ebook img

ERIC EJ585978: A Review of the Woodcock Reading Mastery Test-Revised (WRMT-R). PDF

8 Pages·1999·0.38 MB·English
by  ERIC
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview ERIC EJ585978: A Review of the Woodcock Reading Mastery Test-Revised (WRMT-R).

A Review of the Woodcock Reading Mastery Test-Revised (WRMT-R) Marybeth De Rose The WoodcockReadingMastery Test-Revised (1997)isthelatesteditionofthe Woodcock ReadingMastery Test, which was originallypublishedin1973.The revisedversion,firstpublishedin1987,isnowinits12thyearofserviceand remainsapopularchoiceamongavailablereadingtests.Theobjectiveofthis review is to take a renewed critical look at the test in general, and in par ticular to inquire whether test results mightbe useful in makinginferences aboutESLlearnerreadingability. The WRMT-R was designed to assess the reading levels of test-takers within an age range of4 to 75 years of age and over. The manual suggests that the WRMT-R may be used with ESL (English as a Second Language) students, although provisions must be made for "ensuringthat the subject understands the instructions, and providingadditional practice opportuni ties, or reviewing the instructions" in order to "provide a valid measure of theEnglishreadingskills" (p.18). TheWRMT-R,likeitspredecessortheWRMT,consistsoftwoforms that facilitate retesting at short intervals. The original test consisted offive sub tests: Letter Identification, Word Identification, Word Attack, Word Com prehension, and Passage Comprehension across two parallel forms (A and B). The revised WRMT has retained the two forms, but they are no longer parallel. Form H has been reduced to four subtests by eliminating Letter Identification from the original, and FormG has beenexpanded to six sub tests to improve readiness assessment. Form G consists ofthe five subtests from the WRMT, plus an additional one takenfrom the Woodcock-Johnson Psycho-EducationalBattery(Woodcock&Johnson,1977).Thesixsubtestsof FormG,fourofwhichcompriseFormH,areasfollows. First, the new subtest Visual-Auditory Learning assesses the ability to learn26symbols and theirnames, aswellas twosymbolsfor wordendings (-ing,-s),thenreadtheminsentenceformacrosssevenbriefteststories. The Letter Identification subtest involves reading alphabet characters in several fonts, including cursive writing. Supplementary lists have been added to the revised WRMT, in upper and lower case, which also contain severaldiphthongsanddigraphsforthepurposeofassessingsound-symbol association. Thesetestsarenotnormedandarethereforeofdiagnosticvalue only. 86 MARYBETHDEROSE TheWord Identification subtestentails readingand pronouncingwords in isolation from lists of increasing difficulty, "even though [the test taker mayhave] hadnopreviouspersonalexperiencewiththeword." Essentially the listed words must be decoded and pronounced in a manner consistent withthearticulationguidelinesinordertobecountedascorrect. The Word Attack subtest assesses phonetic decoding skills through the reading aloud of increasingly complex nonsense words. Nonsense words have been used in order to eliminate the possibility of test-takers knowing howtopronouncetheminadvance,asmightbethecasewiththeoccasional realstimulusword,nomatterhowobscure. TheWordComprehensionsubtestconsistsofthreedistinctsections.The first twosections entailreadingstimulus words and respondingwitheither a synonym or an antonym, and the third section requires the test-taker to read thestimuluswords, ascertainthe relationship betweenthem, and then continuetheanalogytocompleteanotherstimuluspairofwords.Thevocab ulary itemshave beencategorizedas eithergeneral reading,science-mathe matics, social studies, or humanities. Separate category scores can be calculatedfor diagnosticpurposes,butarenotnormed. The last subtest, Passage Comprehension, resembles a cloze exercise, which requires test-takers tofill inaseries ofblanks accordingto themean ing of the surrounding sentences or phrases. Picture clues have been in cludedfor approximatelyonethirdoftheeasiestitems. The WRMT-R provides norms for subjects from kindergarten to college seniorsintermsofgrade,andfor agesup to andbeyond75years. The original normative data were gathered from 6,089 subjects in 60 geographically diverse United States communities. The kindergarten to grade12samplecomprised4,201subjects,andthecollege/universitysample contained 1,023 subjects. The adult sample of subjects aged 20-80 but not enrolled in college comprised 865 subjects. The norming sample was struc tured to resemble the distribution of variables included in the 1980 US Census. Individual subject weighting was applied during data analysis to obtain"exact"proportionmatchingwiththecensusdata. The WRMT-R was renormed in 1995-1996. In Canada these norms have justbeenmadeavailablecommerciallyinrecentmonths. The current norms (WRMT-R NU) are based on a somewhat smaller samplepopulationofapproximately3,700,varyingsomewhatfor eachsub test/cluster, whose ratios are reflective of the 1994 American census. The samplecontrolledfor age, gender, race, geographicregion,SES,parentedu cation,andcommunitysize(adultnormsonly).A representativeproportion of special education students were also included in the norm sample. The new technicalchapterofthemanualstatedthatindividualsnotproficientin Englishwerenotincludedinthenormsample;however,anoperativedefini tionofproficiencywasnotincluded. TESLCANADAJOURNAULAREVUETESLDUCANADA 87 VOL.16,NO.2,SPRING1999 TheWRMT-Rrenormingprocesswaspartofaprogramaimedatrenorm ing the PIAT-R, KeyMath-R, and the K-TEA in addition to the WRMT-R. Accordingto theupdatedWRMT-Rtechnicalchapter,thefour achievement batteries included in the normative updating project measure a number of achievement domains in common: word reading, reading comprehension, mathematics computation, mathematics applications, and spelling. The Vi sual-AuditoryLearning,WordAttack,andWordComprehensiontestsofthe WRMT-R are not measures of any of the five shared domains; therefore, these tests were normed individually with sample sizes controlled for throughtheplan'stestingassignments. The advantage of the domain-norming procedure is primarily the in creasedcomparabilityoftestbatteriesintermsoftheircommoncross-battery domains. Inaddition,smallersamplesizesareacceptable. Raschscalingwasusedtocalibrateitemdifficultyofthesubtestsincluded inthecommondomains. However,normsfortheVisual-AuditoryLearning, Word Attack, and Word Comprehension tests were constructed from raw scoreswithnoRaschanalysis. Criticisms ofthe WRMT-R From its debut in 1973 the WRMT has met with controversy. Critics were eitherimpressedbyitspurportedprecisionandeaseofadministration(Ban natyne, 1974;Allington, 1976),ordisappointedbyitstheoreticalframework and questionable psychometric properties (Houck & Harris, 1976; Dwyer, 1978). Current critical review continues to raise many of the same issues left outstanding after the revision process. Both editions were criticized for the Eurocentric and gender-stereotype-linked artwork (Dwyer, 1978; Jaeger, 1989). Ofcontinuingconcernis the assessmentoffragmented skillsrather than the reading act as a whole. The test author apparently views reading as a collectionofsubskillsactinginconcert.Inreaction,Cooter(1989)arguesthat "instruments like the WRMT-Rare useless and representa bygone age that viewedthereadingactinafractured orsplinteredway" (p.910). Jaeger (1989) and Eaves (1990) debated whether readiness skills such as Letter Identification should be included in a reading test per se. The recent renorming project pointed out that Letter Identification tended to correlate positivelywithWord Recognitionscoresinallage groupsbuttheyoungest. The lack of correlation in the early grades was attributed to a curriculum effect, in that young childrenhave a greater portion oftheir program dedi cated to the development of letter recognition skills as opposed to word recognition. As well, both authors had difficulty accepting whether the Auditory-Visual Learning subtest (part of the readiness cluster) actually generalizes to the auditory-visualskills requiredbywrittenEnglish. Forthe ss MARYBETHDEROSE purposes of the renorming project, Auditory-Visual learning was not con sideredpartofeitherthewordreadingorreadingcomprehensiondomains. Althoughtheauthorofthe testasserts thattheWordIdentificationitems represent the culmination of a variety ofhigh frequency word lists, Jaeger (1989)isoftheopinionthat"thewordscontainedinthetestrangefromthose found in early elementary school readers to polysyllabic tongue-twisters likely tobe familiar only to personswhose leisurehours arespentperusing the Oxford English Dictionary [sic]" (p. 914). In addition, he states that reading lists of words in isolation has been faulted as a useful measure of sightvocabulary,becauseexamineescannotusesemanticcontextcuesfound in authentic reading experiences. Tuinman (1978) adds that some subjects willmanagetoreadanumberofthewordscorrectlyassightwords,whereas others will rely on phonetic word attack skills. Consequently, this subtest measuresdifferentskillswithdifferenttest-takers.Onemightanticipatethat the person with superior decoding skills would have an advantage on this subtest. Both editions of the WRMT have drawn criticism for the Word Com prehension subtest. Tuinman (1978) decided that this was more a test of reasoning than word knowledge because of the necessity to ascertain the relationship between the stimulus words in order to complete the analogy. As well, Tuinman felt thatpoor readers were penalized twice because they may miss additional items due to poor decoding skills, not necessarily be cause they lackthe word knowledgenecessary to complete the item.Jaeger (1989) wasunsure ofthe diagnosticandprescriptivevalueofalowscoreon this subtest because of the number of possible constructs being measured. This subtest was not considered part of the word reading or reading com prehensiondomainsin therenormingproject. Houck and Harris (1976), Dwyer (1978), and Tuinman (1978) were par ticularly critical of the Passage Comprehension subtest because of the modified cloze procedure. Houck and Harris (1976) revealed that the test items do not appear to meet Bormuth's (1969) criteria for determining the location of deletions as stated in the manual. In their opinion, "the items resemble more closely a completion test where deletions are made using subjectiveconceptssuchas keywords" (Houch&Harris,1976,p. 77). Bachman (1985) concluded thatnotalldeletionsina givenclozepassage measureexactlythesameabilities.AmongL2clozetest-takers,Turner(1989) found that other factors in addition to knowledge in a second languagedo contribute to successful performance on the cloze procedure. Therefore, changing the type of deletions in any given cloze test affects the construct validity of the test. Straying from Bormuth's (1969) criteria may have sig nificantly altered what the Passage Comprehension subtest purports to be measuring. TESLCANADAJOURNAULAREVUETESLDUCANADA 89 VOL.16,NO.2,SPRING1999 Although the WRMT-Rhasbeenrenormed, itis importantto remember thatmanyschools and institutionsmaybeslowin updatingtheir testpack ages as a result of current funding limitations. It is, therefore, prudent to considerthecriticismsofbothnormeditions. Eaves (1990) and Jaeger (1989) have expressed concerns regarding the sampleused for normative purposes in the originalWRMT-R. In particular they had difficulty with the author's statementin the manual that selective weightingwasusedinordertohavethesamplereflectthevariablesoutlined inthe1980USCensus,butdidnotspecifytheprocedureorofferanydetails. Eaves(1990)hadparticulardifficultywithhowWoodcock"indicatedthat communities were selected according to socio-economic characteristics ina mannerthatwouldobtainasamplematchingthe [1980]USCensusdata,yet no evidence was provided to verify the outcome ofhis selection procedure for communities or individuals" (p. 288). This evidence was similarly not forthcominginthebackgroundinformationaccompanyingthenewnorms. As the readermayrecall, theUSCensusin1980restricteditselftodistin guishing between "white" and "non-white" ethnic groups. Only with the 1988 US Census did further details regarding ethnic background become available. Therefore, the author used only age, socioeconomic status, and white-non-whitevariablesinhisoriginalsample.Thecurrentnormsampleis more inclusive with regard to race, gender, and community size (adult normsonly).However,thereadermustbearinmindthatthe1994USCensus was still limited to the categories White, Black, Hispanic, and other. It re mains questionable whether thenormsampleis representative oftheCana dianpopulationwithwhomitiswidelycompared. MorerecentworkbyKlesmer(1994)hasfoundstrongevidencetosuggest thattheacademic/linguisticdevelopmentofESLstudentsfollows adistinct pattern. Atleastsix years are required for ESL students to approach native Englishspeakers'norms ina variety ofareas, and itappears thateven after sixyears,fullcomparabilitymaynotbeachieved.Withthisinmind,itwould seem prudent to develop separate achievement criteria for these students based on their age and length of residence. Klesmer's (1994) attempts to renorm the Peabody Picture Vocabulary Test illustrated the difficulty and complexity involved in establishing norms appropriate for individual ESL populations. Given Klesmer's (1994) findings, it would appear that the crude unidimensional norms presented in the WRMT-R are hardly adequate to account for the potential variability in performance among ESL test-takers. Indeed, itis questionable whether the ESLpopulationis representedinany meaningful way thatwould allow Woodcockto advocate the use ofthe test withonlyminorprovisions. Alsoofcurrentconcernis the validityevidencepresentedin themanual. As with theformer edition, there is lack ofevidence ofcontentvalidity. No 90 MARYBETHDEROSE information is offered regarding the possible match between the test and variouscurricula,orthe "itemsand thepsychologicalprocessesinvolvedin the readingactfor eachgrade/agelevel" (Cooter, 1989,p. 912). Themanual suggests that the WRMT-R items were developed with the assistance of "outside experts" and experienced teachers, but does notelaboratebeyond this overtstatementregardingthespecificinputofthesesources. Toseveral critics this information is far from sufficient for claiming that the WRMT-R hasanykindofcontentvalidity(Cooter,1989;Jaeger,1989;Eaves1990). Intermsofconcurrentvalidity(howthistestcompareswithothersinthe field for measuring various reading skills), the authors compared the WRMT-R only with the Woodcock-Johnson Reading Tests in the original manual. "It seems implausible that this exercise constitutes a comparison between independent criterion measures" (Cooter, 1989, p. 912). Jaeger (1989)andEaves (1990)weresimilarlyunimpressedwithevidencesupplied inthemanual. The originalWRMT-Rmanual did present a table comparing the earlier editionwith various other tests, arguing that it remained pertinentbecause the "psychometric characteristics of the original WRMT (1973) and the WRMT-Raresosimilarthatmanygeneralizationsfrom oneto theothercan be made" (p. 100). This statementbegged the question: Why bother with a neweditionthen?Theauthortookgreatpainsinthefirstchaptertoillustrate thechangeseffectedfromtheoriginaltest,yethesuggestedatthatpointthat thetwoweresosimilarthattheysharedpsychometricproperties. By using the domain-norming approach in the renorming project, WRMT-RconcurrentvalidityisestablishedwiththePIAT-Rand theK-TEA insofaras thesubtests inthecommondomainareconcerned. In response to the criticisms regardingvalidity levelled at the WRMT-R, Eaves (1990) reported that the test's publisher,AmericanGuidance Service, made available intercorrelations between the WRMT-R Form G and the K-ABC (Kaufman AssessmentBatteryfor Children) and WISC-R (Wechsler Intelligence Scale for Children-Revised) intelligence scales. In both cases correlationswerefairlyhigh(range.75to.96).Eaves(1990)alsoreportedthat the studies were without date. One might surmise that in making these particular intercorrelations available, the publisher was making a brilliant marketing move. Both the WISC and the K-ABC are widely used ineduca tionandpsychology. ByaligningtheWRMT-Rwiththesepopulartools,the publisher implies that they measure the same constructs. This makes the WRMT-R especially attractive for use in diagnosing a variety of learning difficultiesthatentailtheuseofanachievementmeasure. However,without adateonthestudy,onehastoquestionthe credibilityofthestatistics. Curiously,itmustbenotedthatseveralrecentstudiesusedtovalidatethe WISC-III(WechslerIntelligenceScaleforChildren-ThirdEdition)involved the WRMT-R, to establish credibility (Hishinuma & Yamakawa, 1993; TESLCANADAJOURNAULAREVUETESLDUCANADA 91 VOL.16,NO.2,SPRING1999 Newby, Recht, Caldwell, & Schaefer, 1993; Schwean, Saklofske, Yakulic, & Quinn, 1993). Itwould appear that the validation arrangement is mutually beneficial. The K-ABC and the Wechsler tests are not reading tests per se. They purport to measure intelligence; therefore, the only conclusion that can be drawnfromtheaboveinformationisthattheK-ABC,WISC-R,andWISC-III maymeasurethesameundefinedconstructsasdoestheWRMT-RFormG. The construct validity of the WRMT-R is not discussed in the original manual. The renorming project, in definingthe domains and selectingsub tests appropriate for those domains, effectively establishes the constructs beingmeasured. However,threeoftheWRMT-Rsubtestsdidnotfallwithin the confines ofthe definitions ofword readingand readingcomprehension and therefore are measuring other, undefined constructs. The paucity of validity evidence should again raise questions regarding whether the test shouldbeusedasthebasisfordecision-makingofanykind. WithregardtoESLpopulations,itshouldbeapparentthatreadingskills might not be effectively assessed using this tool. The original manual does notdefine"Englishreadingskills";therefore,anynumberofconstructsmay bemeasuredbythis instrument. Letter Identification might be of some use with persons whose L1 is encoded with characters other than those of English; however, this skill should not be interpreted as reading. The Word Identification subtest may notbeavalidmeasureofacquiredEnglishsightwords iftheindividualhas developedsufficientphoneticdecodingskills.WordComprehensionmayat best define some sense of the test-taker's knowledge of synonyms and an tonyms. However, becausewhattheanalogiessubtestis measuring(decod ing?verbalreasoning?wordknowledge?)isunclear,itmaybedangerousto ascribe the construct word comprehension to this subtest. The vocabulary chosenforthetestasawholemaynotrealisticallyreflecteverydaylanguage at a variety of age levels or lengths of residence. Last, the Passage Com prehension subtest may not be measuring true comprehension as much as the ability to fill in a blankadequately. Itwould appear that the US sample used to establishnorms for the WRMT-Rmaynotadequately represent the Canadian and/or North American newcomer; therefore, an ESL test-taker may not be adequately represented in the norm population. Essentially, undertheseconditions,theESLtest-taker'sperformancewouldbecompared withamajoritywhiteAmericannorm. Thisishardlyfairtestingprocedure. Toconclude, the Woodcock ReadingMaster Test-Revised is "unsafeatany speed,"especiallyfor ESLtest-takers. 92 MARYBETHDEROSE The Author MarybethDeRoseisan18-yearveteranofOntarioclassrooms;forthepastdecadeshehasbeen aspecialeducationteacherinapredominantlyESLcommunity.Sheiscurrentlycompletingthe finalyearofherPhDprogramattheOntarioInstituteforStudiesinEducation. References Allington,RL.(1976).Woodcockreadingmasterytestsreview.JournalofReading,20,162-163. Bachman,L.F.(1985).Performanceonclozetestswithfixed-ratioandrationaldeletions. TESOLQuarterly,19,535-556. Bannatyne,A.(1974).Woodcockreadingmasterytestsreview.JournalofLearningDisabilities,7, 398-9. Bormuth,J.(1969).Factorvalidityforclozetestsasmeasuresofreadingcomprehensionability. ReadingResearchQuarterly,4,358-365. Cooter,R (1989).ReviewoftheWoodcockReadingMasteryTests-Revised.InJ.CConoley&J.J. Kramer(Eds.),The10thmentalmeasurementsyearbook(pp.910-912).HighlandPark,NJ: Gryphon. Dwyer,CA.(1978).WoodcockReadingMasterytestsreview.InO.K.Buros(Ed.),The8th mentalmeasurementsyearbook(pp.1303-1305).HighlandPark,NJ:Gryphon. Eaves,RC(1990).WoodcockReadingMasteryTests-revised(WRMTR).Diagnostique,15, 277-297. Hishinuma,E.S.,&Yamakawa,R (1993).Constructandcriterionrelatedvalidityofthe WISC-IIIforexceptionalstudentsatrisk.JournalofPsychoeducationalAssessment:Advances inPsychoeducationalAssessmentMonographSeries,94-104. Houck,C,&Harris,L.(1976).WoodcockReadingMasteryTestsreview.JournalofSchool Psychology,14(1),77-79. Jaeger,RM.(1989).ReviewoftheWoodcockReadingMasteryTests-revised.InJ.CConoley &J.J.Kramer(Eds.),The10thmentalmeasurementsyearbook(pp.913-916).HighlandPark, NJ:Gryphon. Klesmer,H.(1994).AssessmentandteacherperceptionsofESLstudentachievement.English Quarterly,26(3),8-11. Newby,RF.,Recht,D.R,Caldwell,J.,&Schaefer,J.(1993).ComparisonofWISC-IIIand WISC-RIQchangesoveratwo-yeartimespaninasampleofchildrenwithdyslexia. JournalofPsychoeducationalAssessment:AdvancesinPsychoeducationalAssessmentMonograph Series,87-93. Schwean,V.L.,Saklofske,D.H.,Yakulic,RA.,Quinn,D.(1993).WISC-IIIperformanceof ADHDchildren.JournalofPsychoeducationalAssessment:AdvancesinPsychoeducational AssessmentMonographSeries,56-60. Tuinman,J.J.(1978).Woodcockreadingmasterytestsreview.InO.K.Buros(Ed.),The8th mentalmeasurementsyearbook(pp.1306-1308).HighlandPark,NJ:Gryphon. Turner,CB.(1989).TheunderlyingfactorstructureofL2clozetestperformancein Francophone,university-levelstudents:Causalmodelingasanapproachtoconstruct validation.LanguageTesting,6,172-197. Woodcock,RN.(1973).WoodcockReadingMasteryTests.CirclePines,MN:AmericanGuidance Service. Woodcock,RN.(1987).WoodcockReadingMasteryTests-revisedexaminer'smanual.Circle Pines,MN:AmericanGuidanceService. Woodcock,RN.(1997).WoodcockReadingMasteryTests-revised/normativeupdate.Circle Pines,MN:AmericanGuidanceService. Woodcock,R,&Johnson,M.(1977).Woodcock-JohnsonPsychologicalBattery.Allen,TX:DLM TeachingResources. TESLCANADAJOURNAULAREVUETESLDUCANADA 93 VOL.16,NO.2,SPRING1999

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.