Journal of Clinical and Experimental Neuropsychology

ISSN: 1380-3395 (Print) 1744-411X (Online) Journal homepage: https://www.tandfonline.com/loi/ncen20

Executive functions for reading and writing in

typical literacy development and dyslexia

Leah E. Altemeier , Robert D. Abbott & Virginia W. Berninger

To cite this article: Leah E. Altemeier , Robert D. Abbott & Virginia W. Berninger (2008) Executive
functions for reading and writing in typical literacy development and dyslexia, Journal of Clinical and
Experimental Neuropsychology, 30:5, 588-606, DOI: 10.1080/13803390701562818

To link to this article: https://doi.org/10.1080/13803390701562818

Published online: 05 Jun 2008.

Submit your article to this journal

Article views: 3172

View related articles

Citing articles: 49 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ncen20
JOURNAL OF CLINICAL AND EXPERIMENTAL NEUROPSYCHOLOGY
2008, 30 (5), 588–606

Executive functions for reading and writing in typical

NCEN

literacy development and dyslexia

Leah E. Altemeier, Robert D. Abbott, and Virginia W. Berninger

Executive Functions in Literacy and Dyslexia

University of Washington, WA, USA

Experiment 1: Hierarchical linear modeling of growth trajectories of three executive functions (inhibition; rapid
automatic switching, RAS; and combined inhibition/switching) in typical readers and writers showed steady
improvement of inhibition but leveling of RAS and inhibition/switching about fourth grade. In multiple regres-
sions, RAS, entered after inhibition, contributed uniquely to literacy outcomes at every grade. Improvement of
executive functions over the first four grades predicted literacy outcomes at fourth grade. Experiment 2: For children
with dyslexia, executive functions explained less variance in literacy outcomes, and boys were worse in inhibition
and inhibition/switching. Developmental, educational, and clinical significance of findings are discussed.

Definitional issues component processes across time and space (Roberts

Conceptualization of executive functions has
shifted from a unitary construct of supervisory
attention to a constellation of functions that enable
an individual to self-regulate and engage in goal-
directed behavior (Lyon & Krasnegor, 1996). Fac-
tor analytic studies have validated at least three
executive functions: inhibition, shifting, and
updating/monitoring (Hughes, 1998; Miyake et al.,
2000; Pennington, 1997a, 1997b). These executive
functions are both intercorrelated and separable
(Miyake et al., 2000). Most researchers agree that
inhibition is the primary executive function that
precedes and allows development of other execu-
tive functions (Barkley, 1997; Carlson & Moses,
2001). Inhibition supports other executive func-
tions such as mental set shifting, which requires
switching attention among stimuli or tasks (Blair,
Zelazo, & Greenberg, 2005; Hughes, 1998; Miyake
et al., 2000; Pennington, 1997a, 1997b). Other rele-
vant executive functions include self-monitoring,
updating working memory, setting goals, and mak-
ing plans (Barkley, 2003; Denckla, 1996; Miyake
et al., 2000; Pennington, Bennetto, McAleer, &
Roberts, 1996). Executive functions also coordinate
& Pennington, 1996).
Defined broadly, executive functioning is self-
government for regulating mental functions. Exec-
utive functions are important in any problem-solving
activity that is not automatic because conscious,
reflective problem solving requires inhibiting irrel-
evant or overlearned responses when generating
response options, devising and applying strategies
during planning, and maintaining these strategies
and shifting them as needed when pursuing a goal
(Zelazo, Carter, Reznick, & Frye, 1997). However,
executive functions may also facilitate the develop-
ment of efficient, automatic processing of information
(Denckla & Cutting, 1999).
Inhibition
In order to engage in conscious, reflective prob-
lem solving, one first needs to inhibit automatic
responses in order to engage strategic processes in
favor of a long-term goal. Inhibition delays a pre-
potent (automatic, overlearned) response in order
to achieve a goal and may protect that delay even
in the face of interference (Barkley, 2003). When
prepotent responses are inhibited, a temporal

Grants HD25858 and P50 33812 from the National Institute of Child Health and Human Development (NICHD) supported this
research.
Address correspondence to Leah E. Altemeier, 5506 33rd Ave NE, Suite D, Seattle, WA 98105, USA (E-mail: leaha@
u.washington.edu).

© 2007 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business
http://www.psypress.com/jcen DOI: 10.1080/13803390701562818
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA
pause occurs that allows for development and/or
implementation of self-directed or self-regulatory
actions. According to Barkley, inhibition first
appears in development around age three or four,
but continues to develop through adolescence. The
other self-regulatory functions follow behavioral
inhibition in development. Inhibition can occur at
the behavioral level (response control) and/or at the
cognitive level (attentional inertia). Theories may
emphasize one over the other, but usually the bound-
aries between them are blurred as cognitive disinhibi-
tion often leads to behavioral disinhibition. At a
cognitive level, lack of inhibition leads to representa-
tional inflexibility rather than a response control
problem (Diamond, 2002). According to Diamond,
Carlson, and Beck (2005), cognitive set shifting is an
outgrowth of being able to exert inhibitory control.
Set shifting and rapid automatic switching
(RAS)
Set shifting is the ability to flexibly shift atten-
tion as task demands change. Set shifting requires
inhibiting an activated set in order to shift to and
activate another set. Both inhibition and attention-
switching mechanisms (among stimuli and tasks)
are probably involved in set shifting (Barkley, 2003;
Denckla, 1996; Miyake et al., 2000; Pennington
et al., 1996). Perseveration in repeating the same
response may be due to an inability to switch atten-
tional focus from one target to another one (Lezak,
1995). Rapid automatic switching, which Wolf
(1986) developed to study normal reading and dys-
lexia, involves set shifting of linguistic information.
RAS taps executive functions that regulate scan-
ning across lines, sequencing within lines, and inte-
gration of visual and verbal codes (Wolf & Bowers,
1999). Rapid automatic naming (RAN) tasks
involve naming stimuli (e.g., letters; Denckla &
Rudel, 1974; Wolf, Bally, & Morris, 1986), but
RAS or rapid automatic switching involves nam-
ing alternating categories of stimuli (e.g., letters
and numbers). RAS therefore increases the role of
executive functioning over RAN by requiring more
disengaging from the preceding stimulus in order
to switch attention to the next stimulus.
Relationship of executive functions
to reading and writing
Executive functioning may be related to normal
reading and writing development, but the relation-
ship of executive functioning to typical reading
development has received scant research attention.
Executive functions may govern the integration of
visual and linguistic information and the automatic
589
retrieval of linguistic information from memory
while learning to read. Inhibition is thought to
affect memory at both the encoding and retrieval
stages (de Ribaupierre, 2002). Efficient retrieval of
the phonological codes for letters and letter units,
for example, may be affected by the ability to sup-
press irrelevant codes and quickly search and
retrieve the relevant response.
To explain the basis for dyslexia (a reading and
spelling disorder), researchers have primarily
focused on language-based processes such as pho-
nology (Morris et al., 1998; Wagner & Torgesen,
1987). More recently, however, executive functions
(Swanson, 2000) have been explored as a contrib-
uting factor to dyslexia because affected individuals
have difficulty with organization, automatization,
and integration of multiple processes and perform
poorly on executive functions of inhibition and
shifting/cognitive flexibility (Berninger et al.,
2006b; Brosnan et al., 2002; Denckla, 1996;
Helland & Asbjornsen, 2000; Kelly, Best, & Kirk,
1989; Purvis & Tannock, 2000; Reiter, Tucha, &
Lange, 2005). To separate processing speed from
inhibition performance, Reiter and colleagues
(2005) showed that children with dyslexia were not
only slower in their processing time on the color
naming, word reading, and the inhibition condi-
tions of the Stroop test, but also demonstrated a
small to medium effect size in the number of errors
and corrections on the interference (inhibition)
condition. This finding suggests that children with
dyslexia exhibit a tendency to engage in prepotent
(i.e., disinhibited) responses.
Executive functions are correlated with writing
tasks in normally developing populations (Hooper,
Swartz, Wakely, de Kruif, & Montgomery, 2002),
influence handwriting (Berninger et al., 2006a) and
overall written output (Hooper et al., 2002), and
add unique variance to models of integrated read-
ing–writing tasks such as note-taking and report
writing (Altemeier, Jones, Abbott, & Berninger,
2006). From the perspective of writing, executive
functions have been defined as control processes
that guide the self-initiation of thoughts, affect,
and behaviors used to attain writing goals (Zim-
merman & Risemberg, 1997). Executive functions
monitor recursive planning, translating, and
reviewing/revising processes in the problem-solving
process of writing (Hayes & Flower, 1980). Zelazo
and colleagues’ (1997) sequential framework for
executive functions begins with problem represen-
tation, followed by plan generation and then exe-
cution (intending/rule use), and finally evaluation
(error detection/correction). This framework is
remarkably similar to Hayes and Flower’s (1980)
influential model of the cognitive processes in writing.
590 ALTEMEIER, ABBOTT, BERNINGER

Individual differences in executive functions for were no main effects for gender on executive tasks
self-regulation of the writing process may affect and no interactions between age and gender in
both the high-level composing processes (Graham, their sample of children aged 3 to 12 years (Welsh,
1997; Singer & Bashir, 1999) and lower level hand- Pennington, & Groisser, 1991). Few studies have
writing and spelling processes (Berninger & investigated gender differences in both children
Amtman, 2003). Lower level writing processes with and those without reading or writing disabil-
include automatic letter writing of the alphabet ity as was the aim of the current study.
from memory, which is predicted by executive
functions (Berninger, Nielsen, Abbott, Wijsman, &
Research questions
Raskind, in press-b) and, in turn, predicts high-
level composing (Berninger & Amtman, 2003).
The first study, which investigated the develop-
Executive functions also affect dysgraphia (hand-
ment of executive functioning and its contribution
writing and/or spelling disability) that occurs alone
to literacy outcomes for typically developing read-
and/or in combination with dyslexia, which is a writ-
ers and writers, addressed three research questions:
ing, as well as reading disorder (Berninger et al., in
(a) What are the developmental trajectories of inhi-
press-b). Handwriting automaticity requires execu-
bition, rapid automatic switching, and inhibition/
tive control for the integration of multiple processes,
switching? (b) What are the concurrent relation-
including motor planning, orthography, ortho-
ships between those three executive functioning
graphic-motor integration via the orthographic loop
and literacy outcomes in typically developing first
of working memory, and processing speed, any one
through fifth graders? (c) What is the longitudinal
of which may be impaired in dysgraphia (e.g.,
relationship between the development (slope) of
Berninger et al., in press-b). Individuals with dys-
those three kinds of executive functioning across
lexia are impaired in executive functions involving
Grades 1 to 4 and literacy outcomes at Grade 4?
inhibition, switching set, and updating verbal work-
Three hypotheses were tested: (a) The three execu-
ing memory (Berninger et al., 2006b).
tive functions would improve over the course of
elementary school; (b) the three executive func-
tions would contribute differentially to concurrent
Gender differences in executive functioning
literacy skills; and (c) longitudinal improvement in
executive functions would lead to increased liter-
Research findings on gender differences in dyslexia
acy achievement. The second study addressed two
and in executive functions have been mixed. For
research questions across samples of children with
example, most epidemiological studies have found
and without dyslexia: (a) Is the relationship
more males than females to have dyslexia, but
between executive functioning and literacy the
genetics research suggests that the gender differ-
same or different for typically developing fifth-
ences observed in dyslexia are actually due to the
grade readers and writers and children with dys-
associated writing problems that are more severe
lexia of the comparable age? (b) Are there gender
in boys than girls (Berninger, Nielsen, Abbott,
differences in executive functioning in typically
Wijsman, & Raskind, in press-a). Some research
developing readers and writers, and/or children
evidence documents a gender difference with girls’
with dyslexia? Two hypotheses were tested: (a)
performance superior to boys in executive func-
Executive functions would contribute more to
tions. Preschool boys may show less response inhi-
reading and writing achievement in children with-
bition than girls (Berlin & Bohlin, 2002; Carlson &
out than in those with dyslexia; and (b) gender dif-
Moses, 2001). In a sample of typically developing
ferences would be found such that boys performed
children, boys outperformed girls on tasks of oral
more poorly on the same three measures of execu-
fluency, while girls outperformed boys on tasks of
tive function as those used in Experiment 1.
written fluency (Berninger & Fuller, 1992); execu-
tive functions are thought to control the memory
search processes underlying fluency. American
EXPERIMENT 1
Caucasian girls outperformed their male counter-
parts on a Stroop test (Lezak, 1995; Wolff, Hur-
Method
vitz, Imamura, & Lee, 1983). Well-documented
gender differences exist in the prevalence of atten-
Participants
tion-deficit/hyperactivity disorder (ADHD; male-
to-female ratio of 3:1; Barkley, 2003), a disorder in For all of the research questions in the first
which deficits in executive functioning are primary. study, students from a longitudinal study of normal
However, in a study of typical development, there reading and writing development participated.
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA
This sample was recruited from local schools in a
large metropolitan urban school district in the
Northwest region of the United States. All parents
of children who would be entering first grade or
third grade in the fall received a letter explaining
the opportunity to participate in a research project
and procedures for contacting the research coordi-
nator for further information if interested. The
research coordinator obtained informed consent
from those parents who decided to enroll their
child and bring their child to the university annu-
ally for five years to complete comprehensive
assessment of writing, reading, and related proc-
esses. Altogether 128 first graders (71 girls and 57
boys) participated in Year 1 Cohort 1 when they
were in first grade, and 113 third graders (57 girls
and 56 boys) participated in Year 1 Cohort 2.
Overall, the attrition rate was low and mainly due
to families moving to another state. For Cohort 1,
124 participated in Year 2, 122 in Year 3, 119 in
Year 4, and 114 in Year 5. For Cohort 2, 110 par-
ticipated in Year 2, 106 in Years 3 and 4, and 99 in
Year 5. However, at the time this research was con-
ducted, data were available for only Years 1 to 4
when Cohort 1 was in Grades 1 to 4, and Cohort 2
was in Grades 3 to 6.
Self-reported ethnicity and parents’ level of edu-
cation (indicators of socioeconomic background)
documented the diversity in the sample in Year 1:
(a) Asian-American (23.4%, Cohort 1; 21.2%,
Cohort 2), African-American (6.3%, Cohort 1;
9.7%, Cohort 2), European American (64.8%,
Cohort 1; 65.5%, Cohort 2), Hispanic (1.6%,
Cohort 1; 0.9%, Cohort 2), Native American
(1.6%, Cohort 1 only), and other (2.3%, Cohort 1;
2.7%, Cohort 2); and (b) lower than high-school
education or graduated from high school (7%
mothers and 12.5% fathers, Cohort 1; 7.1% moth-
ers and 7.1% fathers, Cohort 2); higher than a
high-school education but lower than a college
education (11.7% mothers and 7.8% fathers,
Cohort 1; 11.5% mothers and 14.2% fathers,
Cohort 2); undergraduate education (45.3% moth-
ers and 39.8% fathers, Cohort 1; 50.4% mothers
and 36.3% fathers, Cohort 2); graduate degrees
(33.6% mothers and 32.0% fathers, Cohort 1;
30.1% mothers and 35.4% fathers, Cohort 2); miss-
ing for 2.4% of the mothers and 7.9% of the fathers
in Cohort 1 and 0.9% of the mothers and 7.2% of
the fathers in Cohort 2.
Measures
Inhibition. The Delis–Kaplan Executive Func-
tion System (D-KEFS; Delis, Kaplan, & Kramer,
2001) Inhibition score is derived from performance
591
on the Color–Word Interference Test, based on
the classic Stroop (1935) test. The Color–Word
Interference test consists of four conditions: nam-
ing color patches, reading words, naming the color
in which words are printed (instead of reading the
words), and switching between naming the color
in which words are printed and reading words that
are printed within a box. The score from the third
condition, Inhibition, is used in this study, and
the score is based on the time students take to
complete the task. Test–retest reliability for this
subtest is .90.
Inhibition/Switching. The D-KEFS (Delis et al.,
2001) Inhibition/Switching score is derived from
the fourth condition of the Color–Word Interfer-
ence Test, in which the individual is required to
switch quickly between naming the color in which
words are printed and reading words that are
printed within a box. Similar to the Inhibition
score, Inhibition/Switching is based on the time
students take to complete the task. Average test–
retest reliability coefficient for this subtest is .80.
Rapid automatic switching. The Process Assess-
ment of the Learner (PAL; Berninger, 2001) Rapid
Automatic Naming of Words and Letters subtest
(PAL RAS) requires the individual to alternate
between rapidly naming a high-frequency struc-
ture/content word (article, pronoun, conjunction,
preposition) that is not completely predictable in
phonological decoding (e.g., the, one) and a double-
digit number presented in five rows of 10 items
each. Z-scores for grade were computed based on
means and standard deviations for grade in the test
manual. Average test–retest reliability coefficient
for this subtest is .92.
Reading accuracy. The Wechsler Individual
Achievement Test–Second Edition (WIAT-II; Psy-
chological Corporation, 2001) Word Reading
and Pseudoword Decoding subtests were used to
measure untimed real-word reading accuracy and
pseudoword decoding accuracy. Average test–
retest reliability coefficients for both subtests are
.97. In addition, the scaled score for Accuracy on
the Gray Oral Reading Test, Third Edition (GORT-
III; Wiederholt & Bryant, 1992) was included. The
average test–retest reliability coefficient for
GORT-III Accuracy is .90.
Reading rate. The Sight Word Reading Effi-
ciency and Phonemic Decoding Efficiency subtests
from the Test of Word Reading Efficiency
(TOWRE; Torgesen, Wagner, & Rashotte, 1999)
were used to measure reading rate. The TOWRE
592 ALTEMEIER, ABBOTT, BERNINGER

consists of two timed measures of real-word read- Association’s guidelines for safeguarding the rights
ing and pseudoword decoding. The score reflects of participants in research.
accuracy within a 45-second time limit. Test–retest
reliability coefficients for these subtests are .91 and Data analysis
.90, respectively. In addition, the scaled score for
First research question. Hierarchical linear mod-
Rate from the GORT-III (Wiederholt & Bryant,
eling (HLM) was used to model the trajectories of
1992) was included. The average test–retest relia-
the three measures of executive functioning: D-
bility coefficient for GORT-III Rate is .87.
KEFS Inhibition, PAL RAS, and D-KEFS Inhibi-
tion/Switching. Trajectories from both the younger
Reading comprehension. The WIAT-II (Psycho-
(Grades 1 through 4) and older (Grades 3 through
logical Corporation, 2001) Reading Comprehen-
6) cohorts were modeled. The two trajectories were
sion subtest requires students to read and answer
evaluated statistically to determine whether they
questions pertaining to different passages within a
could be combined into a single trajectory mode-
grade-appropriate item set. Questions about the
ling the development of each executive function
passages involve the content of the passages, such
from Grade 1 through Grade 6 for Inhibition and
as detecting the main idea and supporting details,
Inhibition/Switching and for Grade 1 through
making inferences, and defining vocabulary. The
Grade 5 for RAS (because this measure was not
test–retest reliability coefficient is .95.
given to Cohort 2 in Year 4 when they were in
Grade 6). This trajectory was then examined for
Written expression. The WIAT-II (Psychologi-
developmental trends.
cal Corporation, 2001) Written Expression score
is based on combined performance on Word Flu-
Second research question. Multiple regressions
ency and Combining Sentences in the primary
were performed separately for each grade level to
grades and on Word Fluency, Combining Sen-
identify the combination of variables that together
tences, and Paragraph tasks in the upper grades.
account for the most variance and how much addi-
For Word Fluency, students are given 60 seconds
tional variance is explained by systematically adding
to write as many words as possible that share a
predictor variables one at a time according to a the-
common characteristic. For Combining Sen-
ory-driven model. For the regressions, data from the
tences, students are asked to combine sentences
two cohorts were combined for Grade 3. These ana-
into one meaningful sentence and to generate a
lyses were restricted to Grades 1 to 5 because RAS
sentence from cues. The Paragraph task requires
was not available in Grade 6 (Year 4, Cohort 2).
the student to write a paragraph based on a story
starter format that uses good spelling, punctua-
Third research question. HLM was used to create
tion, organization, and vocabulary. The test–
values representing the development (indexed by
retest reliability coefficient is .86.
slope) of each executive function measure over time,
which was then entered as an independent variable in
Spelling. The WIAT-II (Psychological Corpora-
regression analyses to determine the correlation of
tion, 2001) Spelling subtest requires the individual
the degree of longitudinal development of executive
to spell real words in writing from oral dictation.
function with literacy outcomes at Grade 4. This
The test–retest reliability coefficient for this subtest
analysis focused on the effect of growth of executive
is .94.
functions during Grades 1 to 4 on fourth-grade read-
ing and writing achievement because Grade 4 is a
Procedure transition period in writing development from early
childhood to middle childhood when curriculum
All measures were individually administered
places increasingly more complex demands on devel-
during an annual visit of the participating children
oping writers than in the earlier grades.
to the university by supervised research assistants.
Age norms were used for all measures except the
PAL for which means and standard deviations for
EXPERIMENT 2
grade in the national norming sample were used to
compute z-scores; the scaled scores for PAL II
Method
RAS were not available at the time this research
was conducted. This study was deemed exempt
Participants
by the institutional review board for human
subjects, but was conducted in keeping with the The sample of typically developing children has
university guidelines and the American Psychological already been described. Children with dyslexia had
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA
been recruited for participation through mailings to
directors of special education, contact with school
psychologists, special and general educators, news-
paper and radio announcements, self-referred par-
ents, and referrals from the medical, psychological,
and educational professions. These children quali-
fied their family for participation in the genetics
study if they met the following inclusion criteria: (a)
prorated Verbal IQ (based on Information, Similar-
ities, Vocabulary, and Comprehension) was at 90
or above (lower limit of average range to the rest of
the top 75% of the population); (b) accuracy or rate
of single-word reading, spelling, or text reading was
below the population mean and at least one stand-
ard deviation below their Verbal IQ, which has
been shown to be the best predictor of reading and
writing achievement; and (c) at least five family
members including both biological parents agreed
to participate.
The lower limit for Verbal IQ was set to limit, to
the extent possible, reading problems due to neuro-
genetic disorders other than dyslexia that are more
prevalent in children whose IQ falls below the 25th
percenttile. Although a minimum aptitude–
achievement discrepancy was set to avoid excluding
children who had had considerable instructional
treatment, but were not fully compensated because
of underlying dyslexia, the sample on average
showed much more discrepancy. Affected children
had reading and spelling achievement scores that
were on average 1.67 standard deviations below
their Verbal IQs and also had associated process-
ing deficits in phonological, orthographic, rapid
automatic naming, and executive functions
(Berninger et al., 2006b). Children were excluded if
they had a positive history or diagnosis for devel-
opmental, neurological, sensory impairment, or
other disorder that would explain difficulty in
reading. All participants in the family genetics
study gave informed consent using procedures
approved by the institutional review board for
human subjects.
The 122 affected children (80 boys, 42 girls) had
a mean of 6.0 (SD 2.8) deficits on the 10 reading
measures used to determine inclusion criteria
(accuracy and rate of single real word and pseu-
doword reading, including prepublication and
published versions of the rate measures; accuracy
and rate of oral reading of text; and two spelling
measures) and a mean of 4.1 (SD 1.6) deficits on
the 6 writing measures (3 measures of alphabet
writing, 2 spelling measures, and written expres-
sion). Thus, they tended to be severely impaired in
writing as well as reading skills.
The majority of the affected children were
European-American (88.5%) but 5.7% were minority
593
children (3.3% Asian-American; 1.6% African-
American; 0.8% Native American; and 3.3%
other); ethnicity was not reported for 2.5%. Their
mothers’ level of education ranged from high
school (5.7%) to community college/vocational
training (22.1%) to college (52.1%) to graduate
degree (19.8%); this information was missing for
0.8%. Their fathers’ level of education ranged from
high school (13.3%) to community college/voca-
tional training (24.2%) to college (36.7%) to gradu-
ate degree (25.8%); this information was missing
for 0.8%. Only four of the children with dyslexia
also had ADHD (2 inattentive only subtype and 2
mixed inattentive and hyperactive subtype).
First research question. Children with dyslexia
(80 boys and 42 girls) in a family genetics study of
dyslexia were compared to the fifth graders in the
longitudinal study of typically developing readers
and writers in Experiment 1. The same multiple
regression tests of the three models for contribu-
tions of executive function(s) to concurrent liter-
acy outcomes were used. The children with
dyslexia had a mean age of 138.3 months (11
years 6 months; SD = 20.6 months). The fifth
graders in the third year of the longitudinal study
had a mean age was 127. 56 months (10 years 7
months; SD = 3.75). Of all the groups in the longi-
tudinal study, the fifth graders had a mean age
that was closest to that of the dyslexics. Although
the range of ages was somewhat different for the
two groups (typically developing fifth graders
from 119 months to 140 months and children with
dyslexia 94 months to 183 months), it was the case
that the mode as well as the mean for the dyslexics
best matched that of the fifth graders in the longi-
tudinal study. Also, the children with and without
dyslexia were compared on standardized tests
with age or grade norms that control for age- or
grade-related differences.
Second research question. The mean level of per-
formance of boys and girls in each of the same two
groups (children with dyslexia and children with-
out dyslexia) were compared on the three measures
of executive functions—inhibition, rapid auto-
matic switching, and inhibition/switching.
Measures
Only measures that differed across studies for
measuring the same reading or writing skills are
included in this section. All other measures, includ-
ing the three measures of executive functions, are
described in Experiment 1 methods and were the
same across studies.
594 ALTEMEIER, ABBOTT, BERNINGER
Reading accuracy measures. The untimed Word
Identification and Word Attack subtests on the
Woodcock Reading Mastery Test (WRMT; Wood-
cock, 1987) were given, which require participants
to read single words and pseudowords, respec-
tively. The average internal consistency reliability
coefficients for both subtests are .97 and .87,
respectively.
Reading comprehension. The Woodcock–John-
son Revised (WJ-R; Woodcock & Johnson, 1990)
Passage Comprehension subtest, which employs a
cloze procedure in which the participant provides
the word that best completes a sentence, was given.
The average internal consistency reliability coeffi-
cient for this subtest is .90.
Rapid automatic switching (RAS). The Wolf Rapid
Automatic Switching (RAS)—Letters and Num-
bers subtest (Wolf, 1986), which had been used to
assess rapid automatic switching for the sample
with dyslexia instead of PAL RAS, requires the
individual to alternate between rapidly naming a
letter and a single-digit number presented in five
rows of 10 items each. Test–retest reliability over a
9-month intervention was .81 for this measure of
RAS.
Data analysis
Multiple regressions were performed to test the
same three models as those in Experiment 1 for the
children with dyslexia. The results were then com-
pared across Experiment 1 and Experiment 2 to
determine the differences in the relationship of
executive functioning to literacy outcomes between
children with and without dyslexia. In addition, a
two-way multivariate analysis of variance was per-
formed for gender on each executive function
measure to determine whether there were mean dif-
ferences in executive functioning for children with
and without dyslexia.
EXPERIMENT 1
Results
Developmental trajectories of executive
functions
For the first research question about typical
development of executive functions, HLM was
used to model the developmental trajectories of
three measures of executive functions: D-KEFS
Inhibition, PAL RAS, and D-KEFS Inhibition/
Switching. Separate trajectories were modeled for
the younger and older cohorts. Data from the first
cohort were analyzed for Grades 1–4. Data from
the second cohort were analyzed for Grades 3–6.
Note that PAL RAS data were only collected from
Grades 1–5. Table 1 shows the means and standard
deviations for each measured executive function at
each grade. The values are coded in number of sec-
onds to complete the task. A decrease in value over
the grades represents improvement in perform-
ance. As shown in Table 1, both cohorts improved
over time; that is, the number of seconds taken to
complete the task decreased for each of the three
executive functioning measures. In order to deter-
mine whether the two cohorts could be combined,
HLM was used to measure the difference between
the intercepts and slopes of each cohort. Because
data for both cohorts overlapped at Grade 3, the
Grade 3 value of time for both cohorts was cen-
tered at zero, representing Grade 3 for both
cohorts.
Inhibition. For D-KEFS Inhibition, intercepts
for each cohort at Grade 3 were not significantly
different (p=.750; see Table 2). The difference
between the two cohorts at the intercept was 0.93.
The slopes were not significantly different (p=.388).
The overall value for the slope was −11.9, reflecting
the decrease in time (i.e., improvement in perform-
ance) associated with each grade. Thus, the cohorts
were combined to form an overall trajectory of the
development of D-KEFS Inhibition from Grades
1–6. The trajectory of the development of inhibition
showed steady improvement in inhibition over the
grades, as measured by a decrease in time it took
the children to complete the task across the grades.
This result is consistent with the means and standard
deviations summarized in Table 1. The trajectory is
shown in Figure 1.
Inhibition/Switching. For D-KEFS Inhibition/
Switching, the intercepts for each cohort at Grade
3 were not significantly different. The difference
between the two cohorts at the intercept was −2.40
(p=.365). There was, however, a significant differ-
ence between the cohorts for slope (p ≤ .001). The
overall value for the slope was −10.39, reflecting
the decrease in time (i.e., increase in performance)
associated with each grade. So, with time centered
at Grade 3, for Cohort 1 (Grades 1–4) the develop-
mental trajectory was 92.95+(−10.39) (grade)
while for Cohort 2 (Grades 3–6) the developmental
trajectory was 92.95+(−4.94) (grade). Because the
values were coded in number of seconds to com-
plete the task, the developmental trajectory
reflected less improvement between Grades 3 and 6
(i.e., there was relatively less decrease in time taken
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA 595

TABLE 1
Mean times for Inhibition, Inhibition/Switching, and Rapid Automatic Switching

Cohort 1 Cohort 2

Grade Mean SD Mean SD

D-KEFS Inhibition 1 119.62 28.93

2 107.56 31.56
3 99.72 28.09 97.77 22.95
4 84.29 23.41 84.17 25.22
5 77.13 25.71
6 63.83 19.03
D-KEFS Inhibition/Switching 1 121.87 24.14
2 113.96 31.92
3 98.23 23.82 93.46 22.46
4 86.16 21.94 86.03 20.33
5 81.05 24.33
6 72.03 19.09
PAL Rapid Automatic Switching (RAS) 1 69.35 22.92
2 62.25 24.03
3 50.60 17.12 49.21 18.64
4 44.65 13.96
5 38.52 13.21

Note. Times are completion times in seconds. D-KEFS=Delis–Kaplan Executive Function System. PAL=Process Assessment of
the Learner. RAS=Rapid Automatic Switching.

TABLE 2
Development of D-KEFS Inhibition and D-KEFS, Grades 1–6, and PAL Rapid Automatic Switching, Grades 1–5

Fixed effect Level Coefficient SE T-ratio df p

D-KEFS Inhibition One Intercept for younger cohort 92.92 1.47 63.42 239 .001
Δ in intercept for older cohort 0.93 2.92 0.32 239 .750
Two Slope for younger cohort −11.90 0.56 −21.11 239 .001
Δ in slope for older cohort 0.96 1.11 0.87 239 .388
D-KEFS Inhibition/Switching One Intercept for younger cohort 92.95 1.27 73.10 239 .001
Δ in intercept for older cohort −2.40 2.54 −0.95 239 .345
Two Slope for younger cohort −10.39 0.56 −18.53 239 .001
Δ in slope for older cohort 5.46 1.10 4.98 239 .001
PAL Rapid Automatic Switching (RAS) One Intercept for younger cohort 51.07 1.15 44.41 239 .001
Δ in intercept for older cohort 0.13 2.31 0.06 239 .956
Two Slope for younger cohort −9.96 0.62 −16.14 239 .001
Δ in slope for older cohort 7.21 1.19 6.05 239 .001

Note. D-KEFS=Delis–Kaplan Executive Function System. PAL=Process Assessment of the Learner. RAS=Rapid Automatic
Switching.

to complete the task) than in the earlier grades.
This result may represent a change in the rate of
development of inhibition/switching between the
early elementary grades and the later elementary
grades; it does not mean that development came to
a standstill. The trajectory is shown in Figure 2.
Rapid automatic switching (RAS). For RAS
Words and Digits, the intercepts for each cohort
at Grade 3 were not significantly different. The
difference between the two cohorts at the intercept
was 0.13 (p=.956). There was, however, a signifi-
cant difference between the cohorts for slope (p ≤
.001). The overall value for the slope was 10.39,
reflecting the decrease in time (i.e., increase in per-
formance) associated with each grade. So, for
Cohort 1, the developmental trajectory was
51.07+(−9.96) (grade) while for Cohort 2 the
developmental trajectory was 51.07+(−2.75)
(grade). Thus, similar to the results from D-KEFS
Inhibition/Switching, the developmental trajec-
tory for PAL RAS in Grades 3–6 reflected less
improvement than that in the earlier grades (1–3).
This change in slope over time may represent a
change in PAL RAS development between the
early elementary grades and the later elementary
grades, with much improvement occurring in the
early primary grades but some still occurring in
596 ALTEMEIER, ABBOTT, BERNINGER

Older Cohort Older Cohort

Younger Cohort Younger Cohort
140
Mean Completion Time in Seconds

80

Mean Completion Time in Seconds

120
70
100 60
80 50

60 40

40 30

20
20
10
0
1 2 3 4 5 6 0
Grade 1 2 3 4 5
Grade
Figure 1. Trajectory of Delis–Kaplan Executive Function
System (D-KEFS) Inhibition in typically developing readers Figure 3. Trajectory of Process Assessment of the Learner,
and writers. Rapid Automatic Switching (PAL RAS) in typically developing
readers and writers.

Older Cohort of the correlations among all variables are availa-

Younger Cohort ble from the first author.
140 Next multiple regressions were conducted to test
Mean Completion Time in Seconds

three sequential, theory-driven models: entering

120
100
80
60
40
20
0 Moses, 2001; Roberts & Pennington, 1996). PAL
1 2
Figure 2. Trajectory of Delis–Kaplan Executive Function Sys-
tem (D-KEFS) Inhibition/Switching in typically developing
readers and writers.
the intermediate grades. The trajectory is shown in
Figure 3.
Concurrent relationship of executive
functions to specific reading or writing
skills
The three executive functions were entered as
predictor variables into hierarchical multiple
regressions for literacy outcomes at each grade.
Data from both cohorts were combined for Grade
3, while data from the younger cohort were ana-
lyzed for Grades 1 and 2, and data from the older
cohort were analyzed for Grades 4 and 5. Summaries
inhibition alone, entering inhibition followed by
rapid automatic switching, and entering inhibition,
followed by rapid automatic switching, and fol-
lowed by inhibiting/switching for each of the read-
ing or writing outcomes in Table 3 that summarizes
the findings. D-KEFS Inhibition was entered first
into the regressions based on theories emphasizing
the importance of inhibition in the development of
executive functioning (Barkley, 2003; Carlson &
3 4 5 6
RAS was entered next based on its significant corre-
Grade
lation with literacy outcomes, its use of alphanu-
meric stimuli encountered in reading and writing
materials, and research reviewed in the Introduction
that emphasized the importance of flexible set
switching in development of executive functions.
The D-KEFS condition that combined inhibition
and switching was entered last to evaluate whether a
direct measure combining both of the executive
functions would contribute uniquely beyond inhibi-
tion alone or rapid automatic switching alone. A
summary of the quantitative results for each of these
multiple regressions is also available from the first
author. The following sections are organized by spe-
cific reading and writing skills because the multiple
regression results depended on the literacy outcome.
Reading real words and decoding pseudowords.
When entered alone, D-KEFS Inhibition uniquely
predicted, at p ≤ .01, untimed and timed real-word
 TABLE 3
Comparing three models of executive functions in uniquely predicting concurrent literacy outcomes

WIAT2 TOWRE GORT3 WIAT2

Pseudoword Word reading Phonemic decoding Reading

Grade Model Word reading decoding efficiency efficiency Accuracy Rate comprehension Spelling Written expression

1 1 Inhib Inhib Inhib Inhib na na Inhib Inhib Inhib

2 RAS RAS RAS RAS na na RAS Inhib
3 RAS I/S RAS RAS RAS I/S na na I/S RAS
2 1 Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib
2 RAS RAS Inhib RAS RAS Inhib RAS Inhib RAS RAS RAS RAS
3 RAS RAS RAS RAS RAS RAS RAS RAS RAS
3 1 Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib
2 Inhib RAS Inhib RAS Inhib RAS Inhib RAS Inhib RAS Inhib RAS RAS Inhib RAS Inhib RAS
3 RAS RAS Inhib RAS Inhib RAS RAS I/S Inhib RAS I/S RAS RAS Inhib RAS I/S
4 1 Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib
2 RAS RAS Inhib RAS Inhib RAS Inhib RAS RAS Inhib (RAS) Inhib RAS Inhib RAS
3 RAS RAS RAS RAS Inhib RAS RAS RAS Inhib RAS
5 1 Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib Inhib
2 RAS RAS RAS RAS RAS RAS RAS RAS
3 RAS RAS RAS RAS RAS RAS RAS

Note. Model 1: Inhibition (Inhib) only. Model 2: Inhibition+RAS. Model 3: Inhibition+RAS+Inhibition/Switching[I/S]. RAS=Rapid Automatic Switching. Significant predictors in each model for
each literacy outcome listed in table. Scores that approached statistical significance have been included and are in parentheses.

597
598 ALTEMEIER, ABBOTT, BERNINGER
reading and pseudoword decoding at all grades.
However, when PAL RAS was entered second into
the model after D-KEFS Inhibition alone, RAS
was significant at p ≤ .001 for all word-level read-
ing tasks at Grades 1 and 5, except rate of reading
real words at Grade 2 level, and both rate measures
at Grade 4. In this second model combining Inhibi-
tion and RAS, both Inhibition and RAS contrib-
uted significantly to all word-level reading
measures at Grade 3 level, which had more power
because the two cohorts were combined. When D-
KEFS Inhibition/Switching was added to Inhibi-
tion and RAS in the third model, it was generally
not statistically significant at the p ≤ .05 level, with
the exception of WIAT-II Word Reading (p=.028)
and TOWRE Phonemic Decoding Efficiency in
Grade 1 (p=.046). Regarding the proportion of
variance explained by the models, R2 for WIAT-II
Word Reading ranged from a low of .18 in Grade 5
to a high of .31 in Grade 1. R2 followed a similar
trend on the WIAT-II Pseudoword Decoding task.
R2 values in models for timed tasks were higher,
with TOWRE Word Reading Efficiency ranging
from .47 (Grades 2, 4, and 5) to .55 (Grade 1), and
TOWRE Phonemic Decoding Efficiency ranging
from .35 (Grade 1) to .51 (Grade 4). Notably,
beyond Grade 1, D-KEFS Inhibition/Switching
generally did not add more than .01 to the R2 when
following PAL RAS in the models.
Reading text orally. D-KEFS Inhibition was sta-
tistically significant at p ≤ .001 when entered alone
in Model 1 for accuracy and rate of oral reading
outcomes at all grades for which it was available
(Grade 2 and above). When PAL RAS was added
to the models, it added unique variance for both
tasks at p ≤ .001 at all grades, but D-KEFS Inhibi-
tion also contributed uniquely to accuracy and rate
of oral reading in Grades 2, 3, and 4. When D-
KEFS Inhibition/Switching was added to the
model, it was significant for accuracy and rate of
oral reading only in Grade 3. Regarding the pro-
portion of variance explained by the models of
text-level oral reading outcomes, R2 for GORT
Accuracy ranged from .29 (Grade 1) to .43 (Grades
3 and 5). In general, D-KEFS Inhibition/Switching
added little value to the R2 in models of text-level
oral reading.
Reading comprehension. D-KEFS Inhibition was
statistically significant at p ≤ .009 when entered
alone to predict WIAT-II Reading Comprehension
at all grades. When PAL RAS was added to D-
KEFS Inhibition in Model 2, RAS contributed
uniquely in Grades 2, 3, and 5 and approached sig-
nificance in Grade 4 (p=.052). When D-KEFS
Inhibition/Switching was entered third in Model 3,
it contributed uniquely in Grade 1 only. The pro-
portion of variance explained by the models was
lower overall than it was for the reading tasks
measuring accuracy and efficiency. R2 ranged from
.10 (Grade 3) to .18 (Grade 4).
Spelling. D-KEFS Inhibition uniquely predicted
WIAT-II Spelling at all grades (p ranged from
≤.001 to .007) when entered alone in Model 1. It
continued to contribute unique variance in Grades
3 and 4 when PAL RAS was entered in Model 2,
although it no longer contributed a unique effect at
any grade when D-KEFS Inhibition/Switching was
entered in Model 3. PAL RAS was significant at
p ≤ .001 when entered second in Model 2 at all
grades. D-KEFS Inhibition/Switching did not con-
tribute unique variance at any grade level in Model
3. The proportion of variance explained by the
final models (R2) was variable and ranged from .16
(Grade 5) to .32 (Grade 4).
Written expression. When D-KEFS Inhibition
was entered alone, it uniquely predicted WIAT-II
Written Expression (p ranged from .001 to .004).
When PAL RAS was entered second in Model 3,
D-KEFS Inhibition continued to contribute a
unique effect in Grades 3 and 4. PAL RAS con-
tributed uniquely in Grades 2, 3, and 4. When D-
KEFS Inhibition/Switching was entered third in
Model 3, all three executive functions assessed
contributed uniquely only in Grade 3, which had
the most power of all the grade levels because
cohorts were combined at that grade level. How-
ever, Inhibition/Switching did not otherwise con-
tribute uniquely, and none of the executive
functions contributed uniquely at Grades 1 or 5.
The proportion of variance explained by the
models ranged from R2 = .11 (Grade 5) to .26
(Grade 3).
Table 3 summarizes the executive functions that
contributed unique variance in models of concur-
rent literacy outcomes for each grade. A psycho-
metric measure of Inhibition significantly
predicted every literacy outcome at each grade
level when entered alone in the regressions in
Model 1, but for many literacy outcomes in Model
2, when RAS was added to Inhibition, only RAS
contributed a unique effect after controlling for
shared variance with D-KEFS Inhibition. D-
KEFS Inhibition/Switching contributed a unique
effect to only a few literacy outcomes after control-
ling for shared variance with Inhibition and RAS.
Frequently, adding RAS to Inhibition in Model 2
or Inhibition/Switching to RAS and Inhibition in
Model 3 resulted in RAS contributing uniquely to
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA 599

reading and writing outcomes in typically developing to .11; D-KEFS Inhibition/Switching added to D-
readers and writers. KEFS and PAL RAS R2 ranged from .05 to .12).

Longitudinal relationships of executive EXPERIMENT 2

functions to literacy outcomes
To determine whether the development of
executive functions, as measured by slope, from
Grades 1 to 4 significantly predicted literacy out-
comes at Grade 4, the three executive functions
was modeled using HLM. Within HLM, a residual
file containing data representing the mean value
for the slope (fitted value) as well as each individ-
ual’s deviation from that value was created and
exported to SPSS. Each individual’s deviation
from the mean was calculated using the empirical
Bayes’ estimate for each individual’s change over
time (fitted value + deviation value). Regressions
were performed using that value as a predictor for
several academic outcomes.
Results for each of the three executive functions,
which were analyzed separately, are displayed in
Table 4. The slope of D-KEFS Inhibition pre-
dicted performance on all reading and writing out-
comes with the exception of WIAT-II Written
Expression, which was marginally significant
(p=.08). The slope of PAL RAS predicted each
reading and writing outcome (p ≤ .001). The slope
of D-KEFS Inhibition/Switching performance also
predicted each reading and writing outcome (p ≤
.05). The model that had PAL RAS and D-KEFS
Inhibition also had the highest R2 of the three exec-
utive functions (PAL RAS R2 ranged from .21 to
.43; D-KEFS Inhibition only R2 ranged from .03
Results
Comparing children with and without
dyslexia
As in Experiment 1 for children without dys-
lexia, measures of each executive function were
entered into hierarchical multiple regressions for
nine literacy outcomes for children with dyslexia.
Summaries of the correlations among all the
measures and the quantitative results of the multi-
ple regressions for the children with dyslexia are
available from the first author. Findings are sum-
marized in Table 5 to facilitate comparison across
groups.
Reading words and pseudowords. In contrast to
the typically developing Grade 5 readers and writ-
ers analyzed in Experiment 1, D-KEFS Inhibition,
when entered alone, did not contribute uniquely to
untimed measures of real-word or pseudoword
decoding in the children with dyslexia. This meas-
ure of Inhibition was, however, significant at p ≤
.001 for the timed real-word reading and pseudow-
ord decoding efficiency outcomes in children with
dyslexia. Wolf RAS was entered second into
Model 2, and it was significant at p ≤ .005 for all
word-level reading outcomes whether accuracy or
rate was involved. In none of the tasks did

TABLE 4
Predicting Grade 4 literacy outcomes from slope of executive functions in Grades 1–4

R2 B SE ß t p

Slope of D-KEFS Inhibition predictor TOWRE Phonemic Decoding Efficiency .11 2.96 0.80 .32 3.69 .001
TOWRE Word Reading Efficiency .06 2.13 0.82 .23 2.60 .011
WIAT-II Pseudoword Decoding .05 1.48 0.60 .23 2.49 .014
WIAT-II Word Reading .07 1.90 0.63 .27 3.01 .003
WIAT-II Spelling .08 2.71 0.83 .29 3.26 .001
WIAT-II Written Expression .03 1.62 0.93 .16 1.75 .084
Slope of PAL RAS predictor TOWRE Phonemic Decoding Efficiency .43 1.29 0.14 .66 9.32 .001
TOWRE Word Reading Efficiency .27 1.01 0.16 .52 6.50 .001
WIAT-II Pseudoword Decoding .26 0.71 0.11 .51 6.36 .001
WIAT-II Word Reading .26 0.77 0.12 .51 6.40 .001
WIAT-II Spelling .27 1.04 0.16 .52 6.62 .001
WIAT-II Written Expression .21 0.97 0.18 .46 5.52 .001
Slope of D-KEFS Inhibition/Switching predictor TOWRE Phonemic Decoding Efficiency .12 1.57 0.39 .35 4.01 .001
TOWRE Word Reading Efficiency .05 0.97 0.41 .22 2.39 .018
WIAT-II Pseudoword Decoding .05 0.75 0.29 .23 2.56 .012
WIAT-II Word Reading .09 1.01 0.31 .29 3.29 .001
WIAT-II Spelling .05 1.02 0.42 .22 2.46 .015
WIAT-II Written Expression .06 1.16 0.45 .24 2.60 .011

Note. TOWRE=Test of Word Reading Efficiency. WIAT-II=Wechsler Individual Achievement Test–Second Edition.
600 ALTEMEIER, ABBOTT, BERNINGER

D-KEFS Inhibition/Switching contribute a unique

Note. Scores that approached statistical significance have been included and are in parentheses. TOWRE=Test of Word Reading Efficiency. WIAT2=Wechsler Individual Achievement Test–Second
WIAT2 Written
effect when entered third into Model 3. For chil-

Expression
dren with dyslexia, the proportion of variance
explained by the final models was higher in general

(Inhib)
Inhib
Inhib

RAS
RAS
for timed tasks than untimed word-level reading
tasks, as it had also been for children without dys-
Spelling
WIAT2 lexia. TOWRE Word Reading Efficiency and Pho-
Summary of executive functions that significantly predicted literacy outcomes for children with dyslexia and typically developing readers and writers

Inhib
Inhib
nemic Decoding Efficiency R2 values were .36 and

RAS
RAS

RAS
RAS
.20, respectively, while WRMT-R Word ID and
Word Attack R2 values were .16 and .08, respec-
Comprehension

(Inhib) RAS
(Inhib) RAS
tively. However, R2 values for all word-level read-
Reading
WIAT2

ing outcomes were less for the children with than

Inhib
RAS
(I/S) for those without dyslexia. These differences were
found during word reading and pseudoword
decoding tasks that were both untimed (R2 for chil-
GORT Rate

dren with dyslexia=.08; for typical children=.20)

and timed (R2 for children with dyslexia=.20; for
Inhib
Inhib

RAS
RAS

RAS
RAS

typical children=.44).
Accuracy
GORT

Reading text orally. Similar to the results for the

Inhib
Inhib

RAS
RAS

RAS
RAS

typical readers and writers, D-KEFS Inhibition

was statistically significant (p ≤ .007) for GORT
(Inhib) RAS
(Inhib) RAS

(Inhib) RAS

Accuracy and Rate when entered alone in the

Efficiency
Phonemic
Decoding
TOWRE

regression models. However, when Wolf RAS was

Inhib
Inhib

RAS

added, D-KEFS Inhibition did not contribute a

unique effect. Wolf RAS was significant (p ≤ .001)
for both tasks in the children with dyslexia, while
TOWRE Word
TABLE 5

Efficiency

D-KEFS Inhibition/Switching did not add unique

Reading

variance. Similar to the word-level reading tasks,

R2 values for accuracy and rate of oral reading of
Inhib

Inhib

Edition. GORT=Gray Oral Reading Test, Third Edition. RAS=Rapid Automatic Switching.
RAS
RAS

RAS
RAS

passages were less in children with than in those

without dyslexia. For GORT III Accuracy, R2 =
WIAT2 Pseudo-
word Decoding

.17 in children with dyslexia and .43 in children

without dyslexia. For GORT III Rate, R2 = .29 in
children with dyslexia and .46 in children without
Inhib
RAS
RAS

RAS
RAS

dyslexia.
Word Reading

Reading comprehension. Results for children with

(Inhib) RAS
WIAT2

dyslexia differed from those for children without

dyslexia: D-KEFS Inhibition was not significant
Inhib
RAS

RAS
RAS

when entered alone into the regression equation

(p=.941). However, similar to the findings for
Model

children without dyslexia, Wolf RAS added a

significant unique effect (p ≤ .001), but D-KEFS
1

1
2
3

2
3

Inhibition/Switching did not add a unique effect.

Typically developing readers and writers

R2 values for reading comprehension were compa-

rably low in the children with and without dyslexia
(R2 =.18 and .13, respectively).

Spelling. Similar to results for children without

Children with dyslexia

dyslexia, D-KEFS Inhibition was significant when

entered alone into the regression equation
(p=.014), but did not contribute a unique effect
when Wolf RAS was entered second. Wolf RAS
added a unique effect (p ≤ .003) but D-KEFS Inhi-
bition/Switching did not. R2 values for spelling
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA 601

were comparably low in the children with and
without dyslexia (R2 =.13 and .16, respectively).
Written expression. For the children with dys-
lexia, when D-KEFS Inhibition was entered alone,
it contributed uniquely to WIAT-II Written
Expression (p=.032). When Wolf RAS was entered
second, D-KEFS Inhibition did not contribute a
unique effect while Wolf RAS contributed
uniquely (p ≤ .001). D-KEFS Inhibition/Switching
was entered third and did not add unique variance
(p=.379). The R2 for the children with dyslexia was
higher than for the children without dyslexia,
although values for both were relatively low
(R2 =.18 and .11, respectively).
Gender differences in executive
functioning
Means and standard deviations for each execu-
tive function are listed separately by gender in
Table 6. A main effect for gender was found for the
children with dyslexia on two of the three measures
of executive functions: D-KEFS Inhibition F(1,
118)=7.16, p ≤ .05, with girls scoring higher
(M=8.73, SD=2.86) than boys (M=7.18,
SD=3.06); and D-KEFS Inhibition/Switching F(1,
118)=9.50, p ≤ .05, again with girls scoring higher
(M=9.61, SD=2.52) than boys (M=7.88,
SD=2.97). Gender differences were not found for
the children with dyslexia on Wolf RAS or PAL
RAS. Gender differences were not found at any
grade in the typically developing sample.
DISCUSSION
Development of executive functions
Results from the hierarchical linear modeling anal-
yses showed that children continue to develop their
executive functions through middle childhood,
thus confirming the first hypothesis. However, not
all executive functions showed the same pattern of
longitudinal growth. Development of inhibition
increased steadily from Grade 1 through Grade 6,
with no change in the direction of the slope over
time. In contrast, development of rapid automatic
switching and combined inhibition and switching
slowed down about the middle of elementary
school, although still showed some change.
Together, these findings suggest that inhibition,
which is considered to be fundamental to other
executive functions (see Introduction), may have a
longer developmental trajectory than the other
executive functions studied—rapid automatic
switching alone or combined with inhibition.
Considering that development often shows discon-
tinuities, although development of rapid automatic
switching and inhibition/switching appeared to

TABLE 6
Means for executive functions by gender

Males Females

Mean SD Mean SD

Children with dyslexia D-KEFS Inhibition 7.18 3.06 8.73 2.86

Wolf RAS 2.98 2.34 2.66 2.55
D-KEFS Inhibition/Switching 7.88 2.97 9.61 2.52
Grade 1 children without dyslexia D-KEFS Inhibition 0.32 0.85 0.44 1.14
PAL RAS −0.49 0.65 −0.52 0.56
D-KEFS Inhibition/Switching 0.40 0.83 0.63 0.92
Grade 2 children without dyslexia D-KEFS Inhibition 9.78 4.26 9.60 3.61
PAL RAS 0.35 1.51 0.18 1.19
D-KEFS Inhibition/Switching 9.96 3.92 9.12 3.47
Grade 3 children without dyslexia D-KEFS Inhibition 10.44 3.02 10.47 3.12
PAL RAS 0.22 1.39 0.05 1.23
D-KEFS Inhibition/Switching 10.81 2.79 11.41 2.76
Grade 4 children without dyslexia D-KEFS Inhibition 11.04 3.10 11.89 3.00
PAL RAS 0.01 1.23 0.07 1.14
D-KEFS Inhibition/Switching 10.92 2.96 11.57 2.42
Grade 5 children without dyslexia D-KEFS Inhibition 11.00 2.74 11.13 3.01
PAL RAS −0.19 0.88 −0.05 1.06
D-KEFS Inhibition/Switching 10.77 2.89 10.70 3.21

Note. D-KEFS=Delis–Kaplan Executive Function System. PAL=Process Assessment of the Learner. RAS=Rapid Automatic
Switching. RAS scores are reported in z-scores, with a lower number indicating better performance. first-grade D-KEFS Inhibition
and D-KEFS Inhibition/Switching scores are reported in scaled scores (M=10, SD=3), with a higher number indicating better
performance.
602 ALTEMEIER, ABBOTT, BERNINGER
slow down from Grades 3 to 6, either or both may
show a growth spurt later in development. Longi-
tudinal studies of middle-school and high-school
students could determine whether inhibition alone
has a protracted course of development. Brocki
and Bohlin’s (2004) study of the development of
executive functioning using a similar Stroop-like
inhibition task showed that executive functions
may have a protracted course of development into
adolescence.
Concurrent contributions of executive
functions to reading and writing
The hypothesis that executive functions would
contribute differentially to literacy learning for
typically developing readers and writers was
confirmed to some degree. However, across
tasks and grades, for the most part Inhibition
alone contributed uniquely, but adding RAS to
Inhibition showed that RAS contributed
uniquely over and beyond its shared variance
with Inhibition. This finding should not be inter-
preted that RAS alone is the unique predictor, as
this finding held when Inhibition was entered
first and RAS second into the model. Thus, the
results provided concurrent and construct valid-
ity for the theory that both Inhibition and RAS
contribute to the reading and writing achieve-
ment of typically developing readers and writers.
However, one measure that combined Inhibition
and RAS did not uniquely explain the concur-
rent reading and writing achievement better than
separate measures of Inhibition and RAS.
Including RAS in Model 2 or Model 3 also
appeared in general to result in accounting for a
larger proportion of the variance in the reading
and writing outcomes.
Results also showed that the proportion of vari-
ance explained by executive functions in the read-
ing and writing outcome models varies by grade
and by task. For reading measures without a timed
component, executive functions predicted literacy
outcomes the most during the early elementary
grades, as evidenced by a general decline in R2 val-
ues over the grades. One possible explanation is
that as early readers are learning to decode, execu-
tive functions guide the process of learning to
relate spoken and written words. At this time, inhi-
bition and rapid automatic switching are needed to
suppress irrelevant codes during phonological
retrieval of sounds for letters or names for the
whole written word and to switch among the
constantly changing letters and written words,
respectively. Conversely, the amount of variance
accounted for increased and held steady during the
later grades for reading measures that are timed.
Once children acquire the procedures for accu-
rately decoding and recognizing familiar words,
executive functioning may play a role in the effi-
ciency with which these processes occur in time. If
executive functions are necessary to integrate and
precisely time multiple, sequential processes, then
individual differences in executive functions may
influence how well reading fluency develops in the
upper elementary grades. However, it was not the
case that executive function measures, all of which
were timed, contributed uniquely only to timed
measures of reading and writing. As shown in
Table 3, executive functions contributed to both
untimed (WIAT II Word Reading and Pseudow-
ord Reading, Reading Comprehension, and Writ-
ten Expression) and timed (TOWRE and GORT 3
Rate) measures of reading or writing. Moreover, at
all grade levels measures of executive functioning
contributed uniquely to both untimed and timed
real-word and pseudoword reading in the typically
developing readers and writers.
An unexpected finding was that the predictive
ability of the executive-functioning measures used
was relatively weak for reading comprehension. It
may be that the kinds of executive functions
assessed in this study contribute more to lower
level word decoding and word reading than to
reading comprehension. For example, making
inferences, predicting, extracting main ideas, and
reflecting during reading comprehension may
require planning, perspective taking, and strategy
implementation, which are not the kinds of execu-
tive functions assessed in this study. In the decod-
ing process, inhibition and set shifting occur
moment to moment in linking letters and sounds.
Planning and perspective taking involve projecting
one’s thinking across longer stretches of time and
space. Likewise, the kinds of executive functions
assessed in this study may contribute more to
lower level spelling than to higher level written
composition. Altemeier et al. (2006) found that a
higher order executive function—planning as
assessed by a Tower Task—contributed to exposi-
tory written composition in third and fifth graders.
Thus, executive functions such as inhibition and
set shifting may be qualitatively distinct from some
other types of executive functions required for
higher level representation, planning, and problem
solving. Put another way, just as there are lower
level and higher level literacy skills, there may also
be lower level and higher level executive functions.
For writing, results for models of spelling and
written expression show that executive functions
do explain some of the variance in writing skills.
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA
Nevertheless, results were not easily interpreted
because there was not a linear progression of vari-
ance explained through the grades for the writing
outcomes. For example, the predictive ability of
executive functioning in written expression
increased during Grades 3 and 4 and then declined
in Grade 5. Again, the kinds of executive functions
that are important in writing may not have been
assessed. Conclusions about the relationships
between executive functions and written expression
across development and nature of writing require-
ments require further research.
Longitudinal relationships between executive
functions and reading and writing
The development of each of the three executive
functions from Grade 1 through Grade 4 predicted
literacy outcomes by Grade 4, thus confirming the
initial hypothesis that the executive functions
directly influence literacy learning over the first
four grades. The growth trajectories showed that
executive functions develop over time; the multiple
regressions showed that individual differences in
executive functions are related to individual differ-
ences in reading and writing achievement; and the
results showing that longitudinal change in each of
the executive functions is related to all literacy out-
comes in Grade 4, except for inhibition and written
composition, support a contribution of executive
functions to literacy learning. Overall longitudinal
change in RAS accounted for the most variance in
Grade 4 literacy outcomes.
Comparison of children with and without
dyslexia
The hypothesis that executive functions were more
related to literacy achievement in children without
than in those with dyslexia was partly confirmed.
Executive functions predicted literacy outcomes
for children with and without dyslexia. For both
children with and those without dyslexia, executive
functions contributed substantially to timed tasks.
This latter finding may reflect the role of executive
functions in coordinating the precise timing for
coordination of orthographic and phonological
codes (Wolf & Bowers, 1999). Also, RAS, entered
after Inhibition in Model 2 or before Inhibition/
Switching in Model 3, was the most consistently
unique predictor of various literacy outcomes for
both children with and those without dyslexia. The
major difference between the two groups was that
the multiple regression models tended to explain
relatively less variance in literacy outcomes in children
603
with than in those without dyslexia. One possible
explanation for this finding is that individual dif-
ferences in executive function contribute more to
normal reading and writing acquisition. Another
possible explanation for this finding is that chil-
dren with dyslexia may not engage or apply their
executive functions to reading and writing as those
without dyslexia are able to do. Effective treatment
of dyslexia may, therefore, require more explicit
instruction on self-regulation strategies for engag-
ing and applying executive functions during read-
ing and writing. Several studies of reading
comprehension and written expression show that
metacognitive and self-regulation strategies influ-
ence outcomes and, moreover, can be taught
(Gersten & Baker, 2001; Graham, 1997; Pressley,
2000; Swanson & Hoskyn, 1998; Wong, Ficzere, &
Kuperis, 1997). Yet another possible explanation is
that the executive functions of children with dys-
lexia fall outside the normal range (Berninger
et al., 2006b) and thus display more restricted
variance that does not explain the individual differ-
ences in attaining different levels of literacy
achievement. In other words, children with dys-
lexia may indeed have significant impairment in
executive functions even if such impairment does
not contribute any or as much unique variance to
their literacy outcomes.
Gender differences
The hypothesis that boys with dyslexia would per-
form more poorly on executive functions than girls
with dyslexia was confirmed for executive func-
tions involving inhibition—alone or combined
with RAS—but not for RAS alone. This finding
suggests that the executive dysfunction related to
writing (Berninger et al., in press-a) and the gen-
der-related writing problems of dyslexics (Berninger
et al., in press-b) may be related to impaired inhibi-
tion; further research is needed on this issue. No
gender differences were found on any of the three
executive functions studied in the children without
dyslexia, consistent with the previous findings of
Welsh et al. (1991). It follows that impaired execu-
tive functions in boys are most likely to explain any
observed reading and writing problems if the boys
also meet diagnostic criteria for a developmental
reading and/or writing disability.
Clinical significance
The results have significance for both diagnosis
and treatment. Executive functions should be
assessed for any student who struggles with learning
604 ALTEMEIER, ABBOTT, BERNINGER
to read and write. On the one hand, level of
executive-function ability may have implications
for the amount of teacher-directed, explicit instruc-
tion that a student requires to learn self-regulation
during reading and writing (Graham & Harris,
2005). On the other hand, level of executive-func-
tion ability may have implications for prognosis
for response to instruction and improvement in
reading and writing. However, treatment studies
that examine response to instruction are needed to
test this hypothesis further. Researchers might
investigate not only whether executive functions
predict response to reading instruction but also
whether specialized reading instruction improves
executive functions. Both children with and those
without dyslexia may benefit from incorporating
strategies to enhance executive functioning, specifi-
cally inhibition (focus on relevant and suppression
of attention to the irrelevant) and mental set shift-
ing within a comprehensive reading and writing
curriculum.
Limitations and future research
The fact that the single measures of inhibition and
rapid automatic switching and set shifting and the
measure that combined inhibition and switching
did not contribute significantly, or contributed less
to text-level reading comprehension and written
expression than to word-level reading and spelling,
suggests that other executive functions not evalu-
ated in this study may contribute to higher order,
text-level reading and writing skills, as contrasted
to word-level reading and spelling skills. Indeed,
there may be lower level and higher level executive
functions that differentially contribute to literacy
learning. Inhibition and set shifting, which are
lower level executive functions, may support word-
level processing, which in turn supports discourse-
level processing of larger stretches of text; dis-
course-level processing requires higher level execu-
tive functions that support thinking across longer
stretches of time and space as needed in reading
comprehension (e.g., extracting the main idea,
making inferences, generating predictions) and
written composition (e.g., planning, organization,
self-monitoring, revising). Such higher order exec-
utive functions may enable information processing
that is past, present, and future oriented in work-
ing memory.
In addition, inhibition and set shifting as assessed
in the current study involved relatively little or no
semantic information. In contrast, higher level
executive functions engaged in text-level reading
comprehension and written expression probably
involve considerable semantic information. Much
remains to be learned through research about how
different kinds and levels of executive functions
help readers and writers to link language and cogni-
tion through semantics to construct meaning.
Original manuscript received 10 April 2007
Revised manuscript accepted 5 July 2007
First published online 9 November 2007
REFERENCES
Altemeier, L., Jones, J., Abbott, R. D., & Berninger, V.
(2006). Executive functions in becoming writing
readers and reading writers: Note-taking and report
writing in third and fifth graders. Developmental Neu-
ropsychology, 29, 161–173.
Barkley, R. (1997). Behavioural inhibition, sustained
attention, and executive functions. Psychological Bul-
letin, 121, 65–94.
Barkley, R. (2003). Attention-deficit/hyperactivity disorder.
In E. J. Mash & R. Barkley (Eds.), Child psychopathol-
ogy (2nd ed., pp. 75–143). New York: Guilford Press.
Berlin, L., & Bohlin, G. (2002). Response inhibition,
hyperactivity, and conduct problems among pre-
school children. Journal of Clinical Child and Adoles-
cent Psychology, 31, 242–251.
Berninger, V. (2001). Process assessment of the learner
test battery for reading and writing. San Antonio, TX:
Psychological Corporation.
Berninger, V., Abbott, R., Jones, J., Wolf, B., Gould, L.,
Anderson-Youngstrom, M., et al. (2006a). Early
development of language by hand: Composing-, read-
ing-, listening-, and speaking- connections, three let-
ter writing modes, and fast mapping in spelling.
Developmental Neuropsychology, 29, 61–92.
Berninger, V., Abbott, R. D., Thomson, J., Wagner, R.,
Swanson, H. L., Wijsman, E., et al. (2006b). Mode-
ling phonological core deficits within a working
memory architecture in children and adults with
developmental dyslexia. Scientific Studies of Reading,
10, 165–198.
Berninger, V., & Amtmann, D. (2003). Preventing writ-
ten expression disabilities through early and continu-
ing assessment and intervention for handwriting and/
or spelling problems: Research into practice. In H. L.
Swanson, K. R. Harris, & S. Graham (Eds.), Hand-
book of learning disabilities (pp. 345–363). New York:
Guilford Press.
Berninger, V., & Fuller, F. (1992). Gender differences in
orthographic, verbal, and compositional fluency:
Implications for diagnosis of writing disabilities in
primary grade children. Journal of School Psychol-
ogy, 30, 363–382.
Berninger, V., Nielsen, K., Abbott, R., Wijsman, E., &
Raskind, W. (in press-a). Gender differences in sever-
ity of writing and reading disabilities. Journal of
School Psychology.
Berninger, V., Nielsen, K., Abbott, R., Wijsman, E., &
Raskind, W. (in press-b). Writing problems in devel-
opmental dyslexia: Under-recognized and under-
treated. Journal of School Psychology.
Blair, C., Zelazo, P. D., & Greenberg, M. T. (2005). The
measurement of executive function in early child-
hood. Developmental Neuropsychology, 28, 561–571.
 EXECUTIVE FUNCTIONS IN LITERACY AND DYSLEXIA
Brocki, K. C., & Bohlin, G. (2004). Executive functions
in children aged 6 to 13: A dimensional and develop-
mental study. Developmental Neuropsychology, 26,
571–593.
Brosnan, M., Demetre, J., Hamill, S., Robson, K., Shep-
herd, H., & Cody, G. (2002). Executive functioning in
adults and children with developmental dyslexia.
Neuropsychologia, 40, 2144–2155.
Carlson, S., & Moses, L. (2001). Individual differences in
inhibitory control and children’s theory of mind.
Child Development, 72, 1032–1078.
Delis, D. C., Kaplan, E., & Kramer, J. H. (2001). The
Delis–Kaplan Executive Function System: Examiner’s
manual. San Antonio, TX: The Psychological Corpo-
ration.
Denckla, M. B. (1996). A theory and model of executive
function. In G. R. Lyon & N. A. Krasnegor (Eds.),
Attention, memory, and executive function (pp. 263–
278). Baltimore, MD: Brookes.
Denckla, M. B., & Cutting, L. E. (1999). History and sig-
nificance of rapid automatized naming. Annals of
Dyslexia, 49, 29–43.
Denckla, M. B., & Rudel, R. G. (1974). Rapid “autom-
atized” naming of pictured objects, colors, letters,
and numbers by normal children. Cortex, 10, 186–
202.
de Ribaupierre, A. (2002). Working memory and atten-
tional processes across the lifespan. [The 2nd Tsu-
kuba International Conference on Memory.] In P.
Graf & N. Ohta (Eds.), Lifespan development of
human memory (pp. 59–80). Cambridge, MA: The
MIT Press.
Diamond, A. (2002). Normal development of the pre-
frontal cortex from birth to young adulthood: Cogni-
tive functions, anatomy, and biochemistry. In D.
Stuss & R. Knight (Eds.), Principles of frontal lobe
function (pp. 466–503). New York: Oxford University
Press.
Diamond, A., Carlson, S., & Beck, D. (2005). Preschool
children’s performance in task switching on the
dimensional change card sort task: Separating the
dimensions aids the ability to switch. Developmental
Neuropsychology, 28, 689–729.
Gersten, R., & Baker, S. (2001). Teaching expressive
writing to students with learning disabilities: A meta-
analysis. Elementary School Journal, 101, 251–272.
Graham, S. (1997). Executive control in the revising of
students with learning and writing difficulties. Jour-
nal of Educational Psychology, 89, 223–234.
Graham, S., & Harris, K. (2005). Writing better. Effec-
tive strategies for teaching students with learning diffi-
culties. Baltimore, MD: Paul H. Brookes.
Hayes, J. R., & Flower, L. S. (1980). Identifying the
organization of writing processes. In L. W. Gregg &
E. R. Steinbert (Eds.), Cognitive processes in writing
(pp. 3–30). Hillsdale, NJ: Lawrence Erlbaum
Associates.
Helland, T., & Asbjornsen, A. (2000). Executive func-
tions in dyslexia. Child Neuropsychology, 6, 37–48.
Hooper, S. R., Swartz, C. W., Wakely, M. B., de Kruif,
R., & Montgomery, J. W. (2002). Executive functions
in elementary school children with and without prob-
lems in written expression. Journal of Learning Disa-
bilities, 35, 57–68.
Hughes, C. (1998). Executive function in preschoolers:
Links with theory of mind and verbal ability. Devel-
opmental Psychology, 34, 233–253.
605
Kelly, M. S., Best, C. T., & Kirk, U. (1989). Cognitive
processing deficits in reading disabilities: A prefrontal
cortical hypothesis. Brain and Cognition, 11, 275–293.
Lezak, M. (1995). Neuropsychological assessment (3rd
ed.). New York: Oxford University Press.
Lyon, G. R., & Krasnegor, N. A. (Eds.). (1996). Atten-
tion, memory, and executive function. Baltimore, MD:
Brookes.
Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A.
H., Howerter, A., & Wager, T. D. (2000). The unity
and diversity of executive functions and their contribu-
tions to complex “frontal lobe” tasks: A latent variable
analysis. Cognitive Psychology, 41, 49–100.
Morris, R., Stuebing, K., Fletcher, J., Shaywitz, S.,
Lyon, G. R., Shankweiler, D. P., et al. (1998). Sub-
types of reading disability: Variability around a pho-
nological core. Journal of Educational Psychology, 90,
347–373.
Pennington, B. F. (1997a). Dimensions of executive
functions and developmental psychopathology. In N.
A. Krasnegor, G. R. Lyon, & Goldman-Rakic, P. S.
(Ed.), Development of the prefrontal cortex: Evolution,
neurobiology, and behavior (pp. 265–281). Baltimore,
MD: Paul H. Brookes.
Pennington, B. F. (1997b). Executive functions and
developmental psychopathology. Journal of Child
Psychology and Psychiatry, 37, 51–87.
Pennington, B. F., Bennetto, L., McAleer, O., & Rob-
erts, R. J. (1996). Executive functions and working
memory: Theoretical and measurement issues. In G.
R. Lyon & N. A. Krasnegor (Eds.), Attention, mem-
ory, and executive function (pp. 327–348). Baltimore,
MD: Brookes.
Pressley, M. (2000). What should comprehension
instruction be the instruction of? In M. L. Kamil, P.
B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Hand-
book of reading research (Vol. 3, pp. 545–562).
Mahwah, NJ: Lawrence Erlbaum Associates.
Psychological Corporation. (2001). Wechsler Individual
Achievement Test, 2nd edition. WIAT II. San Anto-
nio, TX: The Psychological Corporation.
Purvis, K. L., & Tannock, R. (2000). Phonological
processing, not inhibitory control, differentiates
ADHD and reading disability. Journal of the Ameri-
can Academy of Child and Adolescent Psychiatry, 39,
485–494.
Reiter, A., Tucha, O., & Lange, K. W. (2005). Executive
functions in children with dyslexia. Dyslexia, 11, 116–131.
Roberts, R. J., & Pennington, B. F. (1996). An interactive
framework for examining prefrontal cognitive proc-
esses. Developmental Neuropsychology, 12, 106–126.
Singer, B. D., & Bashir, A. S. (1999). What are execu-
tive functions and self-regulation and what do they
have to do with language-learning disorders? Lan-
guage, Speech, and Hearing Services in Schools, 30,
265–273.
Stroop, J. R. (1935). Studies of interference in serial
verbal reaction. Journal of Experimental Psychology,
18, 643–662.
Swanson, H. L. (2000). Working memory, short-term
memory, speech rate, word recognition and reading
comprehension in learning disabled readers: Does the
executive system have a role? Intelligence, 28, 1–30.
Swanson, H. L., & Hoskyn, M. (1998). Experimental
intervention research on students with learning disa-
bilities: A meta-analysis of treatment outcomes.
Review of Educational Research, 68, 277–321.
606 ALTEMEIER, ABBOTT, BERNINGER
Torgesen, J., Wagner, R., & Rashotte, C. (1999). Test of
Word Reading Efficiency (TOWRE). Austin, TX:
Pro-Ed.
Wagner, R. K., & Torgesen, J. K. (1987). The nature of
phonological processing and its causal role in the
acquisition of reading skills. Psychological Bulletin,
101, 192–212.
Welsh, M. C., Pennington, B. F., & Groisser, D. B.
(1991). A normative-developmental study of execu-
tive function: A window on prefrontal function in
children. Developmental Neuropsychology, 7, 131–
149.
Wiederholt, J., & Bryant, B. (1992). Gray Oral Reading
Test—Third Edition. Odessa, FL: Psychological
Assessment Resources.
Wolf, M. (1986). Rapid alternating stimulus naming in
the developmental dyslexias. Brain and Language, 27,
360–379.
Wolf, M., Bally, H., & Morris, R. (1986). Automaticity,
retrieval processes, and reading: A longitudinal study
in average and impaired readers. Child Development,
57, 988–1005.
Wolf, M., & Bowers, P. (1999). The double-deficit
hypothesis for the developmental dyslexias. Journal
of Educational Psychology, 91, 415–438.
Wolff, P. H., Hurvitz, I., Imamura, S., & Lee, K. W.
(1983). Sex differences and ethnic variation in speed of
automatized naming. Neuropsychologia, 21, 283–288.
Wong, B., Ficzere, S., & Kuperis, S. (1997). Teaching
adolescents with learning disabilities and low achiev-
ers to plan, write, and revise compare-contrast essays.
Learning Disabilities Research & Practice, 12, 2–15.
Woodcock, R. (1987). Woodcock Reading Mastery Tests—
Revised. Circle Pines, MN: American Guidance Service.
Woodcock, R., & Johnson, B. (1990). Woodcock–John-
son Psycho-Educational Battery Revised Tests of
Achievement. Chicago: Riverside Publishing.
Zelazo, P. D., Carter, A., Reznick, J., & Frye, D. (1997).
Early development of executive function: A problem-
solving framework. Review of General Psychology, 1,
198–226.
Zimmerman, R. D., & Risemberg, R. (1997). Becoming
a self-regulated writer: A social cognitive perspective.
Contemporary Educational Psychology, 22, 73–101.