This article was downloaded by: [University of Illinois at Urbana-Champaign]On: 04 October 2014, At: 17:24Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK

Reading PsychologyPublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/urpy20

THE ACQUISITION OF GRAPHEME‐PHONEMECORRESPONDENCES IN A WORD LEARNING TASKBarbara Fox a & Robin Baker ba North Carolina State Universityb National Testing Service , Durham, North CarolinaPublished online: 10 Jul 2006.

To cite this article: Barbara Fox & Robin Baker (1980) THE ACQUISITION OF GRAPHEME‐PHONEME CORRESPONDENCES IN AWORD LEARNING TASK, Reading Psychology, 1:3, 156-164

To link to this article: http://dx.doi.org/10.1080/0270271800010305

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) containedin the publications on our platform. However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of theContent. Any opinions and views expressed in this publication are the opinions and views of the authors, andare not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon andshould be independently verified with primary sources of information. Taylor and Francis shall not be liable forany losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use ofthe Content.

This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

READING PSYCHOLOGY Vol. 1 No. 3 Summer 1980

THE ACQUISITION OF GRAPHEME-PHONEME CORRESPONDENCES IN A WORDLEARNING TASK

Barbara Fox, North Carolina State University; Robin Baker, National TestingService, Durham, North Carolina

ABSTRACT

The relationship between reading ability with receptive vocabulary level andthe ability to induce grapheme-phoneme relationships was investigated usinggrade one students. The ability to induce relationships was assessed by aword learning task in which the stimuli were six consonant-vowel-consonanttrigrams composed of seven letter-like forms with a 1:1 grapheme-phonemecorrespondence. The word learning task was followed by a test of each stu-dent's acquisition of the grapheme-phoneme correspondences that appeared inthe word stimuli. The ability to induce the relationships was found to bemuch more strongly related to reading ability than receptive vocabulary.The results suggest that good and poor readers applied different word learningstrategies, with good readers appearing to apply more readily a principledsolution and poor readers an associative solution. The results are consistentwith the theory of reading acquisition advanced by Gibson and Levin.

INTRODUCTION

For the child learning to read a language such as English, in which there issome degree of correspondence between letters and spoken sounds, the identifi-cation of unknown words involves the transfer of learned associations betweenpatterns of graphemes and phonemes to letter combinations within unknownwords. According to the model of reading acquisition elaborated by Gibson andLevin (1975), the child at an early stage of reading learns simple grapheme-phoneme correspondences and, as reading proficiency improves, learns complexgraphemic patterns which have an invariant pronunciation within particularorthographic contexts. While the child may not be able to state the ruleswhich govern learned grapheme-phoneme patterns, mastery is demonstrated throughthe recognition of letter patterns, and through the transfer of this knowledgeto the decoding of unfamiliar words. In an investigation using children ingrades two through six, Tovey (1980) found that children were able to applygrapheme-phoneme generalizations, but were unable to explain the meaning ofterms used to describe grapheme-phoneme correspondences.

According to Gibson and Levin's analysis, grapheme-phoneme patterns are learnedby the combined presentation of graphemic and phonemic stimuli, and need notentail the specific teaching of letter-sound relationships. In an investiga-tion of word mediation and generalization, Schworm (1979) found that acceler-ated readers at the beginning of grade one who had not received formal instruc-tion in grapheme-phoneme associations were able to pronounce the major soundin vowel-vowel and vowel-consonant patterns. Hillerich (1976) interpretedfirst grade children's ability to recognize nonsense syllables to mean that,while specific instruction resulted in significantly better grapheme-phonemeknowledge, children can develop some knowledge of symbol-sound associationsin the absence of direct instruction. Ehri and Wilce (1979) have shown thatorthographic structure has mnemonic value for beginning readers, and thatbeginning readers can spontaneously use alphabetic images to learn sounds.

Gough (1972) has suggested two possible strategies children can adopt whenconfronted with a word learning task. First, the child could adopt the use

156

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

of a principled strategy which would involve the induction of the grapheme-phoneme correspondence structure in the words presented. For example, regard-less of whether the word is "apt," " p a V or "tap," the child will learn toassociate sounds with the letters "a," "p," and "t." Given one of these threewords, the child should be able to pronounce the other two the first time thatthey are presented. Second, the child could adopt the use of an associativestrategy, which would involve the rote learning of whole words. In the exam-ple just described, the child will learn to associate sounds with the lettercombinations "apt," "pat," and "tap." Given one of these three words, thechild should have no idea about how to pronounce the other two the first timethey are presented.

According to Gough's interpretation, when presented with a complex word learn-ing task in which the application of a principled strategy afforded a solutionto word learning which was simpler, or at least as simple as, an associativesolution, the child would use the more economical, principled solution. Pre-sumably, beginning readers induce the grapheme-phoneme correspondence structureof written language when it provides a learning advantage.

The present study investigated the relationship between reading achievement andthe ability to induce grapheme-phoneme associations. If the inductive learningof grapheme-phoneme correspondences is as crucial a skill as Gibson and Levinsuggest, children classified as good readers should be more proficient thanpoor readers at generalizing letter-sound information from whole-word presenta-tions. Word learning for children using a principled strategy should be moreefficient than for children using an associative strategy, since the principledsolution makes use of redundant grapheme-phoneme information contained in thewords presented.

METHOD

Subjects

The subjects were twenty good and twenty poor readers in grade one, with eightgirls and twelve boys in each group. The children were classified as good orpoor readers according to the results of the Metropolitan Achievement Test,Primary I (MAT) (Durost, et al., 1970). Table 1 shows means and standarddeviations for subjects classed by sex and reading ability for MAT Word Know-ledge and Total Reading scores expressed in grade equivalents. This classifi-cation also agreed with teacher judgment.

Parental occupation was measured as one indicant of socioeconomic status usingthe occupational rating scale developed by Hollingshead (1957). According tothis seven-point scale, category 4 includes sales workers and technicians andcategory 5 includes skilled manual workers. The Peabody Picture VocabularyTest (PPVT) (Dunn, 1965) was used as a measure of receptive or aural vocabulary.Table 1 shows group means and standard deviations for parental occupation,chronological age, and IQ and Mental Age scores from the PPVT.

Procedure

To assess grapheme-phoneme acquisition strategy each child was presented witha word learning task followed by a grapheme-phoneme association task. Thetasks were presented individually during the spring in one thirty-minute ses-sion. The word learning task involved the presentation of six visually unfa-miliar words written in an artificial orthography (Gibson, Gibson, Pick, &

157

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

Table 1GROUP MEANS AND STANDARD DEVIATIONS FOR CHRONOLOGICAL AGE,

RECEPTIVE VOCABULARY, PARENTAL OCCUPATION, AND READING ACHIEVEMENT

Hollingshead ScaleChronological AgePPVTb IQPPVT Mental AgeMATC Total Reading

MAT Word Knowledge

Male

Mean

4.8

6.8

101.2

82.6

1.31.2

Poor

S.D

1.

.

6.

7..

•

•

1

4

7

8

12

Readers

Female

Mean S.D

4.9 1.

6.8

93.9 6.

77.1 11.

1.3

1.2

•

4

5

9

61

2

Good

Male

Mean

4.3

6.9

103.7

88.3

2.1

2.1

S.D

1.

.14.

16.

•

Readers

•

2

3

6

2

2

3

Female

Mean

4.6

7.0

107.4

97.0

2.12.4

S.

1

11

16

D.

.3

.3

.8

.8

.3

.8

Measure of Parental OccupationPeabody Picture Vocabulary Test; a measure of Receptive Vocabulary

Metropolitan Achievement Test

Table 2WORDS, SOUNDS, AND LETTER-LIKE FORMS

USED IN LEARNING TASK

Letter-Like Forms Words

Sun

Run

Fun

Sat

Rat

Fat

Single Letter-LikeForms

Id

HV

Sounds

S as in Sun

U as in Sun

N as in Sun

F as in Fun

R as in Run

A as in Rat

T as in Rat

158

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

Osser, 1962). The words consisted of consonant-vowel-consonant trigrams usinga total of seven letter-like forms with 1:1 grapheme-phoneme correspondences.Table 2 shows the words presented for the word learning task.

First, each word was shown to the child and pronounced. The child was askedto say the word (thus ensuring that the child could hear and repeat the words),and then the experimenter used each word in a sentence. Next, the words wereshown to the child, one wor*d at a time, in a fixed random order using a pairedassociate anticipation method. When a correct response was given, the childwas told he/she was right; errors were immediately corrected. Each word wasprinted on a 5 X 8" card. The presentation of all six words constituted onelearning trial. The words were presented to a learning criterion of two con-secutive, errorless trials or to a maximum of twenty trials.

Following the word learning task, a grapheme-phoneme association task was pre-sented. The child was asked to associate a sound with each of the sevenletter-like forms comprising the words. Table 2 shows the seven letter-likeforms. The letter-like forms, printed individually on 5 X 8" cards, wereshown in isolation. Each form was shown once and the child was asked to saythe sound associated with the form. The ability to provide sounds associatedwith the seven graphemes required the induction of these relationships fromthe word learning task. To illustrate this grapheme-phoneme association task,the child was first asked to respond with the appropriate sound for each oftwo conventional letters of the alphabet.

Six dependent measures were used to assess how well the children had inducedthe seven grapheme-phoneme correspondences. All measures were felt to reflectthe nature of the strategy used (i.e., to what extent was the strategy eitherprincipled or associative). The measures were: the number of letter-soundcombinations learned as shown by the grapheme-phoneme association task; thenumber of trials to criterion for word learning; the number of errors tocriterion for word learning; the number of errors during word learning in whichthe child responded with the correct initial consonant; the number of errorsduring word learning in which the child responded with the correct phonogram;and, the number of errors during word learning which were unrelated to thegrapheme-phoneme patterns in the words presented.

RESULTS

Differences with respect to parental occupation, chronological age, receptivevocabulary, written vocabulary, and reading achievement were tested for statis-tical significance using analysis of variance and group membership and sex asthe independent variables. No effects attributable to interaction or sex werefound. As expected, the differences between good and poor readers on writtenvocabulary and total reading were significant at the .001 level, while thedifferences with respect to age and parental occupation were nonsignificant(p-values of .302 and .371, respectively). In addition, significant differ-ences with respect to receptive vocabulary did exist between the two groupsat the .049 level for PPVT IQ and at the .001 level for PPVT Mental Age.

The group means and standard deviations for the six measures of grapheme-phoneme association are presented in Table 3. As Table 3 shows, none of thepoor readers were able to demonstrate mastery within twenty trials. Nonsig-nificant multivariate effects were found for sex and for interaction. Themultivariate F-test for differences between the two reading groups was signifi-cant at the .001 level. In addition, univariate analyses of variance were

159

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

Table 3GROUP MEANS AND STANDARD DEVIATIONS

ON ALL DEPENDENT MEASURES

Poor Readers Good Readers F(l,36) eta

Male Female Male Female

Mean S.D. Mean S.D. Mean S.D. Mean S.D.

Number of SoundsLearneda 1.3 2.0 1.1 1.1 5.7 .9 6.3 .7 118.464* .762Trials toh

Criterion0 20.0 0.0 20.0 0.0 11.9 5.4 10.1 5.8 50.892* .579Errors toCriterion 82.8 13.4 71.4 11.5 22.1 15.2 15.8 13.7 183.120* .819Errors InitialConsonant Correct 26.3 13.6 28.8 12.3 9.4 5.6 7.8 6.6 32.941* .475Errors PhonogramCorrect 27.6 15.0 24.4 11.4 8.6 7.1 5.4 5.1 31.258* .459Errors Unrelatedto GraphemicPattern 28.9 16.6 21.3 12.5 4.3 4.3 2.6 5.4 39.148* .505

Seven Sounds PossibleTwenty Trial Maximum*p<.001

performed to test for group differences with respect to each of the six measures.Children considered good readers performed significantly better than poor read-ers with respect to all measures of grapheme-phoneme association, with the uni-variate F (df=l,36) significant at the .001 level for each measure. In addi-tion, the membership in the two reading groups accounted for atpleast 45 percentof the variance in each of the measures as reflected in the eta column inTable 3; for some measures the percentage was in excess of 75. Clearly, thegroup differences were substantial.

Therefore, it was clear that reading ability was associated with both recep-tive vocabulary and grapheme-phoneme associations. In order to assess theirrelative impact on reading ability, commonality analysis (Cooley & Lohnes,1976) was used. This approach employs stepwise multiple regression to identifythe amount of variance in the dependent variable that is uniquely accounted forby sets of independent variables as well as the amount of variance jointlyaccounted for among all the independent variables. Three dependent measureswere analyzed: group membership, which represents an assessment of the child'sreading ability on a two point scale; MAT Word Knowledge, a measure of written

160

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

vocabulary; and MAT Total Reading, a composite measure of written vocabulary,knowledge of sound-letter relationships, and comprehension of written material.

The Word Knowledge subscale on the MAT measures knowledge of written vocabularyby first presenting a pictorial representation of an object and then requiringselection of the written (or graphemic) representation of that object. Pre-sumably, performance of this task required the child to complete two linksof a chain: first, make the connection between an element of oral vocabularyand the picture, and second, connect the oral vocabulary element with itswritten form. The progression, then, is from a visual representation to anaural representation to a written representation.

The PPVT IQ and Mental Age scores provided a measure of receptive vocabularywhich assessed the first link by requiring the child to select a picture repre-senting a spoken word. To more accurately assess the grapheme-phoneme linkand avoid confounding effects of high or low intelligence, only children witha PPVT IQ between eighty-five and one hundred fifteen were included in thisstudy.

The six scores from the artificial orthography tasks provided grapheme-phonemeassociation measures. These measures assessed the second link by ascertainingthe extent to which the child used a principled or associative strategy todevelop grapheme-phoneme associations.

The unique as well as the joint proportions of variance accounted for arepresented in Table 4, along with the F-ratios to test the significance of theunique effects. As indicated in the table receptive vocabulary may be asso-ciated with reading ability; however, its impact relative to that of grapheme-phoneme association is minimal. Furthermore, even the joint impact of thereceptive vocabulary and grapheme-phoneme association measures is relativelysmall.

DISCUSSION

The results of the present study lend support to Gibson and Levin's hypothesesregarding the conditions necessary for learning grapheme-phoneme patterns andthe application of letter-sound knowledge to word learning. Children learnedspecific letter-sound associations by the combined presentation of graphemicand phonemic stimuli in the absence of formal instruction in letter-soundrelations. The six word list presented a learning task of sufficient complex-ity that the induction of grapheme-phoneme structure by good readers, as com-pared with rote whole-word learning, would provide a learning advantage. Theperformance of good readers on the learning task is consistent with Gough'sinterpretation of the application of word learning strategies by beginningreaders.

Good readers learned the word list in fewer trials and errors and demonstratedknowledge of significantly more letter-sound associations than poor readers.This finding may be explained by the ability of good readers to apply moreeffectively a principled solution for word learning. A comparison of errorsduring word learning lends further support to this interpretation. Childrenclassified as good readers made significantly fewer errors unrelated tographeme-phoneme patterns in the words presented than poor readers. Apparentlygood readers were abstracting, and then applying grapheme-phoneme associationsto the recognition of the words presented. The performance of poor readers on

161

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

Table 4COMMONALITY ANALYSIS FOR READING WITH RECEPTIVE

VOCABULARY AND GRAPHEME-PHONEME ASSOCIATION

Proportion ofTotal VarianceAccounted For

Receptive VocabularyUnique Proportion

Receptive VocabularyF(2,30)

Grapheme-PhonemeAssociation

Unique Proportion

Grapheme-PhonemeAssociation

F(7,30)Proportion ofJoint VarianceAccounted For

Good/PoorGroup

.938

.005

1.193

.780

53.917*

.153

Reading Measure

MAT WordKnowledge

.636 '

.020

.824

.488

5.746*

.128

MAT TotalReading

.836

.008

.732

.681

17.796*

.127

*p<.001

the criteria measures suggests that these children may have applied an associa-tive rather than a principled solution.

The acquisition of a written vocabulary is hypothesized to be a process re-quiring prior knowledge of aural vocabulary and letter-sound associations.While the results of this study indicate that both components are associatedwith reading ability, the set of grapheme-phoneme association measuresuniquely accounted for over 75 percent of the total variance explained by allof the measures. The unique variance explained by receptive vocabulary wasnegligible, but what is somewhat surprising is that the joint variance ex-plained was so low. It does not seem unreasonable to speculate that bothreceptive vocabulary and strategy would be strongly related to some underlyingvariable. These results indicate that while such an underlying variable maybe present and accounting for up to 25 percent of the variance, to a greatextent the nature of the strategy adopted seems to be independent of such avariable. Not only does strategy seem to account for most of the explainedvariance in the measures of reading, but it is also tied to a different set ofunderlying variables than those that affect receptive vocabulary.

However, the use of common trigrams as stimulus words may have provided anopportunity for children, in particular good readers, to apply a principled

162

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

strategy through the use of spelling knowledge as a mediator, In this case,children with spelling knowledge could have associated the artificial orthog-raphy with the sounds associated with traditional orthography.

The importance of the ability to apply a principled solution to word learningmay also be related to the type of reading instructional program used. Thechildren in the present study were placed in an instructional program whichemphasized the importance of learning and transferring grapheme-phonemeassociations to reading. The work of Richek (1977-78) and Barr (1975) supportthe hypothesis that word learning strategy and instructional approach arerelated. It is possible that children placed in an instructional programemphasizing whole word association'learning with little phonics instructionmay not show the same differences in performance. Therefore, a further testof the studied hypotheses should include instructional approach as an addi-tional variable.

The generalizability of these results may be questioned in view of the researchby Knafle and Legenze (1978) because an artificial orthography was used in theword learning and grapheme-phoneme association tasks. However, the primarypurpose of the tasks in the present study was to assess the competency ofchildren in developing strategies for learning grapheme-phoneme associations.The results have shown that possession of an efficient strategy is stronglyassociated with good achievement in reading. It is questionable whether ornot the nature of the orthography would have much impact on assessing theefficiency of children's strategy, consequently the results may be independentof the orthography used.

This research leads to a recommendation regarding the diagnosis of a child'sword learning ability. A task such as the one described in this study can beused to assess the adequacy of the grapherne-phoneme induction strategy. Lowachievement on written vocabulary tests may be attributable to an insufficientlevel of acquisition of an aural vocabulary, insufficient acquisition ofgrapheme-phoneme correspondences, or both. Each, therefore, must be assessedseparately to arrive at an adequate assessment of the problem.

REFERENCES

Barr, Rebecca. The effect of instruction on pupil reading strategies. ReadingResearch Quarterly, 1974-75, 4, 555-582.

Cooley, W. W., & Lohnes, P. Evaluation research in education. New York:Irvington Publishers, Inc., 1976.

Ehri, L. C., & Wilce, L. The mnemonic value of orthography among beginningreaders. Journal of Educational Psychology, 1979, 71, 26-40.

Gibson, E. J., & Levin, H. The psychology of reading. Cambridge, Mass.:MIT Press, 1975.

Gibson, E. J., Gibson, J., Pick, H., & Osser, J. A developmental study ofthe discrimination of letter-like forms. Journal of Comparative andPhysiological Psychology, 1976, 6, 897-906.

Gough, P. B7 One second of reading. In J. Kavanagh & I. Mattingly (Eds.),Language by ear and by eye. Mass.: Massachusetts Institute of Technology,1972.

Hillerich, R. L. Vowel generalizations and first grade reading achievement.The Elementary School Journal, 1976, 47 246-250.

Knafle, J. D., & Legenza, A. External generalizability of inquiry involvingartificial orthography. American Educational Research Journal, 1978, 15,331-347.

163

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4

Fox and Baker: Grapheme-Phoneme

Richek, M. A. Readiness skills that predict initial word learning using twodifferent methods of instruction. Reading Research Quarterly, 1977-78,2, 200-222.

Schworm, R. W. Word mediation and generalization in beginning readers.Journal of Reading Behavior, 1979, 11, 139-151.

Tovey, D. R. Children's grasp of phonics terms vs. sound-symbol relationships.The Reading Teacher, 1980, 33, 431-437.

TEST REFERENCES

Dunn, L. M. Expanded Manual: Peabody Picture Vocabulary Test. Minneapolis,Minnesota: American Guidance Service, 1970.

Durost, W. N., Bixler, H. H., Wrightstone, J. W., Prescott, G. A., & Balow,I. H. Metropolitan Achievement Tests. New York: Harcourt BraceJovanovich, 1970.

Hollingshead, A. B. Two-Factor Index of Social Position. New Haven, Conn.:Mimeo, Author, 1957.

When you are old and grey and full of sleep,And nodding by the fire, take down this book.And slowly read, and dream of the soft lookYour eyes had once, and of their shadows deep;How many loved your moments of glad grace,And loved your beauty with love false or true,But one man loved the pilgrim soul in you,And loved the sorrows of your changing face.And bending down beside the glowing bars,Murmur, a little sadly, how Love fledAnd paced upon the mountains overheadAnd hid his face amid a crowd of stars.

William Butler Yeats (1865 - 1939)

164

Dow

nloa

ded

by [

Uni

vers

ity o

f Il

linoi

s at

Urb

ana-

Cha

mpa

ign]

at 1

7:24

04

Oct

ober

201

4