired population has been recognized and researched (Phelps & Ensor, 1986)with few conclusive results. A review of the literature in assessment of NativeIndian and hearing impaired children revealed similar difficulties in testDISCUSSION I 148administration.Pantomime instruction allowed a standard or consistent method of testadministration which can be used with all children regardless of hearing loss andlanguage ability. This alternative should be considered if measures of intelligencecontinue to be administered. The present study investigated verbal andpantomime administration methods. Subjects from the hearing peer group wererandomly assigned to either verbal or pantomime instruction. Resultsdemonstrated pantomime instruction did give subjects the information required tocomplete subtest tasks of the WISC-lIl and SB:FE.In addition to the t-tests and ANOVAs that were completed to answer thehypotheses, item analysis provided additional information regarding theperformance of subjects when the Picture Completion (PC) subtest wasadministered using verbal and pantomime administration. Results indicated 10 ofthe 30 items on the PC subtest were significant (p < .05) for both groups. T-testswere then completed to analyze whether there were difference between groupswhen only the 10 responses were considered.There was not a significant difference between the groups using verbal andpantomime instruction for this subtest of the WISC-Ill giving additional support forDISCUSSION I 149the use of pantomime instruction as a method of administration. Results of theitem analysis were exploratory, and certainly further investigation using itemanalysis with a larger sample size, would be recommended to support use of analternate form of administration such as pantomime instruction. This is very difficultin research with low incidence exceptionalities such hearing impairment and in thiscase Native Indian children.The majority of research investigating test administration has involved deaf childrenfrom residential schools. Subjects have typically used a manual form ofcommunication. Sullivan (1982) reported lower PIQ scores on measures ofintelligence when pantomime instruction was utilized. Results of the present studydid not confirm these results as there was not a significant difference betweenverbal and pantomime administration.It should be noted that previous investigations focused on children with severe toprofound hearing loss. The otitis prone group had mild conductive loss andwere Native Indian. The subjects in the study may have responded differently topantomime instruction than the group with a severe to profound loss.Braden (1992), in a meta analysis of research on assessment of deaf and hard ofhearing persons, found the most popular method of test administration was signDISCUSSION / 150accompanied with speech, followed by oral and pantomime/gesturaladministration. A significant finding of the analysis was that more than half of thestudies failed to describe test administration procedures. In studies whereadministration procedures have not been reported, replication would be verydifficult. Additionally, results should be questioned if standard procedures have notbeen followed.Consistent instruction to a group of diverse subjects is possible via pantomimeinstruction. Pantomime instruction does give the examinee sufficient informationto complete tasks on measures of nonverbal intelligence.J. Limitations of the Study1. Achievement MeasuresThe SAT-HI was chosen for this study for several reasons. Children in theexperimental and control groups were thought to be language impaired, and themajority of the otitis prone subjects (87%) had a mild loss. The SAT-HI is the onlyachievement measure to offer norms for hearing and hearing impaired children.Norms are available for deaf children with severe to profound loss. In addition,DISCUSSION / 151norms for children with a moderate loss (Holt, Traxler, & Allen, 1991) are alsoavailable. Availability of norms for children with a mild to moderate loss was afactor that was considered when the SAT-HI was chosen as the measure ofachievement.Previous research demonstrated math computation and reading comprehensionlevels for Native children are not commensurate (Hynd & Garcia, 1979). When theSAT was standardized for the deaf population, screening tests for mathcomputation and reading comprehension were developed to ensure theappropriate level of the test was administered to subjects in various content areas.To ensure that the level of the test to be administered was reliable, it wasimperative screening tests were administered prior to assignment to a level of thetest. St. John and Krichev (1976) found subjects’ math computation and readingcomprehension abilities to differ, with math skills being superior to reading skills forNative Indian populations.The SAT-HI can be administered as a group measure, which for the purposes ofthe current study (N = 167) was desirable, rather than the administration ofindividual measures of reading comprehension and math computation. The choiceof the SAT-HI reading comprehension and math computation appeared to beDISCUSSION/ 152appropriate, in theory.However, when children were given either the math computation or readingcomprehension subtest to complete independently, they required a great deal ofdirection and constant reinforcement to remain on task. The number of questionson the primary and intermediate levels of the tests were numerous with the readingcomprehension subtest often requiring more than 45 minutes to complete. Thetime required to complete the subtests caused the examiners to question whetherthe length of the reading comprehension tests at the intermediate levels may haveaffected the subjects’ scores and whether they were accurate estimates of readingcomprehension.If subjects were unable to complete several of the initial items, one questionswhether they were indeed “reading” towards the end of the twenty-page protocol.An indMdual measure of achievement would have provided similar informationregarding levels of reading comprehension and math computation, and theexaminer would have had additional control over administration, and could havereported individual differences among examinees.It is important that individual achievement measures are considered for futureresearch. Although the SAT is normed for deaf and hard of hearing students,DISCUSSION / 153individual achievement measures may give additional information as to how a childapproaches a reading or math task.Curriculum based assessment (CBA) may be an alternative which would giveteachers information which is pertinent to their students. Standardized measuresof achievement enable the examiner to evaluate the child by comparingperformance to children of the same age, or grade. The SAT-HI enabled theexaminers to examine students’ performance in relationship to other hearingimpaired students of the same age.Alternatives to standardized assessment require further investigation andconsideration. Qualitative assessment techniques such as CBA could assistteachers and the school based team to determine individual achievement and howstudent’s performance relates to class peers (SaMa & Yssledyke, 1991).Formal assessment may not give information as to what is being learned on a dayto-day basis. Alternate forms of assessment could be used to ascertain whetherstudents have learned and synthesized information. CBA is one alternative whichcould provide information to the teacher which could have a direct effect inprogramme planning, placement of children in instructional groups, and decisionmaking.DISCUSSION / 154Assessment can be viewed as a process, with many dimensions, but with onemain purpose, that is, to give further information about the child. Formalassessment will continue to play a role in the process, as will informal measures.New models (Naglieri & Das, 1988) have been suggested that will examine theprocesses children utilize when given specific tasks. All of these options shouldbe available to those involved in the assessment process. Methods ofadministration to minority groups will continue to require attention if measures areto give reliable and valid information.2. SampleFindings of the present study are limited to the sample, Native Indian children withchronic otitis media and their hearing peers. Results should not be generalized todeaf children or to other Native Indian or minority groups.3. Item AnalysisThis analysis was exploratory in nature and results were likely inflated due to thesmall sample size. It is. recommended item analysis, utilizing a larger sample,would give further information as to the use of pantomime instruction as a viablemethod to administer tests of nonverbal intelligence to children that have limitedDISCUSSION / 155language skills.K. Implications of the Study and Future ResearchMailer and Braden (in press) examined the construct validity of the WISC-llI.Correlations between nonverbal intelligence and achievement were moderate andonly significant for reading and math. Results of the present study did not supportthe use of nonverbal measures of intelligence as “reference points” or indicatorsof achievement for Group I (otitis prone) or Groups II and Ill (hearing peer groups).Regression analyses did not support nonverbal measures of intelligence aspredictors of either reading comprehension or math computation for Groups I- Ill.Correlation between achievement and ability is approximately .50 (Jensen, 1980)in studies with hearing subjects. Results of the current study did find significantcorrelations between intelligence and reading comprehension and mathcomputation, although results were not consistent across Groups I - Ill.Correlations were significant, although, they were lower than correlations reportedfor hearing subjects (Jensen, 1980).Therefore, it is imperative that educators administering nonverbal measures ofintelligence be clear as to WHY achievement measures and tests of intelligence areDISCUSSION/156administered. Results of the present study found language and short termmemory skills to be very low for all three groups. It was expected language skillswould be below average, however, low scores in visual short term memory skillswere not expected.These findings give support as to HOW assessment results can provideinformation which can be valuable in program planning, and choosing appropriateteaching strategies. Results in the present study illustrate this point. Although lowlanguage levels were expected, short term memory was not an area of weaknessthat was expected. Understanding a child’s strengths and weaknesses is valuablewhen attempting to find instructional strategies which will be effective for the learner(Ellis & Deshler, 1991).Long term differential educational outcomes related to the administration ofstandardized measures, such as the WISC-lll, SB:FE and SAT-HI, requires furtherinvestigation and research. If results can be used to determine strengths andweaknesses in a child’s profile, then the assessment process can provide valuableinformation to the teacher and resource personnel involved in providingappropriate programming to children.Results of the study provides evidence to indicate that pantomime instruction couldDISCUSSION/157be an instructional methodology used with otitis prone and hearing Native children.Further research is necessary to determine if standard pantomime instruction canbe utilized to administer nonverbal measures of intelligence to other groups ofchildren with limited or delayed language skills.Placement may not be an issue as the majority of students with special needsreceive service from their neighbourhood or home schools. Appropriate serviceto a child with special needs continues to be an issue in special education. It isinthe delivery of appropriate service that assessment results will continue to providevaluable information./ 158REFERENCESANen, T.E., White, C.S., & Karchmer, M.A. (1983). Issues in the developmentof a special edition for hearing impaired students of the 7th edition of theSAT. American Annals of the Deaf, 128, 34-39.Anastasi, A. (1988). Psychological testing: Sixth Edition. New York: CollierMacmillan Publishing Company.Anderson, R.J., & Sisco, F.H. (1977). Standardization of the WISC-RPerformance Scale for deaf children. (Office of Demographic StudiesPublication Series Y, No.1). Washington, D.C.:Gallaudet College.Bachor, D., & Crealock, C. (1988). Instructional strategies for students withspecial needs. Toronto: Prentice-Hall.Balow, T.A., & Brill, R.G. (1975). An evaluation of reading and academicachievement levels of 16 graduating classes of the California School forthe Deaf, Riverside. The Volta Review, 77, 255-266.Beery, Q., Doyle, W., Cantekin, E.,Bluestone, C., & Wilt, R. (1980). Eustachiantube function in an American Indian population. The Annals of Otology,Rhinology and Laryngology, 89, supp. 68, 28-33.Birch, J.R., & Birch, J.W. (1956). Predicting school achievement in young deafchildren. American Annals of the Deaf, 101, 348-352.Bluestone, C.D., & Klein, J.O.(1983). Otitis media with effusion, atelectasis, andeustachian tube dysfunction. In C.D. Bluestone and S.E. Stool (Eds.),Paediatric Otolaryngology. pp. 356-512, Phi!adelphia:W.B.Saunders.Braden, J.P. (1984). The factorial similarity of the WISC-R Performance Scalein deaf and hearing children. Journal of Personality and IndividualDifferences, 4, 403-410.Braden, J.P. (1985a). WISC-R deaf norms reconsidered. Journal of SchoolPsychology, 23, 375-382.Braden, J.P. (1985b). The structure of nonverbal intelligence in deaf andhearing subjects. American Annals of the Deaf, 131, 496-501.REFERENCES / 159Braden, J.P. (1985c). The criterion validity of the WISC-R performance scale ndther nonverbal tests for deaf children. American Annals of the Deaf,329-332.Braden, J.P. (1985d). Futile gestures: A reply to Courtney et al. regardingpantomimed administration of the WISC-R. Journal ofPsychoeducational Assessment, 3, 375-382.Braden, J.P. (1989). Deafness as a natural experiment: A meta-analytic reviewof IQ research. In D.S. Martin (Ed.), Second International Symposiumon Cognition, Education and Deafness: Working Papers (Vol. 1) (pp.42-60). Washington, DC:Gallaudet University.Braden, J.P. (1992). The assessment of intelligence in deaf and hard of hearingpeople:Quantitative and qualitative research analysis. SchoolPsychology Review, 21, 82-94.Bragman, R. (1982). Review on research on test instruction for deaf children.American Annals of the Deaf, 127, 337-346.Brill, R.C. (1962). The relationship of the Wechsler l.Q.’s to academicachievement among deaf students. Journal of Exceptional Children,28, 315-321.Brody, J.A. Overfield, T., & McAlister, R. (1965). Draining ears and deafnessamong Alaskan Eskimos. Archives of Otolaryngology, 81, 29-33.Brooks, C.R., & Riggs, S.T. (1980). WISC-R, WISC and reading achievementrelationship among hearing impaired children attending public schools.Volta Review, 82, 96-102.Bryde, J.F. (1968). A new approach to Indian education. IntegratedEducation, September-October.Chrisjohn, R., & Lanigan, C. (1986). Research on Indian intelligence testing:Review and prospects. In selected papers from The First MOKAKITConference, (1984) (pp. 50-57), MOKAKIT Indian Education ResearchAssociation, University of British Columbia.Chrisjohn, R.D., Towson, S.M.J., Pace, D.F., & Peters, M. (1989). The WISC-Rin Native application:lnternal and external analysis. In Beery et al. (Eds.),Proceedings of the First regional North American Conference of theInternational Association for Cross Cultural Psychology.REFERENCES I 160Common, R.W., & Frost, L.G. (1988). The implications of the mismeasurementof Native students’ intelligence through the use of standardizedintelligence tests. Canadian Journal of Native Education, 15, 18-30.Collyer, C.E., & Enns, J.T. (1987). Analysis of variance, the basic designs.Chicago:Nelson-Hall Publishers.Conrad, R. (1979). The deaf school child:Language and cognitive function.London:Harper & Row.Courtney., Hayes, F.B., Couch, K.W., & Frick, M. (1984). Administration of theWISC-R Performance Scale to hearing impaired children using pantomimeinstruction. Journal of Psychoeducational Assessment, 2, 1-7.Crocker, L., & Algina, J. (1986). Introduction to classical and modern testtheory. New York:Holt Rinehart and Winston.Davis, J.M., Elfenbein, J. L., Schum, R. L., & Bentler, R.A. (1986). Effects ofmild and moderate hearing impairment on language, educational, andpsychosocial behaviour of children. Journal of Speech and HearingDisorders, 44, 73-79.Davis, J., Shepard, N.T., Stelmachowicz, P.G., & Gorga, M.P. (1981).Characteristics of hearing impaired children in the public schools: PartIl:Psychoeducational data. Journal of Speech and Hearing Disorders,46, 130-137.Delaney, E.A., & Hopkins, T.F. (1987). Examiner’s Handbook:An expandedguide for fourth edition users. Chicago:Riverside Publishing Company.Dobie, R.A., & Berlin, C.l. (1979). Influence on otitis media on hearing anddevelopment. Annals of Otology, Rhinology, and Laryngology, 88(Supp. 60), 48-53.Downs, M. (1977). Effects of mild hearing loss on auditory processing.Otolaryngologic Clinics of North America, 18, 337-343.Downs, M., & Bergman, F. (1982). Otitis prone children. Journal ofDevelopmental and Behavioral Paediatrics, 3, 106-112.REFERENCES / 161Ellis, E.S., & Deshler, D.D. (1991). An analysis of the critical features of effectivelearning strategies. Focus on Learning Disabilities, 2, 94-107.Everett, F., Proctor, N., & Cartmell, B. (1983). Providing psychological servicesto American Indian children and families. Professional Psychology:Practice and Research,14, 588-603.Feagan,S.L., Sanyal, M., Henerson, F., Collier, A., & Applebaum, M. (1987).Middle ear disease in early childhood and later language skill. Journalof Paediatric Psychology, 12, 581-594.Fraser, S. (1969). Indian language proficiency assessment: Proposedprocedures for Saskatchewan schools. In F. Ahenekew and S. Fredeen(Eds.), Proceedings of the Seventh Annual Native AmericanLanguage Issues. Saskatoon: Indian Languages Institute.Friel-Patti, S. (1990). Otitis media with effusion and the development oflanguage: A review of the evidence. Topics in Language Disorders, 11,11-22.Gerweck, S., & Yssledyke, J. (1975). Limitations of current psychologicalpractices for intellectual assessment of the hearing impaired : A responseto the Levine study. Volta Review, 77, 243-248.Glass, G.V., & Hopkins, K.D. (1984). Statistical methods in education andpsychology. New Jersey:Prentice-Hall, Inc.Goetzinger, M.R., & Houchins, R.R. (1969). The 1947 coloured Raven’s Matriceswith deaf and hearing subjects. American Annals of the Deaf, 114,95-101.Goldstein, C.R. (1988). Cautions concerning assessment of Native Indianchildren. Canadian Journal of School Psychology, 4, 17-20.Graham, M.D. (1989). Prevalence of middle ear disease among the Indianpopulation of coastal B.C. Hearing Instruments, 26, 26.Graham, E.E., & Shapiro, E. (1953). Use of the Performance Scale of the WISCwith deaf children. Journal of Consulting Psychology, 17, 396-398.Gray, C.P. (1983). Effects of small perceptual losses on language and onspeech discrimination. Volta Review, 84, 408-412.REFERENCES / 162Hayes, F.B., & Courtney, A.S. (1985). Commentary in response to Braden.Journal of Psychoeducational Assessment, 3, 187-192.Hime, W.D. (1970). The abilities of partially hearing children. British Journalof Educational Psychology, 40, 1971-1978.Hirshoren, A., Kavale, K., Hurley, O.L., & Hunt, J.T. (1977). The reliability of theWISC-R Performance Scale with the deaf. Psychology in the Schools,14, 412-415.Hirshoren, A., Hurley, O.L., & Kavale, K. (1979). Psychometric characteristicsof the WISC-R Performance Scale with deaf children. Journal of Speechand Hearing Disorders, 44, 73-79.Hiskey, M.S. (1966). Hiskey-Nebraska Test of Learning Aptitude. NB:CollegeView Printers.Hoim, V. A., & Kunze, L.H. (1969). Effects of chronic otitis media on languageand speech development. Paediatrics, 43, 833-839.Holt, J.A., Traxler, C.B., & Allen, T.E. (1992). Interpreting the scores:A user’sguide to the 8th edition Stanford Achievement Test for educators ofdeaf and hard of hearing students. Center for Assessment andDemographic Studies, Washington:Gallaudet Research instituteHorowitz, E.D., & Leake, H. (1979). Effects of otitis media on cognitivedevelopment. The Annals of Otology, Rhinology, and Laryngology,Supplemental 68, 89, 264-268.Huebner, E.S. (1989). Errors in decision makiing:a comparison of schoolpsychologists’ interpretations of grade equivalents, percentiles anddeviation lQs. School Psychology Review, 18, 51-55.Hunter, M.W. (1992). Comparison of the Peabody Picture Vocabulary Test-Revised and the Stanford-Binet intelligence scale, fourth edition withelementary students referred for learning problems. Diagnostique, 17,108-114.Hurley, O.L., Hirshoren, A., Kavale, K., & Hunt, J.T. (1979). Predictive validityof two mental ability tests with black deaf children. Journal of NegroEducation, 48, 14-19.REFERENCES / 163Hynd, G.W., & Garcia, W.I. (1979). Intellectual assessment of the NativeAmerican student. School Psychology Digest, 8, 446-454.Irvine, S.H., & Saunders, J.T. (1972). Logic, language and method in constructidentification across cultures. In L. Cronbach & P. Drenth (Eds.), Mentaltests and cultural adaptation. The Hague: Mouton.Jackson, J. (1984). Psychoeducational testing:Turn out the lights the party’sover. Exceptional Children, 52, 288-299.Jenesma, C. (1978). The relationship between achievement and thedemographic characteristics of hearing impaired children and youth.Series R. Number 2. Washington, D.C.:Gallaudet College, Office ofDemographic Studies.Jensen, A.R. (1980). Bias in mental testing. New York:Free Press.Jensen, A.R. (1991). General mental ability:from psychometrics to biology.Diagnostique, 16, 134-144.Kaplan, G.J., Fleshman, J.K., Bender, T.R., Baum, C., & Clark, D.S. (1973).Long term effects of otitis media: A ten year cohort study of AlaskanEskimo children. Paediatrics, 52, 577-585.Kaufman, A.S. (1982). Verbal-Performance IQ discrepancies on the WISC-R.Journal of Consulting and Clinical Psychology, 44, 739-744.Kelly, M.D., & Braden, J.P. (1990). Criterion related validity of the WISC-RPerformance Scale with the Stanford Achievement Test-Hearing Impairededition. Journal of School Psychology, 28, 147-151.Klein, J., Chase, C., Teele, D., Menyuk, P., & Rosner, B. (1988). Otitis mediaand the development of speech, language and cognitive abilities at sevenyears of age. In D. Lim, C. Bluestone, J.Klein & J. Nelson (Eds.),Recent Advances in Otitis Media, Philadelphia, Penn. : Decker.Kyle, J.G. (1980). Measuring the intelligence of deaf children. Bulletin of theBritish Psychological Society, 33, 54-57.Levine, E.S. (1971). Mental assessment of the deaf child. The Volta Review,33, 80-105.REFERENCES / 164Leviton, A. (1980). Otitis media and learning disorders. Developmental andBehavioral Paediatrics, 1, 58-63.Ling, D., McCoy, R.J., & Levinson, E.D. (1969). The incidence of middle eardisease and its educational implications among Baffin Island Eskimochildren. Canadian Journal of Public Health, 60, 385-390.Mailer, S.l., & Braden, J.P. (in press). The construct and criterion validity of theWISC-IlI.Menyuk, P. (1980). Effect of persistent otitis media on language development.Annals of Otology, Rhinology, and Laryngology, 89, supp. 68.Menyuk, P. (1986). Predicting speech and language problems with persistentotitis media. in J. Kavanagh (Ed.), Otitis Media and ChildDevelopment. Parkton, Md : York.Miller, H.J. (1968). The effects of integration of rural Indian pupils.Unpublished doctoral dissertation, University of North Dakota.Mindel, E., & Vernon, M. (1971). They grow in silence. Silver Spring,Maryland:National Association of the Deaf.Murphy, K.P. (1957). Test of abilities and attainments. In A.W.G. Ewing (Ed.),Educational guidance and the deaf child, Manchester University Press.Mustain, W.D. (1979). Linguistic and educational implications of recurrent otitismedia. Ear Nose and Throat Journal, 58, 62-67.Myklebust, H.R. (1960). The psychology of deafness. New York:Grune &Stratton.MacArthur, R.S. (1968). Assessing intellectual potential of Native Canadianpupils. Alberta Journal of Educational Research, 14, 115-122.MacArthur, R.S. (1969). Some cognitive abilities of Eskimo, White and IndianMetis pupils age 9-12. Canadian Journal of Behavioral Science, 50-59.McQuaid, M.F., & Alovisetti, M. (1981). School psychological services forhearing impaired children in New York and New England. AmericanREFERENCES I 165Annals of the Deaf, 126, 37-42.McShane, D. (1982). Otitis media and American Indians : Prevalence, etiology,psychoeducational consequences, prevention and intervention. In S.Manson (Ed.), New directions in prevention among American Indianchildren and Alaska Native communities. Portland Oregon : OregonHealth Services University.McShane., & Mitchell, J. (1979). Middle ear disease, hearing loss, andeducational problems of American Indian children. Journal of AmericanIndian Education, 19,7.McShane, D., & Plas, J.M. (1984). The relationship of intellectualpsycholinguistic abilities to the achievement gains of American Indianchildren. Journal of Preventative Psychiatry, 1, 66-79.Naglieri, J.A., & Das, J.P. (1988). Planning-arousal-simultaneous-successive(PASS):A model for assessment. Journal of School Psychology, 26,35-48.Needleman, H. (1977). Effects of hearing loss from early recurrent otitis mediaon speech and language development. In B.F. Jaffee(Ed.), Hearing lossin children, Baltimore, MD :University Park Press.Neuhaus, M. (1967). Modifications in the administration of the WISC-RPerformance subtest for children with profound hearing losses.Exceptional Children, 33, 573-574.Nober, E., & Nober, L. (1977). Effects of hearing loss on speech and languagein the post babbling stage. In B. Jaffe (Ed.), Hearing loss in children.Baltimore:University Park Press.Northern, J. L., & Downs, M. P. (1978). Hearing in children (2nd Ed.).Baltimore:Williams & Williams.Nozza, R. (1988). Auditory deficit in infants with otitis media infusion: Morethan a “miId hearing loss. In D. Lim, C. Bluestone, J. Klein, & J. Nelson(Eds.), Recent advances in otitis media with effusion: Proceedingsof the fourth international symposium. Philadelphia, Penn : Decker.Paradise, J.(1981). Otitis media during early life: How hazardous todevelopment? A critical review of evidence. Paediatrics, 68, 869-871.REFERENCES / 166Paradise, J., Kaplan, G.L., & Bender, T.R. (1976). Management of middle earinfections in infants. Annals of Otolarnagology, Rhinology, andLaryngology, 85, 285-289.Perley-McField, J. (1990). A comparison of WISC-R and SB:FE scores withthe SAT-HI for a group of residential deaf children. Unpublishedmasters thesis, University of British Columbia, Vancouver, B.C.Phelps, L., & Branyan, B.J. (1988). Correlations among the Hiskey, K-ABCNonverbal Scale, Leiter, and WISC-R Performance scale with publicschool deaf children. Journal of Psychoeducational Assessment, 6,354-358.Phelps, L., & Branyan, B.J. (1990). Academic achievement and nonverbalintelligence in public school hearing impaired children. Psychology inthe Schools, 27, 21 0-21 7.Phelps, L., & Ensor, A. (1986). Concurrent validity of the WISC-R using deafnorms and the Hiskey Nebraska. Psychology in the Schools, 23,138-141.Pintner, R. (1924). Results obtained with the nonlanguage group tests.Journal of Educational Psychology, 15, 473-483.Pintner, R., & Patterson, D. (1915). The Binet scale and the deaf child. Journalof Educational Psychology, 6, 201-210.Pollard, G., & Oakland, T. (1985). Variables associated with the educationaldevelopment of residential deaf children. Special Services in theSchools, 1, 67-82.Porter, L.J., & Kirby, E.A. (1986). Effects of two instructional sets on the validityof the Kaufman Assessment Battery for Children-Nonverbal Scale with agroup of severely hearing impaired children. Psychology in theSchools, 23, 37-43.Ray, S. (1982). Adapting the WISC-R for deaf children. Diagnostique, 7,147-157.Reed, M. (1970). Deaf and partially hearing children. In P. Mittler (Ed.), Thepsychological assessment of mental and physical handicaps (pp.403-441). London: Methuen.REFERENCES 1167Reynolds, C.R. (1981). The fallacy of “two years below grade level for age” asa diagnostic criterion for reading disorders. Journal of SchoolPsychology, 19, 350-358.Reynolds, H.N. (1978). Correlates of reading achievement in deaf children. InC. Williams (Ed.), Proceedings of the Gallaudet Conference onReading in Relation to Deafness. Washington, D.C:Gallaudet CollegePress.Roberts, J., Sanyal, M., Burchinal, M., Collier, A., Ramey, C., & Henderson, F.(1986). Otitis media in early childhood and its relationship to later verbaland academic performance. Paediatrics, 78, 423-430.Roberts, J.E., & Schuele, C.M. (1990). Otitis media and later academicperformance:The linkage and implications for interventions. Topics inLanguage Disorders, 11, 43-62.SaMa, J., & Yssledyke, J. (1991). Assessment in special and remedialeducation, (3rd ed.), Boston: Houghton Mifflin Company.Sattler, J. M. (1988). Assessment of children, (3rd ed.), San Diego, CA:Jerome M. Sattler.Scaldwell, W.A. (1989). Effects of otitis media upon reading scores of Indianchildren in Ontario. Journal of American Indian Education, 28, 36-43.Scaldwell, W. A., & Frame, J.E. (1985a). Prevalence of otitis media in Cree andOjibway school children in six Ontario communities. Journal ofAmerican Indian Education, 25, 1-5.Scaldwell, W.A., & Frame, J.E. (1985b). WESDIAND project report. Toronto,Ontario: Indian and Northern Affairs, Canada.Scaidwell, W., Frame, J.E., & Cookson, D. (1984). IndMdual intellectualassessment of Chippewa, Mucey and Omeida children using the WISC-R.Canadian Journal of School Psychology, 1, 15-21.Seyfort, B., Spreen, 0., & Lahmer, V. (1980). A critical look at the WISC-R withNative children. The Alberta Journal of Educational Research, 26,14-24.REFERENCES /168Shavelson, R.J. (1988). Statistical reasoning for the behavioral sciences.University of California:Santa Barbara.Shurin, P.A., Pelton, S.I., Donner, A., & Klein, JO. (1979). Persistence of middleear effusion after acute otitis media. New England Journal of Medicine,300, 1121.Sisco, F.H., & Anderson, R. J. (1978). Current findings regarding theperformance of deaf children on the WISC-R. American Annals of theDeaf, 123, 115-121.Skinner, P.N. (1988). The relationship of non verbal intelligence measures toacademic achievement among deaf adolescents. Journal ofRehabilitation of the Deaf, 21, 8-11.Smith, M.D. (1974). Recategorized WISC-R scores of disabled readers.Journal of Learning Disabilities, 10, 48-53.Sternberg, R.J. (1984). Intelligence applied: Understanding and increasingyour intellectual skills. San Diego :Harcourt Brace Jovanowich.St. John, J., & Krichev, A. (1976). Northwestern Ontario Indian children and theWISC. Psychology in the Schools, 13, 4-16.Sullivan, P.M. (1982). Administrative modifications on the WISC-R Performancescale with different categories of deaf children. American Annals of theDeaf, 127, 780-788.Sullivan, P.M., & Schulte, L.E. (1992). Factor analysis of the WISC-R with deafand hard-of-hearing children. Psychological Assessment, 4, 537-540.Sullivan, P. M., & Vernon, M. (1979). Psychological assessment of hearingimpaired children. School Psychology Digest, 8, 271-290.Tall, P. (1991). Auditory temporal perception, phonics and reading disabilitiesin children. Brain and Language, 9, 182-198.Thorndike, R. L., Hagen, E.P., & Sattler, J.M. (1986). Technical manual,Stanford Binet Intelligence Scale: Fourth Edition. Chicago: RiversidePublishing.REFERENCES / 169Tomusiak, R. (1983). Summary report of the northern Saskatchewannorms project (Report No. 96). Regina, Saskatchewan: SaskatchewanSchool Trustees Association.Trybus, R.J., & Karchmer, J.A. (1977). School achievement scores of hearingimpaired children: National data on achievement and growth patterns.American Annals of the Deaf, 122, 62-69.Ulissi, S.M., Brice, B.M., & Gibbins, S. (1989). Use of the Leiter InternationalPerformance Scales and the Wechsler Intelligence Scale forChildren-Revised with hearing impaired children. Diagnostique, 9,142-153.Vernon, M. (1968). Fifty years of research on intelligence of the deaf and hardof hearing children: A review of literature and discussion of implications.Journal of Rehabilitation of the Deaf, 1, 1-12.Vernon, M., & Alles, B.F. (1986). Psychoeducational assessment of the deafand hearing impaired children in adolescents. In P.J. Lazarus & S.Strichart (Eds.), Psychoeducational Evaluation of Children andAdolescents with Low Incidence Handicaps, New York: Grune andStratton.Vernon, P.E. (1966). Educational and intellectual development among CanadianIndians and Eskimos part 1 and 2. Educational Review, 18, 79-91.Vernon, M.C., & Brown, D.W. (1964). A guide to psychological tests and testingprocedures in the evaluation of deaf and hard of hearing children.Journal of Speech and Hearing Disorders, 29, 41 4-423.Vonderhaar, W.F., & Chambers, J.F. (1975). An examination of deaf students’WISC performance subtest scores. American Annals of the Deaf, 120,540-543.Wallace, H.M. (1973). The health of American Indian children. AmericanJournal of Diseases of Children, 125, 449-454.Watson, B., Goldgar, D.E., Kroese, J., & Lotz, W. (1986). Nonverbalintelligence and academic achievement in the hearing impaired. TheVolta Review, 88, 151-158.REFERENCES / 170Watson, B., Sullivan, P.M., & Jensen, J. K. (1982). Nonverbal intelligence andEnglish language ability in deaf children. Journal of Speech andHearing Disorders, 2, 199-204.Wechsler, D. (1939). The measurement of adult intelligence. Baltimore:Williams and Wilkins.Wechsler, D. (1991). Manual for the Wechsler Intelligence Scale forChildren- Third edition. New York: Psychological Corporation.Weinburg, R.A. (1989). Intelligence and IQ:Landmark issues and great debates.American Psychologist, February, 98-104.White, A.H., & Stevenson, V. (1975) The effects of total communication, manualcommunication, oral communication on the learning of factual informationon residential school children. American Annals of the Deaf, 120, 48-57.Wilgosh, L., Mulcahy, R., & Watters, B. (1986). Assessing intellectualperformance of culturally different Inuit children with the WISC-R.Canadian Journal of Behavioral Science, 17, 270-277.Wrightstone, J., Aronow, M., & Moskowitz, S. (1963). Developing reading testnorms for deaf children. American Annals of the Deaf, 108, 311-316.Ysseldyke, J.E. (1979). Implementing the protection in evaluation proceduresprovision of P.L. 94-142. In P.E.P.:Developing criteria for theevaluation of protection in evaluation procedures provisions (pp.143-194). Philadelphia:Research for Better Schools, Inc.Zachary, R.A. (1990). Wechsler’s intelligence scales:Theoretical and practicalconsiderations. Journal of Psychoeducational Assessment, 3,276-289.Zinkus, P.W. (1986). Perceptual and academic deficits related to early chronicotitis media. In J.F. Kavanagh (Ed.), Otitis media and childdevelopment (pp. 107-116). Maryland:York Press.Zinkus, P.W., & Gottleib, Mi. (1980). Patterns of perceptual and academicdeficits related to early chronic otitis media. Paediatrics, 66, 246-253.REFERENCES/171Zinkus, P.W., Gottleib, M.I., & Shapiro, M. (1978). Developmental andpsychoeducational sequelae of chronic otitis media. AmericanJournal of Diseases of Children, 132, 1100-1104.Zimmerman, l.L., & Woo-Sam, J. (1972). Research with WISC :1960-70.Psychology in the Schools, 9, 232-271./ 172APPENDIX AModified Pantomime Instructions for Administering the WISC-Ill (adapted fromSaltIer, 1988)Picture CompletionShow the child card 2 from the Wechsler Intelligence Scale for Children (WISC).This item depicts a table with three legs. Count the legs by pointing to each withindex finger. Hold up three fingers, then point to missing leg by tracing itsdesignated outline with index finger. Hold up four fingers. Summarize directionsby counting each leg with index finger, holding up three fingers, pointing to themissing area, and holding up four fingers.Present age appropriate intital item for the WISC-lll Picture Completion booklet.Point to the child, point to the picture (not to the missing part) and point to thechild. Follow the instructions in the WISC-IlI manual regarding failure anddiscontinuation procedures. If child fails to indicate the missing part on Cards 1and 2 within 20 seconds, point to the missing part.Picture ArrangementPlace sample item (CAN) in front of child in numerical order indicated in theWISC-lII manual. Point to pictures in general. Arrange pictures in correct order.Point to Picture A and hold up one finger, point to Picture B and hold up twofingers, and point to Picture C and hold up three fingers, thereby designating first,second, and third. Rearrange pictures in original administration order. Point toseries in general sweeping motion, and then point to child.Present age appropriate initial item from WISC-lll Picture Arrangement subtest.Point to the stopwatch to indicate this is a timed test. Point to series of picturesin general sweeping motion, and then point to the child. If the child does notrespond or does not arrange the cards in the correct order for Item 1 (FUN),and/or Item 2(DOG), arrange the cards in the correct sequence. Then point toeach card, designating first, second and third. Allow the child to look at thecorrect sequence for about 10 seconds, and then put the cards in the originalAPPENDIX A I 173numerical order. Point to series in general sweeping motion, and then point to thechild. If the child fails to arrange Item 3(WALK) and/or Item 4(RUSH) in the correctsequence, put the cards in the original numerical order.In Items 5 to 12, arrange cards in numerical sequence, point to cards in generalsweeping motion, and point to child. Follow instructions in WISC-IlI manualregarding timing and discontinuation procedures.Block DesignDesign 1Place four blocks in front to the child and turn each block to show the differentsides. Point to each different side of each block during this demonstration.Arrange the four blocks into Design 1. Give the child four blocks. Point to thechild, point to the child’s blocks, and point to the model. If the child fails,assemble child’s blocks to match modeled design. Point to the child’s blocks.Point to child, point to child’s blocks, and point to model.Design 2Assemble Design 2 behind a screen. Present model to child in completed form.Point to the stopwatch to indicate this is a timed test. Point to the child, point tothe child’s blocks and point to the model. If the child fails on the first trial, followprocedures given for Design 1.Design 3Display card depicting Design 3. Assemble blocks of Design 3 in full view of thechild. Scramble the blocks. Point to the stimulus card and to the blocks. Point tothe stopwatch, the child, the child’s blocks, and then to the stimulus card. If childfails on first trial, repeat the above administration procedure. If child is 8 or olderand subtest begins with this item, follow procedure for designating color patternsand similarities of blocks given at the beginning of section on Design 1.Designs 4 to 11Display card showing each respective design. Point to child, point to blocks, andpoint to child. Follow instructions in WISC-lll manual regarding timing anddiscontinuation procedures.APPENDIX A 1174Object AssemblyArrange the pieces of sample item (APPLE) behind shield. Expose array andassemble pieces together. Rearrange pieces in original presentation sequence.Point to the the stopwatch to indicate that this is a timed test. Point to the childand point to the array. If the child does not respond, repeat the above procedure.Items 1 to 5, arrange the pieces behind the shield. Expose array, point to child,and point to pieces. Correct errors on Item 1 only. Follow instructions in WISC-lllmanual for timing and scoring.CodingCoding APoint to star, circle, triangle, cross, and box in general sweeping motion. Point toeach mark in each indMdual geometric shape. Point to blank sample items andfill in the first two shapes. Then, give the child a pencil, point to the child, andpoint to the remaining sample items. Stop the child after completion of the lastsample item. Point to the stopwatch to indicate this is a timed test. Point to theremaining items in a general sweeping motion for each row. Point to the child andpoint to the first item. If the child discontinues work after completing the first row,point to the next row. Follow the timing instructions in the WISC-lll manual.Coding BPoint to numbered boxes in general sweeping motion. Then point to each numberand its respective symbol individually in the entire array. Point to blank sampleitems and fill in the first two. Give child pencil, point to child, and point to theremaining items in the sample.Then point to the stopwatch to indicate this is a timed test. Point to the child, andpoint to the first item. If the child discontinues work after completing the first row,point to the next row. Follow timing instructions in the WISC-lll manual.MazesGive the child the test protocol. Point to sample item. Point to figure in center ofmaze and to opening that leads to exit. Demonstrate sample item. On reachingopening to outside of center, pause, and without lifting the pencil, point to blindalley. Then point to correct route and finish the tracing.APPENDIX A / 175Items 1 to 9, point to center of maze, to exit opening, and to child. Followdiscontinuation procedure in WISC-IIl manual. If child does not begin to work incenter of maze, stop work, place pencil in center of maze, and point for child tocontinue. Follow timing and scoring instructions in WISC-Ill Manual.Symbol SearchSymbol Search AGive the child the test protocol. In a sweeping motion point to both the sampleitems. Point to the stimulus shape, the shapes in the row, the shapes that match,and then the word “YES”. In sample 2, point to the stimulus shape, point to therow, point to each stimuls shape and shape in the item and then to the word “NO”to indicate there is not a match. Administer the Practice items by pointing to thefirst item, point to the child, and then pointing to the first item. If the child doesnot respond repeat the procedure as stated for admnistration of the Sample Items.Point to the stopwatch to indicate this is a timed test, give the child the pencil, andpoint to the first item. Administer the subtest according to the time and scoringinstructions as stated in the manual.Symbol Search BIn a sweeping motion point to both the sample items. Point to the stimulusshapes, and the shapes in the row. Point to each sample item and each of theshapes in the item. Indicate when there has been a match by pointing to the word“YES”. In sample 2, follow the same procedure, except there is no match bypointing to the word “NO”. Administer the practice items and the test items asoutlined above in Symbol Search A./ 176APPENDIX BPantomime Instructions for Administration of the Stanford-Binet:Fourth Edition(SB:FE) (Adapted from Testing Examinees Who Have Limited English Proficiencyor Who Are Non-Language Proficient, Delaney & Hopkins, 1987)Bead MemoryBegin by showing the examinee the box of beads. Show the examinee each of thebeads shapes and colors. Display the picture of the bead layout to examinees atEntry Levels A-G. Administer the pretest, if appropriate, by using gestures.Continue testing with the examinee’s first entry level item, using gestures toindicate the task. For examinees continuing with item 11 (Level H) or entering atLevels H through K, display the bead stick and show each of the bead shapes andcolors. Then administer Sample Item 1 using gestures. Demonstrate the task ifnecessary. Proceed with the examinee’s first entry level item. For examineesentering the test at Level L or higher, display the bead stick and beads and thenadminister the appropriate sample item. Continue with the examinee’s first entrylevel.Pattern AnalysisIf administering Entry Levels A-C, present the form board with its pieces in place.For each item, gesture to the examinee to watch as the pieces are removed andlaid out. Follow the exact directions given in the item book. Then indicate that theexaminee is to replace the pieces in the form board.If administering Entry Levels D-Q, give the examinee one cube and keep one cube.Show the child each side of the cube and gesture that the examinee shouldsimultaneously find and match each side. Administer the appropriate sample itemby pointing to the stimulus and then to the child’s blocks. If the sample is failed,demonstrate again. Even if the examinee fails again, continue with this test andattempt at least six items before discontinuing.APPENDIX B /177Memory for ObjectsTo demonstrate Sample Items 1 and 2 point to the first picture and raise onefinger, point to the second picture and raise a second finger. On the next pagepoint to the first picture, and then the second picture. If an incorrect response isgiven, demonstrate the item again and have the examinee point to the correctresponses.Continue testing with the examinee’s first entry level item. Discontinue as indicatedin the manual.MatricesFor Entry Levels l-M, administer Sample Item , pointing to the empty blank, andthen to each choice in turn, one by one. Point to the examinee to select aresponse. If incorrect, point out the correct response. Have the examinee againselect a response before administering Sample Item 2 in a similar manner. Thengive the examinee’s first entry level item.For Entry Level 0 or higher, administer Sample Items 3 and 4 in a similar manner.Begin testing with the examinee’s first entry level item.APPENDIX C /178ANOVA results comparing method of administration (verbal, n=40 andpantomime, n=38) for hearing Native peersWISC-lll F ratio Bartlett FPIQ.0001 .581 P= .446P0.12 .141 P= .707PS.008 2.073 P= .950PC 1.53 .210 P= .647COD.56 1.883 P= .670PA.06 2.115 P= .646OA.05 4.949 P= .656BD.25 .084 P= .772ss.57 .740 P= .390SB:FEPattern Analysis.92 .201 P= .647Matrices.03 .733 P= .693Bead Memory.01 .788 P= .775Memory for Objects .59 .099 P= .753Abstract/Visual Reasoning .83 .046 P= .829Short Term Memory .01 .065 P= .800SBTOT.002 .040 P= .841APPENDIX C /179ANOVA results comparing pantomime administration for otitis prone (n=87) andhearing Native peers (n=78)WISC-III F ratio Bartlett FPIQ 5.01 .772 P= .380PC 4.29 .127 P= .721PS 1.31 .039 P= .843PC 5.82 .118 P= .928COD 1.66 .000 P= .983PA 3.54 .118 P= .732CA 1.33 3.02 P= .832BD.74 .024 P= .87755.73 .319 P= .573SB:FEPattern Analysis 6.93 1.12 P= .843Matrices 7.41 .186 P= .271Bead Memory 4.82 2.16 P= .667Memory for Object 1.72 .018 P= .893Abstract/Visual Reasoning 10.08*.521 P .471Short Term Memory 5.13 .773 P= .379SBTOT 8.7 .959 P= .385*p>01APPENDIX C /180ANOVA results comparing pantomime administration for otitis prone (n=87) andhearing peers (n=40)WISC-lll F ratio Bartlett FPIQ 3.65 .008 P= .766P0 2.30 .012 P= .913PS .84 1.05 P= .305PC 1.58 .092 P= .761COD 2.36 .749 P= .387PA 2.10 .373 P= .541CA .77 .100 P= .752BD .17 .005 P= .943SS 1.04 .91 P=.34SB:FEPattern Analysis 2.51 .377 P= .539Matrices 4.20 .776 P= .378Bead Memory 3.8 .451 P= .502Memory for Objects .37 .004 P= .949Abstract/Visual Reasoning 5.93 .248 P= .619Short Term Memory 3.75 .367 P= .545SBTOT 6.09 .521 P=.47BIOGRAPHICAL INFORMATIONNAME: Ly4dc A ?/,7MAILING ADDRESS: 7 3 Y 7Va, kr,2gPLACE AND DATE OF BIRTH: ,4 2j,EDUCATION (Colleges and Universities attended, dates, and degrees):JQ M v//y __/979 d,/Y&S7POSITIONS HELD:Ed’ 4//,7 I46/7z y 41/ /A0/ycAo/; /PUBLICATIONS (if necessary, use a second sheet):AWARDS:Complete one biographical form for each copy of a thesis presentedto the Special Collections Division, University Library.OE.5


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items