predictive influence of phonological processing, morphological/syntactic skill, and naming speed on...
TRANSCRIPT
Brain and Cognition 55 (2004) 368–373
www.elsevier.com/locate/b&c
Brain and Cognition 55 (2004) 368–373
Predictive influence of phonological processing,morphological/syntactic skill, and naming speed
on spelling performanceq
Monique Plazaa,* and Henri Cohenb,*
a Laboratoire Cognition et D�eveloppement, Paris, Franceb Centre de Neurosciences de la Cognition, UQAM, Montreal, Canada
Accepted 12 February 2004
Available online 5 May 2004
Abstract
This paper focuses on the predictive influence of phonological awareness, morphological/syntactic skill, and naming speed on
spelling. The retrospective study correlated spelling performance in a group of 199 French-speaking children at the end of grade 2
with earlier capacities for phonemic manipulation, morphological/syntactic correction, and naming speed, assessed at the end of
grade 1. The results are consistent with an integrative model that challenges the unitary phonological disorder hypothesis and
confirmed that in French, as in other languages, naming speed is an independent predictor of reading performance.
� 2004 Elsevier Inc. All rights reserved.
Keywords: Reading; Spelling; Naming speed; Phonology; Morphological/syntactic skill; Longitudinal study
1. Introduction
An earlier study of a group of 267 children at the end
of grade 1 established that phonological awareness,morphological/syntactic skills, and processes underlying
naming speed strongly correlate with reading and
spelling (Plaza & Cohen, 2003). While phonological
awareness was the most potent factor, naming speed,
and morphological/syntactic skill accounted for a sig-
nificant proportion of variance above and beyond
phonological awareness. Each of the three variables
correlated with written language independently of thecontribution made by the other two.
The principal question addressed in this paper is
whether the contemporaneous correlation observed at
the end of grade 1 would remain significant from grade
1 to grade 2. The course of reading acquisition is
qA longitudinal study from grade 1 to grade 2. This research was
aided in part by a grant from SSHRC.*Corresponding authors.
E-mail addresses: [email protected] (M. Plaza),
[email protected] (H. Cohen).
0278-2626/$ - see front matter � 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bandc.2004.02.076
variable, depending on a range of factors such as
cognitive maturity, the family�s social and economic
status, motivational features, and instructional and
teaching methodology (Plaza, Chauvin, Lanthier, &Rigoard, 2002; Plaza & Touzin, 2002). At the end of
grade 2, children have benefited from two years of
reading and spelling instruction. If difficulties observed
at the end of grade 1 persist at the end of grade 2, the
diagnosis of written language disorder is stronger. The
longitudinal dimension allows one to determine whe-
ther correlated variables can be assigned predictive
status.The secondary question was whether in French, as in
English, German, Finnish, Dutch, and Spanish, naming
speed is a predictor of reading and spelling performance,
independent of phonological awareness (Wolf et al.,
2002). Like any other linguistic task, naming speed ev-
idently involves accessing a phonological code. Never-
theless, the processes underlying speed also require
attentional, perceptual, conceptual, memory, lexical,and articulatory processes (Greig Bowers & Newby-
Clark, 2002; Wolf & Greig Bowers, 1999; Wolf & Obr-
egon, 1992, 1997).
M. Plaza, H. Cohen / Brain and Cognition 55 (2004) 368–373 369
To answer these two questions, a group of 254children was given three spelling tasks at the end of
grade 2. The spelling tasks tapped grapheme–phoneme
correspondence rules (pseudo-word dictation), lexical
knowledge (real word dictation), and integrated ortho-
graphic skill (text dictation). Of the 254 children, 76.7%
obtained average scores, 11.8% obtained weak scores
()2 SD) and 11.4% obtained very weak scores. From
this group of children, 199 had been individually testedwith five tasks requiring phonological awareness, mor-
phological/syntactic skill, and naming speed at the end
of grade 1. The study presented here compares the
performance of the three groups of children (average,
weak, and very weak spellers), establishes a correlation
between spelling in grade 2 and phonological, morpho-
logical/syntactic and naming speed skills in grade 1, and
conducts a series of multiple hierarchical regressionanalyses.
2. Method
2.1. Subjects
A group of 254 French-speaking children was givencollective spelling tasks in June 2002 at the end of grade
2; 199 children in the group had been individually as-
sessed at the end of grade 1 in June 2001.
2.2. Procedure at the end of grade 2
The spelling tasks collectively administered at the end
of grade 2 consisted of a pseudo-word spelling task (20items), a word spelling task (20 items), and a text dic-
tation.
(a) The pseudo-word spelling task requires mastery of
an assembly strategy: phoneme–grapheme corre-
spondence rules and sequential processing. The
items involved simple graphemes, contextual gra-
phemes which vary according to the vocalic context
(/g/, /s/, /c/), and complex graphemes (/ouil/, /on/, /ou/, /eu/, /eil/, /in/, /ille/). Subjects� responses were
scored 1 or 0, for correct and incorrect, respectively.
(b) The word spelling task requires mastery of both an
assembly strategy and an addressed lexical strategy.
The items were regular, except for one (femme,
‘‘woman’’) and involved simple, contextual, and
complex graphemes. Subjects� responses were dou-
bly scored, for (1) phonetics (1 point if phoneme/grapheme correspondences were respected) and (2)
usage (1 point if the word was correctly spelled).
(c) The text dictation task requires the assembly strat-
egy, the addressed strategy, and morphological
knowledge: word boundaries, plural, and gender.
Each word received a different score for phonetics,
usage, and grammar.
Five scores were calculated: (1) grapheme–phonemecorrespondence (GPC) in pseudo-words (maximum
score ¼ 20); (2) GPC in isolated words (maximum score
¼ 20); (3) lexical usage in isolated words (maximum
score ¼ 20); (4) global phonetics, involving GPC mas-
tery in all three tasks (maximum score ¼ 72); and (5)
global usage, involving the lexical orthographic use of
words in the real word spelling and text dictation tasks
(maximum score ¼ 44).
2.3. Procedure at the end of grade 1
The tasks administered at the end of grade 1 included
individual assessment of reading, spelling, auditory
verbal memory, oral comprehension, phonological
processing, morphological/syntactic skill, and naming
speed. This paper focuses only on phonological aware-ness, morphology, and naming speed, which were the
variables with the strongest influence.
Phonological awareness was assessed by means of an
initial phoneme deletion task (12 items) (Chevrie-
Muller & Plaza, 2001). The child was asked to delete
the initial phoneme of pseudo-words and to pronounce
the resulting pseudo-words (e.g. /pouk/ without /p/
becomes /ouk/). The list included 6 monosyllabic and 6disyllabic pseudo-words, with 9 vowels and 3 conso-
nants in the initial position. The experimenter gave
two examples. The child had to repeat each item be-
fore giving a response. Subjects� responses were scored
1 or 0.
Morphological/syntactic skill was assessed by means
of a judgment/correction task used in previous studies
(Plaza, 2001; Plaza & Cohen, 2003). The child was askedto listen to 12 sentences and required (a) to decide
whether or not each sentence is grammatical and (b) to
correct any mistakes. Errors in these sentences con-
cerned determiners, subject–verb agreement, gender,
adverbs, prepositions, conjunctions, and word order.
Subjects� responses were scored 1 or 0. Correct responses
involved (a) correct judgment of grammaticality, and (b)
production of the expected responses (substitution ofone word in 11 sentences, and rearrangement of word
order in one sentence).
Naming speed was assessed using three tasks involv-
ing 48 stimuli (pictures, digits, and letters), derived from
the RAN Test (Denckla & Rudel, 1974).
Task 1: Picture naming. The child was asked to rap-
idly name a visual array of 48 stimuli, consisting of
four pictures (glass, dog, heart, and bed) that are pre-sented 12 times in random order.
Task 2: Digit naming. The child was asked to rapidly
name a visual array of 48 stimuli, consisting of four
digits (7, 2, 1, and 9) that are presented 12 times in
random order.
Task 3: Letter naming. The child was asked to rapidly
name a visual array of 48 stimuli, consisting of four
370 M. Plaza, H. Cohen / Brain and Cognition 55 (2004) 368–373
letters (U, A, C, and B) that are presented 12 times inrandom order.
For each naming speed task, a percentage score was
calculated:
Number of correctly named items
Total time� 100
Inasmuch as strong correlations between the three taskshad been observed in previous studies, a global RAN
score was calculated, including picture, digit and letter
naming.
3. Results
3.1. Children’s spelling skills at the end of grade 2
Table 1 shows that the three groups exhibited sig-
nificantly different performances, both for the assembly
strategy and for the addressed strategy.
The average spellers displayed GPC mastery in 78.5%
of the pseudo-words, in 84% of the isolated words and,
globally, in 83.7% of the words in the three spelling
tasks. On the other hand, they respected lexical ortho-graphic usage in 66.8% of real words.
The weak spellers exhibited GPC mastery in 56.5% of
the pseudo-words, in 59% of the isolated words and,
globally, in 63.8% of the words in all three tasks. They
respected lexical orthographic usage in 34.7% of real
words.
The very weak spellers exhibited GPC mastery in
34.5% of the pseudo-words, in 39% of the isolated wordsand, globally, in 41.1% of the words in the three tasks.
They respected lexical orthographic usage in 21.8% of
real words.
Statistically speaking, the most significant differences
between the three groups concern the GPC mastery as-
sessed in the three spelling tasks (F ¼ 251, p < :0001).
3.2. Early performance at the end of grade 1 on
phonological awareness, naming speed, and morphologi-
cal/syntactic skill
The three groups of children were compared based on
the three tasks they had been given one year before.
Table 1
Spelling skills at the end of grade 2 in the average, weak, and very weak spe
Pseudo-word dictation Word us
Average spellers (N ¼ 159) 15.7 (2.8) 12.8 (3.6
Weak spellers (N ¼ 23) 11.3 (2.6)a 6 (2)b
Very weak spellers (N ¼ 17) 6.9 (4.1)a 3.6 (3.6
a F ð1; 198Þ ¼ 122, p < :0001.b F ð1; 198Þ ¼ 121, p < :0001.c F ð1; 198Þ ¼ 183, p < :0001.d F ð1; 198Þ ¼ 251, p < :0001.e F ð1; 198Þ ¼ 119, p < :0001.
Compared with the average spellers, the weak and veryweak spellers exhibited a significant deficit affecting all
three tasks. They were significantly lower for the pho-
neme deletion task (F ¼ 25, p < :0001), the morpho-
logical/syntactic task (F ¼ 11:8, p < :0001) and the
naming speed tasks (F ¼ 9:2, p < :0001) Table 2.
3.3. Correlation matrix
The purpose of the correlation matrix was to examine
the correlation between individual differences in pho-
nological awareness, morphological/syntactic skill, and
naming speed at the end of grade 1, and spelling per-
formance at the end of grade 2. As shown in Table 3,
GPC mastery significantly correlated with naming speed
(.42, <.0001), morphological/syntactic skill (.43,
p < :0001), and phonological awareness (.61, p < :0001).Lexical orthographic strategy also significantly corre-
lated with naming speed (.39, p < :0001), morphologi-
cal/syntactic skill (.35, p < :0001), and phonological
awareness (.56, p < :0001). The most significant corre-
lation concerned GPC mastery and phonological
awareness.
3.4. Hierarchical multiple regression analyses
To determine whether phonological awareness, mor-
phological/syntactic skill, and naming speed in grade 1
could account for spelling ability in grade 2, several
hierarchical multiple regression analyses were per-
formed; these are shown in Table 4. To determine
whether one variable significantly predicts GPC mastery
and lexical orthographic usage after the other twovariables have been statistically controlled for, each
variable was, respectively, entered at Step 1, Step 2, and
Step 3.
As Table 4 shows, when this was done, the three
variables accounted for a significant independent pro-
portion of variance in GPC mastery and lexical ortho-
graphic skill. It was observed that, when entered at the
last step, phonological awareness accounted for 15% ofvariance for phonological strategy and 14% for ad-
dressed strategy; morphological/syntactic skill ac-
counted for 4% of variance for phonological strategy
llers: mean scores and SDs
age Word phonetics Global phonetics Global usage
) 16.8 (2.3) 60.3 (6) 29.4 (7)
11.8 (2.4)c 46 (4.7)d 15.3 (3.2)e
)b 7.8 (4)c 29.6 (13.9)d 9.6 (6.7)e
Table 4
Hierarchical multiple regression analysis for grade 1 variables predicting phonetic spelling (PHSP) and usage spelling (USP) in grade 2
Step Variable R2 PHSP R2 change PHSP p value R2 USP R2 change USP P value
1 Phoneme deletion .37 <.0001 31 <.0001
2 Morpho-syntax .40 .3 <.002 33 3 <.05
3 Naming speed .46 .6 <.0001 38 7 <.0001
2 Naming speed .43 .6 <.0001 37 6 <.0001
3 Morpho-syntax .47 .4 <.001 38 1 <.04
1 Morpho-syntax .18 <.0001 12 <.0001
2 Phoneme deletion .40 .22 <.0001 33 21 <.0001
3 Naming speed .47 .7 <.0001 38 5 <.0001
2 Naming speed .32 .14 <.0001 24 12 <.0001
3 Phoneme deletion .47 .15 <.0001 38 14 <.0001
1 Naming speed .18 <.0001 15 <.0001
2 Phoneme deletion .44 .16 <.0001 37 22 <.0001
3 Morpho-syntax .47 .3 <.001 38 1 <.04
2 Morpho-syntax .32 .14 <.0001 24 9 <.0001
3 Phoneme deletion .47 .15 <.0001 37 13 <.0001
Table 3
Correlation between performance on the tasks in grade 1 (Gr1) and spelling performance in grade 2 (Gr2)
Variable Global phonetics Gr2 Global usage Gr2 Morpho-syntax Gr1 Phonology Gr1 Naming speed Gr1
Global phonetics Gr2 1 .83 .43 .61 .42
Global usage Gr2 1 .35 .56 .39
Morpho-syntax Gr1 1 .45 .14 ns
Phonology Gr1 1 .29
Naming speed Gr1 1
Note. Correlations greater than .19 are significant at the .05 level, greater than .25 are significant at the .01 level, and correlations greater than .32
are significant at the .001 level.
Table 2
Performances of the three groups of spellers at the end of grade 1 on phonological awareness, naming speed, and morphological/syntactic skill: mean
scores and SDs
Phonological awareness Naming speed Morphological/syntactic task
Average spellers (N ¼ 159) 9.9 (2.5) 37 (7.2) 8.6 (2.1)
Weak spellers (N ¼ 23) 6.7 (4)a 33 (5.8)b 7.1 (2.1)c
Very weak spellers (N ¼ 17) 5.8 (3.7)a 30 (4.9)b 6.3 (2.4)c
a F ð1; 198Þ ¼ 25:9, p < :0001.b F ð1; 198Þ ¼ 9:2, p < :0001.c F ð1; 198Þ ¼ 11:8, p < :0001.
M. Plaza, H. Cohen / Brain and Cognition 55 (2004) 368–373 371
and 1% for addressed strategy; and naming speed ac-
counted for 6% of variance for phonological strategy
and 5% for addressed strategy. Taken as a whole, the
three tasks accounted respectively for 47% of the vari-
ance for phonological strategy and 37% for addressed
strategy.
4. Discussion
The principal aim of the study was to determine
whether the synchronic correlation between written
language, morphology, phonology, and naming speed at
the end of grade 1 demonstrated in our previous study
(Plaza & Cohen, 2002) would remain significant at the
end of grade 2. To answer this question, a group of 254
children was assessed on three spelling tasks at the end
of grade 2. Based on their spelling performances, three
groups of children were constituted: the ‘‘average spell-
ers,’’ ‘‘weak spellers,’’ and ‘‘very weak spellers.’’ Among
the 254 children, 199 had been given tasks requiring
phonology, morphology, and naming speed at the end of
grade 1. The retrospective study explored the links
between these children�s spelling skills at the end of grade2 and their performance profile at the end of grade 1.
With regard to spelling performance in grade 2, the
analysis of variance between the groups revealed that all
three groups of children better mastered the assembly
strategy than the addressed phonological strategy
(pseudo-words versus words; GPC mastery versus lexi-
cal orthographic usage in isolated words and text dic-
tation). These results confirm that, in alphabetic
372 M. Plaza, H. Cohen / Brain and Cognition 55 (2004) 368–373
languages, the orthographic strategy crucially dependson the assembly strategy, which allows the lexicon to
become increasingly redundant and stable (Perfetti,
1997; Share, 1995, 1999).
Nevertheless, it was observed that the discrepancy
between GPC mastery and orthographic lexical strategy
was less marked in average spellers (16.9%) than in weak
and very weak spellers (respectively, 34.1 and 19.3%).
Children who had thoroughly mastered the GPC rules,i.e., the ‘‘phonological route,’’ harmoniously developed
their orthographic lexicon, i.e., the ‘‘addressed route.’’
By contrast, children who experienced difficulties with
GPC rules also suffered from orthographic lexicon re-
striction. The insufficient automaticity of GPC does not
allow an ‘‘amalgamation’’ (Ehri, 1995) between the
phonological, graphic, morphological, semantic, and
orthographic features of the words. The children whodid not use sufficient grapho-phonemic cues to quickly
identify a word when they read it and, a fortiori, when
they spelled it, at the beginning of the learning process
did not create high-quality representations of words in
long-term memory.
The retrospective analysis from grade 2 to grade 1
revealed that, compared with the average spellers, the
weak, and very weak spellers had performed signifi-cantly worse one year before on phonological aware-
ness, naming speed, and morphological/syntactic tasks.
The comparative data were confirmed by the correla-
tion matrix and, overall, by the multiple hierarchical
regression analysis. The importance of GPC mastery
at the earliest stages of written language learning
probably explains why the three variables predicted
47% of variance for the phonetic features of words,and 38% for the orthographic lexical features. Im-
pairment of the phonological route is assumed to
be the best developmental marker of written language
difficulties (Bryant, Nunes, & Bindman, 1998; Morris
et al., 1998).
Among the three factors entered in the regression
equation, phonological awareness appeared as the more
potent predictive variable, which was not surprising.The phoneme deletion task clearly reflects mastery of the
GPC rules and automaticity of the assembly strategy. In
this task, the child was required to encode and hold
phonological information (pseudo-words) in auditory
working memory, access phonemic units, manipulate
them (segmentation, elision, and assembly) and articu-
late a new pseudo-word. Similar processing is required
when a child encodes a written or dictated word in orderto read or spell it.
Naming speed was the second predictive factor for
spelling performance. The tasks that assessed this skill
required the child to encode visual information (pic-
tures, digits, and letters), retrieve lexical labels from
long-term memory (i.e., phonological codes), and ar-
ticulate them. Although naming speed and phonological
awareness tasks share the same emphasis on retrievaland articulation of phonological codes, they differ since
naming speed tasks, like reading activities, require cross-
modal (visual/verbal) processing and automatized re-
trieval speed. The involvement of naming-speed tasks is
in accordance with neuropsychological findings indi-
cating that a slower naming speed may be an index of
lower-level problems disrupting the development of
fluency in word identification (Wolf & Greig Bowers,1999). Reading and writing require efficient, high-speed
connections to be made between visual stimuli (printed
words) and their phonological counterparts (Breznitz,
1997).
Morphological/syntactic skill was the third predictive
factor for spelling performance. The morphological/
syntactic task required the child to encode verbal in-
formation (sentences) in auditory working memory, fo-cus on morphemic units (gender and number markers;
topological and temporal function words), retrieve the
correct item from verbal long-term memory, and artic-
ulate it. The ability to be aware of morphemic features,
which is an indicator of a fine-grained linguistic capac-
ity, is probably also a facilitator during the process of
word identification (Joanisse, Manis, Keating, & Se-
idenberg, 2000). Morphology allows the child to definethe smallest units of meaning and expression, and store
words in the orthographic lexicon on the basis of roots,
derivations and inflections. Some longitudinal studies
have indicated that morphological awareness in spoken
language assessed in kindergarten and first grade is
predictive of reading achievement in second grade
(Carlisle, 1995).
Taken as a whole, the results reported here providesupport for an integrative hypothesis about the inter-
actions between phonological processing, morphologi-
cal/syntactic skill and naming speed as predictors of
reading and spelling (dis)ability.
Our data also answer our secondary question, con-
firming that in French, as in English, German, Finnish,
Dutch, and Spanish, naming speed is indeed a predictor
of reading and spelling performance, independent ofphonological awareness.
Finally, while phonological awareness, morphologi-
cal/syntactic skill, and naming speed represent inde-
pendent sources of difficulties, there are clinical
implications for diagnosis. The data highlight the ne-
cessity of including naming-speed measures and mor-
phological/syntactic skill in dyslexia evaluation, in
addition to phonological measures.
References
Breznitz, Z. (1997). Enhancing the reading of dyslexic children by
reading acceleration and auditory masking. Journal of Educational
Psychology, 89, 103–113.
M. Plaza, H. Cohen / Brain and Cognition 55 (2004) 368–373 373
Bryant, P., Nunes, T., & Bindman, M. (1998). Awareness of language
in children who have reading difficulties. Historical comparisons in
a longitudinal study. Journal of Child Psychology and Psychiatry,
39, 501–510.
Carlisle, J. F. (1995). Morphological awareness and early reading
achievement. In L. B. Feldman (Ed.), Morphological aspects of
language processing. Hillsdale, NJ: Lawrence Erlbaum Associates.
Chevrie-Muller, C., & Plaza, M. (2001). Nouvelles �epreuves pour
l�examen du langage. Paris: Les Editions du Centre de Psychologie
Appliqu�ee.
Denckla, M. B., & Rudel, R. G. (1974). Rapid automatized naming of
pictured objects, colors, letters and numbers by normal children.
Cortex, 10, 186–202.
Ehri, L. C. (1995). Phases of development in learning to read words by
sight. Journal of Research in Reading, 18, 116–125.
Greig Bowers, P., & Newby-Clark, E. (2002). The role of naming speed
within a model of reading acquisition. Reading and Writing: An
Interdisciplinary Journal, 15, 109–126.
Joanisse, M., Manis, F., Keating, P., & Seidenberg, M. (2000).
Language deficits in dyslexic Children: Speech perception, phonol-
ogy and morphology. Journal of Experimental Child Psychology,
77, 30–60.
Morris, R., Stuebing, K., Fletcher, J., Shaywitz, S., Lyon, R.,
Shankweiler, D., Katz, L., Francis, D., & Shaywitz, B. (1998).
Subtypes of reading disability: A phonological core. Journal of
Educational Psychology, 90, 347–373.
Perfetti, C. (1997). Psycholinguistique de l�orthographe et de la lecture.
In L. Rieben & C. Perfetti (Eds.), L�apprenti lecteur. Neuchatel:
Delachaux et Niestl�e.
Plaza, M. (2001). The interaction between phonological processing,
syntactic awareness and reading. First Language, 21, 3–24.
Plaza, M., Chauvin, D., Lanthier, O., & Rigoard, M.-T. (2002).
Validation longitudinale d�un outil de d�epistage des troubles du
langage �ecrit. Etude d�une cohorte d�enfants d�epist�es en fin de CP et
r�e�evalu�es en fin de CE1. Glossa, 81, 22–35.
Plaza, M., & Cohen, H. (2003). The interaction between phonological
processing, syntactic sensitivity and naming speed in the reading
and spelling performance of first-grade children. Brain and Cogni-
tion, 53, 287–292.
Plaza, M., & Touzin, M. (2002). Exp�erience de d�epistage des troubles
du langage �ecrit dans un groupe de 267 enfants scolaris�es en CP.
A.N.A.E., 66, 20–28.
Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua
non of reading acquisition. Cognition, 55, 151–218.
Share, D. L. (1999). Phonological recoding and orthographic learning:
A direct test in self-teaching hypothesis. Journal of Experimental
Child Psychology, 72, 95–129.
Wolf, M., Goldberg O�Rourke, A., Gidney, C., Lovett, M., Cirino, P.,
& Morris, R. (2002). The second deficit: An investigation of the
independence of phonological and naming-speed deficits in devel-
opmental dyslexia. Reading and Writing: An Interdisciplinary
Journal, 15, 43–72.
Wolf, M., & Greig Bowers, P. (1999). The double-deficit hypothesis for
the developmental dyslexias. Journal of Educational Psychology, 91,
415–438.
Wolf, M., & Obregon, M. (1992). Early naming deficit, developmental
dyslexia and a specific deficit hypothesis. Brain and Language, 42,
219–247.
Wolf, M., & Obregon, M. (1997). The double-deficit hypothesis:
Implications for diagnosis and practice in reading disabilities. In L.
Putnam (Ed.), Readings on language and literacy. Boston: Brooklyn
Books.