Академический Документы
Профессиональный Документы
Культура Документы
AND
WOLFGANG SCHNEIDER
One robust idea among those investigating reading acquisition has been
that children’s metalinguistic awareness affects their reading performance.
This one idea has set the stage for many valuable lines of inquiry into the
This research was supported by a fellowship to the first author from the Alexander von
Humboldt Stiftung in Bonn, Germany, and data was collected at the Max Planck Institute for
Psychological Research in Munich, Germany. We thank Prof. Franz E. Weinert, Herman Schnei-
der, and the professional technical assistants at the Max Planck Institute for Psychological
Research for their support in completing this study. Address correspondence and reprint requests
to Jan Carol Näslund, University of New Mexico, College of Education, Program in Educational
Linguistics, Albuquerque, NM 87131.
30
0022-0965/96 $18.00
Copyright q 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.
exact nature of these linguistic processes and their interaction with other
factors in children’s reading performance. Many studies suggest that metalin-
guistic awareness, especially the ability to analyze and segment speech into
phonological units, begins to develop about the time children acquire skills
for reading. There is also evidence that the level of metalinguistic awareness
a child displays before entering formal education has a significant effect on
reading acquisition, especially in acquiring the ability to decode letter strings
into pronounceable syllables and words (Bradley & Bryant, 1985; Bryant,
Bradley, MacLean, & Crossland, 1989; Lundberg, 1987, 1989; Lundberg &
Høien, 1991; Lundberg, Frost, & Petersen, 1988; Lundberg, Olofsson, &
Wall, 1980; MacLean, Bryant, & Bradley, 1987). Thorough reviews of this
research over the past two decades are available (Sawyer & Fox, 1991;
Wagner & Torgesen, 1987).
Evidence for levels of metalinguistic skill in the development of literacy
is related to the first concern of this study: the comparison of various tasks
which are intended to measure preschool phonological awareness in their
predictability of later reading performance. Mann (1991), Treiman (1991),
and Tunmer and Rohl (1991) suggest that phonological processing is not a
unitary skill. There is likely a hierarchy of skills which arise over the course
of a child’s development. Phonological awareness refers to sensitivity to
multiple levels of spoken language. This includes sensitivity to language
components in the broad sense, such as rhyme and alliteration detection, and
the ability to segment words into syllables. In the narrower sense, this includes
more explicit segmentation skill, as demonstrated in the analysis and synthesis
of individual phonemic units (i.e., onset consonants and vowel/coda rime
components) and individual phonemes.
The study reported here was conducted in southern Germany, where chil-
dren are seldom given reading instruction before entering first grade. Parents
are not encouraged to provide any reading instruction before their children
enter formal schooling. Lack of preschool literacy in Germany does not reflect
lack of educational opportunity or a disadvantaged preschool environment.
This situation in German kindergartens provides the opportunity to compare
performance of children with various levels of letter knowledge and phonolog-
ical awareness. One objective was to determine if performance of kindergarten
children in a variety of phonological awareness tasks differentially predicts
later reading performance. The phonological awareness tasks used in this
study were expected to span various levels of development, with some tasks
expected to be at ceiling level for most children (those concerning linguistic
operations at the level of the syllable), and some expected to be difficult for
preliterate children (requiring intrasyllable operations, such as those pertaining
to onset/rime units). Yopp (1988) provides a comparison of validity and
reliability for a wide range of tasks commonly used to test phonological
processing. She administered these tasks to English speaking kindergarten
children who had received some instruction in the alphabet (and possibly
the present study, phonological awareness and memory span are measured
over time. This allows the assessment of a possible direction of influence
between memory span and phonological processing before and during reading
acquisition. In addition, phonological recoding interference effects in memory
are compared in kindergarten (age 6) among children controlling for letter
knowledge. The presence of interference effects, independent of letter knowl-
edge, and their correlation with phonological awareness and later reading
would provide some support for development of phonological recoding in
working memory prior to formal reading instruction.
METHOD
Subjects
One hundred thirty-four children from the southern region of former West
Germany were mean age 4.1 years (SD Å 0.5 years) at the start of the study.
These children were participants in the Munich Longitudinal Study of the
Genesis of Individual Competencies (LOGIC; Weinert & Schneider, 1989).
Parents of participants responded to an advertisement in a local newspaper
for voluntary participation in a longitudinal study. Screening was intended
to determine parents’ and children’s commitment to a longitudinal study.
Researchers intended to secure a varied group of children. About half of the
children were from the city of Munich, and the other half were from near
suburbs. Only those children who remained during the first 4 12 years of the
study were included in the analyses. Twenty-two children were promoted
from kindergarten to first grade 1 year later than the rest of the group. Their
performance was not included in all the analyses. It should be noted that
children usually begin kindergarten in Germany at age 3 or 4 and remain
until age 6 or 7. The decision when a child enters first grade is based on
parents’ and teachers’ assessments of a child’s readiness for the structured
environment of elementary school.
Design
Most of the tasks reported in the present study were administered through-
out a day visit to the Max-Planck Institute or satellite testing center in a
suburb near Munich. These visits occurred three times during each year of
the study. The emergent literacy tasks were a subset of the total tasks per-
formed during the early summer visit of the last kindergarten year for most
children. The verbal ability and word discrimination tasks were administered
during the first grade visits. The decoding tasks were administered during the
first and last visits during the second grade. The reading comprehension and
spelling tests were administered in children’s classrooms as part of a larger
study of children in the Munich area schools. A list of measures used in this
study and the time(s) administered are found in Table 1. Descriptions of
measures follow.
TABLE 1
Measures and Age of Administration
Tasks at Age 6
Word span. The German version of Case et al.’s (1982) word span task
was repeated. A second memory span task was given using phonologically
similar words. Nine sets of four words each were used. Three of these sets
contained words sharing the same consonant cluster ending. Three sets shared
the same middle vowel sound, and three sets had all the same vowel, but in
addition, three of the words in each set also had the same consonant cluster
ending.
Preschool letter knowledge phonological and tasks. The following letter
knowledge and phonological awareness tasks were devised by Jansen, Knorn,
Mannhaupt, Marx, Beck, and Skowronek (1986) and Brügelmann (1986) for
use in a large scale longitudinal study in northern Germany. All tasks pre-
sented verbal material via a reel-to-reel tape player to assure clarity. A female
voice was used. These tasks were presented when children were in the last
month of kindergarten (June/July; mean age Å 6.1 years). All tasks described
were developed by Jansen et al. (1986), unless otherwise indicated.
Letter knowledge. Letter knowledge was assessed from a subset of answers
in a task measuring preschool number and letter knowledge. The children
were first asked if they knew what numbers and letters were. Numbers (1–
9) and letters (all but x, ä, ö, and ü) were presented in a fixed random order
on five cards. Children were to point to each of the characters on the cards
and say if they recognized it or not. The idiomatic German instructions trans-
late almost literally as how the letter or number ‘‘calls.’’ This could mean
what it is called or how it sounds. Only responses to the letters are used in
the analyses here. Responses were classified as letter name, letter sound, word
(e.g., saying ‘‘Hans’’ when identifying ‘‘H’’), or wrong letter. Letter names
and letter sounds were considered correct responses. Parents were also inter-
viewed about their children’s letter and word reading abilities in kindergarten.
Rhyme detection. Children were presented 6 practice and 10 experimental
trials of two words each. Half of the word pairs rhymed. The child was asked
to say whether each of the pairs rhymed.
Syllable count. Children first practiced saying two and three syllable words
while clapping once for each syllable. Three practice and 10 experimental
trials were given. Five trials each were given for two and three syllable words.
Sound-in-word detect. Children were presented with a vowel sound and
asked to listen to a word and say whether the sound was in the word they
heard. All target words had the vowel sound in the initial position. Word
initial vowels are very consistent in pronunciation in German. The vowels
sounds used were ÉaÉ as in ‘‘sock,’’ ÉeÉ as in ‘‘able,’’ ÉiÉ as in ‘‘easy,’’ ÉoÉ
as in ‘‘only,’’ ÉuÉ as in ‘‘through,’’ and ÉaiÉ as ‘‘eye.’’ The three additional
German vowel sounds, ÉäÉ, ÉöÉ, and ÉüÉ were not used. Three practice and
10 experimental trials were administered.
Syllable segment (Brügelmann, 1986). The child was asked to play ‘‘ro-
bot,’’ imitating how a robot says words with breaks (e.g., BA-NA-NE, ‘‘ba-
nana’’). After three practice trials, the child was presented with 10 familiar
words to say in robot talk.
Pseudoword repeat. Children were presented with two practice and 10
experimental trials in repeating ‘‘Zauberwörter’’ (magic words). Seven words
had four syllables, two had five syllables, and one had six syllables. Four of
the words (including the one with six syllables, ‘‘pi-ni-ku-ra-ku-la’’) had a
simple consonant vowel (CVCV) structure. The rest of the words included
at least one CVC syllable. Stress and other prosodic variations were kept to
a minimum. Pseudowords did not contain morphemes common to German.
Syllable blend (Brügelmann, 1986). This task was presented by a fictitious
robot that spoke mechanically. The child was made familiar with the way
the robot spoke and was told that the mechanical parts that made him talk
malfunctioned. The child was to help the experimenter identify the words the
robot was trying to say. The purpose of the robot language was to present
items without stress or other prosody cues, which is expected of robots.
Three practice and 10 experimental trials were given. Familiar words were
pronounced by the robot without stress and with elongated vowels in each
syllable (e.g., ‘‘ZUUU-KEEER,’’ meaning ‘‘sugar’’), and the child was to
identify what word the robot was attempting to say.
Onset/rime blend. The child heard a monosyllable word pronounced in two
parts. Four of the trials separated the first onset/rime unit from the ending
consonant (vowels followed one glottal, one aspirated, one stop, and one
fricative consonant). Six trials separated the initial onset consonant or conso-
nant clusters from the final rime/consonant segment (four fricative, and two
nasal consonants). After hearing each presentation, the child was to tell the
experimenter what word had been broken apart (e.g., Sch-uh, pronounced the
same as the English word ‘‘shoe’’). Two practice and 10 experimental trials
were presented.
Phoneme oddity. A German version of Bradley and Bryant’s (1985) pho-
neme oddity task was devised for this research. Three sets of items were
presented according to Bradley and Bryant’s procedure, with the same expla-
nations and games used in practice trials to assure that children understood
the ‘‘odd man out’’ rules of the game. Each child had 10 trials in each of
the three oddity detection tasks. The first task, middle-sound-oddity, required
detecting which of four monosyllable words did not share the same middle
sound (all words presented in each trial shared the same ending consonant
or consonant cluster; e.g., Hahn Sohn Lohn Mohn). The second task, end-
sound-oddity, required detecting which of four monosyllable words did not
rhyme with the other three (all words presented in each trial shared the same
middle vowel or diphthong; e.g., Haus Baum Maus Laus). The third task,
onset-sound-oddity, required detecting which of four presented words did not
share the same consonant onset (all words presented in each trial shared the
same middle vowel or diphthong; e.g., Tal Tat Tag Rad).
First Grade Tasks
Verbal IQ. The verbal section of the Hamburg–Wechsler Intelligence Test
for Grade School Children (HAWIK-R, Tewes, 1983) was administered to
children at the end of first grade. This test includes word usage, verbal compre-
hension, and general knowledge items.
Word discrimination. This task was administered at the end of first grade.
Children were presented with three sets of four pictures and one set of three
pictures and were asked to identify these (i.e., ‘‘Apfel’’-apple, ‘‘Zug’’-train,
‘‘Auge’’-eye, ‘‘Baum’’-tree). Words printed in large letters were presented
for 1 s on a portable slide projector screen. Each set of trials with four pictures
had 12 words, 4 to 6 of which were the words represented in the pictures.
The rest were distractors (i.e., ‘‘Bauen’’ and ‘‘Bau’’ as distractors for
‘‘Baum’’). In the trial with three pictures, only 9 words were presented. Only
real words were used. Distractors were chosen for phonemic similarity. The
mean number of errors was low for word discrimination (mean, 4.4; standard
deviation, 3.9).
Second Grade Tasks
Word span. The German version of Case et al.’s (1982) word span task
was repeated.
Phoneme recoding. Two phonological tasks were administered in the sec-
ond grade (when most children were 8 years old). These tasks were more
difficult than those administered in preschool and required phonemically re-
coding information. The phoneme recognition task required children to listen
to nine pairs of prerecorded three syllable pseudowords and determine if these
were the same or if these differed. They were told the words may differ just
slightly. There were two pairs of pseudowords that were identical. If they
answered yes, they were instructed to listen carefully again to the two pseu-
dowords and say which sounds differ between the two. Responses were con-
sidered correct if one or both sounds were correctly identified. The phoneme
manipulation tasks required replacing and switching phonemes within two
syllable pseudowords. One practice and six experimental trials were adminis-
tered for each of these tasks. Children were first asked to listen to then
repeat a prerecorded pseudoword (e.g., ‘‘ANI’’). They were then instructed
to replace one sound with another. For example, after saying ‘‘ANI,’’ they
were asked to replace the ÉiÉ with ÉuÉ, resulting in ‘‘ANU.’’ Vowels were
always replaced with vowels and consonants with consonants.
The last task also had one practice and six experimental trials. The proce-
dure was the same, except that after repeating the prerecorded word, children
were asked to switch two sounds within the word (e.g., switching the ÉoÉ
and ÉmÉ sound in ‘‘OMT’’ to produce ‘‘MOT’’). It is assumed that an
orthographic representation strategy can be used in these tasks in addition to
phonological awareness.
A combined score for both of these tasks is used. This combined measure
is referred to as phoneme recoding.
Word and nonword decoding speed. Four letter words and pseudowords
(30 of each) were adapted from Rott and Zielinski (1984). This list was
developed according to specifications by Hogaboam and Perfetti (1978) in
order to test decoding speed. Items were presented on a computer screen. An
internal timing device measured children’s responses from the moment of
presentation on the screen. Letters were about 2.5 in. high. The experimenter
pressed the computer keyboard space bar as soon as the child pronounced
the last sound of each item.
In pilot studies, it was found that German children tend to pronounce the
first letter of each item while decoding (e.g., ‘‘rrrrrrr-rot’’). This may reflect
a ‘‘phonetic attack’’ decoding approach taught in many schools. Therefore,
voice onset was not appropriate for the speed measurement. A second concern
was comparability of word decoding and nonword decoding tasks. Voice
onset time is not justified in measuring nonword decoding, given that each
grapheme is decoded serially in German. It is also assumed that word recogni-
tion interacts with speed of production in word decoding responses. The
analyses with word and nonword decoding in this study were at the level of
speech production, given that recognition could not be accurately timed.
Nonword decoding is used in this study primarily as a measure of grapheme–
phoneme decoding in comparison to word decoding, which includes both
grapheme–phoneme recoding and lexical access processes. The decoding
data used here are from the fall of second grade (E-early) and early summer
(L-late) in second grade. The schools are in session in this region from late
September to early July. Very few children made errors in either decoding
task during pilot trials of the tasks and in the study itself. All items were
only four letters, and no irregular or unusual spellings were used. The tasks
were designed to minimize variance due to errors so that variance in speed
among subjects could be better estimated.
Reading comprehension. This test was developed by the first author to be
used in a large scale longitudinal study of over 1200 children in the Munich
area and suburbs. This test was administered in the fall (E-early) and the
early summer (L-late) during second grade. The first set of 18 items tested
comprehension within the context of individual sentences. The vocabulary
was drawn from textbooks being used by the Bavarian school system. Six of
these items tested word discrimination in addition to comprehension. Syn-
onyms were sought in four items, and the remaining items were word usage
and comprehension (e.g., translation: ‘‘You cannot read ,’’ (a) a newspa-
per, (b) a radio, (c) stories, (d) a book). The second part of the test included
five short stories, also with familiar vocabulary, each followed by two or
three multiple choice questions. Questions required children to infer from
part or all of the text for their answer choice.
Analyses with the larger group of 1200 children indicated that the more
difficult questions were those requiring inferences from more than one sen-
tence in the story. Chronbach as for test item reliability in each of the two
test parts at both testing times ranged from 0.81 to 0.88. During the last phase
of the study reported here, children were enrolled in 52 different second grade
classrooms in the region. Some children did not take the reading comprehen-
sion test, but did take part in all other tasks. Not all teachers allowed these
tests to be administered in their classrooms, and the researchers could not
justify giving these children a different test situation as the others. Therefore,
analyses for the reading measures include only 89 children. Composite literacy
measures (described later) include only children with scores from the reading
comprehension tests.
Spelling. The 18 items for this test were taken in part from Brügelman’s
TABLE 2
Distribution of Letter Knowledge the Summer before Entering First Grade
0 28 20.9 20.9
1 17 12.7 33.6
2 17 12.7 46.3
3 11 8.2 54.4
4 10 7.5 62.0
5 9 6.7 68.7
6–9 15 11.2 79.9
10–13 8 6.0 85.9
14–17 6 4.5 90.4
18–21 1 0.7 91.1
22–26 12 8.9 100.0
(1986) spelling test items. Additional items were taken from the official list
of vocabulary for the second grade distributed by the Ministry of Education
in the region of Bavaria. The spelling test was given twice, once in the fall
(E-early) and in the early summer (L-late) in the second grade.
TABLE 3
Descriptive Statistics Reported in Proportions for Phonological Tasks
Kindergarten
Rhyme detection .86 .23 0.40 1.0 (guessing controlled)
Syllable counting .82 .18 .30 1.0
Sound-in-word .77 .36 01.0 1.0 (guessing controlled)
Syllable segment .76 .21 .10 1.0
Pseudoword repeat .72 .20 .10 1.0
Syllable blend .63 .14 0 1.0
Onset/rime blend .55 .18 0 1.0
Onset-oddity .25 .28 0.20 1.0 (guessing controlled)
Middle-oddity .64 .29 0.06 1.0 (guessing controlled)
Rime-oddity .28 .28 0.20 1.0 (guessing controlled)
Second grade
Phoneme recognition .57 .23 0 .86
Phoneme manipulation .57 .17 .08 .83
Note. Proportions reporting minimums under zero are corrected for guessing. All other propor-
tions with minimums of zero and above indicate tasks not associated with guessing responses.
AID
TABLE 4
Correlations of Predictor Measures with Reading and Spelling Measures
JECP 2294
Comprehension Word decode P-word decode Spelling
Word
discrimination E L E L E L E L
Word span (4) .40** .24* .13 .01 0.04 0.02 0.08 .21* .31**
Letter know. (6) .27** .33** .32** 0.36** 0.04 0.36** 0.10 .16 .31**
Word span (6) .33** .38** .28** 0.41** 0.14 0.37** 0.17 .30** .48**
/ ad01$$2294
Rhyme detection (6) .32** .34** .38** 0.21* 0.37** 0.23* 0.20* .28** .36**
Syllable count. (6) .07 .12 .54** 0.01 0.02 0.07 0.15 .14 .36**
Sound-in-word (6) .27** .32** .27** 0.11 0.11 0.10 0.09 .23* .25*
Syllable segment (6) .12 0.07 0.12 0.02 0.02 0.09 0.14 0.01 0.12
P-word repeat (6) .14 .17 0.01 0.20* 0.33** 0.25* 0.27* .28** .35**
Syllable blend (6) .15 .05 .21* 0.08 .33** 0.07 0.04 0.12 .18
Onset/rime blend (6) .22* .26* 0.01 0.22* .10 0.16 0.01 .06 0.01
Phoneme oddity (6) .38** .50** .45** 0.35* 0.21* 0.26* 0.21* .38** .60**
NÄSLUND AND SCHNEIDER
Verbal IQ (7) .28** .37** .38** 0.13 0.21* 0.17 0.02 .23* .15
05-16-96 00:24:15
Word span (8) .12 .31** .29** 0.10 0.01 0.27** 0.06 .26* .11
Phon recode (8) .31** .48*** .31** 0.24* 0.07 0.36** 0.03 .44** .25*
jecpa
E, Early in second grade; L, late in second grade.
* p õ .05.
** p õ .01.
*** p õ .001.
AP: JECP
KINDERGARTEN PHONOLOGICAL AWARENESS 45
TABLE 5
Canonical Redundancy Analysis with Early Literacy and Kindergarten Predictor Measures
Note. Canonical r (Adjusted) Å .58; Canonical r 2 Å .42; likelihood ratio standard error Å .06,
degrees of freedom Å 40, p õ .002; variance of early literacy explained by: literacy variables
Å .54 and kindergarten predictor variables Å .28.
variables in the other set of variables. Table 5 provides the multiple correla-
tions and r 2 values of each of the predictor measures as it correlates and
subsequently explains the variance of the combined literacy measures. Pho-
neme-oddity and letter knowledge appear to explain the largest proportion of
the variance in the literacy variables. The proportion of variance accounted
for in early literacy by the linear combination of individual measures that
make up that variable is 54%. The amount of variance accounted for in the
literacy variable by the linear combination of the predictor measures is 28%.
The canonical correlation analyses only answers part of our question. The
significance of predictor measures in explaining the variance in our early
literacy measures will now be addressed.
Hierarchical regression models for early second grade. In order to deter-
mine the relative strength of these predictor variables, hierarchical regression
analyses were performed, using a weighted sum of the literacy variables
(multiplying standardized scores for the literacy measures by their standard-
ized canonical coefficients for each of the analyses). Table 6 provides the
results of these analyses. Only the 89 children who completed all reading tasks
are included in these analyses. Separate regression analyses were performed in
order to determine differences in results depending upon which phonological
processing measure is used. There are five separate models presented in
Table 6, each including only one of the kindergarten phonological awareness
measures. The redundancy analysis (see Table 5) already indicated the relative
strength of predictability of phoneme-oddity in comparison to the other phono-
logical measures. However, our purpose is, in part, to demonstrate relative
influences of letter knowledge and various phonological tasks. The change
TABLE 6
Hierarchical Regression for the Kindergarten Predictors on Early Literacy
* p õ .05.
** p õ .01.
*** p õ .001.
TABLE 7
Canonical Redundancy Analysis with Late Literacy and Kindergarten Predictor Measures
Note. Canonical r (Adjusted) Å .77; Canonical r 2 Å .59; likelihood ratio standard error Å .04,
degrees of freedom Å 32, p õ .0001; variance of early literacy explained by: literacy variables
Å .35 and predictor variables Å .21.
TABLE 8
Hierarchical Regression for the Kindergarten Predictors on Late Literacy
* p õ .05.
** p õ .01.
*** p õ .001.
was tested in the first year of kindergarten, then again along with the kinder-
garten phoneme-oddity, and in second grade along with phoneme recoding
(the combined score for the phonological tasks administered in the second
grade). Correlations among the word span and phonological tasks can be
found in Table 9. Phoneme-oddity was used as the kindergarten phonological
measure given its high predictability in later reading measures in comparison
with the other kindergarten phonological measures used in this study. All 134
children are included in the correlations and subsequent analyses of memory
and phonological awareness.
All correlations are significant (p õ .05 level), with the exception of word
span (4) and phoneme recoding (8). In order to test the relative proportion
of variance accounted for in word span (6 and 8) by the phonological measures
(6 and 8), and the relative proportion of variance in phonological measures
(6 and 8) accounted for by word span (4, 6, and 8), a series of forced step
hierarchical regressions were performed. The unique variance accounted for
TABLE 9
Correlations among the Phoneme-Oddity, Phoneme recode and Word Span Measures
* p õ .05.
** p õ .01.
*** p õ .001.
TABLE 10
Hierarchical Regressions for Word Span (6) and Phoneme-Oddity (6) as Dependent Measures
Change Change
Predictor step entered in r 2 Predictor step entered in r 2
* p õ .05.
** p õ .01.
*** p õ .001.
TABLE 11
Hierarchical Regressions for Word Span (8) and Phoneme recode (8) as Dependent Measures
Change Change
Predictor step entered in r 2 Predictor step entered in r 2
* p õ .05.
** p õ .01.
*** p õ .001.
in phoneme-oddity (6), even after controlling for letter knowledge (K) in each
model. Phoneme-oddity (6) predicts a significant proportion of the variance
in word span (6) after being placed second in the model, either after letter
knowledge (6) or after word span (4). Word span (4) is a significant predictor
of phoneme recoding (8), even when controlling for either letter knowledge
(6) or for phoneme-oddity (6). Word span (6) is a significant predictor of
phoneme recoding (8) after letter knowledge (6), but not after phoneme-
oddity (6). Word span (8) was not a significant predictor of phoneme recoding
(8). Neither phoneme-oddity (6) nor phoneme recoding (8) were significant
predictors of word span (8).
These analyses suggest an interaction of verbal memory span and phonolog-
ical awareness by kindergarten for most of our German subjects, given the
reciprocal strength of word span (4 and 6) and phoneme-oddity (6). By the
second grade, it is likely that the influence of memory span and phonological
processing are mutually influential to such a high degree that their effects
might be considered redundant. This may prevent attempts to statistically
partial out the unique effects of phonological processing and verbal memory
measures. Recall the regression analyses for the early and late composite
literacy variables (see Tables 6 and 8). Word span (6) is significant in all but
the rhyme detection models when placed last after the letter knowledge,
phonological, and verbal IQ measures in predicting early literacy. For the
late composite literacy variable, word span (6) is significant when placed last
in all of the models, except those including rhyme detection and phoneme
oddity. This indicates that word span (6) is itself a reliable predictor of the
early and late literacy measures in this study, and probably shares a large
proportion of the variance with the phonological measures.
What this pattern of results also may suggest is that by the age of six,
spontaneous phonological recoding and rehearsal strategies develop and in-
fluence verbal memory for some children, such that phonological recoding
becomes an integral component of verbal memory (Gathercole et al., 1991).
An additional analysis was performed to test the interaction of phoneme
oddity (6) and word span (6). A significant partial correlation (controlling for
letter knowledge) was found between intra individual differences in memory
span for lists of phonologically similar and phonologically dissimilar words
(indicating an interference effect for phonologically similar lists of words)
and phoneme-oddity (r Å .57, p õ .001). The correlation between word span
(K) and phoneme-oddity (K) was .33 (p õ .001). The lists of phonologically
similar words shared either onset, middle sound, or rime. Larger discrepancies
in children’s recall for phonologically similar and dissimilar words suggests
larger phonological interference effects and spontaneous phonological recod-
ing in working memory (see Brady et al. 1983; Mann et al., 1980; Mann &
Liberman, 1984). The significant positive correlation between phonological
interference effects and phoneme-oddity performance further suggests that
memory span and phonological awareness are covaring processes for many
children by the age of 6.
GENERAL DISCUSSION
Previous studies have indicated that among beginning readers, phonological
processing skills are highly reliable predictors of reading performance. For
example, for the 5-year-old children in a study conducted by Liberman, Shan-
kweiler, Fisher, and Carter (1974), reading performance was significantly
predicted by syllable segmentation. The 6-year-old preliterate children in this
study demonstrated a ceiling effect in the syllable segmenting and counting
tasks. Due to their age, the German sample was likely more developed meta-
linguistically than children in the Liberman et al. (1974) study. However,
when compared to the semiliterate 6-year-old children in Perfetti et al. (1987)
and Yopp (1988) studies, the German children at age 6 would appear to be
less proficient in some of the more advanced phonological tasks, especially
tasks that require mental operations at the level of individual phonemes.
The development of phonological processes involved in recoding informa-
tion in working memory in the preschool years needs to be investigated in
more detail. Gathercole et al. (1991) have conducted a longitudinal investiga-
tion with young children, introducing a unitary variable they call ‘‘phonologi-
cal working memory,’’ which is measured by tasks of pseudoword repetition.
They argue that phonological processing and verbal memory are integrated
cognitively to the point where these cannot be theoretically or practically
p õ .0001). The low level of errors for this German sample in both decoding
tasks may be due to these orthographic-specific factors. However, longer letter
strings or presenting lists rather than single items in the decoding tasks may
elicit more errors.
A study recently completed by Näslund, Schneider, and van den Broek
(1994), which compared the relationships among reading, phonological, mem-
ory, and lexical access measures between German and American children in
the first and second grade, used multiple lists of four words or nonwords
each, instead of single words for the decoding task. The lists were also of
increasing complexity. More errors, and thus measurable variance, were elic-
ited using this method, compared to the easier decoding tasks used in the
present study. The number of errors for nonwords (but not words) was signifi-
cantly related (just at the .05 level) to reading comprehension. The speed of
word and nonword decoding was more highly related to comprehension for
both German and American children compared to decoding errors. Differential
relationships among phonological awareness, memory span, lexical access,
decoding (using longer strings of words and nonwords), and reading compre-
hension, and differences in types of errors and strategies between German
and American children will provide further information about variability in
reading associated with orthographic and strategic differences in these two
populations of beginning readers.
REFERENCES
Bowey, J. A. (1994). Phonological sensitivity in novice readers and nonreaders. Journal of
Experimental Child Psychology, 58, 134–159.
Bradley, L., & Bryant, P. (1985). Rhyme and reason in reading and spelling. Ann Arbor, MI:
Univ. of Michigan Press.
Brady, S., Mann, V., & Schmidt, R. (1987). Errors in short term memory for good and poor
readers. Memory and Cognition, 15(5), 444–453.
Brady, S., Shankweiler, D., & Mann, V. (1983). Speech perception and memory coding in
relation to reading ability. Journal of Experimental Child Psychology, 35, 345–367.
Brügelman, H. (1986). Lese- und schreibaufgaben für schulanfänger. Universität Bremen: Stu-
diengang Primarstufe.
Bryant, P., Bradley, L., MacLean, M., & Crossland, J. (1989). Nursery rhymes, phonological
skills and reading. Journal of Child Language, 16, 407–428.
Byrne, B., & Fielding-Barnsley, R. (1990). Phonemic awareness and letter knowledge in the
child’s acquisition of the alphabetic principle. Journal of Educational Psychology, 82, 805–
812.
Byrne, B., & Fielding-Barnsley, R. (1991). Evaluation of a program to teach phonemic awareness
to young children. Journal of Educational Psychology, 82, 451–455.
Case, R., Kurland, D. M., & Goldberg, J. (1982). Operational efficiency and the growth of short-
term memory span. Journal of Experimental Child Psychology, 33, 386–404.
Ehri, L. (1989). The development of spelling knowledge and its role in reading acquisition and
reading disability. Journal of Learning Disabilities, 22(6), 356–365.
Ellis, N., & Large, B. (1988). The Early Stages of Reading: A Longitudinal Study. Applied
Cognitive Psychology, 2(1), 47–76.
Frith, U. (1985). Beneath the surface of developmental dyslexia. In K. E. Patterson, J. C.
Marshall, & M. Coltheart (Eds.), Surface dyslexia: Neuropsychological and cognitive studies
of phonological reading (pp. 301–330). Hillsdale, NJ: Erlbaum.
Gathercole, S. E., Willis, C., Emslie, H., & Baddeley, A. D. (1991). The influences of number
of syllables and wordlikeness on children’s repetition of nonwords. Applied Psycholinguis-
tics, 12, 349–367.
Goswami, U. (1988). Orthographic analogies and reading development. Quarterly Journal of
Experimental Psychology, 40A, 239–268.
Harris, R. J. (1985). A Primer of Multivariate Statistics (2nd ed.). San Diego: Harcourt Brace
Jovanovich.
Hogaboam, T. W., & Perfetti, C. A. (1978). Reading skill and the role of verbal experience in
decoding. Journal of Educational Psychology, 70, 717–729.
Hulme, C. (1984). Developmental differences in the effects of acoustic similarity on memory
span. Developmental Psychology, 20, 650–652.
Hulme, C., & Tordoff, V. (1989). Working memory development: The effects of speech rate, word
length and acoustic similarity on serial recall. Journal of Experimental Child Psychology, 47,
72–87.
Jansen, H., Knorn, P., Mannhaupt, G., Marx, H., Beck, M., & Skowronek, H. (1986). Bielefelder
screening zur vorhersagen von lese-rechtschreibschwierigkeiten. Universität Bielefeld, SBF
227, F. R. Germany.
Leather, C. V., & Henry, L. A. (1994). Working memory span and phonological awareness tasks
as predictors of early reading ability. Journal of Experimental Child Psychology, 58, 88–
111.
Liberman, I. Y., Shankweiler, D., Fisher, M. F., & Carter, B. (1974). Explicit syllable and
phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18,
201–212.
Lundberg, I. (1987). Are letters necessary for the development of phonological awareness?
Cahiers de Psychologie Cognitive, 7, 472–475.
Lundberg, I. (1989). Lack of phonological awareness. A critical factor in dyslexia. In C. von
Euler, G. Lennerstrand, & I. Lundberg (Eds.), Brain and reading (pp. 221–231). New York:
Macmillan Co.
Lundberg, I., Frost, J., & Petersen, O. P. (1988). Effects of an extensive program for stimulating
phonological awareness in preschool children. Reading Research Quarterly, 23, 263–284.
Lundberg, I., Olofsson, Å., & Wall, S. (1980). Reading and spelling skills in the first school
years predicted from phonemic awareness skills in kindergarten. Scandinavian Journal of
Psychology, 21, 159–173.
Lundberg, I. & Høien, T. (1991). Initial enabling knowledge and skills in reading acquisition: Print
awareness and phonological segmentation. In D. J. Sawyer & B. J. Fox (Eds.), Phonological
awareness in reading: The evolution of current perspectives (pp. 73–96). New York:
Springer-Verlag.
MacLean, M., Bryant, P., & Bradley, L. (1987). Rhymes, nursery rhymes and reading in early
childhood. Merrill-Palmer Quarterly, 33, 255–282.
Mann, V. A. (1991). Phonological awareness and early reading ability: One perspective. In
D. J. Sawyer & B. J. Fox (Eds.), Phonological awareness in reading: The evolution of
current perspectives (pp. 191–215). New York: Springer-Verlag.
Mann, V. A., & Liberman, I. Y. (1984). Phonological awareness and verbal short-term memory.
Journal of Learning Disabilities, 17, 592–598.
Mann, V. A., Liberman, I. Y., & Shankweiler, D. (1980). Children’s memory for sentences and
word strings in relation to reading ability. Memory and Cognition, 8(4), 329–335.
McDougall, S., Hulme, C., Ellis, A., & Monk, A. (1994). Learning to read: The role of short-
term memory and phonological skills. Journal of Experimental Child Psychology, 58, 112–
133.
Morais, J. (1991). Phonological awareness: A bridge between language and literacy. In D. J.
Sawyer & B. J. Fox (Eds.), Phonological awareness in reading: The evolution of current
perspectives (pp. 31–72). New York: Springer-Verlag.
Morais, J., Cary, L., Alegria, J., & Bertelson, P. (1979). Does awareness of speech as a sequence
of phones arise spontaneously? Cognition, 7, 323–331.
Näslund, J. C., Schneider, W., & van den Broek, P. W. (1994, October). Beginning reading in
Germany and the U.S.: A comparison of phonological segmentation, decoding, lexical access
and comprehension. Paper presented at the NATO Advanced Study Institute for Cognitive
and Linguistic Bases of Reading, Writing, and Spelling, Alvor-Algarve, Portugal.
Perfetti, C., Beck, I., Bell, L., & Hughes, C. (1987). Phonemic knowledge and learning to read
are reciprocal: A longitudinal study of first grade children. Merrill-Palmer Quarterly, 33,
283–319.
Read, C., Zhang, Y., Nie, H., & Ding, B. (1986). The ability to manipulate speech sounds
depends on knowing alphabetic reading. Cognition, 24, 31–44.
Rott, C., & Zielinski, W. (1984). Entwicklung der lesefertigkeit in der grundschule. Zeitschrift
für Entwircklungspsychologie und Pädagogische Psychologie, 18(2), 165–175.
Sawyer, D. J., & Fox, D. J. (1991). Phonological awareness in reading: The evolution of current
perspectives. New York: Springer-Verlag.
Smith, P. T. (1986). The development of reading: The acquisition of a cognitive skill. In
P. Fletcher & M. Garman (Eds.), Language acquisition: Studies in first language acquisition
(pp. 475–493). Cambridge, UK: Cambridge Univ. Press.
Snowling, M., Chiat, S., & Hulme, C. (1991). Words, nonwords, and phonological processes:
Some connections on Gathercole, Willis, Emslie, and Baddeley. Applied Psycholinguistics,
12, 369–373.
Stanovich, K. E., Cunningham, A. E., & Feeman, D. J. (1984). Intelligence, cognitive skills, and
early reading progress. Reading Research Quarterly, 19, 278–303.
Tewes, U. (1983). Hamburg-Wechsler Intelligenztest für Kinder (HAWIK-R). Bern, Switzerland:
Huber.
Treiman, R. (1985). Onsets and rimes as units of spoken syllables: Evidence from children.
Journal of Experimental Child Psychology, 39, 161–181.
Treiman, R. (1991). Phonological awareness and its roles in learning to read and spell. In D. J.
Sawyer & B. J. Fox (Eds.), Phonological awareness in reading: The evolution of current
perspectives (pp. 191–215). New York: Springer-Verlag.
Tunmer, W. E., & Rohl, M. (1991). Phonological awareness and reading acquisition. In D. J.
Sawyer & B. J. Fox (Eds.), Phonological awareness in reading: The evolution of current
perspectives (pp. 1–30). New York: Springer-Verlag.
Van Orden, G. C. (1987). A rows is a rose: Spelling, sound, and reading. Memory and Cognition,
15, 181–198.
Wagner, R. W., & Torgesen, J. K. (1987). The nature of phonological processing and its causal
role in the acquisition of reading skills. Psychological Bulletin, 101(2), 192–212.
Wagner, R. W., & Torgesen, J. K., & Rashotte, C. A. (1994). The development of reading-
related phonological processing abilities: New evidence of bi-directional causality from a
latent variable longitudinal study. Developmental Psychology, 30, 73–87.
Weinert, F. E., & Schneider, W. (1989). The Munich longitudinal study on the genesis of
individual competencies (LOGIC). Report No. 5. Results of wave three. Max–Planck Insti-
tute for Psychological Research, Munich, Germany.
Wimmer, H., & Hummer, P. (1990). How German-speaking first graders read and spell: Doubts
on the importance of the logographic stage. Applied Psycholinguistics, 11, 349–368.
Yopp, H. K. (1988). The validity and reliability of phonemic awareness tests. Reading Research
Quarterly, 23, 159–177.
RECEIVED: May 15, 1992; REVISED: March 7, 1995.