Differences in Language Performance among High ... - CiteSeerX

8 downloads 6642 Views 2MB Size Report
repeated failures in FL courses in college (20;. 35; 48). A number ...... Of the initial 501 students, then, 174 fit the criteria for inclusion in this data set. the upper left ...
Differences in Language Performance among High-, Average-, and Low-Anxious College Foreign Language Learners Author(s): Leonore Ganschow, Richard L. Sparks, Reed Anderson, James Javorshy, Sue Skinner and Jon Patton Source: The Modern Language Journal, Vol. 78, No. 1 (Spring, 1994), pp. 41-55 Published by: Wiley on behalf of the National Federation of Modern Language Teachers Associations Stable URL: http://www.jstor.org/stable/329251 . Accessed: 14/04/2013 20:11 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Wiley and National Federation of Modern Language Teachers Associations are collaborating with JSTOR to digitize, preserve and extend access to The Modern Language Journal.

http://www.jstor.org

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

Differencesin Language Performance among High-, Average-, and Low-Anxious College Foreign Language Learners

LEONORE GANSCHOW Department ofEducational Psychology Miami University OH 45056 Oxford,

RICHARD L. SPARKS EducationDepartment CollegeofMountSt.Joseph 5701 DelhiRoad Cincinnati,OH 45051 REED ANDERSON Department ofSpanishand Portuguese Miami University Oxford,OH 45056

JAMES JAVORSHY Department ofEducationalStudies PurdueUniversity W Lafayette, IN 47907 SUE SKINNER GraduateProgram Department ofEducational Psychology Miami University Oxford,OH 45056

JONPATTON

ApplicationConsultant Miami University OH 45056 Oxford,

FOREIGN LANGUAGE (FL) TEACHERS AND special educators have long been puzzled by findingsthat some studentsare able to learn a foreignlanguage withrelativeease while others have repeated failuresor learn only with great Among FL educators,otherwisesucdifficulty.1 cessfulstudentswho have difficulty withFLs are often referredto as underachieversor as lacking motivation (22; 24; 25; 44; 47). In special education manyof these at-riskFL learnersare identified as having language learning disabilities (LLD), and this identificationsometimes occurs only after a student has experienced repeated failures in FL courses in college (20; 35; 48). A number of explanations have been offered to account for why students have difficulties with FL as it is typicallytaught in school settings. Intelligence,for the most part, has been ruled out as having much significance in the determinationof FL learning potential (4; 26; TheModern 78,i (1994) Language Journal,

0026-7902/94/41-55 $1.50/0

?1994TheModern Language Journal

46). However, other factors such as affective variables (motivation, attitudes, anxiety) and native language skills (oral and written language, listening, speaking) are said to influence FL learning.Among the affectiveexplanations, recent attention has been given to the role of anxiety (30; 31; 34; 37; 49). Proponents of the anxiety hypothesis suggest that FL learnershave a mentalblock, similarto thatexperienced by some students in math, testtaking, and speech-making.As for the native language factorsthatmay affectlearning a FL, difficultieswith phonology and syntax,rather than with semantics of the language (vocabularyand reading comprehension,in particular), have been found to differentiate good and poor FL learners (18; 55; 56). A recent theory that postulates a relationship between problems in oral and written performance in native language and problems with the acquisition of a second language is called the "Linguistic Coding Deficit Hypothesis" (57; 59). In a previous paper Sparks and Ganschow (57) proposed the possibilitythattheremay be a confoundinginteractionbetween anxietyand receptive/expressivelanguage skills.They sug-

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

42

TheModernLanguageJournal78 (1994)

gested thatmanyof the questions on theForeign Language ClassroomAnxietyScale (FLCAS) (31) involve language skills. Over eighty-sevenpercent of the questions (twenty-nine out of thirtythree) were found to involveproblemstypically associated with difficultiesin listening,speaking, reading, writing,memory for language, and speed of language processing.They urged FL educators to investigateperformancein native oral and writtenlanguage among students who experience difficultyin FL learning as an alternativeexplanation to affectivevariables. FL educators have stated the need to explore the relationshipbetween anxiety systematically and language learning (49). Horwitz,Horwitz, and Cope suggest that there is likely to be a strongnegativecorrelationbetween FL anxiety and measures of FL proficiency,i.e., less anxiety is associated with higher grades and vice versa. They cite listening and speaking, especiallyimpromptuspeaking,as the main sources of anxiety. The relationship between anxiety and FL proficiencyis unclear,however.MacIntyreand Gardner (39) suggestthatas FL proficiencyincreases,anxietydecreases, thusraising the question of causality;i.e., does anxietyinterfere with pre-existing (language) ability and impair FL performance,or does poor FL performancelead to anxietyas a consequence? To date, the relationshipsamong nativelanguage skills,FL aptitude, and FL anxietyhave not been explored empirically.This exploration is difficultwithoutappropriate instrumentsto measure these relationships. Though the authors of the FLCAS have published reliability and validityinformationon their instrument (29), theyhave not validated it on large numbers of students,nor have theyprovided a system for determiningwhat constituteshigh or low anxiety levels. Standardized measures do exist formeasuringnative oral and writtenlanguage skills in college learners.2 Likewise, thoughdated, the ModernLanguageAptitudeTest (MLAT) (6) provides a method for determining a student'saptitude for learning a FL. The present studywas designed to explore FL classroom anxietyin relation systematically to nativeoral and writtenlanguage and to FL aptitude. To do this the authors developed a scoring systemfortheFLCAS using local norms and devised a batteryof nativelanguage instruments based on earlier studies they had conducted with at-riskFL learners. In this paper theyexamine the native oral and writtenlanguage and FL aptitude performance of students with differentlevels of anxiety.They be-

gin with a briefreviewof the literatureon the effectsof anxietyand of native language skills on FL learning. Then theydescribe methodology and results of the study they conducted using the FLCAS (for which they devised a method of classifyingstudentsinto anxietycategories), plus a varietyof native oral and written language measures and a measure of FL aptitude-the MLAT Last, they make recommendationsfortestingand accommodatingstudents who exhibit high levels of anxiety. EFFECTS OF ANXIETY ON FL LEARNING Over the past few years FL educators have hypothesizedthatanxietyplaysa role in success or failurein the FL classroom.A recentbody of researchsuggeststhathigh levelsof anxietycan interferewith FL learning (31; 37; 38). FL researchershave speculated thatanxietyin the FL classroom is specific to language learning and have proposed that languageanxietyis distinct fromother formsof anxiety (31; 37). Daly, for example, describes language in termsof "communication apprehension," or the fear that an individualhas about oral communication.Horwitzet al. (31) and Young (68) suggestthatanxiety in the FL classroom occurs primarilybecause the studenthas to speak theFL in frontof a group. MacIntyre and Gardner (40) distinguish between general anxietyand communicative anxiety. MacIntyre and Gardner (39) also propose thatanxietyplayslittlepart in the learner'sfirst experiences withthe FL and thatlanguage aptitude and motivationare the dominant factors in the early stages of FL learning. FL anxiety develops if the student's subsequent experiences withthe FL are not positive.Poor FL performance,in turn,reinforcesFL anxiety.Other studies (e.g., Gardner, Smythe,Clement, and Gliksman) have suggested negative correlations between language classroom anxietyand second language proficiency. Recently,Horwitz,Horwitz,and Cope developed a self-reportinstrument,the ForeignLanguage ClassroomAnxietyScale (FLCAS). The items in their instrumentinclude thirty-three test anxiety,communication about questions apprehension, and fear of negative evaluation in the FL classroom. When the instrumentwas administeredto studentsin FL classrooms,the findings suggested that FL anxiety is experienced by manystudentsin various facetsof FL learning. FL anxiety appears to be negatively correlated with FL grades (29). Other studies,

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Ganschow,Sparks,Anderson, Skinner, Javorshy, however, have not shown a relationship between anxietyand specific aspects of FL learning. Although he did not referto studies that used specific measures of language anxiety, Scovel found the resultsof studies he reviewed on the subject of anxietyto be equivocal. Young (69) found nonsignificantcorrelationsbetween the resultsof a "practice" (unofficial) Oral ProficiencyIndex (OPI) and anxietymeasures and concluded that abilityin the FL may have been a more importantvariable. She states that test anxietywould most likelyaffectthose withlow levels of oral proficiencymore than those with high levels of proficiency,and she furthersuggests thattestanxietymayneed to be viewed in conjunction withan individual's language proficiency.Sparks and Ganschow (57) have suggested that affectivedifferences,such as anxiety, might be the consequenceof FL learning problems,which themselvesbear a strongrelationship to a student'sproblems in his/her nativelanguage. Support forthisspeculation is examined next. EFFECTS OF NATIVE LANGUAGE SKILLS ON FL LEARNING The impact of native language on FL learning was examined in the 1960s byJohn Carroll and by Paul Pimsleur,who speculated that success or failurein FL learning was primarilythe result of language-based difficulties.Through factor analytic studies, Carroll (5) found four variables related to language to be important for FL learning: phonetic coding, grammatical inductive language learning ability, sensitivity, and rote memory.Pimsleur (46) studied FL underachievers,studentswho did less well in a FL than in theirothercourses. He found thataudiwiththe sound toryproblems,thatis, difficulty discrimination and sound-symbol aspects of language, were largelyresponsibleforthose FL learning problems which were not explainable by low intelligence or poor motivation. Both Carroll and Pimsleurdeveloped testsof FL aptitude, the ModernLanguageAptitudeTest(MLAT) (6) and the Language AptitudeBattery(LAB) (45), which contained items measuring these language-based skills. The relationship between FL learning and dyslexia (a formof learningdisability)was first alluded to in 1971 when Kenneth Dinklage, a counselor at Harvard, reported his conclusions from extensive interviewshe had conducted over the years with studentswho experienced FL learningproblems.Dinklage ruled out affec-

43

tivedifferencessuch as low motivationas causal factorsand instead suggested that these bright students had problems similar to dyslexia, or reading disabilities,in thattheyhad difficulties with auditory discrimination of sounds and sound-symbollearning. In a recent article Carroll (3) also mentions that students with FL learning problems may experience difficulties similar to studentswith dyslexiawho have particular difficultywith phonetic coding in their native language. In an examination of physiological, social, cultural, and other factorsaffectingFL learning,Spolksycites dyslexiaas an example of a physiological,i.e., language, variable. Skehan suggests that individualswho develop quicklyin theirnative language have relativelyhigh levels of FL aptitude and suggests thatFL learning aptitude is the "second or foreign language equivalent of a first language learning capacity" (pp. 200-201). In 1987 Gajar reported the first empirical studyof aptitude for FL in studentswith and withoutlearning disabilities (LD). As mightbe expected, she found significantdifferencesbetween the two groups' scores on the MLAT; students with LD scored especially poor on subtests involvingthe syntacticand rote memory aspects of FL learning. Ganschow,Sparks,and colleagues (17; 18; 21; 33; 54; 55; 56) have carried out a number of studies on high school and college students withFL learning problemswho were havingserious difficulty withnative language learning.3 Analysesof these studies have led them to propose the Linguistic Coding Deficit Hypothesis (LCDH) (57; 59). In theirLCDH, derived from theworkofVellutinoand Scanlon withchildren who have reading disabilities,the authorsspeculated that FL learning problems are linked to native language learning difficultiesin mastering the phonological, syntactic,and/or semantic "codes" of language. Vellutino and Scanlon found that poor readers have particular difficultywith the phonological and syntactic,but not semantic, codes of language.4 Empirical data have been generatedto supportthe LCDH. For example, significantdifferenceshave been found between the native language phonological skillsof good and poor FL learners,whereas these learners have not exhibited semantic differenceson testsof native language vocabulary or reading comprehension (17; 55; 56).5 Sometimes syntactic differences between these learners have been found as well (17). Sparks and Ganschow (57) have speculated that the affectivedifferences(e.g., lack of motivation,

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

44

TheModernLanguageJournal78 (1994)

anxiety)proposed byFL educators as a cause of FL learningproblemsmayinstead be the result of native language learning problems. In their work with studentswith and withoutlearning disabilities, Javorsky,Sparks, and Ganschow found that studentswithLD werejust as motivated to learn a FL as were their nondisabled peers. In a recentlycompleted studywith lowand high-riskstudents and students with LD (52), the same authors found that all three groups experienced equal amounts of anxiety when taking FL classroom tests.As a result of their early studies, Sparks and Ganschow have developed a language-based model of assessment for FL learning problems (16; 53; 56; 59) that includes the administrationof both oral and writtennative language measures and FL aptitude tests to evaluate students' phonological, syntactic,and semantic skills. PURPOSE OF STUDY In thisstudythe authorsexamine differences in FL anxietyand native oral and writtenlanguage skills,and FL anxiety and aptitude for learning a FL among college students identified as high-anxious (HI-ANX), averageanxious (AVE-ANX), and low-anxious (LOANX) in FL classes. A battery of oral and writtennative language instrumentsmeasuring phonological, syntactic,and semantic skills is used to determine students' native language skilllevels.New to theirnativelanguage battery are a testof oral language, the TestofLanguage Edition(TLC-E) (65), and a Competence-Expanded basic memorytest, the Memory for Sentences subtest of the Woodcock-Johnson Psycho-Educational Battery:MemoryCluster(WJPB) (67). These tests were added because of the body of literature indicating that poor language learners (such as poor and disabled readers) have verbal memory and oral language problems. The MLAT (6) is used to measure aptitude for learning a FL and the FLCAS (31) to identifyanxietylevels.In order to quantifyanxiety levels, the authors of this studydevised a scoringmethodfortheFLCAS, described in the Methodsection. Overall correlationsamong the measures are determined. Average FL grades are included in the correlation matrix.6 The followinghypothesesare based on previous studies of at-riskFL learners and the authors'speculation thatstudentswithhigh levels ofFL anxietyhave related nativelanguage difficulties and weaker aptitude for learning a FL than students with low levels of FL anxiety.7

They hypothesizedthatHI-ANX studentswould perform significantlymore poorly than LOANX studentson severaloral and writtennative language measures: all of the phonological measures, one semantic measure (oral language), and the written language measures. They also predictedthattherewould be significant differencesbetween HI- and LO-ANX students on the measure of FL aptitude (MLAT) and the verbal memorysubtest.On the semantic measure of reading comprehension, however, the authors anticipated that there would be no differences,as several previous studies have shownthatthismeasure does not discriminate between good and poor FL learners. Included in theirsubject sample was an AVE-ANX group, whichwas expected to score somewhere in between the HI- and LO-ANX groups. In the correlation matrix, they hypothesized that there would be a significantnegative correlation betweenaverage FL grades and theFLCAS, as previousstudies byHorwitz (29) have shown thathigherlevelsof anxiety,as measured bythe FLCAS, are associated withlower grades. METHOD college stuSubjects.Subjects were thirty-six dents (ten males and twenty-six females) enrolled in introductorySpanish classes at a medium-sized midwesternuniversity.Prior to the selection of subjects for the study, the FLCAS was administered during the seventh and eighthweeks of the semesterto 501 volunteersin twenty-two Spanish classes during their scheduled instructional hour.8 In order to make an accurate estimateof the students'anxietylevels,local normswere established for the FLCAS. This procedure involved determining the extentto whicha givenstudent'sscore deviated froman "ideal" anxiety(ANX) score. The ideal ANX score was determined by counting up the ideal answersfor each question on the FLCAS and calculating the ideal average (mean) for the 501 Spanish-enrolledstudents. items and uses a The FLCAS has thirty-three five-pointratingscale and a forced-choice,balanced design format;the student responds to each question with a single answer: strongly agree; agree; neither agree nor disagree; disagree; stronglydisagree. The ideal answer was either "stronglyagree/agree" or "stronglydisagree/disagree," dependilg upon the direction of the question. (See note 9 forillustrative examples.) Studentswhose mean score was one or more standard deviations below the mean

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Skinner, Ganschow,Sparks,Anderson, Javorshy,

45

TABLE I Studentsin Three AnxietyCategories Numberand Percentof Spanish-Enrolled Male Female Total Percent* AnxietyCategory 48 58 106 30% (HI-ANX) High Anxiety 69 151 43% 82 AverageAnxiety(AVE-ANX) 48 97 45 27% Low Anxiety(LO-ANX) 175 179 354** TOTAL * roundedoffto nearestwholenumber ** This totalrepresentsonlystudentswhoseANX scorefellintothe threeanxietycategories.It does not includestudentswhosescoresfellbetween+.5 to +.99 or -.5 to -.99 (N=147). TABLE II SATs,OverallGPAs,and CognitiveScoresbyAnxietyLevel Group

N

SAT Mean S.D.

HI-ANX AVE-ANX LO-ANX

15 10 11

1075 1069 1140

119.7 106.5 88.8

were judged to be HI-ANX; those whose mean score was .5 below the mean to .5 above the mean werejudged to be AVE-ANX;thosewhose mean score was one or more standard deviations above the mean were judged to be LOANX. Three hundred and fifty-four Spanishenrolled studentswere identified in the three anxietycategories (Table I). Letters were sent to all of the HI-ANX and LO-ANX students and to a random sample of one hundred of the AVE-ANX students.10Because of the extensivetime commitment(three and one-halfto fourhours per student) and the nature of the study (a testingsituation),it was difficult to obtain volunteers; only about a dozen studentsresponded to the initialwritten invitation.Follow-upphone calls to about two hundred students resulted in thirty-six volunteers who offered to participate in the study: fifteenstudents identified as HI-ANX (mean age = nineteen; five males, ten females); ten students identified as AVE-ANX (mean age = nineteen; threemales, seven females);and eleven students identified as LO-ANX (mean age = nineteen; two males, nine females). Mean anxiety level for HI-ANX was -1.58 (range = -1.03 to -2.40); forAVE-ANX,+.16 (range = -.49 to +.46); and for LO-ANX, +1.19 (range = +1.08 to +1.45). Table II showsdemographic information on each group, including SAT scores, overall college GPAs, and scores on a general measure of cognitiveability(IQ), the WoodcockJohnsonPsycho-Educational Battery(WJPB):Brief Scale Cognitive Cluster.A Multiple Analysisof

OverallGPA Mean S.D. 3.0 2.8 3.2

8.3 8.2 4.5

Cognitive Mean S.D. 117 116 123

8.6 9.5 6.8

Variance (MANOVA) procedure showed no significantdifferencesamong HI-ANX,AVE-ANX, and LO-ANX groups on the three measures (Wilks's Lambda = .827; F(6,62) = 1.03; p = n.s.). Oral and writtennativelanguage Instruments. instrumentsincluded measures of phonology, syntax, semantics, and verbal memory. The measure of FL aptitude was the MLAT The measure of FL performancewas the students' overall average of theirFL grades. A brief descriptionof each of these instrumentsand what theymeasure is found in Figure I. The measure of anxietywas the FLCAS (31), one of two instrumentsspecificallydesigned to measure FL anxiety (34).11Items include questions such as "I alwaysfeel that the other students speak the foreignlanguage betterthan I do" and "It frightensme when I don't understand what the teacher is sayingin the foreign language." Its authorshave conductednumerous validityand reliabilitystudieson the instrument (29; 31). They indicate that pilot studies show internalconsistency,consatisfactoryreliability, structvalidity,and test-retest reliability.Furthermore, theyhave found that their measure appears to be ". . . independentof theconfounding effectsof (general) testanxiety"(29: p. 39). For purposes of thisstudy,the students'individual risk factor scores (overall range = 2.40 standard deviations below mean to 1.45 standard deviations above mean) were used. (See Subjectsforexplanation of how riskfactorswere derived and range and mean by group.) Procedure. Testingwas conducted individually

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

46

The Modern Language Journal 78 (1994)

FIGURE I and WhatTheyMeasure AlphabeticalListingof Instruments ModernLanguage AptitudeTest(MLAT) (6):

testsforeignlanguageaptitudein simulated has fivesubtests: PartI (NumberLearning);PartII format; (PhoneticScript);PartIII (SpellingClues);PartIV (Wordsin Sentences);and PartV (PairedAssociates). PartsIII, IV,and V comprisetheShortFormand PartsI-V,theLong Form.The Long Formis used in thisstudy.

NelsonDennyReading Test(N-DEN) (2):

Subtest:testssilentreadingcomprehension in a timedtestformat. Comprehension

TestofLanguage Competence-Expanded Edition (TLC-E) (65):

testsoral expressiveand receptivelanguageproficiency; twosubtestsused in thisstudyare Recreating Sentences(RS), an expressive languagetest,and Figurative Language (FS), a receptivelanguagetest; a RS score. and FS are used in thisstudy. they give together ScreeningComposite

WideRange Achievement Test-Revised (WRAT-R) (Spelling) (32):

testsstudents'performance on writing wordsfromdictation.

Battery(WJPB)(67): Psycho-Educational Woodcock-Johnson

forlanguageand numbers;twosubtests are MemoryforSentences(MS) MemoryCluster:testsmemory and NumbersReversed(NR); togethertheygivea MemoryClusterscore.OnlyMS is used in thisstudy.

Woodcock (WRMT-R) (66): ReadingMasteryTest-Revised

testsstudents'readingability;twosubtestsused in thisstudyare WordAttack(WA),whichtests on readingpseudowords, and WordIdentification (WI), whichtestsabilityto read single performance words;togethertheygivea Basic SkillsCluster(BSC) score.WAand WI are used in thisstudy.

WritingSample*:

T-units(TU): countstotalnumberof mainclausesplus attachedor embeddedsubordinateclauses rulesused accordingto (27). ContextualStyle(CS): testsnumberof punctuationor capitalization a maturity (TOWL-2)(28). ratingscale devisedbytheauthorsof the TestofWritten Language-2 * Writingsampleresultsare notincludedin theMANOVAor in Table III, as one subjectfailedto in note twelve. completethewritingsample.Resultsare describedseparately

or in groupsof twoor threestudents(depending upon the test) over a threeand one-halfto four hour period by the authors.The subjects were invitedto take breaksbetweentestsas needed. A light snack was provided and each studentreceived a ten dollar giftcertificateupon completion of the testing.Gradeswere obtained by permissionof the studentthroughhis/herrecords. AnalysisofData. A Multiple Analysisof Variance (MANOVA) procedure was used to determine whether there would be overall differences in group performanceamong the eight variables, which included: a) Testof Language Competence-ExpandedEdition, TLC-E (Recreating

Sentences [RS] and FigurativeLanguage [FL]

subtests); b) Nelson Denny Reading Test,N-DEN; c) Wide Range AchievementTest-Revised,WRAT-R (Spelling only); d) Woodcock Reading Mastery Test-Revised, WRMT-R (Word Identification

[WI]

and Word Attack [WA] subtests);

(ANOVA) procedure was used to compare HIANX, AVE-ANX, and LO-ANX groups on the variables.A level of .05 was used as the criterion of significance. To reduce the possibility of Type 1 error,a Scheffeprocedure was used in comparing the individualgroup differences.To handle the unbalanced design due to unequal group sizes, the General Linears Models procedure in SAS was used. Because of missing data from one subject, the measures of syntax-a writing sample scored for T-units(TU) and Contextual Style (CS)-were removed from the MANOVA. A briefdiscussionof separateANOVA findingson these measures is provided in note twelve. Pearson Product Moment correlations were used to examine the degree of the relationship among all the variables. Of particular interest here was the correlation between average FL grade and the FLCAS.

e) Woodcock-Johnson Psycho-Educational Battery,

RESULTS

and f) the Modern Language Aptitude Test,MLAT

Results of the MANOVA procedure showed overall significantdifferencesamong HI-ANX, AVE-ANX, and LO-ANX groups on the variables (Wilks' Lambda = .243; F(20,46) = 2.37;

WJPB(Memory for Sentences [MS] subtest);

(Long Form). (See Figure I for explanation of instruments.)In the event that the MANOVA was significant,a one-wayAnalysisof Variance

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Skinner, Ganschow,Sparks,Anderson, Javorshy,

47

TABLE III Means and StandardDeviationson Language and MemoryTestsand FL AptitudebyAnxietyLevel HI-ANX Mean S.D.

Testsand Subtests Semantic TLC-E RecreatingSentences FigurativeLanguage N-DEN Phonological WRAT-R Spelling WRMT-R WordAttack WordIdentification Memory WJPB MemoryforSentences FL Aptitude MLAT: Long Form

AVE-ANX Mean S.D.

LO-ANX Mean S.D.

10.0 11.2 106.9

1.5 1.9 13.8

8.7 10.5 102.0

1.9 1.9 10.8

13.0 13.1 113.5

2.0 1.0 7.6

107.5

10.0

107.3

7.3

112.1

7.5

98.3 100.6

9.4 6.8

99.9 97.4

9.3 10.8

110.3 109.0

3.5 6.4

104.1

10.4

104.2

16.3

114.0

14.5

103.2

12.7

109.0

24.5

117.4

11.3

p < .008). Groups comparisons were then run on the individual measures. Table III presentsmeans and standard deviations on measures of oral and writtennative language (phonology and semantics),memory, and FL aptitude. Results on testsof native language skillsshow significantdifferencesamong HI-ANX, AVE-ANX, and LO-ANX groups on both subtestsof the TLC-E: RS, F(2,33) = 16.61, p < .0001 and FL, F(2,33) = 6.71, p < .004. On these measures both HI-ANX and AVEANX subjects scored significantlylower than LO-ANX students. No differenceswere noted between HI-ANX and AVE-ANX subjects. No significant differences were found on the measure of reading comprehension, N-DEN: F(2,33) = 2.68, p < .08. Findingssupport the authors' hypothesis that significant differences would exist between HI- and LO-ANX students in oral language but not in reading comprehension. These findingswere expected, in light of research on other at-risklanguage populations, in particular, poor readers, which has shown thatstudentswithphonological difficultiesmay also have subtle (or overt) difficulties with speech perception and production.'3 Findings are also consistentwithresearchon FL learning whichsuggeststhatoral communication,in particular,poses problems for at-riskstudents;HIANX and AVE-ANX studentsin this studyhad relativelyweaker oral expression and listening comprehension skillsthan LO-ANX students. Phonological measures reveal significantdif-

ferences among HI-ANX, AVE-ANX, and LOANX studentson both subtestsof the WRMT-R: WA, F(2,33) = 7.64, p < .002; and WI, F(2,33) = 6.08, p < .006. No differenceswere noted on the spelling measure, WRAT-R, F(2,33) = 1.12, p = n.s. In all cases, both HI-ANX and AVE-ANX subjects scored significantlylower than LOANX subjects.Findingspartiallysupportthe authors'hypothesisthattherewould be significant differenceson measures of phonology,in that studentsdifferedin the basic skillsof phonetic analysis of words, measured by the WRMT-R. This findingis similarto previousfindingswith at-riskFL learners in high school (55; 56) and both general at-riskand FL-at-risklearners in college (17; 59). However,the findingof a lack of a significantdifferencein spellingwas unlike previous studies showing that spelling lags significantlybehind in at-riskcollege learners (17; 27; 64). Results on the memorymeasure indicate no significant differences on the WJPB: MS, F(2,33) = 2.03, p = n.s. This finding does not support the authors' hypothesis.Based on their studies of high school at-risknative language learners, the authors had expected to find significantdifferencesin verbal memory. On the foreign language aptitude measure (MLAT: Long Form), results show significant differences:F(2,33) = 4.07, p < .03. HI-ANX subjects scored significantlylower than LO-ANX subjects. No differenceswere found between AVE-ANXand HI-ANX and between AVE-ANX

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

48

TheModernLanguageJournal78 (1994)

TABLE IV CorrelationsamongMeasuresin the Study TLC-RS TLC-FL

TLC-RS TLC-FL ND x .49

x

ND

.44 (.007)

TU

.23 (NS)

.42 (.01) .15 (NS)

CS

-.03 (NS)

WRAT WA

(.002)

.31

TU

CS

WRAT WA

WI

MEM GRADE MLAT FLCAS

x

-.21 (NS)

.11 (NS) .04 (NS)

x .22 (NS)

x .13

(NS)

(NS) (NS) (NS) (NS)

.07

.61

.12

.45 (.005)

.26 (NS)

.38 (.02)

.15 (NS)

x

.005 .49 x (NS) (.002) WI .62 .56 .08 -.01 .42 .50 x .70 (.000) (.01) (.000) (NS) (NS) (.003) (.000) MEM .39 .08 -.13 .20 .06 -.02 x .28 .27 (NS) (.02) (NS) (NS) (NS) (NS) (NS) (NS) GRADE .37 .11 .27 -.01 -.09 .31 .43 -.25 -.15 x (.03) (NS) (NS) (NS) (NS) (NS) (.01) (NS) (NS) MLAT .54 .55 .07 -.03 .12 .59 .50 .50 .10 .47 x (.000) (NS) (.000) (NS) (NS) (.000) (.002) (.002) (NS) (.004) FLCAS -.40 -.08 -.17 .09 -.13 -.45 -.27 -.16 -.32 -.36 -.43 (.02) (.06) (NS) (NS) (NS) (NS) (.006) (NS) (NS) (.03) (.008) For explanationof each Testor Subtest,see FigureI. TLC-RS = Testof Language Competence-ExpandedEditionSubtest:RecreatingSentences TLC-FL = Testof Language Competence-ExpandedEditionSubtest:FigurativeLanguage

x

ND = Nelson Denny

TU = WritingSample,T-units CS = WritingSample,ContextualStyle WRAT = Wide Range Achievement Test:Spelling WA= WoodcockReadingMasteryTest-Revised:WordAttackSubtest WI = WoodcockReadingMasteryTest-Revised:WordIdentification Subtest MEM = Woodcock-Johnson Psycho-Educational Battery:MemoryforSentencesSubtest GRADE = averageforeignlanguagegrade MLAT = ModernLanguage AptitudeTest,Long Form FLCAS = ForeignLanguage ClassroomAnxietyScale and LO-ANX. This findingsupportsthe authors' hypothesisthattherewould be significantdifferences betweenHI- and LO-ANX studentsin aptitude forlearninga FL and is similarto findingson otherat-riskpopulations (17; 55; 56; 59). Pearson Product Moment correlations were run among all the variables (Table IV). The finding of particular interestto this studywas the significantlynegative correlation between anxiety and FL grades. Horwitz (29) found a significant negative correlation of r = -.49, p < .003, N = 35, for beginning Spanish classes

and r = -.54, p < .001, N = 32, for beginning

French classes. The correlationin thisstudywas somewhat lower, though still significant (r = -.36, p < .03, N = 35). Here the overall average of FL grades over the students' entire college FL history (mean = 2.25 courses, range = 1-3

courses) was used as the basis for the correlation, rather than one end-of-semesterfinal grade as in the Horwitz study,which may have accounted for the lower correlation. Other studies,however,have shown considerablevariation in the correlationbetween anxietyand FL grades (37; 40). The authors noted that in the presentstudy,of the total numberof FL grades

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Skinner, Ganschow,Sparks,Anderson, Javorshy, (over severalsemesters)among the HI-ANX subjects, twenty-eight percent (eight out of twentynine grades) wereDs, Fs,and withdrawals, as comto eleven out of (three twenty-seven pared percent grades) of the AVE-ANXsubjectsand four percent (one out of twenty-five grades) of the LOANX subjects.(See Table V forFL GPAsbyanxiwerenot countedin the etylevel.) As withdrawals overallGPA,the2.5 GPAfortheHI-ANXstudents maybe somewhatunderstated. TABLE V FL GPAsbyAnxietyLevel GPA A

B

C D F/Withdraw

HI-ANX

AVE-ANX

(N=15) 4.0 4.0 4.0 3.5

(N=10) 4.0 4.0 4.0 3.5

3.0 3.0 3.0 3.0 2.5 2.5

3.0

1.7 2.0 0 0 W

2.5 2.5 2.5 1.5 .5

LOANX (N=11) 4.0 4.0 3.7 3.7 3.6 3.5 3.5 3.0 3.0 3.0 3.0

DISCUSSION Findingsof thisstudylend support to the authors'overallhypothesisthattherewould be differences in native oral and writtenlanguage and FL aptitude performanceamong students withdifferentlevels of FL anxiety.Though HIANX, AVE-ANX, and LO-ANX students had similarSAT scores,overallgrades,and cognitive ability,theirperformancedifferedon testsmeasuring oral language (listening and speaking), phonological processing (phonetic analysis,single word recognition), and aptitude for learning a FL. In general, LO-ANX studentsdemonstrated superior native language skills and FL aptitude; HI-ANX students exhibited average skillson these same measures. In thisstudystudents were grouped solely on the basis of anxiety level. Yet, among the HI-ANX group, language differenceswere noted thatwere similar

49

to previous studies in which students were grouped according to theirhistoryofsuccess or failure in FL classes (51; 54; 55; 56; 59). What is interestingabout these findingsoverallis thatit is "language variables"-native oral and written language and FL aptitude-that discriminate in HI- fromLO-ANX subjects,despitesimilarities and in overall academic cognitiveability performance in college. The findingssuggestthatHIANX studentsdo not have weak (poor) language skills,as their performanceon native language and FL aptitudemeasureswaswellwithintheaverage range. Rather,LO-ANX students"look relativelystrong"because theiroverallnativeoral and writtenlanguage and FL aptitudeskillslie in the above averageto superiorrange. Previous studies have shown that at-riskFL high school and college students, including those withidentifiedlanguage learning disabilities (LLD), have particular problemswith the phonological code of language. The present studyshowsthatHI-ANX studentsalso differentiated themselves from LO-ANX students on phonological tasks.Upon closer examination of the HI-ANX population, however,the authors found that a number of these students performedwell on the phonological tasksand also did well in FL classes. In order to explore this findingfurther,the authors of the present studydecided to examine more closelythe relationshipbetween phonological performance and grades. To do so, theyregrouped the subjects according to their WRMT-R: Basic Skills Cluster (BSC) score (a combination of two phonological measures) and then compared FL grades. Two groups were formed: those with BSC scores < 100 (N = 13) and those with BSC scores = > 100 (N = 23). Their rationale forusing the phonological measure was that in previous studies atrisk studentsscored particularlypoorlyon this measure (17; 51; 54; 55; 56). Having grouped subjects in thismanner,the authorsnow found significant differencesbetween HI- and LOBSC groups on SAT scores, F(1,34) = 7.31, p < .01, but not on overall college GPA, F(1,34) = 1.35, p = n.s. Importantly,significant differenceswere found in the FL GPAs of the two groups, F(1,34) = 8.21, p < .007. Mean FL grade of the LO-BSC group was 2.2 (range = 0 to 4.0); of the HI-BSC group, the mean FL grade was a fullpoint higher,3.3 (range = 1.7 to 4.0). The mean BSC standard score of the LOBSC group was 94.3, over one standard deviation (nineteen standard score points) below their mean cognitive performance (mean cog-

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

50

TheModernLanguageJournal78 (1994)

nitive score = 113,S.D. = 5.8, range = 105-126). The findingthatstudentswithrelatively low phonology scores tend to do more poorly in FL courses lends furthersupportto earlierfindings on the impactof phonologyon FL performance. In this studythe authorsaddress an area not previouslyexamined in theirresearch on FLoral language. HI-ANX and LO-ANX students were found to exhibitsignificantdifferenceson oral language, as measured by the TLC-E. Again, it should be emphasized that the group differencesare accounted for by the superior performanceof the LO-ANX students,in contrastto the average performanceof the HI-ANX students.A differenceof one standard deviation (fifteen standard score points) on the TLC-E: Screening Composite was seen between the two groups. Evidence reviewedearlier from native language reading literaturesuggeststhat early difficultiesin the primary grades with phonology may have long-termeffectson the skillsunderlyingefficientoral language performance (listeningand speaking). This phenomenon has been termed Matthew Effects (i.e., "rich get richer,poor get poorer") (62). Here it describes a set of circumstancesin which less skilled readers do not read as much as skilled readers, therebylessening theirexposure to vocabulary, grammar, and general knowledge. This lack of exposure, in turn,leads to less welldeveloped skillsin oral expressionand listening comprehension,which are importantskillsfor FL learning. FL researchers have speculated that studentscome to the task of FL learning with preconceived beliefsabout how to learn a language and thatthese beliefscan affecta student's achievement in the FL classroom (30). However, a proponent of the phenomenon of MatthewEffectswould suggestthatthe preconceived beliefsabout language learning of a less skilled student entering a FL classroom have some basis in fact;thatis, the beliefshave been acquired because the student'srelativelyweaker oral and writtennative language skillshave not permittedhim/herto achieve as well in school as his/her peers who have strongernative language skills. (Their beliefs about language learning may often be reinforced by lower grades.) The findingsof thisstudyalso suggest thatthe higherFL anxietylevelsof the HI-ANX and AVE-ANX groups may be related to their relativelyweaker oral language skills. In order to analyze and describe furtherthe relationship between anxiety and FL grades, the authors displayed the grades of the three ANX groups subject by subject (Table V). This

examination of individual FL profiles of subjects shows that among the HI-ANX students, over half received average FL grades in the A and B ranges. The findingsuggeststhatin this studysome of the HI-ANX students,whose overall cognitive,native language, and FL aptitude scores were in the above average range, were good language learners. Giventheabove findings,the authorsspeculate that there mightbe a new and different way to explain the role of anxietyin FL learning.Howback to theorigiever,thiswouldrequirereferring nal sampleofstudentswho tooktheFLCAS. What followshere is a descriptionof our speculations derivedfromthe originaldata used to screenthe populationforthe presentstudy. In a previousstudyGanschow and Sparks (18) had found student perceptions of ease of learning a FL to be one of the best questions for discriminating HI-risk from LO-risk FL learners on a screening instrumentdevised for identifyingFL learning problems. In the present studythe authorswere able to examine the interactionbetween ease of learning a FL and anxiety because theyhad included along with the FLCAS several background questions, one of which was a request for informationabout the student's perceived difficulty with FL courses. In order to examine the relationship between ease of learninga FL and anxiety,they referredback to theiroriginal pool of 501 identified HI-, AVE-, and LO-ANX students.From this pool, theyfirsteliminated all of the AVEANX studentsin order to select only those students at HI- and LO-ANX extremes.Then they identifiedstudentswho indicated on the "ease of learning a FL" question (a one to fivepoint scale) that the learning was either "very to somewhateasy" or "somewhatto verydifficult," i.e., those students at high and low "ease of learning" extremes.Of the initial 501 students, 174 fitthe criteriaof being identifiedas HI- or LO-ANX and as findingFL eitherrelativelyeasy or relatively difficult. Findings (see Table VI) show that close to seventypercent of the studentswho were identified as either HI- or LO-ANX responded as one mightexpect on the ease of learningquestion: those withhigh anxiety found the course difficult (N = 61) and those with low anxietyfound it easy (N = 59). However, a number of students (twenty-three percent,N = 40) foundthe course easy but were stillanxious, and some (7.5%, N = 14) found the course difficultbut were relativelylow in anxiety. In examining the four quadrants in Table VI the authorsspeculate thatthe individualsin

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Ganschow,Sparks,Anderson, Skinner, Javorshy,

51

TABLE VI RelationshipbetweenEase of Learninga FL and Anxiety(N = 174)* Ease of Learning

HI ANX

AnxietyLevel

(= /> -1.00)

veryto somewhateasy somewhatto verydifficult

LO-ANX (= /> +1.00)

N

Percent

N

Percent

40 61

23 35

59 14

34 7.5

*The ease of learningscale had a five-point range: (1) veryeasy; (2) moderatelyeasy; (3) average; (4) somewhatdifficult;(5) verydifficult.Studentswho reporteda (3) on ease of learningwere eliminatedfromthispartof the study,as were studentswhoseanxietyscoreswerein theAVE-ANX range.Of the initial501 students,then,174fitthe criteriaforinclusionin thisdata set. the upper leftquadrant-the twenty-three percent of the 174 who found the course easy but are nevertheless anxious-may be the most likely to benefit from the anxiety reduction seminars advocated by Horwitz and her colleagues (29-31, 34). Those who found learning a FL difficult and were also highly anxious (thirty-five percent of the 174) may be the students for whom psychoeducational and language assessmentsmay be appropriate before proceeding withfurtherFL study. SUMMARYAND RECOMMENDATIONS In thispaper the authorsexamine differences in FL anxiety,native oral and writtenlanguage, and aptitude for learning a FL among three groups of college learners: HI-, AVE-, and LOANX students.They speculate thatHI-ANX students mayexhibitpoorer language skillsand FL aptitude than LO-ANX students. Results support this hypothesis. LO-ANX learners as a group had above average to superior language skills and FL aptitude, with mean standard scores rangingfrom109 to 118. In contrast,HIANX students' mean scores, though still in the average range,were substantiallylower,ranging from97 to 107. Significantdifferencesbetween HI- and LO-ANX studentswere found in oral language, two phonological tasks,and FL aptitude. Despite these significantdifferences,it is clear that there are highlyanxious FL learners whose language skills are commensurate with theiroverallintellectualability.Twenty-five percent of the HI-ANX learners in this study received an A average in FL classes. There are others, however,whose language skills are not commensuratewiththeirintelligence,and most of these students tend to do poorly in FL classes. Among the HI-ANX group, twentyeight percent of the FL grades were Ds, Fs, or

withdrawals.When subjectswere compared according to their performanceon tests of phonology, FL GPAs of the HI-phonology group were found to be significantly higherthan those of the LO-phonologygroup. The mean phonology score of the LO-phonology group was over one standard deviation lower than theirmean cognitiveabilityscore. Level of one's perceived withFL studyalso appeared to be redifficulty lated to anxiety.Since the subject sample was small and representativeof a population of student volunteers at only one university,inferences mustbe interpretedcautiously.However, based on the present findings and results of previous studies, the authors make two recommendations. First,a studentwho expresses anxietyabout FL learning and experiences persistent difficultyin passing FL courses should be referred for a psychoeducational evaluation, which should include tests of oral and writtennative language (phonology, syntax,and semantics) and FL aptitude. Generally,the Office of Handicapped Student Services, the Developmental Office,or the StudentLearning AssistanceCenter has a specialistwho can assistthe studentin the evaluation process. Should a comprehensive language evaluation be advised, instruments that measure oral language (oral expression and listening comprehension) also should be considered. Sample testbatteriesmaybe found in Figure I of thisarticle and in other publications (16; 17; 51; 53; 54; 55; 56; 59). Second, an examination of possible reasons for FL difficulty will result in differentrecommendations for remediation/compensation. Studentswithhigh anxietyand intactlanguage skillsmay be candidates for an anxietysupport group of the type Horwitz,Horwitz,and Cope recommend. Students with high anxiety and subtle or overtlanguage skilldifficultiesmaybe

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

52

The Modern Language Journal 78 (1994)

candidates for compensatory approaches, such as those offered to students with LD. Modest accommodations include untimed tests, modifications in the kinds of tests administered (e.g., written instead of oral or vice versa), and allowances for poor pronunciation, spelling, or grammar.14 More extreme accommodations include provision of special classes for students (10; 12; 16; 58) or FL course waivers/substitutions (11; 14; 19; 43). The use of multisensory structured language approaches focusing on direct teaching of the sounds and symbols of the foreign language is one modification that has shown positive results in preliminary research on its effectiveness, in particular, for students with weak phonological skills (51; 54; 58). The findings from this study suggest that for at least some students, FL anxiety may be related to relative weaknesses in understanding and applying the systems of linguistic codes, and in particular, the phonological code. These phonological problems may be subtle, and students may be able to compensate well for them when performing tasks in their native language. However, when asked to learn a new phonological system (a FL), compensatory mechanisms may break down. Furthermore, these subtle phonological difficulties may also have a negative impact on the students' ability to listen to and speak the FL. In addition to the negative cognitive effects of the phonological difficulty, poor performance in the FL classroom could lead to motivational and anxiety spin-offs, causing even more difficulty with FL learning. While students with FL learning problems may indeed have high levels of anxiety or hold negative beliefs about their language learning or ability to learn a FL, this study provides support for the notion that such affective differences may be due to a history of and current difficulty with the language "codes." The present study provides alternative speculations to existing hypotheses about the complex relationship between anxiety and foreign language performance. It suggests the possibility of identifying "subtypes" of anxious learners-students who are anxious and do well in FL and students who are anxious and do poorly in FL. In general, it would appear that higher FL ability is related to lower anxiety. For those students who are anxious and do poorly in FL classes, further exploration into their basic language skills, in particular, the phonological code, is in order.

NOTES 1 The first two authors, Ganschow and Sparks, made the mostsubstantialcontributionsto the manuscriptand contributedequally in the preparation of this manuscript. 2 Forstandardizedmeasuresofnativeoral and written language used withcollege learners,see Ganschowand Sparks (16) and Sparks,Ganschow,and Pohlman (59). 3 For anecdotal reports about students with LD who have FL learning difficulties,the reader is referred to Ganschow and Sparks (20; 21), Lefebvre, Levine, and Pompian and Thum. 4 A discussion of the phonological difficultiesof poor readersand the lack of semanticdifficultiesmay be found,forexample, in Stanovichand in Spear and Sternberg. does not primarilyreferto the 5 The termphonology abilityto pronounce wordsin eitherthe nativeor foreign language. It mayinclude pronunciation,but specificallyrefersto the abilityto discriminatebetween speech sounds, learn sound/symbolcorrespondences, and identifysound segments (phonemes) in words. The abilityto make explicitreportsabout sound segmentsin words is called "phonemic awareness."It involves a "meta-awareness"of language because one mustbe able to segmentthe phonemes withina word (e.g., the word planthas fivephonemes,p-l-a-n-t). 6 Because one subject dropped the class before receiving a FL grade, FL grades were not included in the overall analysis (MANOVA and individual ANOVAs). AverageFL grades are included in the correlation matrixand are discussed in relation to findings by Horwitz and her colleagues (29; 31). 7 The reader is referred to references 17, 55-57, and 59. 8 Sections fromtwo differentSpanish courses were used for this study.Spanish 101 is an "elementary" language course which is coordinated among all sections, using a common textbook, common syllabus, and withthe exception of a fewminor quizzes, common examinations and a uniformapproach to grading. Spanish 111 (an intensive review of first-year Spanish taken mainlyby freshmen) uses a common textbook in all sections, but does not coordinate other aspects of the course. Testing and grading in Spanish 111may thereforereflecta range of instructional approaches and emphases, withsome instructors favoringgraded oral activitiesand writtenproduction of the language, while others favora more traditionalemphasis on patterned practice for masteringgrammaticalstructuresand vocabulary.Since grades are lateranalyzed in thisstudy,a general word of explanation is necessary.At best, foreignlanguage classroomgrades representan instructor'sevaluation of manydifferentfactors,some of them as subjective as the qualityof classroom participationand "effort." 9 Examples from the FLCAS that illustratea balanced design include the following: "I never feel

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

and Patton Ganschow,Sparks,Anderson, Skinner, Javorshy, quite sure of myselfwhen I am speaking in my foreign language class" (STRONGLY DISAGREE/ DISAGREE is the ideal answer); "I don't worryabout making mistakes in language class" (STRONGLY AGREE/AGREE is the ideal answer). 10Since there was a larger number of AVE-ANX studentsthan HI- or LO-ANX studentsin the subject pool, only 100 AVE-ANX studentswere randomlyselected to receive informationabout the study. 11The other instrumentdesigned specifically to measure FL anxietyis Robert Gardner's FrenchClass AnxietyScale (39). 12 On the syntactic measures (derived from the writing sample; see Figure I), results of individual ANOVAs show no significantdifferencesamong HIANX, AVE-ANX,and LO-ANX subjects on numberof T-units (TU), F(2,32) = .642, p = .53, or on CS, F(2,32) = .481, p = .62. Based on previous findingson good and poor FL learnersin college (17; 18), significant differencesbetween HI- and LO-ANX subjects on both measures were anticipated. 13 Evidence for subtle or overt difficulties with speech perception and production among poor readers may be found, for example, in Catts, Crain, and Mann et al. (41; 42). 14Other ideas for accommodations may be found in such referencesas Bilyeu, Fisher,Ganschow and Sparks (16), and Sparks et al. (53).

53

8. Crain, Stephen. "Why Poor Readers Misunderstand Spoken Sentences." Phonologyand Reading Disability: Solving the Reading Puzzle. Ed. Donald Shankweiler& Isabelle Liberman. Ann Arbor: Univ. of Michigan Press, 1989: 133-65. 9. Daly,John. "Understanding Communication-Apprehension: An IntroductionforLanguage Educators." LanguageAnxiety: FromTheoryand Research to Classroom Implications. Ed. Elaine Horwitz & Dolly Young. Englewood Cliffs,NJ: Prentice-Hall,1991: 3-13. 10. Demuth, Katherine& Nathaniel Smith."The Foreign Language Requirement: An Alternative Program." ForeignLanguage Annals 20 (1987): 67-77. 11. Dinklage, Kenneth. "Inabilityto Learn a Foreign Language." EmotionalProblems oftheStudent.Ed. Graham Burt Blaine & Charles C. McArthur. NY: Appleton-Century,1971: 185-206. 12. Downey, Doris M., Karen Bever & Barbara Hill. "Foreign Language Modification Program." Paper, Orton Dyslexia SocietyAnnual Meeting, Portland,November,1991. 13. Fisher, Elissa. "Learning Disability Specialist Looks at Foreign Language Instruction." Hill4 (1986): 1-3. topSpectrum 14. Freed, Barbara. "Exemptions from the Foreign Language Requirement: A Review of Recent Literature,Problemsand Policy."ADFL Bulletin 18, ii (1987): 13-17. 15. Gajar, Anna. "Foreign Language Learning DisBIBLIOGRAPHY abilities: The Identification of Predictive and Diagnostic Variables." Journalof LearningDis1. Bilyeu,E. E. "Practice Makes Closer to Perfect:Alabilities20 (1987): 327-30. ternativeTechniques forTeaching ForeignLan16. Ganschow,Leonore & Richard Sparks. "'Foreign' guages to Learning Disabled Students in the Language Learning Disabilities: Issues, ReUniversity."Fund for the Improvementof Post search, and Teaching Implications." Successfor Secondary Education Project #116CH 10305. CollegeStudentswithLearning Disabilities. Ed. Susan Vogel & Pamela Adelman. New York: Ellensburg:CentralWashingtonUniv.,1982. 2. Brown,J.,J. M. Bennett & G. Hanna. Nelson-Denny Springer,1993: 283-322. 17. Reading Test.Chicago: Riverside,1981. , Richard Sparks,JamesJavorsky, Jane Pohl3. Carroll,John. "Cognitive Abilitiesin Foreign Lanman & Angela Bishop-Marbury."Identifying Native Language Difficulties among Foreign guage Aptitude: Then and Now." LanguageAptitudeReconsidered. Ed. Thomas Parry& Charles Language Learners in College: A 'Foreign' LanStansfield.Englewood Cliffs,NJ:Prentice-Hall, guage Learning Disability?"JournalofLearning 1990: 1-29. Disabilities24 (1991): 530-41. . "Twenty-Five Years of Research on Foreign 18. & Richard Sparks. "A Screening Instru4., and ment for the Identification of Foreign LanLanguage Aptitude." IndividualDifferences Universalsin Language Learning Aptitude.Ed. guage Learning Problems." ForeignLanguage Karl C. Diller. Rowley,MA: Newbury House, Annals 24 (1991): 383-98. 1981: 83-117. 19. , Bettye Myer & Kathy Roeger. "Foreign 5. . "The Prediction of Success in Intensive Language Policies and Procedures forStudents with Specific Learning Disabilities." Learning Foreign Language Training." Trainingand Research in Education. Ed. Robert Glaser. PittsDisabilitiesFocus 5 (1989): 50-58. burgh: Univ. of PittsburghPress, 1962: 87-136. 20. &Richard Sparks. "The Foreign Lan6. & StanleySapon. ModernLanguageAptitude guage Requirement." LearningDisabilitiesFocus Test(MLAT) Manual. San Antonio,TX: Psycho2 (1987): 116-23. logical Corp., 1959. 21. &Richard Sparks. "Learning Disabilities 7. Catts, Hugh. "Speech Production/Phonological and Foreign Language Difficulties:Deficit in Deficits in Reading Disordered Children."Jourand ListeningSkills?"JournalofReading,Writing, nal ofLearningDisabilities19 (1986): 504-08. International 2 (1986): 306-19. LearningDisabilities

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

54

The Modern Language Journal 78 (1994)

22. Gardner,Robert C., N. Lalonde, R. Moorcroft& F T. Evers. "Second Language Attrition:The Role of Motivation and Use." Journalof Lan6 (1987): 29-47. guage and Social Psychology 23. , R. Clement,P. C. Smythe& C. L. Smythe. TheAttitude/Motivation TestBattery: RevisedManual. Research Bulletin 15. London, Ontario: Language Research Group, Univ. of Western Ontario, 1979. 24. , P. C. Smythe,R. Clement & L. Gliksman. "Second Language Learning: A Social Psychological Perspective."Canadian ModernLanguage Review32 (1976): 198-213. & Wallace Lambert. Attitudeand Motiva25. tion in SecondLanguage Learning.Rowley,MA: NewburyHouse, 1972. 26. Genesee, Fred. "The Role of Intelligence in Second Language Learning." Language Learning 26 (1976): 267-80. 27. Gregg,Noel. "College Learning Disabled Writers: Error Patternsand InstructionalAlternatives." JournalofLearningDisabilities16 (1983): 334-38. 28. Hammill, Donald & Stephen Larsen. TestofWrittenLanguage-2 (TOWL-2). Austin,TX: Pro-Ed, 1988. 29. Horwitz, Elaine. "PreliminaryEvidence for the Reliabilityand Validityof a Foreign Language AnxietyScale." Language Anxiety:From Theory Ed. Elaine and ResearchtoClassroom Implications. K. Horwitz & DollyJ.Young. Englewood Cliffs, NJ:Prentice-Hall,1991: 37-39. 30. . "Attendingto the AffectiveDomain in the Foreign Language Classroom." Shiftingthe InstructionalFocus to the Learner.Ed. Sally S. Magnan. Middlebury,VT: Northeast Conference, 1990: 15-33. 31. , Michael B. Horwitz & Joanne Cope. "Foreign Language Classroom Anxiety." Modern LanguageJournal70 (1986): 125-32. 32. Jastak, Sarah & Gary Wilkinson. Wide Range Achievement Test-Revised (WRAT-R).Wilmington, DE: Jastak,1984. 33. Javorsky, James,Richard Sparks & Leonore Ganschow. "Perceptions of College Students with and without Learning Disabilities about Foreign Language Courses." LearningDisabilities: Researchand Practice7 (1992): 31-44. 34. Language Anxiety:From Theoryand Researchto ClassroomImplications.Ed. Elaine Horwitz & Dolly J. Young. Englewood Cliffs,NJ:PrenticeHall, 1991. 35. Lefebvre, R. Craig. "A Psychological Consultation Program for Learning Disabled Adults." Journal of CollegeStudentPersonnel25 (1984): 361-62. Variationand Learn36. Levine, Melvin.Developmental ingDisorders.Cambridge, MA: Educators, 1987. 37. Madsen, Harold, Bruce Brown & Randall Jones. "EvaluatingAttitudestowardSecond-Language Tests." Language Anxiety:From Theoryand Researchto ClassroomImplications.Ed. Elaine K.

Horwitz & DollyJ.Young. Englewood Cliffs,NJ: Prentice-Hall,1991: 65-86. 38. MacIntyre,Peter & Robert C. Gardner. "Investigating Language Class AnxietyUsing the Focused EssayTechnique." ModernLanguageJournal 75 (1991): 296-304. 39. & Robert C. Gardner. "Methods and Results in the Study of Anxiety and Language Learning: A Review of the Literature." Language Learning41 (1991): 85-117. 40. & Robert C. Gardner. "Anxietyand Second Language Learning: Toward a Theoretical Clarification." Language Learning 32 (1989): 251-75. 41. Mann, Virginia,Elizabeth Cowin & JoyceSchoenheimer. "Phonological Processing, Language Comprehension, and Reading Ability."Cognitiveand BehavioralCharacteristics ofChildrenwith LearningDisabilities.Ed. Joseph Torgesen. Austin, TX: Pro-Ed, 1990: 59-87. 42. , Donald Shankweiler& Suzanne T. Smith. "The Association between Comprehension of Spontaneous Sentences and EarlyReading Ability:The Role of Phonetic Representation."Journal of ChildLanguage 11 (1984): 627-43. 43. Philips, Lois, Leonore Ganschow & Reed Anderson. "The College Foreign Language Requirement: An Action Plan for Alternatives." NACADA (National AcademicAdvisingAssociation)Journal11 (1991): 51-56. 44. Pimsleur, Paul. "Language Aptitude Testing." Language TestingSymposium:A LinguisticApproach. Ed. Alan Davies. New York: Oxford Univ. Press, 1968: 98-106. . PimsleurLanguage AptitudeBatteryand 45. Manual. New York: Harcourt, Brace, 1966. . "PredictingSuccess in High School For46. eign Language Courses." Educational and Psy2 (1963): 349-57. chologicalMeasurement 47. , Donald M. Sunderland & Ruth McIntyre. "Underachievement in Foreign Language Learning." InternationalReviewofAppliedLinguistics2 (1964): 113-50. 48. Pompian, Nancy & Carl Thum. "Dyslexic/ Learning Disabled Studentsat Dartmouth College." Annals ofDyslexia38 (1988): 176-84. 49. Scovel, Thomas. "The Effectof Affecton Foreign Language Learning: A Review of the Anxiety Research." Language Anxiety:From Theoryand Researchto ClassroomImplications.Ed. Elaine Horwitz& DollyJ.Young. Englewood Cliffs,NJ: Prentice-Hall,1991: 15-23. 50. Skehan, Peter. "The Role of Foreign Language Aptitudein a Model of School Learning." Language Testing3 (1986): 188-221. Leonore Ganschow. "The Effects of 51. &MultisensoryStructuredLanguage Instruction on the Native and Foreign Language Aptitude SkillsofAt-RiskHigh School ForeignLanguage Learners: A Replication and Follow-UpStudy." Annals ofDyslexia43 (1993): 194-216.

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

Ganschow, Sparks, Anderson,Javorshy,Skinner,and Patton 52. Sparks, Richard, Leonore Ganschow & JamesJavorsky."Perceptions of Low and High Risk Students and Students with Learning Disabilities about High School Foreign Language Courses. ForeignLanguage Annals (forthcoming). "Di53. , Leonore Ganschow &JamesJavorsky. agnosing and Accommodating the Foreign Language Learning Difficultiesof College Students withLearning Disabilities." LearningDisabilities:Researchand Practice7 (1992): 150-60. 54. , Leonore Ganschow,Jane Pohlman, Marjorie Artzer & Sue Skinner. "The Effectsof a Multisensory,Structured Language Approach on the Native and Foreign Language Aptitude Skillsof High-RiskForeignLanguage Learners." Annals ofDyslexia42 (1992): 25-53. 55. , Leonore Ganschow,JamesJavorsky, Jane Pohlman & John Patton. "IdentifyingNative Language Deficits in High and Low Risk Foreign Language Learners in High School." ForeignLanguageAnnals 25 (1992): 403-18. 56. , Leonore Ganschow,JamesJavorsky, Jane Pohlman & John Patton. "Test Comparisons among Students Identified as High-Risk,LowRisk, and Learning Disabled in High School Foreign Language Courses." ModernLanguage Journal76 (1992): 142-59. 57. & Leonore Ganschow. "Foreign Language Learning Difficulties:Affectiveor Native Language Aptitude Differences?"ModernLanguage Journal75 (1991): 3-16. 58. , Leonore Ganschow, Silvia Kenneweg & Karen Miller. "Use of an Orton-Gillingham Approach to Teach a Foreign Language to Dyslexic/LearningDisabled Students: Explicit Teaching of Phonology in a Second Language." Annals ofDyslexia41 (1991): 96-118. 59. , Leonore Ganschow & Jane Pohlman. "Linguistic Coding Deficits in Foreign Language Learners." Annals ofDyslexia39 (1989): 179-95. 60. Spear, Louise & Robert Sternberg. "An In-

61. 62. 63.

64.

65. 66. 67. 68.

69.

55

formation-Processing Framework for Understanding Reading Disability.,"Handbookof Cognitive,Social, and Neuropsychological Aspectsof Learning Disabilities.Ed. Stephen Ceci. Hillsdale, NJ:Erlbaum, 1987: 3-31. Spolsky,Bernard. Conditions for SecondLanguage Learning.Oxford: Oxford Univ. Press, 1989. Stanovich, Keith. "The Right and Wrong Places to Look for the Cognitive Locus of Reading Disability."Annals ofDyslexia38 (1988): 154-77. Vellutino, Frank & Donna Scanlon. "Linguistic Coding and Metalinguistic Awareness: Their Relationship to Verbal and Code Acquisition in Poor and Normal Readers." Metalinguistic Awarenessand BeginningLiteracy.Ed. Daniel Yaden & Shane Templeton. Portsmouth,NH: Heinemann, 1986: 115-41. Vogel, Susan & Mary Moran. "WrittenLanguage Disorders in Learning Disabled College Students: A PreliminaryReport." Comingof Age: The Best ofACLD. Ed. William Cruickshank & Janet Lerner. Syracuse: Syracuse Univ. Press, 1982: III, 211-25. Wiig, Elisabeth & WayneSecord. TestofLanguage Edition (TLC-E). San AnCompetence-Expanded tonio, TX: PsychologicalCorp., 1988. Woodcock, Richard. WoodcockReading Mastery Tests-Revised(WRMT-R). Circle Pines, MN: American Guidance, 1987. & M. Bonner Johnson. Woodcock-Johnson Psycho-Educational Battery(WJPB).Allen Park, TX: DLM Teaching Resources, 1978. Young,Dolly. "Creating a Low-AnxietyClassroom Environment: What Does Language Anxiety Research Suggest?" ModernLanguageJournal75 (1991): 426-37. . "The Relationship between Anxietyand Foreign Language Oral ProficiencyRatings." Language Anxiety:From Theoryand Researchto ClassroomImplications. Ed. Elaine K. Horwitz & Dolly J. Young. Englewood Cliffs,NJ:PrenticeHall, 1991: 57-63.

This content downloaded from 141.210.2.78 on Sun, 14 Apr 2013 20:11:36 PM All use subject to JSTOR Terms and Conditions

Suggest Documents