Frijters Et Al (2011) Neurocognitive Predictors of Dyslexia

391185
LDX
Neurocognitive Predictors of
Reading Outcomes for Children
With Reading Disabilities
Journal of Learning Disabilities

44(2) 150166
Hammill Institute on Disabilities 2011
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0022219410391185
http://journaloflearningdisabilities
.sagepub.com
Jan C. Frijters1, Maureen W. Lovett2, Karen A. Steinbach2,

Maryanne Wolf3, Rose A. Sevcik4, and Robin D. Morris4
Abstract
This study reports on several specific neurocognitive process predictors of reading outcomes for a sample of 278 children
with reading disabilities. Three categories of response (i.e., poor, average, and good) were formed via growth curve
models of six reading outcomes. Two nested discriminant function analyses were conducted to evaluate the predictive
capability of the following models: (a) an intervention and phonological processing model that included intervention group,
phonological awareness, and rapid naming and (b) an additive cognitive neuropsychological model that included measures
of memory, visual processes, and cognitive or intellectual functioning. Over and above the substantial explanatory power
of the base model, the additive model improved classification of poor and good responders. Several of the cognitive
and neuropsychological variables predicted degree of reading outcomes, even after controlling for type of intervention,
phonological awareness, and rapid naming.
Keywords
treatment, dyslexia, response to intervention, reading, neuropsychological assessment
Two specific language processes have been consistently and
convincingly demonstrated to be a keyand some would
say a causalskill set underlying the successful development of reading skill. These processes are phonological
awareness (PA), or the manipulation of individual sounds in
the speech stream (Liberman & Shankweiler, 1985; Wagner
& Torgesen, 1987), and rapid automatized naming (RAN),
or the rapid naming of serial displays of visual symbols
such as letters, numbers, objects, and colors (Wolf & Bowers,
1999; Wolf, Bowers, & Biddle, 2000). A comprehensive
meta-analysis of correlational research linking these processes and reading achievement is provided by Swanson,
Trainin, Necoechea, and Hammill (2003).
Deficiencies in PA and naming speed have been demonstrated to be characteristic of individuals with reading disabilities (RD) and those who struggle to acquire basic
reading skills. Past research has shown that PA and RAN
are distinct constructs but related in their prediction of reading processes. Providing independent predictive power to
explaining reading skills, deficits in both PA and RAN prior
to being able to read have been related to Grade 2 and Grade
3 reading ability (Manis, Seidenberg, & Doi, 1999; Wimmer,
Mayringer, & Landerl, 1998). The two skills may also have
a different developmental course, with PA exerting more
influence on early, decoding-dependent tasks and RAN
exerting more influence on later, word identification and
fluency-dependent tasks (Kirby, Parrila, & Pfeiffer, 2003;

Wolf et al., 2002).
Several authors have suggested recently, in reviews and
in meta-analyses, that a great deal is known about what constitutes efficacious intervention for reading problems (for
comprehensive reviews, see Block, Parris, Reed, Whiteley,
& Cleveland, 2009; Lovett, Barron, & Benson, 2003; Lovett
et al., 2005; Torgesen, 2005). Despite what is understood
about the process of typical reading development and about
what constitutes effective intervention for reading acquisition failure, the processes that moderate intervention response
for particular individuals with RD may be quite different.
Far less is known of this aspect of intervention for RD.
Two reviews have synthesized a small but growing literature on the factors that predict degree of response to reading
intervention. Al Otaiba and Fuchs (2002) focused on young
1
Brock University, St. Catharines, Ontario, Canada

The Hospital for Sick Children, Toronto, Ontario, Canada
3
Tufts University, Boston, MA, USA
4
Georgia State University, Atlanta, USA
2
Corresponding Author:
Jan C. Frijters, Brock University, Departments of Child and Youth
Studies and Psychology, 500 Glenridge Ave., St. Catharines,
Ontario, L2S 3A1, Canada
Email: jan.frijters@brocku.ca
Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011
151
Frijters et al.
children at risk for RD and reviewed studies in which the
characteristics of children unresponsive to intervention were
investigated. Nelson, Benner, and Gonzalez (2003) reviewed
many of the same studies, but in a quantitative meta-analytic
framework and with a broader focus on intervention efficacy. Across both reviews, PA and rapid naming were consistently associated with the degree of intervention response.
Al Otaiba and Fuchs reported that poor PA predicted nonresponse in 16 of the 21 studies that included this construct;
similarly, slow naming speed predicted nonresponse in five
of six studies that measured this construct. Nelson et al.
reported in their analyses that the mean effect sizes for the
prediction of intervention efficacy from PA (0.42) and rapid
naming (0.51) were equivalent. In addition, these two skills
yielded the strongest effect sizes, substantially higher than
effects for behavior, orthography, memory, and IQ.
Notably absent, or consistently associated with smaller
effect sizes in the studies covered by these reviews, are several specific neurocognitive processes. Of the 23 studies
reviewed by Al Otaiba and Fuchs (2002), 7 included memory measures and 15 included some measure of IQ. In both
cases, there were equivocal results for each construct, with
some studies reporting associations with nonresponse and
some studies reporting none. Within the studies reviewed
by Nelson et al. (2003), effect sizes for memory (M = 0.30)
and IQ (M = 0.26) were substantially smaller than those for
the phonologically based constructs, which consistently
demonstrated stronger evidence of association with reading. Recent studies have tested IQ as a moderator effect but
have not found evidence of differential growth or change
along these dimensions. For example, Lovett et al. (2008)
and Morris et al. (2010) compared growth curves for average versus low IQ participants and found parallel treatment
curves within intervention conditions. Fuchs and Young
(2006) have recently shown mixed and small effects for IQ
as a predictor of intervention responsiveness, especially for
younger children and for word identification outcomes.
There may be several reasons for these patterns of findings. The first is that such factors as IQ are not important to
remedial outcomethat current best practice multiple component interventions are largely complete instruction.
That is, individual differences in skills and knowledge are
compensated for through interventions designed to provide
explicit instruction, repeated practice, and structured attention to facilitating generalization of learning (Braden &
Shaw, 2009). However, were this the case, studies of sufficient power should show moderation effects for IQ, with
steeper growth rates or greater gains for children low on IQ
and related factors. Another possibility in the case of IQ
may be that the measure of IQ predominately analyzed in
studies to date has been the full-scale IQ measure, obscuring and even averaging out critical profile differences that
may predict responsiveness (Fiorello, Hale, & Snyder, 2006).
Finally, Fuchs and Young (2006) have demonstrated in their
review of 13 studies that the contribution of IQ to
explaining responsiveness systematically depends on the

type of instruction (i.e., a stronger relationship when the
instruction is comprehensive), outcome measure (i.e., a
stronger relationship when the outcomes are not single
word identification or nonword reading but more complex
reading tasks), and age of the children involved (i.e., a stronger relationship for children beyond Grade 2). As is argued
below, several cognitive and neuropsychological constructs
have a rich theoretical history, and at least some have demonstrated empirical evidence of association with reading
processesif not as moderators of intervention outcome.
Many cognitive and neuropsychological constructs may
have not been studied as predictors of responsiveness simply because of the focus on the reading-related language
processing factors studied most to date, PA and rapid
naming.
Phonological Coding
The process of encoding or representing linguistic information for later analysis and synthesis is a cognitive skill that
underlies PA, graphemephoneme mapping, and the development of individual word identification. These processes
have been studied at several levels of resolution, including
the individual phoneme and the morpheme. There is strong
evidence that these processes underlie vocabulary development (Gathercole & Baddeley, 1990; Metsala, Stavrinos, &
Walley, 2009) and spoken language comprehension (Baddeley,
1986). Scarborough (1998) showed that verbal memory
measured in second grade substantially predicted reading
achievement in eighth grade, but only for normally develo
ping readers. Gathercole, Alloway, Willis, and Adams
(2005) reported that phonological memory at the phoneme
and word levels was a significant predictor of reading for
children with RD, though weaker in predictive power than
complex working memory tasks. One form of phonological
coding is nonword repetition, and its relationship with reading on theoretical and empirical levels has been systematically reviewed by Gathercole (2006). The key dynamic in
this relationship is that the ability to repeat nonwords is an
index of the overall quality of the phonological storage system, involved in vocabulary, word learning, and PA.
A recent meta-analysis synthesized the literature comparing children with and without RD on measures of working memory and short-term memory (Swanson, Zheng, &
Jerman, 2009). This review included 43 studies between
1963 and 2006 that involved verbal short-term memory comparisons of these groups of children. Across these studies,
phonological memory measured in tasks such as nonword
repetition was significantly impaired for children with RD
when compared to age-matched children (weighted effect
size = 0.39). The parallel effect size for word-level shortterm memory was 0.55. In the fully partial model that controlled for the influence of working memory and attention,
only phonological memory among measures of short-term
152
Journal of Learning Disabilities 44(2)
memory was retained as significantly impaired in the group

with RD. Berninger, Raskind, Richards, Abbott, and Stock
(2008) reviewed evidence that children with RD with the
chromosome 6 DCDC2 deletion allele differ from children
without RD on phonological short-term memory, specifically
on tasks of nonword repetition. In addition, they reviewed
evidence that children with RD with a transcription site
alteration in DYX1C1 also differed from children without
RD. Thus, phonological coding appears to be consistently
impaired among children with RD, with stronger evidence
for impairment at the level of individual phonemes than at
the word level.
However, whether phonological coding is related to
change in reading skill is an open question. When readers
with RD were compared to skilled readers and to a combined RD and calculation-disordered group, Swanson and
Jerman (2007) found that nonword repetition and wordlevel short-term memory did not predict rates of growth in
reading ability. However, Berninger et al. (2008) also reviewed
evidence that differences with and without RD in the left
inferior frontal and parietal regions are evident before but
not after PA training, suggesting that changes in phonological coding may occur in conjunction with the reading growth
associated with intervention.
Visual Processes
Visual-motor integration (VMI) is the process of integrating visual perception with body movements. Although the
act of reading does not require motor output per se, there
is empirical evidence of a relationship between reading
skill and VMI. Across 161 studies, VMI was shown to have
consistently moderate relationships with reading processes,
with stronger relationships for word reading versus other
reading outcomes and for younger children versus older
children (Kavale, 1982; median r = .39). Younger children
were identified as preschool and older as intermediate,
with an average age of 7.88 years and an average grade
of 3.2 in this review. However, in another review covering
more recent studies, smaller relationships with reading
achiev ement were found (Scarborough, 1998; median
r = .28). This latter finding was confirmed in a recent
multivariate dominance analysis (Schatschneider, Fletcher,
Francis, Carlson, & Foorman, 2004) that yielded correlations ranging from .21 to .29 depending on the particular
reading outcome studied and whether VMI was assessed in
April or October of the kindergarten year. Notably, the
earlier studies reviewed in the meta-analyses focused on
cross-sectional relationships with reading skill, whereas the
Schatschneider et al. (2004) study specifically investigated
kindergarten predictors of Grade 1 and 2 reading outcomes.
OMalley, Francis, Foorman, Fletcher, and Swank (2002)
found that children without reading impairment in kindergarten and Grade 1 showed more robust growth from initial
assessment to Grade 2 in the development of VMI itself.
There has been some debate about the nature and contribution of visual sequential memory (VSM) to explaining
variability in reading skill (see Scheiman & Rouse, 2005).
In a longitudinal study of the kindergarten cognitive predictors of Grade 4 reading, Grogan (1995) found a moderate
contribution of VSM to reading skill (i.e., R2 = .05, after
accounting for age, auditory memory, and motor skills).
Kavales (1982) earlier meta-analysis showed substantial
associations between VSM and word reading (rs = .40 to
.49), with stronger relationships found for the younger children. Differences on tasks of VSM have been found when
the performances of normally developing and children with
RD (Watson & Willows, 1995) have been compared. Diff
erences in VSM were substantial when children with RD
were matched with younger successful readers, a readinglevel match that controlled for reading experience. There is
also evidence contextualizing these findings that children
with RD do not have visual memory problems per se but are
less likely to adopt a verbal labeling strategy on such tasks.
For example, children with RD can improve VSM with
training (Hicks, 1980). In addition, given the substantial
contribution of naming speed to reading skill, the rate of
visual stimuli presentation can have a marked impact on
whether relationships with reading skill are found (Huba,
Vellutino, & Scanlon, 1990). Thus, the deficiencies in VSM
that have been observed may be the result of verbal or rate
issues and not VSM per se.
Intelligence Quotient
Across several studies and meta-analyses, no systematic
differences in the reading achievement of children with RD
or LD that are attributable to IQ have been shown to exist
(Fletcher et al., 2002; Gustafson & Samuelsson, 1999;
Hoskyn & Swanson, 2000; Siegel, 1989, 1992; Stanovich &
Siegel, 1994; Stuebing et al., 2002). The vast majority of
previous studies have examined global intelligence for potential direct relationships to reading achievement. More recently,
perhaps because of reduced interest in IQ discrepancy versus
low achievement definitional debates (see Fletcher et al.,
1994; Scruggs & Mastropieri, 2003), more nuanced investigations into the role of intelligence have emerged. Tiu,
Thompson, and Lewis (2003) tested a structural model of
reading across normally developing and samples with RD
and found that performance IQ (PIQ) was related to reading
comprehension, but only for children with RD and only as
mediated by decoding skill. Consistent with these results,
Johnston and Morrison (2007) recently matched poor readers, reading age controls, and chronological age controls and
divided them into high (>101) and low (<100) IQ subsamples to form six comparison groups. Their results showed
that IQ moderated the relationship between reading outcome
and specific phonological deficits, such that high IQ poor
readers manifested more severe phonological reading
deficits.
153
Frijters et al.
Another way in which the IQreading relationship has
been obscured is that many studies have not considered the
well-defined factors that constitute global intelligence scores.
Vellutino, Scanlon, and Lyon (2000) reported correlations
between reading achievement and Verbal and PIQ factors at
several points from Grade 1 through Grade 4. Their results
indicated that verbal IQ (VIQ) was correlated with word
identification for normal readers from Grade 2 onward, but
for difficult to remediate readers the correlation that was
apparent in Grade 1 was nonsignificant by the winter of
Grade 2. In a recent confirmatory factor analysis of the
Wechsler Intelligence Scale for ChildrenFourth Edition
(WISC-IV), Glutting, Watkins, Konold, and McDermott
(2006) found that reading achievement on the Wechsler
Individual Achievement Test was related to a g factor but
that the verbal comprehension index also had a direct relationship with reading skill. In an earlier confirmatory factor
analysis of the WISC-III, Oh, Glutting, Watkins, Youngstrom,
and McDermott (2004) found a similar pattern of relationships, with an additional negative relationship between the
perceptual organization factor and reading processes. Hale,
Fiorello, Kavanagh, Holdnack, and Aloe (2007) have noted
that using g alone, or together with subcomponent scores,
is problematic because of the inherent multicollinearity
between full-scale IQ and these components derived from
it. It is important to contextualize the WISC-IV studies
with reference to other studies that have examined the relationship of CattellHornCarroll factors with reading (e.g.,
Evans, Floyd, McGrew, & Leforgee (2001)). These studies
have demonstrated that a focus on g incurs a loss of profile
information that is contained within subcomponents (Hale
et al., 2007). Floyd, Keith, Taub, and McGrew (2007, for
example, found persistent indirect relationships between g
and decoding, whereas within a two stratum structural
model processing speed and crystallized intelligence (i.e.,
specifically listening ability within a four stratum model)
had direct relationships with reading (see also Evans et al.,
2001).
There exists enough research and knowledge about the
development of reading processes to suggest that short-term
memory, visual memory, and IQ are important factors. Little
research is available to suggest whether any of these factors
moderate degree of response to reading intervention among
struggling readers. The present study investigates the contribution of eight specific neurocognitive process predictors
to understanding degree of intervention response among
young children with RD. Each of these predictors has a
theoretical or, at minimum, an empirical association with
reading skill. After estimating intervention response via
multilevel human growth curve models, three categories of
response were formed. Through a discriminant function ana
lysis of these categories the following research questions
were considered:
1. What is the contribution of intervention condition,

PA, and RAN to explaining degree of intervention
response, as measured by individual growth curves?
2. Beyond the contribution of specific intervention
conditions, PA, and RAN skills, do specific neurocognitive processes make independent contributions to knowing which children with RD show
a poor, average, or good reading outcome?
3. Does the inclusion of specific neurocognitive process measures improve the ability to classify intervention participants into poor, average, and good
responders; more specifically, for which category
does it improve classification the most?
Method
Participants
The current study reports on 278 participants originally
reported in Morris et al. (in press). Each child participated
in a multisite intervention study that compared four smallgroup interventions (described below) that provided 70
hours of intensive instruction. Children between 78 and 102
months whose first language was English were recruited
through referrals from classroom teachers who considered
them to have significant difficulty acquiring reading skills.
Referral was followed by administration of a screening battery consisting of the Kaufman Brief Intelligence Test (K-BIT;
Kaufman & Kaufman, 1990), the Woodcock Reading Mastery
TestsRevised (WRMT-R; Woodcock, 1987), the Wide
Range Achievement Test3rd Edition (WRAT-3; Wilkinson,
1993), and a demographic questionnaire completed by
parents. The screening battery was administered by trained
psychometrists at each of three sites (Atlanta, Georgia; Boston,
Massachusetts; Toronto, Ontario).
Two criteria were used to determine whether the child
had a RD. The low achievement criterion was met with a
K-BIT Composite standard score greater than or equal to 70
and a standard score of 85 or lower on at least one of three
reading indices as follows: (a) the average of the WRAT-3
Reading subtest, the WRMT-R Word Identification, Word
Attack, and Passage Comprehension subtests; (b) the
WRMT-R Basic Skills Cluster Score; (c) the WRMT-R
Total Short Scale Score. The abilityachievement regressioncorrected discrepancy criterion was met with a reading
index one standard error of the estimate below the regressionpredicted reading score. An assumption of an average .60
correlation between the reading indices described above
and the K-BIT Composite Index was used to calculate the
discrepancy scores. A child qualified for inclusion having
met either or both criteria.
Morris et al. (in press) described in more detail the number of children screened and the procedures for defining RD
154
Table 1. Average Preintervention Characteristics of the

Participants
Age at intervention start (months)

WRMT-R Word Attack (SS)
WRMT-R Word Identification (SS)
WRMT-R Passage Comprehension (SS)
K-BIT Matrices (SS)
K-BIT Vocabulary (SS)
K-BIT Composite (SS)
% male
% below 1 SD age expectations on
WRMT-R Word Identification
WRMT-R Passage Comprehension
Word Reading Efficiency
Remedial Intervention Programs
SD
93.59
76.57
73.38
78.49
91.54
94.12
90.60
6.07
11.94
10.39
9.85
11.16
10.61
13.46
61.5
75.8
77.3
94.6
Note: WRMT-R = Woodcock Reading Mastery TestsRevised;

SS =Standard Score; K-BIT = Kaufman Brief Intelligence Test.
status and ensuring representative sampling of socioeconomic levels, IQ, and race. The sampling strategy
resulted in equal numbers and at least 8 children per cell,
with at least 64 per condition and no site contributing
more than 5 participants in any cell. The specific levels of the
three factors were as follows: socioeconomic level, derived
from demographic information provided by parents, as low
or average and above; IQ as low (7090) or average and
above (>90); race as reported by the participants parents,
either Black or Caucasian. Table 1 details the average reading ability, IQ, and age of the 278 children participating in
the present study. Males constituted 61.5% of the sample.
Also contained in Table 1 is the proportion below 1 SD ageexpected performance across the following individual
measures: word reading (WRMT-R Word Identification),
comprehension (WRMT-R Passage Comprehension), and
fluency (Word Reading Efficiency). After confirmation of
RD, the WISC-III (Wechsler, 1991) was administered to all
children along with the cognitive and neuropsychological
measures detailed below.
The study design involved random assignment of small
groups of struggling readers, and their intervention teachers,
to one of four remediation conditions. The factorial design
of the study developed four randomly assigned groups of
struggling readers, such that each group included equal
numbers of Caucasian and Black students, of students with
average and below average family socioeconomic situations, and of students with average and below average IQ.
Groups of four children from each sample were taught one
of four remedial programs (detailed below) by trained
teachers for 70 contact hours during the school year.
Participants with similar single word reading levels (WRMT-R

and WRAT-3 Reading raw scores) were assigned to an
instructional group of four children, and these groups were
randomly assigned to one of four intervention programs.
Children were taught by trained teachers in a pull-out format for 60 minutes a day, four to five days a week, for a
total of 70 intervention sessions. The intervention design
included five components (Phonological Analysis and
Blending/Direct InstructionPHAB/DI, Word Instruction
Strategy TrainingWIST, Retrieval, Automaticity, Voca
bulary Elaboration, OrthographyRAVE-O, Classroom
Survival SkillsCSS, Mathematics ProgramMATH)
that were combined, two at a time, into four different intervention programs. The PHAB/DI component (a phonological decoding program based on Reading Mastery Fast Cycle
I/II by Engelmann & Bruner, 1988), part of each reading
intervention, averaged 30 minutes of instructional time in
every lesson.
The four programs resulting from combinations of these
components were the Phonological and Strategy Training
Program (PHAST), PHAB+RAVE-O, PHAB+CSS, and
MATH+CSS. The first two were new experimental, multiple
component interventions (PHAB/DI+WIST, which became
PHAST, and PHAB/DI+RAVE-O). PHAST (Lovett,
Lacerenza, & Borden, 2000; Lovett et al., 2005) and the
RAVE-O program (Wolf et al., 2009; Wolf, Miller, &
Donnelly, 2000) have been described in detail in separate
publications. PHAST integrates direct instruction of letter
sound mapping and phonological decoding with increasingly autonomous word identification strategy training.
This integration is illustrated in the sequence of five strategies introduced in a skill-based sequence as follows: (a)
Sounding Out, (b) Rhyming, (c) Peeling Off, (d)
Vowel Alert, (e) I Spy. A scaffolded framework of ins
truction incorporates explicit training in strategy selection,
application, and monitoring as guided by the metacognitive Game Plan. The three core goals of RAVE-O are
developing internal word structure pattern analysis at the
phonological and orthographic level, building word retrieval
capacity through strategies and vocabulary development,
and engaging in semantic analysis of words for their roots
and meaning associations. One core component of the program, vocabulary elaboration, emphasizes the following
through games and activities in group and on the computer:
(a) the multiplicity of meanings in each word (e.g., flexibility in word use), (b) semantic breadth (e.g., associations of
word), (c) semantic depth (e.g., contexts and use of word),
and (d) morphological endings (e.g., Ender Benders). Each
of these two intervention combinations devoted equal time
to its two components. The remaining two programs served
as control or contrast conditions: MATH+CSS served as an
155
Frijters et al.
alternative treatment control condition, and PHAB/DI+CSS
provided a phonological-only reading comparison.
Outcome Measures
Six reading and reading-related outcome measures were
included in the present study. These ranged from one phonological skill outcome to several measures of single word
identification, including a speeded measure, and one measure of passage comprehension. The measures for the present analysis were chosen from a much larger set of outcomes
assessed with this sample; the present subset of outcome
measures was selected to include four standardized measures and provide a broad assessment of reading ability.
Outcomes were assessed before intervention began, after
35 hours of intervention, and again after 70 hours of intervention. An overview of each outcome, the construct it
measures, and basic reliability and validity information are
provided below, but see Morris et al. (in press) for more
details on all aspects of the measures, the interventions, and
the overall study design.
Research-based measures. Blending Words from the Com
prehensive Tests of Reading Related Phonological Process
ing (CTRRPP; Torgesen & Wagner, 1996 was the research
version of the test in current publicationthe Comprehen
sive Tests of Phonological Processing (CTOPP; Wagner,
Torgesen, & Rashotte, 1999). This subtest assessed the ability to combine individual orally presented phonemes into
real words. For the age range in the present study, the measures authors report the internal consistency reliability
(coefficient alpha) to be .92 for a sample of normally developing readers and .87 for a sample of 61 children with RD.
Testretest reliability for this age range was .88.
The Test of Transfer is a computer-administered, timed
word identification test consisting of 40 words that are systematic transformations of high-frequency spelling patterns
taught in the intervention programs (Lovett et al., 2000).
For example, the keyword bake is taught in the intervention,
and the Test of Transfer may contain fake, babe, bike, and
baker. These four words represent onset, rhyme, vowel, and
suffix-based transformations of the keyword. Words were
presented singly and displayed on a computer monitor for
6,000 milliseconds, and children responded orally, speaking into a microphone to trigger the next list word. Cirino
et al. (2002) report the testretest reliability of this measure to
be .94 for the present sample. Raw scores of both Blending
Words and the Test of Transfer were used in the present
analysis.
Standardized tests of reading processes. Three subtests from
the WRMT-R, Form G (Woodcock, 1987) were administered. The Word Identification, Word Attack, and Passage
Comprehension subtests represent respectively context-free
assessments of single-word reading, nonword decoding, and
reading comprehension assessed through a cloze procedure.
Internal consistency reliability derived from the normative

sample for these subtests was .98, .94, and .98, respectively.
A measure of word-level fluency, the Word Reading Effi
ciency test, was also administered (Torgesen & Wagner,
1996). This test is the research version of the currently published Test of Word Reading Efficiency (TOWRE) Sight
Word Reading subtest (Torgesen et al., 2001), and the items
overlap substantially with this test and partially with the
WRMT-R Word Identification subtest. Children are asked
to read as many words as quickly and accurately as possible
from a list of 104 real words. The raw number of words read
within 45 seconds represented a measure of single-word
reading fluency. Testretest reliability for children in the
same age range as the present sample ranged from .93 to
.97. Testretest reliability for 29 children with RD ages 6 to
9 was .97.
Neurocognitive Process Predictors

Predictors included in the present study were drawn from a
broad range of cognitive and neuropsychological predictors
originally assessed within the present sample. Selection
of predictors was guided by demonstrated theoretical or
empirical links with either reading processes or with
response to remediation, as reviewed earlier. In total eight
specific neurocognitive process measures were included,
along with two oral language skills.
Oral language skills. The Elision subtest from the CTRRPP
and the Letters array from the Rapid Automatized Naming
task (Denckla & Rudel, 1974, 1976) were chosen for inclusion in the base model described below. These two particular measures were chosen because of the strong history of
empirical association with both reading skills and remedial
response. The Elision subtest is a phonological analysis task
that requires children to orally segment individual phonemes
from within words and resynthesize them. The Rapid Auto
matized Naming Letters task presents five letters, randomly
arrayed in 5 lines of 10 items. Children are required to rapidly name the letters, and time to complete the full array is
the score for this measure.
Memory processes. The CTRRPP subtests Memory for
Words and Nonword Repetition are measures of phonologically mediated short-term memory, at two levels of resolution. The Nonword Repetition task requires children to
repeat a graduated series of 25 nonwords increasing in the
number of phonemes, presented orally in succession via a
prerecorded audiotape. This task is discontinued after five
consecutive errors. The Memory for Words task requires
children to listen to and then repeat orally a word sequence.
Words are presented with increasing list length, until three
consecutive errors are made.
Visual processes. The Beery-Buktenica Developmental Test
of Visual-Motor Integration, 5th Edition (VMI; Beery, &
Buktenica, 1989) is a structured measure of visual-motor
156
Table 2. Growth Curve Models of Outcome Variables

CTRRPP
Blending
Fixed Effects
Par.
Par.
SE
WRMT-R Word
Att.
Par.
SE
Test of
Transfer
Par.
Initial status
Intercept 00 9.76** 0.27 448.62** 0.77 3.76**
Rate of change Intercept 10 2.19** 0.12 10.06** 0.40 4.64**
Variance
components
Level 1
Within
2 4.14** 0.35 69.24** 5.90 10.13**
person
Level 2
In initial
20 16.84** 1.75 105.53** 14.73 32.82**
status
In rate of 21 1.91** 0.38 10.23* 4.82 11.80**

change
Correlations Intercept 201
.20
.32
.08
with
slope
WRMT-R Word
Id.
SE
Par.
SE
WRMT-R P.
Comp.
Par.
SE
TOWRE S.
Word
Par.
SE
0.39 402.84**
0.25 14.83**
1.54 439.92**
0.47
9.30**
1.02 12.25**
0.40 5.01**
0.57
0.19
0.88
3.37
4.18
6.06**
0.51
3.61 622.51** 55.79 249.40** 24.95 83.97**
7.58
1.50
6.92**
0.88
.05
39.63**
42.32**
5.55
.19
49.14**
20.38**
.20
4.36
Note: Par. = Parameter; CTRRPP = Comprehensive Tests of Reading Related Phonological Processing; WRMT-R = Woodcock Reading Mastery TestRevised;
TOWRE = Test of Word Reading Efficiency.
*p < .05. **p < .001.
processes that does not require a verbal response. Children

are presented with a series of increasingly complex geometric shapes to copy into a specific location within a response
booklet. The test has been designed for ages 2 to 18. The
Test of Visual-Perceptual Skills, 3rd Edition, Visual Sequential Memory subtest (TVPS-VSM; Gardner, 1988) is a measure of nonverbal sequential memory. Children are presented
a sequence of several shapes and have 1 minute until the
shapes are removed and a target stimulus is presented with
the original shape and three distracters. Recognition of the
pattern that matches the original sequence of shapes is the
target response.
WISC-III indices. The WISC-III (Wechsler, 1991) was part
of the neurocognitive battery administered once to each
child. The present analysis incorporates the four index
scores that have emerged from factor analyses of the WISCIII subtests (Sattler, 1992). These four indices are as follows: Verbal Comprehension (VC), Perceptual Organization
(PO), Freedom from Distractibility (FDI), and Processing
Speed (PS).
Analyses
The overall analytic plan consisted of three steps. First,
response to intervention was estimated via growth curve
models applied to the six reading outcomes. Second, three
categories of intervention response were formed for each
outcome using the individually estimated growth rates generated by the growth models. Third, a series of discriminant
function analyses (DFAs) was performed on these categories, using a fixed set of cognitive and neuropsychological
measures to predict group membership and evaluate the
relative contribution of individual predictors to classifying

poor, average, and good responders for each outcome.
Results
Estimating Relative Size of
Reading Outcomes
Individual growth curve models were formulated to define
each participants response to intervention. Across the 70 hours
of intervention, three repeated measurements were taken on
each outcome measure, providing sufficient resolution for a
growth model with a random linear slope and random intercept. Visual inspection of individual growth curves did not
provide evidence of curvilinear growth rates. Several quantifications of response to intervention were explored, inc
luding intercept only, intercept and slope together, and
slope alone. In the analysis to follow each led to similar
model results; thus, the slope-only model representing the
number of items gained per 70 hours of intervention was
used to quantify response to intervention. In the case of the
Test of Transfer and the CTRRPP Blending, raw scores were
used, and in the case of the WRMT-R subtests, Rasch-scaled
W scores were used as the scale of measurement. Mean
growth rate per outcome, tests of random slopes and intercepts, and residual within-participant variation are reported
in Table 2.
Three categories representing degree of response to
intervention on each outcome were formed from these data.
Because the current study was designed to investigate factors that contribute to suboptimal and excellent types of res
ponses to intervention, the top (n = 56) and bottom (n = 56)
157
Frijters et al.
Table 3. Gains Over Full Intervention for Three Response Categories
CTRRPP
Blending
(Raw Scores)
Poor
Average
Good
WRMT-R
Word Att. (W
Scores)
Test of
Transfer (Raw
Scores)
WRMT-R
Word Id. (W
Scores)
WRMT-R P.
Comp (W
Scores)
TOWRE S.
Word (Raw
Scores)
SD
SD
SD
SD
SD
SD
1.64
4.30
7.48
0.78
0.89
1.02
2.04
8.82
19.24
0.39
3.73
2.84
14.76
20.23
24.97
1.64
1.52
1.78
16.45
28.38
47.90
2.36
5.10
7.82
10.36
18.20
28.62
2.78
2.80
3.47
4.24
9.63
17.48
1.05
2.35
2.93
Note: Par. = Parameter; CTRRPP = Comprehensive Tests of Reading Related Phonological Processing; WRMT-R = Woodcock Reading Mastery Test -Revised;
TOWRE = Test of Word Reading Efficiency. Poor and good responders n = 56 per group; average responders n = 166.
quintiles were chosen to define these categories and maximize separation between them. The middle three quintiles
were combined to form a category of average responders
(n = 166). Table 3 details the response categories across the
six outcomes. In general, the poor responders gained some
reading skills across the six outcomes; in contrast, the good
responders made consistently 3 to 4 times the gains of the
poor responders. There was substantial overlap across outcomes on how participants were classified. Across the six
outcomes, significant associations, 2(4) ranging from 6.74
to 120.97; M = 40.03, SD = 34.83, were observed when any
two classifications were paired (evaluated at = .003 to
correct for multiple comparisons), with the exception of the
pairing of the Test of Transfer, 2(4) = 6.74, p = .15, with
CTRRPP Blending. Across all 15 combinations of classifications, the Cramers V average effect size was 0.25 (range =
0.11 to 0.47, SD = 0.11), indicating moderate overlap in the
response categories across the six outcomes.
DFA Base Model

Two hierarchically nested discriminant function models
were constructed. To evaluate the utility of intervention condition and phonological and language measures to explain
response categories, the first discriminant function model
incorporated intervention condition, the Elision subtest from
the CTRRPP, and the Numbers subtest from the Rapid
Automatized Naming task to predict membership in the three
categories of response. Intervention condition was entered
into the model using a priori single degree of freedom contrasts as follows: (a) Contrast 1 compared the control condition of MATH+CSS to all three reading interventions together,
(b) Contrast 2 compared single component reading intervention (PHAB/DI+CSS) to multiple component interventions,
and (c) Contrast 3 directly contrasted PHAB/DI+WIST to
PHAB/DI+RAVE-O. Elision and Rapid Automatized Naming
Numbers were entered into the model along with these contrasts as z score conversions based on the prepublication norms
available at the time the tests were administered.
There were no missing data across all outcome measures,
and minimal missing data were present across the predictors
(6.5%). Exploratory analyses comparing participants with

missing and nonmissing data points indicated that the pattern
was likely ignorable and randomly distributed across predictors. Expectation maximization imputation was used to
create a full data set for the DFAs. Evaluation of assumptions for both this model and the additive model described
below revealed moderate violation of the homogeneity of
variances across the three response groups for three out of
six outcomes. As a result, a quadratic discriminant function
was performed for those outcomes. All other DFA assumptions (e.g., linearity, normality, and multicollinearity) were
found to hold with the current data.
For each of the six outcomes, two discriminant functions
were extracted (see Table 4 for 2 and percentage variability
accounted for values); however, in only one case was the
second function significant after the first function was remo
ved (WRMT-R Passage Comprehension), F(4, 272) = 2.89,
p = .02. For all outcomes, examination of the group centroids indicated that the first function equally separated the
three response categories from each other. In both the base
and the additive model described below, semipartial univariate results per predictor were also evaluated, correcting for
multiple comparisons via Sidaks formula. Thus, in the base
model, univariate results were evaluated at an alpha of .01.
For all outcomes, Intervention Contrast 1 significantly separated the three response groups indicating that treatment versus
control status significantly discriminated among groups.
Intervention Contrast 2 also significantly separated the three
response groups on the Test of Transfer and Word Attack,
indicating that multiple component intervention significantly
discriminated among groups for those outcomes. Intervention
Contrast 3 did not significantly separate the three response
groups on any of the outcomes. Elision significantly separated
groups at the univariate level on the all outcomes but Blending
Words and TOWRE Sight Words. Table 4 details the univariate results across the six discriminant function base models.
The standardized discriminant coefficients (dc) were
examined to determine the relative importance of each predictor. The coefficients represent the influence of each variable on the discriminant function, controlling for all other
variables, and were evaluated against a criterion of .30. For
158
Table 4. Intervention and Phonology Discriminant Function Models, Reporting on Only the First Discriminant Function
Reading Outcome
CTRRPP
Blending
WRMT-R
Word Att.
Model F(10, 542)

3.62*
7.79*
Canonical corr.
.32
.47
Eigenvalue
.11
.28
Univariate semipartial tests, F(2, 275)
Contrast 1
14.30*
17.48*
Contrast 2
1.79
7.28*
Contrast 3
0.68
2.37
RAN Let.
1.44
3.67
Elision
0.06
8.00*
Standardized discriminant coefficients (only >.25 reported)
Contrast 1
.97
.83
Contrast 2
.45
Contrast 3
RAN Let.
.29
Elision
.38
Test of
Transfer
WRMT-R
Word Id.
WRMT-R P.
Comp.
TOWRE S.
Word
13.94*
.60
.57
2.16*
.26
.07
4.70*
.34
.13
10.00*
.51
.36
11.54*
6.02*
0.48
20.10*
36.90*
8.15*
1.45
0.38
0.15
0.86
8.38*
3.59*
1.42
6.81*
4.78*
11.89*
2.53
1.41
21.51*
17.80*
.61
.25
.92
.29
.73
.47
.68
.28
.60
.66
.52
.50
.42
Note: CTRRPP = Comprehensive Tests of Reading Related Phonological Processing; WRMT-R = Woodcock Reading Mastery TestsRevised; TOWRE = Test
of Word Reading Efficiency; RAN = Rapid Automatized Naming. The three orthogonal contrasts were as follows: Contrast 1 = intervention vs. control;
Contrast 2 = multiple component vs. single component intervention; Contrast 3 = Phonological and Strategy Training vs. Retrieval, Automaticity,
Vocabulary Elaboration, Orthography.
*p < .01.
all outcomes, Intervention Contrast 1 was consistently an

important predictor based on this criterion, particularly for
the Blending Words, Word Identification, and Passage Com
prehension outcomes (dc ranging from .60 to .97). For the
Test of Transfer, Passage Comprehension, and TOWRE
Sight Words outcomes, rapid naming was an important and
unique predictor (dc ranging from .47 to .66). For these
same outcomes and for Word Attack, Elision was an important and unique predictor (dc ranging from .38 to .68). Post
hoc comparisons of mean levels of Elision and Letter
Naming per outcome indicated an overall pattern of poor
responders having lower scores compared to average and
good responders, who were not different from each other. In
contrast, good responders on the Test of Transfer and the
TOWRE Sight Words had better Elision scores and faster
Letter Naming speeds than average responders.
DFA Cognitive and Neuropsychological

Additive Model
Nested models relative to the base model described above
were formed by including eight neurocognitive process
predictors along with the previously evaluated intervention
contrasts, Elision, and Blending Words. The goals of the
nested models were as follows: (a) to evaluate which cognitive and neuropsychological predictors contributed to the
discriminant function, controlling for intervention, PA, and
rapid naming; and (b) to establish if predictive power could
be enhanced by the inclusion of this class of predictors and

specifically how this improvement functions.
In the additive model, univariate results were evaluated
at an alpha of .004, using Sidaks formula to correct for multiple tests. As can be seen in Table 5, a range of cognitive
and neuropsychological predictors significantly explained
membership in the response categories. The standardized
discriminant coefficients (dc) were examined to investigate
the relative importance of each predictor, controlling for all
other additive predictors including those in the base model.
Even with all additive predictors in the model, the coefficients for Contrast 1 remained significant, indicating that
treatment status continued to explain membership in the
response categories. Rapid naming and elision remained
unique and important predictors across WRMT-R Word
Attack, Passage Comprehension, Test of Transfer, and TOWRE
Sight Words.
Among the neurocognitive process predictors, the WISCIII VC factor explained unique variance in response categories across the six outcomes (dc ranging from .25 to .75). A
similar pattern of post hoc differences among the response
categories was observed, with poor responders having lower
VC scores than average and good responders, who were not
different from each other. In contrast, good responders on
the Test of Transfer and the TOWRE Sight Words had better VC scores than average responders. The VMI remained
a significant predictor of response on the Blending and
Word Attack outcomes, in both cases with high VMI scores
159
Frijters et al.
Table 5. Final Additive Discriminant Model Overall and Univariate Predictor Results
Reading Outcome
CTRRPP
Blending
Model F(26, 526)

2.45*
Canonical corr.
.40
Eigenvalue
.19
Univariate semipartial tests, F(2, 275)
Beery VMI
4.46
Memory for Words
1.26
Nonword Rep.
1.82
TVPS VSM
1.24
WISC-III FDI
2.58
WISC-III PO
1.61
WISC-III PS
3.03
WISC-III VC
0.34
Standardized discriminant coefficients (dc >.25 reported)
Contrast 1
.75
Contrast 2
Contrast 3
RAN Let.
Elision
.46
Beery VMI
Memory for Words
Nonword rep.
TVPS VSM
WISC-III FDI
.27
WISC-III PO
WISC-III PS
.27
.75
WISC-III VC
WRMT-R
Word Att.
Test of
Transfer
WRMT-R
Word Id.
WRMT-R P.
Comp.
4.67*
.54
.41
6.57*
.65
.72
1.71*
.33
.13
7.27*
1.96
7.65*
1.07
2.42
5.82*
1.77
5.89*
6.84*
1.85
3.34
7.38*
10.33*
14.29*
3.91
14.02*
1.77
2.70
1.88
2.67
3.29
3.10
0.34
4.74
5.38
5.09
3.40
3.88
6.01
5.36
10.1*
4.90
.80
.33
.61
.77
.60
.34
.54
.57
3.06*
.43
.23
.32
.36
.40
.30
.33
.25
.61
.42
.34
.60
TOWRE S.
Word
4.93
.55
.44
2.39
4.75
3.52
4.27
8.85*
5.38
5.85*
7.68*
.59
.66
.35
.59
Note: CTRRPP = Comprehensive Tests of Reading Related Phonological Processing; WRMT-R = Woodcock Reading Mastery TestsRevised; TOWRE = Test
of Word Reading Efficiency; VMI = visual-motor integration; TVPS-VSM = Test of Visual-Perceptual Skills Visual Sequential Memory subtest; WISC-III =
Wechsler Intelligence Scale for ChildrenThird Edition; FDI = Freedom from Distractibility Scaled Score; PO = Perceptual Organization Scaled Score;
PS = Processing Speed Scaled Score; VC = Verbal Comprehension Scaled Score. The three orthogonal contrasts were as follows: Contrast
1 = intervention vs. control; Contrast 2 = multiple component vs. single component intervention; Contrast 3 = Phonological and Strategy Training vs.
Retrieval, Automaticity, Vocabulary Elaboration, Orthography.
*p < .004.
predicting better response across all response categories.

WISC-III FDI and PS predicted responses on Blending, but
only separating good responders from average and poor,
who did not differ from each other on these cognitive variables. For the remaining effects of particular cognitive or
neuropsychological predictors explaining response, the following pattern was observed: Poor responders had low
scores on the predictor, separating them from good and average responders, who did not differ from each other. This
pattern held for Nonword Repetition predicting Test of
Transfer response, PO predicting Word Attack response,
VSM predicting Word Identification response, and WISCIII PS predicting Passage Comprehension response.
Follow-Up Analysis of Change in

Classification Errors From Base
to Additive Model
The goal of the present analysis was not classification per
se. However, an examination of the classification error
rates was conducted to further characterize the utility of the
cognitive and neuropsychological predictors in explaining
the relative size of the reading outcome. The first error rate
analysis consisted of comparisons between models in the
overall improvement in reclassification. The proportion of
misclassifications unique to the base model minus misclassifications unique to the additive model squared, divided by
160
Table 6. Comparison of InterventionPhonology and CognitiveNeuropsychological Additive Model

Reading Outcome
Misclassifications unique to base

Misclassifications unique to additive
2difference (1, .05)
CTRRPP
Blending
WRMT-R
Word Att.
Test of
Transfer
62
20
21.51*
75
20
31.84*
43
18
10.25*
WRMT-R
Word Id.
33
25
1.10 (ns)
WRMT-R
P. Comp.
TOWRE
S. Word
59
16
24.65*
39
13
13.00*
Base model reclassification error rates
Poor
Average
Good
0.54
0.19
0.91
0.39
0.39
0.62
0.42
0.23
0.48
0.53
0.21
0.87
0.57
0.18
0.80
0.38
0.18
0.70
Additive model reclassification error rates
Poor
Average
Good
0.33
0.16
0.35
0.20
0.21
0.36
0.27
0.18
0.30
0.87
0.05
0.98
0.34
0.11
0.37
0.28
0.12
0.46
Improvement in reclassification error rates

base to additive model
Poor
Average
Good
0.22
0.03
0.57
0.20
0.18
0.26
0.15
0.05
0.17
0.34
0.16
0.11
0.23
0.06
0.43
0.11
0.05
0.24
Note: CTRRPP = Comprehensive Tests of Reading Related Phonological Processing; WRMT-R = Woodcock Reading Mastery TestsRevised; TOWRE = Test of
Word Reading Efficiency. Poor and good responders n = 56 per group; average responders n = 166.
*p < .01.
the sum of these two is distributed as a 2 with one degree

of freedom. When applied to each outcome, significantly
greater correct reclassification was observed for every outcome model, except WRMT-R Word Identification. Table 6
details the 2 tests for each outcome. These analyses indicate
that the additive model improved classification over the base
model.
The second error rate analysis consisted of examining
changes in the errors of classification across the three res
ponse categories from the base to the additive model. The
base model was more effective in classifying the poor res
ponders, with average error rates across the six outcomes
25% below the error rate of classifying good responders.
With the addition of cognitive and neuropsychological predictors, classification improved markedly with lower error
rates across the three response categories. On Blending
Words, WRMT-R Passage Comprehension, and TOWRE
Sight Word reading, the improvement in classifying good
responders was twice the improvement of classifying poor
responders. On WRMT-R Word Attack and Test of Transfer,
roughly equal improvement was seen in classifying poor
and good responders. Table 6 details the base and additive
model error rates, along with tests of improvement to classification error rates.
Discussion
The present study was conducted to establish whether reliable predictors of reading processes are also predictive of
relative size of reading outcomes in the context of intensive

remedial reading instruction. Among the potential neurocognitive processes that have been shown to relate to the
process of reading, the present study considered whether
short-term memory, visual memory, and specific components of IQ moderated degree of response to reading intervention among young struggling readers. A particular focus
of this study was to evaluate the profile of neurocognitive
processes that explained poor, average, or good relative sizes
of reading outcomes.
Results from the present study suggest that phonological
memory, cognitive, and visual processing predictors do have
predictive value in explaining response to intervention among
children with RD. These effects were demonstrated above
and independent of the contributions of multiple-component
intervention, PA, and rapid naming skill. As supported by
extensive empirical and theoretical work on RD, PA and
rapid naming significantly distinguished the three degrees
of response. Slow naming speed and poor PA in particular
separated poor responders from both average and good
responders. This pattern held across all outcomes except the
two requiring rapid responses (i.e., Test of Transfer and
TOWRE Sight Words), on which faster naming and better
PA separated the three groups in a continuous function.
The specific contribution of the additive predictors was
to improve classification accuracy of both poor and good
responders, with a notable advantage in classifying who
was a good responder. Excluding WRMT-R Word
Identification, for which the additive predictors did not
161
Frijters et al.
improve classification, incorporating the cognitive and
neuropsychological variables improved classification of
poor responders on average by 18.2%. Across a broad
range of reading skillsphonological blending, comprehension, and fluencythe cognitive and neuropsychological
predictors improved classification of good responders by
twice the amount for poor responders, an average of 41.3%.
On the three measures of single word reading or nonword
decoding, the improvement to classifying good responders
was equivalent to the improvement in classifying poor
responders.
These results are consistent with Vellutinos (2001) suggestion that cognitive processes may be correlated more
strongly with reading comprehension while less strongly correlated with more basic reading skills, such as nonword reading and word identification. On the surface, the present results
may be viewed as inconsistent with those of Johnston and
Morrison (2007), who found that high-IQ poor readers were
more phonologically impaired than low-IQ poor readers and
had relatively more difficulty decoding new words and nonwords. However, in the context of phonologically intense
instruction, higher IQ versus low IQ was associated with
greater responsethat is, benefit. This finding highlights the
need to consider carefully past correlational results linking
cognitive and neuropsychological factors with reading acquisition. Static predictors of reading ability or the reading process may not demonstrate relationships, whereas the same
predictors may explain reading outcomes when related to
response or rate of growth. Results from the present study
are consistent with this perspective, given that the degree of
response on the two phonologically based outcomes
Blending Words and WRMT-R Word Attackwas uniquely
predicted by the greatest number of neurocognitive process
predictors (see standardized discriminant coefficients reported
in Table 5). The addition of neuropsychological predictors
increased the classification accuracy for poor and good
responders on phonological blending skill by 22 and 57%
respectively, consistent with Torgesen and Davis (1996).
At the univariate level, the pattern of significant predictors was linked to the underlying reading process that each
outcome represents. These relationships were supported by
interpretation of the standardized discriminant coefficients,
which fully partial out the effect of intervention, PA, and
rapid naming skills. For example, growth in phonological
blending skills was explained by VMI and the FDI of the
WISC-III. Vellutino et al. (2000) have shown that there
may be a stronger relationship for VIQ than PIQ with reading outcomes, with the strongest relationship for younger
children in their sample. This is consistent with the present
results, with the WISC-III VC index a statistically significant and notable predictor of degree of response across all
outcomes. As indicated by the standardized discriminant
coefficients in Table 5, VMI, Nonword Repetition, and the
WISC-III indices of FDI, PS, and VC each played a role in
predicting degree of response for these outcomes. The present
sample was a younger sample, with participants ranging from
Grade 1 to Grade 3. As such, the results highlighting the

importance of naming speed are consistent with the current
literature, which shows a greater influence of naming speed on
early reading development, with less influence after Grade 3
(Wolf, Bowers, et al., 2000). The present sample also had
significantly impaired reading skills, and thus the present
results are consistent with a recent large-scale and longitudinal study that found RAN to be a more important predictor
for the poorest readers (Lervg, Brten, & Hulme, 2009).
The results of the present study are consistent with
Vellutino et al.s (2000) reanalysis of Vellutino et al. (1996;
see also Scanlon & Vellutino, 1997), who found that IQ
did not predict growth in reading comprehension or word
identification for children in the good or very good growth
rate groups. The two exceptions to this in the present study
involve the measures of transfer of learning (i.e., Test of
Transfer) and fluency (i.e., TOWRE Sight Word). Both are
measures of single word identification with a significant
speeded component. In the former, children are presented
with single words that persist a limited amount of time on a
computer display; in the latter, children are instructed to
read a list of words as quickly as possible. Both of these
tasks are demanding on a skill integration level and tax multiple and independent cognitive and perceptual systems. In
the present study, additional cognitive capability did distinguish between the best responders and the average or lower
responders. This finding is also consistent with the review
by Fuchs and Young (2006), which suggested that cognitive
and neuropsychological predictors were more predictive of
intervention response when the instructional focus was on
more complex reading tasks.
Limitations
Given the current state of research into the factors that predict responsiveness to intervention, the present analysis was
conducted in an exploratory manner, preserving individual
outcomes and the individual predictors. At the outcome
level, this decision is supported by the tests of association
across classifications derived from the six reading measures. Although the association from measure to measure
on who was a good, average, or poor responder was statistically significant and of moderate effect size, the association
was not perfect. Though some of the disagreement is attributable to the reliability of these outcome measures, all outcome measures demonstrated excellent psychometric properties
as previously reviewed. This finding suggests that each
reading measure had unique ability to separate degree of
response and suggests that studies on intervention responsiveness should take care to incorporate multiple dimensions of reading process.
At the predictor level, it is likely that these predictors are
components of a smaller number of dimensions, which may
or may not relate to intervention response. An additional
argument for retaining the individuality of the predictors is
that those included in the present study are a select group
162
that notably excluded orthographic processing (e.g., Mano &

Osmon, 2008), executive function (Reiter, Tucha, & Lange,
2005), and working memory (e.g., Swanson & Jerman, 2007),
among others. A dimensionality analysis in the present context would inevitably result in a measure-dependent oversimplification of the variable set. Regardless, an exploratory
post hoc exploratory factor analysis of the present measures
indicated three factors: a phonological processing factor
that consisted of Elision and Letter Naming, a memory factor that consisted of Nonword Repetition and Memory for
Words, and a general cognitive factor that included all of
the other study predictors. A comprehensive structural
model could be envisioned that would combine such a
dimensionality analysis with a structural analysis of differing relationships between neurocognitive predictors and the
type of reading outcome involved. Such a model would
explicitly model the suggestions within the present analysis
that neurocognitive predictors share a stronger relationship
with the more multidetermined reading processes (e.g.,
comprehension, fluency-based outcomes, etc.).
To preserve the richness of exploration across multiple
predictors and outcomes, the present analysis did not consider interactions of cognitive and neuropsychological predictors with the particular interventions involved. Half of
the sample participated in multiple component interventions that emphasized direct training of phonological skills
and orthography combined with either metacognitive strategy instruction (PHAST; Lovett et al., 2000) or fluency and
vocabulary building instruction (RAVE-O; Wolf, Miller,
et al., 2000). With 70 participants in each of the control and
PHAB/DI conditions, incorporating interactions between
cognitive predictors and intervention type would have
necessitated dropping several predictors.
It is important to note that the present study considered
the association of cognitive and neuropsychological measures with the degree of response to intervention from a
variable-centric perspective. This perspective emphasizes
overall relationships among constructs and speaks very
minimally to the dynamics for particular individuals. Par
ticular component skills that have been demonstrated to
moderate response to intervention may be targets for pre- or
simultaneous intervention. Gathercole and Alloway (2006)
recently described how short-term and working memory are
amenable to intervention, with an emphasis on both strategies to compensate for deficits in these areas and training to
expand capacity. Thus, at the level of practice, the present
study supports the use and inclusion of a broader range of
cognitive and neuropsychological factors as targets for generating individualized hypotheses about intransigent reading difficulties (for a description of such an approach, see
Fiorello et al., 2006).
Implications for Intervention Architectures

Overall, the present analysis suggests that predictors of rea
ding outcomes for young children with RD may need to be
conceptualized more broadly than phonological analysis

and rapid naming skill alone. Such a broader conceptualization has been illustrated by the recent evaluation of the
convergent skills model by Vellutino, Tunmer, Jaccard, and
Chen (2007). This type of integrated and theory-driven
model that includes language-based processes along with
component memory and visual skills may help expand a
full understanding of who does and does not respond to
generally efficacious interventions. In the context of formal response to intervention (RTI) programs and implementations, research-based interventions that focus on
decoding and fluency processes have been shown to be
effective for early readers and for struggling readers. On
the weight of the present results and of past research, Tier
1 and Tier 2 RTI protocols are well justified in implementing these types of interventions.
As recent neuroimaging work has shown, phonological
and word-rate-focused interventions can lead to a shift in
processing for children with RD from the right to the left
hemisphere (see Richards & Berninger, 2008; Richards
et al., 2007, Simos et al., 2007a, 2007b). In this research,
children with RD displayed an overactive profile in areas
typically associated with visual processing prior to intensive reading intervention. After intervention, both functional
magnetic resonance imaging and magnetic source imaging
scans showed a normalization pattern, with decreased right
hemispheric (occipitotem-poral) and increased left hemispheric (e.g., temporoparietal) processing areas used by fluent, nonimpaired readers for phonologically based tasks.
Consistent with the present findings that showed an important role for visual, verbal, and memory neurocognitive processes, this neuroimaging work suggests that reading outcomes
are multidetermined and dynamic. This may be especially
true when the focus shifts to poor or even nonresponders, as
shown in the work by Simos and colleagues (2007a, 2007b).
In these studies, children who did not make clinically significant gains to intensive phonological intervention did not
show the same normalization, continuing to rely on alternate brain regions to accomplish decoding and phonological tasks. In these cases, both recent imaging studies and
the present results suggest that a comprehensive evaluation
of neurocognitive processes may be needed to support the
development of more individualized interventions (Hale,
Kaufman, Naglieri, & Kavale, 2006).
Declaration of Conflicting Interests
The authors declared no potential conflicts of interests with respect
to the authorship and/or publication of this article.
Financial Disclosure/Funding
The authors disclosed receipt of the following financial support
for the research and/or authorship of this article:
The research reported here was supported by a National
Institutes of Child Health and Human Development Grant
(HD30970) to Georgia State University, Tufts University, and
The Hospital for Sick Children/University of Toronto.
163
Frijters et al.
References
Al Otaiba, S., & Fuchs, D. (2002). Characteristics of children who are unresponsive to early literacy intervention: A
review of the literature. Remedial and Special Education, 23,
300316.
Baddeley, A. D. (1986). Working memory. Oxford, UK: Oxford
University Press.
Beery, K. E., & Buktenica, N. A. (1989). Developmental Test of
Visual-Motor Integration. Odessa, FL: Psychological Assessment Resources.
Berninger, V. W., Raskind, W., Richards, T., Abbott, R., & Stock, P.
(2008). A multidisciplinary approach to understanding developmental dyslexia within working-memory architecture:
Genotypes, phenotypes, brain, and instruction. Developmental
Neuropsychology, 33, 707744.
Block, C. C., Parris, S. R., Reed, K. L., Whiteley, C. S., &
Cleveland, M. D. (2009). Instructional approaches that signi
ficantly increase reading comprehension. Journal of Educa
tional Psychology, 101, 262281.
Braden, J. P., & Shaw, S. R. (2009). Intervention validity of cognitive assessment: Knowns, unknowables, and unknowns.
Assessment for Effective Intervention, 34, 106115.
Cirino, P. T., Rashid, F. L., Sevcik, R. A., Lovett, M. W.,
Frijters, J. C., Wolf, M., & Morris, R. (2002). Properties of
measures of beginning decoding and related abilities in poor
readers. Journal of Learning Disabilities, 35, 526539.
Denckla, M. B., & Rudel, R. G. (1974). Rapid automatized naming of pictured objects, colors, letters, and numbers by normal
children. Cortex, 10, 186202.
Denckla, M. B., & Rudel, R. G. (1976). Naming of objects by
dyslexic and other learning-disabled children. Brain & Lan
guage, 3, 115.
Engelmann, S., & Bruner, E. C. (1988). Reading Mastery I/II fast
cycle: Teachers guide. Chicago, IL: Science Research Associates.
Evans, J. J., Floyd, R. G., McGrew, K. S., & Leforgee, M. H. (2001).
The relations between measures of Cattell-Horn-Carroll (CHC)
cognitive abilities and reading achievement during childhood
and adolescence. School Psychology Review, 31(2), 246-262.
Fiorello, C. A., Hale, J. B., & Snyder, L. E. (2006). Cognitive
hypothesis testing and response to intervention for children with
reading problems. Psychology in the Schools, 43, 835853.
Fletcher, J. M., Lyon, G. R., Barnes, M., Stuebing, K. K.,
Francis, D. J., Olson, R. K., . . . Shaywitz, B. A. (2002).
Classification of learning disabilities: An evidenced-based
evaluation. In R. Bradley, L. Danielson, & D. Hallahan (Eds.),
Identification of learning disabilities: Research to practice
(pp. 185239). Mahwah, NJ: Erlbaum.
Fletcher, J. M., Shaywitz, S. E., Shankweiler, D. P., Katz, L.,
Liberman, I. Y., Stuebing, K. K., . . . Shaywitz, B. A. (1994).
Cognitive profiles of reading disability: Comparisons of discrepancy and low achievement definitions. Journal of Educa
Floyd, R. G., Keith, T. Z., Taub, G. E., & McGrew, K. S. (2007).

Cattell -Horn -Carroll cognitive abilities and their effects on
reading decoding skills: g has indirect effects, more specific
abilities have direct effects.School Psychology Quarterly, 22,
(2) 200233.
Fuchs, D., & Young, C. L. (2006). On the irrelevance of intelligence in predicting responsiveness to reading instruction.
Exceptional Children, 73, 830.
Gardner, M. E. (1988). Test of Visual-Perceptual Skills. San
Francisco, CA: Health Publishing.
Gathercole, S. E. (2006). Nonword repetition and word learning:
The nature of the relationship. Applied Psycholinguistics, 27,
513543.
Gathercole, S. E., & Alloway, T. P. (2006). Practitioner Review:
Short-term and working memory impairments in neurodevelopmental disorders: diagnosis and remedial support. Journal
of Child Psychology and Psychiatry, 47, (1), 415.
Gathercole, S. E., Alloway, T. P., Willis, C., & Adams, A.-M.
(2005). Working memory in children with reading disabilities.
Journal of Experimental Child Psychology, 93, 265281.
Gathercole, S. E., & Baddeley, A. D. (1990). The role of phonological memory in vocabulary acquisition: A study of young
children learning new names. British Journal of Psychology,
81, 439454.
Glutting, J. J., Watkins, M. W., Konold, T. R., & McDermott, P. A.
(2006). Distinctions without a difference: The utility of obs
erved versus latent factors from the WISC-IV in estimating
reading and math achievement on the WIAT-II. Journal of
Special Education, 40, 103114.
Grogan, S. C. (1995). Which cognitive abilities at age four are
the best predictors of reading ability at age seven? Journal of
Research in Reading, 18, 2431.
Gustafson, S., & Samuelsson, S. (1999). Intelligence and dyslexia:
Implications for diagnosis and intervention. Scandinavian
Journal of Psychology, 40, 127134.
Hale, J. B., Fiorello, C. A., Kavanagh, J. A., Holdnack, J. A., &
Aloe, A. M. (2007). Is the demise of IQ interpretation justified? A response to special issue authors. Applied Neuropsy
chology, 14, 3751.
Hale, J. B., Kaufman, A., Naglieri, J. A., & Kavale, K. A. (2006).
Implementation of idea: Integrating response to intervention
and cognitive assessment methods. Psychology in the Schools,
43, 753770.
Hicks, C. (1980). The ITPA visual sequential memory task: An
alternative interpretation and the implications for good and
poor readers. British Journal of Educational Psychology, 50,
1625.
Hoskyn, M., & Swanson, H. L. (2000). Cognitive processing of
low achievers and children with reading disabilities: A selective meta-analytic review of the published literature. School
Psychology Review, 29, 102119.
Huba, M. E., Vellutino, F. R., & Scanlon, D. M. (1990). Auditory
and visual retention in poor and normal readers when verbal
164
encoding is disrupted. Learning and Individual Differences,

2, 95112.
Johnston, R. S., & Morrison, M. (2007). Toward a resolution of
inconsistencies in the phonological deficit theory of reading
disorders: Phonological reading difficulties are more severe
in high-IQ poor readers. Journal of Learning Disabilities, 40,
6679.
Kaufman, A. S., & Kaufman, A. L. (1990). Kaufman Brief Intel
ligence Test. Circle Pines, MN: American Guidance Service.
Kavale, K. A. (1982). Meta-analysis of the relationship between
visual perceptual skills and reading achievement. Journal of
Learning Disabilities, 15, 4251.
Kirby, J. R., Parrila, R. K., & Pfeiffer, S. L. (2003). Naming speed
and phonological awareness as predictors of reading development. Journal of Educational Psychology, 95, 453464.
Lervg, A., Brten, I., & Hulme, C. (2009). The cognitive and lin
guistic foundations of early reading development: A Norwegian
latent variable longitudinal study. Developmental Psychology,
45, 764781.
Liberman, I. Y., & Shankweiler, D. (1985). Phonology and the
problems of learning to read and write. Reading and Special
Education, 6(6), 817.
Lovett, M. W., Barron, R. W., & Benson, N. J. (2003). Effective
remediation of word identification and decoding difficulties in
school-age children with reading disabilities. In H. L. Swanson,
K. R. Harris, & S. Graham (Eds.), Handbook of learning dis
abilities (pp. 273292). New York, NY: Guilford.
Lovett, M. W., De Palma, M., Frijters, J. C., Steinbach, K.,
Temple, M., Benson, N., & Lacerenza, L. (2008). Interventions
for reading difficulties: A comparison of response to intervention by ELL and EFL struggling readers. Journal of Learning
Disabilities, 41, 333352.
Lovett, M. W., Lacerenza, L., & Borden, S. L. (2000). Putting
struggling readers on the PHAST track: A program to integrate
phonological and strategy-based remedial reading instruction
and maximize outcomes. Journal of Learning Disabilities, 33,
458476.
Lovett, M. W., Lacerenza, L., Borden, S. L., Frijters, J. C.,
Steinbach, K. A., & De Palma, M. (2000). Components of
effective remediation for developmental reading disabilities:
Combining phonological and strategy-based instruction to
improve outcomes. Journal of Educational Psychology, 92,
263283.
Lovett, M. W., Lacerenza, L., Murphy, D., Steinbach, K. A., De
Palma, M., & Frijters, J. C. (2005). The importance of
multiple-component interventions for children and adolescents who are struggling readers. In J. Gilger & S. Richardson
(Eds.), Research-based education and intervention: What we
need to know (pp. 67-102). Baltimore, MD: International Dyslexia Association.
Manis, F. R., Seidenberg, M. S., & Doi, L. M. (1999). See Dick
RAN: Rapid naming and the longitudinal prediction of reading subskills in first and second graders. Scientific Studies of
Reading, 3, 129157.
Mano, Q. R., & Osmon, D. C. (2008). Visuoperceptualorthographic reading abilities: A confirmatory factor analysis
study. Journal of Clinical and Experimental Neuropsychology,
30, 421434.
Metsala, J. L., Stavrinos, D., & Walley, A. C. (2009). Childrens
spoken word recognition and contributions to phonological
awareness and nonword repetition: A 1-year follow-up. App
lied Psycholinguistics, 30, 101121.
Morris, R. D., Lovett, M. W., Wolf, M., Sevcik, R. A.,
Steinbach, K. A., Frijters, J. C., & Shapiro, M. (in press).
Multiple-component remediation for developmental reading
disabilities: IQ, SES, and race as factors on remedial outcome.
Journal of Learning Disabilities. Advance online publication.
doi: 10.1177/0022219409355472.
Nelson, J. R., Benner, G. J., & Gonzalez, J. (2003). Learner characteristics that influence the treatment effectiveness of early
literacy interventions: A meta-analytic review. Learning Dis
abilities Research & Practice, 18, 255267.
Oh, H.-J., Glutting, J. J., Watkins, M. W., Youngstrom, E. A., &
McDermott, P. A. (2004). Correct interpretation of latent versus observed abilities: Implications from structural equation
modeling applied to the WISC-III and WIAT linking sample.
Journal of Special Education, 38, 159173.
OMalley, K. J., Francis, D. J., Foorman, B. R., Fletcher, J. M., &
Swank, P. R. (2002). Growth in precursor and reading-related
skills: Do low-achieving and IQ-discrepant readers develop
differently? Learning Disabilities Research & Practice, 17,
1934.
Reiter, A., Tucha, O., & Lange, K. W. (2005). Executive functions
in children with dyslexia. Dyslexia: An International Journal
of Research and Practice, 11, 116131.
Richards, T. L., & Berninger, V. (2008). Abnormal fMRI connectivity in children with dyslexia during a phoneme task:
Before but not after treatment. Journal of Neurolinguistics,
21, 294304.
Richards, T. L., Berninger, V., Winn, W., Stock, P., Wagner, R.,
Muse, A., & Maravilla, K. (2007). Functional MRI activation
in children with and without dyslexia during pseudoword aural
repeat and visual decode: Before and after treatment. Neuro
psychology, 21, 732741.
Sattler, J. M. (1992). Assessment of children (3rd ed.). San Diego,
CA: Jerome M. Sattler.
Scanlon, D. M., & Vellutino, F. R. (1997). A comparison of the
instructional backgrounds and cognitive profiles of poor, average, and good readers who were initially identified as at risk
for reading failure. Scientific Studies of Reading, 1, 191215.
Scarborough, H. S. (1998). Predicting the future achievement
of second graders with reading disabilities: Contributions of
phonemic awareness, verbal memory, rapid naming, and IQ.
Annals of Dyslexia, 48, 115136.
Schatschneider, C., Fletcher, J. M., Francis, D. J., Carlson, C. D.,
& Foorman, B. R. (2004). Kindergarten prediction of reading
skills: A longitudinal comparative analysis. Journal of Educa
165
Frijters et al.
Scheiman, M., & Rouse, M. W. (2005). Optometric management
of learning-related vision problems. New York, NY: C. V.
Mosby.
Scruggs, T. E., & Mastropieri, M. A. (2003). Issues in the identification of learning disabilities. In T. E. Scruggs & M. A. Mastropieri
(Eds.), Identification and assessment: Advances in learning and
behavior disabilities (pp. 136). Oxford, UK: Elsevier.
Siegel, L. S. (1989). IQ is irrelevant to the definition of learning
disabilities. Journal of Learning Disabilities, 22, 469479.
Siegel, L. S. (1992). An evaluation of the discrepancy definition of
dyslexia. Journal of Learning Disabilities, 25, 618629.
Simos, P. G., Fletcher, J. M., Sarkari, S., Billingsley-Marshall, R.,
Denton, C. A., & Papanicolaou, A. C. (2007a). Altering the
brain circuits for reading through intervention: A magnetic
source imaging study. Neuropsychology, 21, 485496.
Simos, P. G., Fletcher, J. M., Sarkari, S., Billingsley-Marshall, R.,
Denton, C. A., & Papanicolaou, A. C. (2007b). Intensive ins
truction affects brain magnetic activity associated with oral
word reading in children with persistent reading disabilities.
Journal of Learning Disabilities, 40, 3748.
Stanovich, K. E., & Siegel, L. S. (1994). Phenotypic performance
profile of children with reading disabilities: A regression-based
test of the phonological-core variable-difference model. Jour
nal of Educational Psychology, 86, 2453.
Stuebing, K. K., Fletcher, J. M., LeDoux, J. M., Lyon, G. R.,
Shaywitz, S. E., & Shaywitz, B. A. (2002). Validity of IQdiscrepancy classifications of reading disabilities: A meta-analysis.
American Educational Research Journal, 39, 469518.
Swanson, H. L., & Jerman, O. (2007). The influence of working memory on reading growth in subgroups of children with
reading disabilities. Journal of Experimental Child Psychol
ogy, 96, 249283.
Swanson, H. L., Trainin, G., Necoechea, D. M., & Hammill, D. D.
(2003). Rapid naming, phonological awareness, and reading:
A meta-analysis of the correlational evidence. Review of Edu
cational Research, 73, 407444.
Swanson, H. L., Zheng, X., & Jerman, O. (2009). Working memory, short-term memory, and reading disabilities: A selective
meta-analysis of the literature. Journal of Learning Disabili
ties, 42, 260287.
Tiu, R. D., Jr., Thompson, L. A., & Lewis, B. A. (2003). The role
of IQ in a component model of reading. Journal of Learning
Disabilities, 36, 424436.
Torgesen, J. K. (2005). Recent discoveries on remedial interventions for children with dyslexia. In M. J. Snowling & C. Hulme
(Eds.), The science of reading: A handbook (pp. 521537).
Oxford, UK: Blackwell.
Torgesen, J. K., Alexander, A. W., Wagner, R. K., Rashotte,
C. A., Voeller, K. K. S., & Conway, T. (2001). Intensive
remedial instruction for children with severe reading disabilities: Immediate and long-term outcomes from two
instructional approaches. Journal of Learning Disabilities,
34, 3358.
Torgesen, J. K., & Davis, C. (1996). Individual difference variables

that predict response to training in phonological awareness.
Journal of Experimental Child Psychology, 63, 121.
Torgesen, J. K., & Wagner, R. K. (1996). The comprehensive test
of reading related phonological processes. Tallahassee, FL:
Author.
Vellutino, F. R. (2001). Further analysis of the relationship
between reading achievement and intelligence: A response to
Naglieri. Journal of Learning Disabilities, 34, 306310.
Vellutino, F. R., Scanlon, D. M., & Lyon, G. R. (2000). Differentiating between difficult-to-remediate and readily remediated poor readers. Journal of Learning Disabilities, 33,
223238.
Vellutino, F. R., Scanlon, D. M., Sipay, E. R., Small, S. G.,
Pratt, A., Chen, R., et al. (1996). Cognitive profiles of difficult-to-remediate and readily remediated poor readers:
Early intervention as a vehicle for distinguishing between
cognitive and experiential deficits as basic causes of specific reading disability. Journal of Educational Psychology,
88, (4)601638.
Vellutino, F. R., Tunmer, W. E., Jaccard, J. J., & Chen, R. (2007).
Components of reading ability: Multivariate evidence for a
convergent skills model of reading development. Scientific
Studies of Reading, 11, 332.
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, 192212.
Wagner, R., Torgesen, J. K., & Rashotte, C. A. (1999). Comprehen
sive Tests of Phonological Processing (CTOPP). Austin, TX:
Pro-Ed.
Watson, C., & Willows, D. M. (1995). Information processing patterns in specific reading disability. Journal of Learning Dis
abilities, 28, 216231.
Wechsler, D. (1991). Wechsler Intelligence Scale for Children
Third Edition (WISC-III). New York, NY: Psychological
Corporation.
Wilkinson, G. S. (1993). The Wide Range Achievement Test 3. San
Antonio, TX: Harcourt Assessment.
Wimmer, H., Mayringer, H., & Landerl, K. (1998). Poor reading:
A deficit in skill-automatization or a phonological deficit? Sci
entific Studies of Reading, 2, 321340.
Wolf, M., Barzillai, M., Gottwald, S., Miller, L., Spencer, K., Norton, E., Lovett, M. W., & Morris, R. D. (2009). The RAVEO intervention: Connecting neuroscience to the classroom.
Mind, Brain, and Education, 3, 8493.
Wolf, M., Bowers, P. G., & Biddle, K. (2000). Naming speed processes, timing and reading: A conceptual review. Journal of
Learning Disabilities, 33, 387407.
Wolf, M., Goldberg-ORourke, A., Gidney, C., Lovett, M.,
Cirino, P., & Morris, R. (2002). The second deficit: An investigation of the independence of phonological and naming-speed
deficits in developmental dyslexia. Reading and Writing: An
Interdisciplinary Journal, 15, 4372.
166
Wolf, M., Miller, L., & Donnelly, K. (2000). Retrieval, Automaticity, Vocabulary Elaboration, Orthography (RAVE-O): A
comprehensive, fluency-based reading intervention program.
Journal of Learning Disabilities, 33, 375386.
Woodcock, R. W. (1987). Woodcock Reading Mastery Tests
Revised (WRMT-R). Circle Pines, MN: American Guidance
Service.
About the Authors

Jan C. Frijtersis an Associate Professor at Brock University,
Departments of Child and Youth Studies and Psychology. His
primary research areas are modeling treatment outcomes, the role
of motivation in reading disability and reading interventions, and
the development of motivation for reading throughout childhood.
Maureen W. Lovett is a Senior Scientist in the Neurosciences
and Mental Health Program at The Hospital for Sick Children, a
Professor of Pediatrics, Medical Sciences, and Human
Development and Applied Psychology at the University of
Toronto. She is Founder and Director of the hospitals Learning
Disabilities Research Program, a clinical research unit that develops and evaluates interventions for developmental reading disabilities. Her research focuses on questions about the effective
remediation of decoding, word identification, fluency, and reading
comprehension deficits in struggling readers of elementary, middle, and high school age.
Karen A. Steinbach is Clinical Research Project Manager of the
Learning Disabilities Research Program at The Hospital for Sick
Children. Ms. Steinbach has overseen the implementation of several large multi-site grant funded studies. She supervises and trains
psychology staff, and coordinates and directs all assessment and
programming offered in our elementary-level research
classrooms.
Maryanne Wolf is the John DiBiaggio Professor of Citizenship
and Public Service, Director of the Center for Reading and
Language Research, and Professor in the Eliot-Pearson
Department of Child Development at Tufts University. Her recent
research interests include reading intervention, early prediction,
fluency and naming speed, cross-linguistic studies of reading, the
relationship between entrepreneurial talents and dyslexia, and the
uses of brain imaging in understanding dyslexia and treatment
changes.
Rose A. Sevcik is Professor of Psychology at Georgia State
University, Atlanta. Her research has focused on the language,
communication, and reading acquisition of children and youth
with developmental disabilities.
Robin D. Morris is the Vice President for Research and
Regents Professor of Psychology at Georgia State University.
He also holds a joint appointment in the Department of
Educational Psychology and Special Education in the College
of Education. His current research is focused on reading and
language development, reading disabilities and dyslexia, bilingual language and reading development, and neuroimaging of
the developing brain.

Frijters Et Al (2011) Neurocognitive Predictors of Dyslexia

Uploaded by

Copyright:

Available Formats

Frijters Et Al (2011) Neurocognitive Predictors of Dyslexia

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Frijters Et Al (2011) Neurocognitive Predictors of Dyslexia

Uploaded by

Copyright:

Available Formats

391185

Journal of Learning Disabilities

Jan C. Frijters1, Maureen W. Lovett2, Karen A. Steinbach2,

fluency-dependent tasks (Kirby, Parrila, & Pfeiffer, 2003;

Brock University, St. Catharines, Ontario, Canada

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

explaining responsiveness systematically depends on the

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

memory was retained as significantly impaired in the group

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

1. What is the contribution of intervention condition,

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

Table 1. Average Preintervention Characteristics of the

Age at intervention start (months)

Remedial Intervention Programs

Note: WRMT-R = Woodcock Reading Mastery TestsRevised;

Participants with similar single word reading levels (WRMT-R

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Internal consistency reliability derived from the normative

Neurocognitive Process Predictors

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

Table 2. Growth Curve Models of Outcome Variables

In rate of 21 1.91** 0.38 10.23* 4.82 11.80**

3.61 622.51** 55.79 249.40** 24.95 83.97**

processes that does not require a verbal response. Children

relative contribution of individual predictors to classifying

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

DFA Base Model

(6.5%). Exploratory analyses comparing participants with

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

Model F(10, 542)

all outcomes, Intervention Contrast 1 was consistently an

DFA Cognitive and Neuropsychological

be enhanced by the inclusion of this class of predictors and

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Model F(26, 526)

predicting better response across all response categories.

Follow-Up Analysis of Change in

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

Table 6. Comparison of InterventionPhonology and CognitiveNeuropsychological Additive Model

Misclassifications unique to base

Base model reclassification error rates

Additive model reclassification error rates

Improvement in reclassification error rates

the sum of these two is distributed as a 2 with one degree

relative size of reading outcomes in the context of intensive

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Grade 1 to Grade 3. As such, the results highlighting the

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Journal of Learning Disabilities 44(2)

that notably excluded orthographic processing (e.g., Mano &

Implications for Intervention Architectures

conceptualized more broadly than phonological analysis

Downloaded from ldx.sagepub.com at UNIV OF OTTAWA LIBRARY on July 13, 2011

Floyd, R. G., Keith, T. Z., Taub, G. E., & McGrew, K. S. (2007).

3.61 622.51 55.79 249.40 24.95 83.97**