reading fluency children read individual but not whole words
TRANSCRIPT
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
Jpn. J. Spec. Educ,, 46 (6), 405 416, 2009.
Original Article
Reading Fluency in Young Children Who
Can Read Individual Letters Fluently But Not Whole Words:
A Longitudinal Study
Shino SAKONo* and Tomohiko ITo'"
The purpose of the present study was to investigate longitudinally the develop-ment of fiuency in reading words and nonwords in typically developing children
who could read all the letters in the words but could not read thc whole words
fiuently, The participants were 5 children aged 4 to 5. Words and nonwords were
used as stimuli in a reading task. The number of correct and fiuent readings
increased more rapidly for the words than for the nonwords in 3 of thc children.
aorrect and fluent rcadings of both the words and the nonwords by the fourth
child increased rapidly. The fifth child could only 1 word and 1 nonword by thc
final measurement day. At thc beginning ofthe study, the mean reading time for
thc words by 2 of the 5 children was shorter than for the nonwords. The presentresults suggest that a variety of rates or proc¢ sses exist in the dcvclopment of
reading fiuency for words and nonwords in typically developing children.
Key Words: reading, Iluency, dcvclopment, longitudinal study, preschool chil-
dren
Introduction
The Japanese language uses two kinds of orthographic systems: kana and kanji,Each kana character represents a speech unit roughly equivalent to a syllable
(Sasanuma, 1986), The relationship between orthography and phonology in kana is
perfectly regular and rule-governed. Each kana character essentially always repre-
sents one and the same mora with no context sensitivity (Sasanuma, Ito, Patterson,
& Ito, 1996). There{bre, accuracy of reading kana words can be acquired easily in the
Japanese language by using letter-by-letter reading. As a result, in languages like
Japanese in which the relationship between orthography and phonology is regular,
dilliculty of reading is reflected in reading fluency, Thus, in research on the develop-
ment and disorders of reading in Japanese, it is important to study the development
of reading fiuency.
*United
Graduate Scheol of Education, Tokyo Gakugei University**Tekyo
Gakugei University
- 4D5
-
NII-Electronic Mbrary
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
S. Sakono & T. Ito
Many researchers have studiecl the development of reading in typically develop-
ing children. For example, Amano (I986) condueted a series of studies on the
acquisition of the reading of kana letters. Takahashi (1993, 1996) investigated thedevelopment of reading ability after children had learned the kana letters, Recently,many studies have investigated the relationship between the ability to read kana
characterst and phonological awareness in young children (Amano, 1970i Arriano,1986; Hara, 2001; Takahashi, 1999), In addition, numerous studies have fbcused on
the segmentation of words into moras (e.g,, Inagaki, Hatano, & 0take, 2000; Ito,
2006; Ito & Tatsumi, 1997).
However, few studies havc lbcused on the development of reading fiuency intypically developing childrcn. In order to determine the characteristics of reading
fiuency in Japanese children with developmental dyslexia, we must first study the
development ot' reading fiuency in typically developing children,
In order to teach children with developmental dyslexia successfuIIy, it is impor-tant to find deficits early, The later that deficits are identified, the more serious is thedelay of learning (Shaywitz, 2003). If characteristics of the development of reading
fiuency could be found in typically developing children, that may provide a way to
identify children who might have diMculty reading,
Children with developmental dyslexia have been shown to have diNculty read-
ing nonwords (Rack, Snowling, & Olson, 1992; Snowling, 200e). Many studies have
demonstrated that Japanese children with developmental dyslexia have dirnculty
reading nonwords (Matsumoto, 20061 Oishi, 2007; Tatsurni, 2007; Yamada, 1997).
Shaywitz (2003) tbund that the results from children's reading of nonwords were
important in the diagnosis of developmental dyslexia. Therefore, the development ofreading fluency should be investigated, in typically developing ehildren, in an early
stage of reading development, using nonwords to identify dyslexic children.
Sasanuma (1995) propesed that each word carries infbrmation about ortho-
graphy, phonology, and semantics, and that such intbrmation and its interactiverelationship are decoded in the mental lexicon, On the basis of this, we assume that
it is easier lbr children to read real words than nonwords fluently, because they can
use their mental lexicon when they read the real words. Theretbre, we predict that
children should be able to acquire reading fiuency earlier for words than fbr
nonwords. However, as we have indicated above, few studies have focused on the
development of reading fluency in typically developing chlldren.
The purpose of the prcsent study was to investigate the development of fluency
in reading words and nonwords in typically developing chi!dren who eould read each
letter in the words, but could not read whole words fluently,
Method
Stimulus Letters and vaorcts
1) Letters used in the stimultts words. Nine letters were used as the stimulus words
and nonwords. These letters were employed to determine whether participants could
-406-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
Reading Fluency in Young Japanese Children
TABLE 1 Stimulus Words
2 moras 3 moras 4 moras 5 moras
WordsNonwords kumatsusu tsumiki
sutarl
kutsushita
miritashi
kurisurnasu
kimamikusu
TABLE 2 Participants
Participant
Agc
Child A
4:4
Child B
4:6
Child C
4: 11
Child D
5:1
Child E
5:2
Atbtes, Age when the present longitudinal study was started.
Child E was a boy; the other four participants were girls,
read all the letters that were used to tbrm the stimulus words and nonwords.
According to the National Institute fbr Japanese Language (1972), such words
present the easiest and the second-easiest degree of difficulty in acquisition.
2) Stimulas wortis. The stimulus words, 4 words that are used frequently
(Iwabuchi & Muraishi, I976) and 4 nonwords, were from two moras to five moras in
length. The stimulus words and nonwords are listed in Table 1.
ParticiPants
The originally seiected participants were eight children who could not read the
stimulus words correctly and fluently in three trials from among 38 participants in anearlier study by the present authors (Sakono & Ito, 2007). We wanted to inciude al1eight of the children as participants. However, two o{' the children could not
participate in the present study because they had graduated from nursery scheol, In
addition, one child was exciuded because of stuttering. As a consequence, the
participants in the present study were five out of the eight participants in the presentauthors' earlier research (Sakono & Ito, 2007). Child E was a boy; the other four were
girls. Further information about the participants is provided in Table 2.
Procednre
The children were tested individually. First, nine hiragana letters that were
included in the stimulus words were presented to the children one by one on a card,
and the children were asked to read each letter. Next, those children who could read
all the letters were asked to read the four hiragana words and the four nonwords as
fast as possible, The stimulus words were written in hiragana on cards, If the children
did not notice a reading error, they were asked to read that word or nonword again.
If a child read letter-by-letter, the child was then asked again to read that word, up
to a total of two additional trials.
The present longitudinai study continued fbr about one year. The data were
collected on four occasions: when the study started, and at 7, 9, and 11 months
thereafter.
-407-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
S. Sakono & T. Ito
The children's responses were tape-recorded using a digital audio tape recorder
(TCD-DIO) and a microphone (ECM-959DT).
Method of Analysis The children's correct and fluent readings were analyzed. When a stimulus word
was read by a child without letter-by-letter reading or disfiuencies such as repetitions
or revisiQns, the word was judged to have been read fiuently. The rate of agreement
between the two graduate students who judged fiuency was 93.6%, calculated by
dividing the number of agreements by the total of the number oi' agreements plusdisagreements, and multiplying by 100.
In addition, reading time (the time between the presentation of a stimu}us word
or nonword and the end of the child's reading of that word or nonword) was
measured by a digital stop watch (WATER 5BAR RESIST, Q & Q).
Results
Developmental Change in Correct and Fluent Reading
Figure 1 shows the developmental changes in the five children's correct and
fluent reading of the stimulus words and nonwords over the year ofthe present study.
The abscissa shows the time in months from the start of the study. The ordinate
shows the number of words read correctly and fiuently. The results for the fbur words
were analyzed as a whole, as were the results fbr the four nonwords.
As shown in Fig, 1, three of the children, A, C, and D, showed a rapid increase
in the number of werds read correctly and fiuently, but a slower increase in the
number of n,onwords read correctly and fiuently. The number of words read correctly
and fiuently by Child B increased rapidly fbr both the words and the nonwords. In
contrast, Child E read correctly and fluently only one word starting with the second
test (at 7 months from the start of the study) and, in addition, by the final day of
measurement, oniy one nenword, Child D was absent from the nursery school on the
day of the 9-month tests, so her data fbr that day are missing.
Developmental Changes in the Mean Reading Time
Figure 2 shows the developmental changes in the five children's mean time to
read the fosTords and nonwords. The abscissa shows the time in months from the start
of the study. The ordinate shows the mean time to read the stimulus words and
nonwords. Child D was absent from the nursery school on the day of the 9-month
measurement, so her data tbr that day are missing.
At the start of the study, the mean reading time was quite diflerent across the
five children. However, all the children's reading times tended to decrease up to the
Iast day of measurement. The mean times for Child A and Child E to read the
nonwords was longer than fbr the words on the first day of measurement. Child A'smean tirne to read the nonwords was longer than for reading the words throughout
the study. The difference between Child E's mean reading time for the words and the
-
4e8 -
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
Reading FIuency in Young Japanese Children
nonwords became smaller and smaller over the year ofthe study, and few differencesin reading time between the words and the nonwords were observed on the last day
of the study. Throughout the year, few diflerences were observed between Child B
and Child D;s mean reading times for the words and the nonwords, Curiously, at the
start of the study, Child C's mean reading time for the words was longer than for the
nonwords.
mtu-[vtuocr-c
¢
nLT[a-e
¢
too-oLvAE=z
d
3
2
1
o
Child A
o
Child C
q
3
79 11 U
Child D-----o
'
2
''''
l
'te
oo
79 ll
T9 11 o 79 11
O/ WoTds
e: Nonwordsme / Words and Nonwords
't
3
2
I
o
Child E
o T9 li
Time From the Start of the Study (Months)FIG. 1 Developmental Changes in Correct and Fluent Reading
-409-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
S, Sakono & T. Ito
DeveloPmental Chauges in Correct and Fluent Reading in Relation to the
rvumber of Moras of the Stimuli
Figure 3 shows the relationship between the developmental changes in correct
and fluent reading and the moras of the stimuli by the five children. The abscissa
shows the time in months from the start of the study. Only one word was used as a
stimulus word with 2, 3, 4, and 5 moras. The ordinate shows the total number of
correct and fiuent readings by the five children. The data at the 9-month assessment
AcoEvoEFtnrEvamxccaoE
9UOOBOOO700U6eoo5ooe4000300020001000
o
9000sooo70UU6000sooe400030002oeolveo
o
o 79 il
o T A 11
v T 9 Tl
o T・ 9 ]l
O: Words
e/ Nonwords
op : Words aiid Nonwords
9000souoTOOO6000500040003oon20001ono
o
Child C
O 79 11
Time From the Start of the Study (Months)FIG. 2 Developmental Changes in the Children's Mean Reading Time
- 41e-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation ofSpecial Education
Reading Fluency in Young Japanese Children
were from 4 children, because Child D was absent f'rom the nursery school on the day
of testing.
The total number of correct and fluent readings of the 2-, 3-, and 4-mora words
tended to increase from the first to the last day of measurement. However, the total
number of correct and fiuent readings of the 5-mora words tended to be smaller and
morc unstable than those with 2-, 3-, and 4-moras. While the total number of correct
and fluent reading of the 2-, 3-, and 4-mora nonwords increased from the first to the
last day of measurement, even by that time, only one of the 5-mora words was read
correctly and fluently.
mun=vNocrvcU=tav[asPootoo-oLoAE-z-eE,-oH
5
4
3
2
1
o
5
4
3
2
1
o
o 7 9t11
o 7 9 11
---O---2moras
+3moras
+4meras
+ S moras
-ee-2 mDras and 3 moras
-4moras
+ S moras
FIG.Mtes.
Time From Start of the Study (Months)
3 Developmental Changcs in Correct and Fluent Reading and
The data at 9 months arc from 4・ children, because one of the c
from the nursery school on thc day of testing.
Moras
hildrenof
Stimuliwas absent
411
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation ofSpecialEducation
S. Sakono & T. Ito
DeveloPmentat Changes in Mean Reading Time in Relation to the JVumber
of Moras of the Stimuli
Figure 4 shows the relationship between the developmental changes in rnean
reading time by the five children and the number of moras of the stimulus words and
nonwords. The abscissa shows the time in months from the start of the study. The
ordinate shows the mean reading times. Child D was abs,ent from the nursery school
on the day of the 9-month measurement, so her data for that day are missing.
The mean readlng time for the words regardless of the nuinber of rnoras
gradually became shorter and shorter. The mean reading time of the words with 4
and 5 moras tended to be longer than for those words with 2 and 3 moras throughout
FIG.
Mtes.
GEWUEFtu.!vtudicr=tooE
7000
6000
5000
4000
3000
2eoe
1000
o
7eoo
6000
5000
4eoo
3eoo
2000
lOOO
o
o 7 9*11
H)-2mnras
+3moras
-4 moras
+5 rnoTas
--O-2moras
+3 moTas
-O-4nwras
+ 5 moras
e 7 9 11
Time From Start of the Study (Months)
4 Developmental Changes in Mean Reading Time and Number of Moras
of the Stimuli
The data from the 9-month measurement were frem 4 children, because
Child D was absent from the nursery school on the day of testing.
-412-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
Reading Fluency in
the study. Although the mean reading
gradually became shorter and shorter,
to be longer and unstable.
Young Japanese Children
time of' the nonwords with 2, 3, and 4 moras
the reading time of the 5-mora words tended
Discussion
In order to achieve early identification of children who have diMculty in reading,
it is necessary to know the characteristics of developmental changes of fluent reading
in typically developing children, The purpose of the present study was to investigate
the devclopment of reading fluency for words and nonwords in typically developing
children who could read each letter in the stimulus words but could not read the
whole words fluently. The participants were five children who could read each letter
in those words but could not read the whole words fluently within three trials in the
present authors' earlier research (Sakono & Ito, 2007). The data were obtained on
fbur occasions; viz., when the present study started, and 7, 9, and 11 months later.
Developmental Changes in Correct and Fluent Readiug
Each word carries information about orthography, phonology, and semantics,
and this information and its interactive relationship are clecoded in the mental lexicon
(Sasanuma, 1995). Theretbre, it is easier for children to read words fluently thannonwords, because they can use their mental lexicon when reading words but not
when reading llonwords. We could predict that the fiuency of reading the stimulus
words would develop rapidly, but that the fiuency of reading the nonwords would
develop slowly.
As expected, the number of fluently read stimuli increased faster for the words
than the nonwords in 3 out of 5 children (A, C, and D), The results fbr these three
children agreed with our prediction that reading fluency would be acquired easier for
the words than the nonwords. Thc reason why these three children could read the
words more fluently than the nonwords is, we hypothesize, that they could use their
mental lexicon while reading the words.
However, the number of correct and fluent readings increased rapidly for both
the words and the nonwords in one of the children (B). The development of her
reading fluency was different from our expectation, probably tbr the tbllowing reason.
For her, although, as we expected, reading the nonwords was more diMcult than
reading the words, she acquired both the words and the nonwords within 7 months
from the start of the present study. Her results suggest that some children can acquire
fluent reading of both words and nonwords in a pcriod of less than 7 months, so that
it is necessary to examine developmental changes within shorter intervals than 7
months when the development of reading fluency in typically developing children isbeing investigated.
Child E was able to read correctly and fiuently only one word and one nonword,
even by the last measurement day of the present study. The characteristics of his
development in reading were also different from our expectations. Althouglt we
-rt13-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation of Special Education
S. Sakono & T. Ito
cannot explain why Child E:s fluency of reading the words did not develop within the
time limit of the present study, his results suggest that there are some children of this
age who cannot acquire fluent reading of words and nonwords within a year's time.
That is, Child E might have some reading diMculty.
The present results suggest that there is a variety of rates or processes from
letter-by-letter reading to fiuent reading. The present study should be repeated on a
larger sample in future studies of the development of reading in children between 4to 6 years old, because since only five children participated in this study, individual
diffbrences may have been an impertant l'actor.
Develqpmental Changes in Mean Reading 71me
Developmental changes in reading fiuency are refiected in reading time. We
predicted that the time to read the words would be shorter than that to read the
nonwords at the beginning of the development of reading. Although the results forChildren A and E agreed with our predictien, the results from Children B, C, and D
did not. As we expected, Child A's mean time to read the nonwords was longer than
the time to read the words from the first to the last day of the present study. Child
E;s mean reading time tbr the nenwords was longer than for the words on the first
day of measurement. Child E read only one word and one nonword fluently even on
the last day of measurement. These results suggest that Child E's speed of letter-by-
letter reading gradually increased.
Few differences were observed between in Child B's mean reading time fbr the
words and the nonwords. Her reading time tbr both was very short, and although the
number of words and nonwords read eorrectly and 'Huently
was O on the first day of
this study, her correct and fiuent reading ofboth increased simultaneously. This may
be because Child B read both the words and the nonwords at a very high speed of
letter-by-letter reading on the first day of measurement.
Little difltirence was observed in Child D's mean reading times between the
words and the nonwords from the first to the last day of measurement. The number
of words she read correctly and fluently increased faster than that did the number
of nonwords, probably because her reading time in fiuent reading was not very
diflerent from that in letter-by-letter reading.
Contrary to our expectations, on the first day, Child C's mean reading time on
the words was longer than on the nonwords. Shc was first presented with the werds,
followed by the nonwords, and took longer to read the words than the nonwords. One
possible explanation for these results is that it took her a long time to become
accustomed to the experimental task.
Taken together, the results of the present longitudinal study suggest that the
decreases in reading time for words and nonwords difEer ameng children.
Moras of Stimuli in Relation to Reading Fluency
In the present study, the relationship between developmental changes in correct
and fiuent reading and the number of moras of the stirnuli was investigated, The
--4J4・-
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation ofSpecial Education
Reading FJuency in Young Japanese Children
correct and fluent reading of the words tended to increase from the first to the lastday of thc present study in the 2-, 3-, and 4-mora words, However in the 5-mora
words, it tended to be smallcr and more unstable, While thc total number of correct
and fiuent reading' of the 2-, 3-, and "-mora nonwords increased from the first to the
last day of Tneasurement, even on the last day, only one of the 5-mora nonwords was
read correctly and fluently. These results suggest that the developmental changes in
correct and fluent reading are influenced by the number ol' moras in stimu]i,
especially with more than 4 moras.
While the mean reading time fbr the words gradually became shorter and
shorter regardless of the number of moras, the mean reading time for the words with
4 or 5 moras tended to be longer than for those with 2 or 3 moras from the first to
the last day of measurement. The mean reading time lbr the nonwords with 2, 3, or
4 moras gradually became shorter and shorter, whereas it tended to be longer and
unstable in the 5-mora words. These results suggest that developmental changes in
mean rcading timc iarcrc influcnced by the number of moras in the stimuli, cspecially
those with more than 4 moras,
tNote
In the Japanesc literaturc en rcading disordcrs, ,Japancsc kanu charactcrs arc referred
to as "letters".
Acknowledgments
We express our gratitude to the children and caretakers at the Hanakoganei
Nursery School.
References
Amano, K. (1970) Formation of the act of analyzing phonernic structure of words and its
re!ation to learning .Japanese syllabic characters (kanamoji). .lbPanese
Jbumal of' Eduvational Psychology, 18, 76-88, (in Japanese)Amano, K, (1986) Precess of acguisition qf kana tetters in Joung .IZvPanese children. Akiyama
Shoten, Tokyo. (in Japanese)Hara, K. (2001) The clevelopment of phonological awareness in Japanese children.
.laPanese
.lburnal of Cbmmunication Disorders, 18, 10-18. (i] Japanese)Inagaki, K,, Hatano, G,, & Otake, T, (2000) The effect of kana Hteracy acquisition on the
spccch scgmcntation unit uscd by Japancsc young children. .lbumal of ExPerimental enild A]ehology, 75, 72-91.
International Dyslexia Association (2005) wwat is cipslexia.P http:lwww.interdys, org/serv-
let/compose?sectionid=5&pageid=95#What%20is9)620dyslexia?
Ito, T. (2006) Development of prosodic units in Japanese: Focusing on the segmentation of
words into mora. .laPanese
Jbarnal of SPecial Education, 44, 191-196. (in Japanese)
415・
The Japanese Association of Special Education
NII-Electronic Library Service
The JapaneseAssociation ofSpecial Education
S. Sakenu & T, Ite
Ito, T, & Tatsumi, I. (1997) The clevelopment of metalinguistic awareness of mora-
phonemes in Japan¢ se young children. JIltPan lburnal of Lagopedies and Rheniatri6s, 88,
196-203. <in Japanese)Iwabuchi, E. & Muraishi, S. (1976) 7'brms and examples in Joung children. Japan Broadcast Publishing aompany, Tokyo. (in Japanese)Matsumotu, T. (2006) Semanti ¢ and phonological processes in Teading hiragana characters:
Young adult with devclopmental dyslexia, .14Panese Jbumal qf' .EPecial Education, 44,
I03-1 IS. (in Japanese)National Institute for Japanese Language. (1972) ne abiEly qf reading and zvn'ting' in .young
children, Tokye Shoseki, Tokyo, (in Japanese)Oishi, K. (2007) Developmental dyslexia in Japanese. In S. Sasanuma (Ed.), A new Poinl of eview and intewentien theot:y of speeah and language communication ddserders in chiidren. Igaku
Shoin, Tokyo, I13-13I. (in Japanesc)Rack, J, P., Snowling, M. J., & Olson, R, K. (1992) The nonword reading deficit in
developmental dyslexia: A Teview. Reading Research Q!tarterly, 27, 29-53.Sakono, S, & Ito, T. (2007) The relationship betwecn 1,ctter by letter reading and fiuent
reading in youngJapanese children: Using words and nonwords. Proceedings of the 45th Cbnjlarence of' the .lkPanese
Association of' SPeq'at Edueation, 585. (in Japanese)Sasanuma, S. (1986) Universal and language-specific symptomatology and trcatment of
aphasia, Ploha Phoniatn'ca, 38, I21-175.Sasanuma, S. (1995) Univcrsalities and Ianguage-specificities of reading processes. In Y,
Otsu (Ed.), Cbgnitieve psyahelqgy llI, Laagacag'e. Tokyo Daigaku Shuppankai, Tokyo,
193-208. (in Japancse)Sasanuna, S., Ito, II,, Pattersen, K., & Ito, T, (l996) Phonologieal alexia in Japanese: A case
study. Cbgnitieve ,7Vliump.tyahology, IS, 823"848.Shaywitz, S. E. (2003) Overcoming dyslaxia: A nezc) and comptete science-based Pragium ,fbr reading
ptvblems at aay level. Alfred A, KnopflRandom House, New York.
Snowling, M, J. (2000) Qysimia, Blackwell, Oxford, UK.
Takahashi, N. (199S) Bccomipg skillf'ul at rcading in bcginners, JT4Panese JIburnal of Educatienal RsJchotQgy, 41, 264-274, (in Japanese)Takahashi, N. (1996) Aequisition of reading ability: A two-year longitudinal analysis in
preschool and lst grade years, .14Panese .lbu,veal of Educational RsJahology, 44, l66-175,
(in Japanese)Takahashi, N. (1999) Acguisition of reading ability in children, Kazama Shobo, Tokyo. (in Japanese)Tatsumi, I. (2e07) Cognitive neuropsychology of deve!opmental dyslexia. In S, Sasanurna
(Ed.), A new Point of view and intervention theony), of speeth and laagaage communicatien
dsonters in chiidren. Igaku Shoin, Tokyo, 93-1I2. (in Japanese)Yamada, J. <l997) Developmental dys!exia in Japanese children.
.litPan .lburnal of Lagopedics
and Rhoniatn'es, 38, 287-29e. (in Japanese)
-Receivcd June 20, 2008i Accepted February 21, 2009- -
-416-