Professional Documents
Culture Documents
Abstract
A synthesis and meta-analysis of the extant research on the effects of storybook read-aloud interventions for children
at risk for reading difficulties ages 3 to 8 is provided. A total of 29 studies met criteria for the synthesis, with 18 studies
providing sufficient data for inclusion in the meta-analysis. Read-aloud instruction has been examined using dialogic reading;
repeated reading of stories; story reading with limited questioning before, during, and/or after reading; computer-assisted
story reading; and story reading with extended vocabulary activities. Significant, positive effects on childrens language,
phonological awareness, print concepts, comprehension, and vocabulary outcomes were found. Despite the positive effects
for read-aloud interventions, only a small amount of outcome variance was accounted for by intervention type.
Keywords
read aloud, story book, dialogic reading, early childhood, reading
Reading aloud to young children has been a commonplace 2004, 2006) have led to an increased focus on the early iden-
practice in homes and schools for years. Parents, educators, tification and prevention of reading difficulties in the early
policy makers, and politicians have promoted read alouds grades, including preschool (Catts & Hogan, 2003; Lyon, 1998).
with the common belief that reading to children makes a The focus on early intervention for children at risk for reading
difference in childrens literacy development. Indeed, adults difficulties has led early childhood and elementary educators
reading to children under varying conditions may have social- and researchers to examine early literacy instructional practices,
emotional benefits for the children (Bus, 2001) as well as including the commonplace practice of storybook read alouds
positive impacts on childrens language and literacy develop- (National Early Literacy Panel, 2009). A corpus of research
ment (Lonigan, 1994; National Research Council, 1998; van investigating the effects of read-aloud practices in preschool
Kleeck, 2004; Wasik & Hendrickson, 2004) and motivation and the early elementary grades has developed over the past
to read (Gambrell & Marinak, 2009).
Many children learn to read effortlessly; however, an esti- 1
The Meadows Center for Preventing Educational Risk, Austin, TX, USA
mated 10% of children have difficulty acquiring reading skills 2
The University of Texas at Austin, TX, USA
and need additional support or specialized instruction (Catts 3
Florida State University, Tallahassee, FL, USA
4
& Hogan, 2003). Conditions that place children at risk for Kennesaw State University, Kennesaw, GA, USA
5
Educational Consultant
reading difficulties include poverty, cultural and linguistic 6
American Institutes for Research, Washington, DC, USA
differences, neurologically based problems, inadequate
instruction, limited development-enhancing opportunities, Corresponding Author:
or familial history of reading disabilities (McCoach, Elizabeth Swanson
OConnell, Reis, & Levitt, 2006; Neuman & Dickinson, The Meadows Center for Preventing Educational Risk
The Unviersity of Texas at Austin
2001). Knowledge of these early risk factors as well as the 1 University Station D4900
negative impact of reading difficulties on future success in Austin, TX 78712
school and life (National Center for Educational Statistics, Email: easwanson@mail.utexas.edu
several decades. Previous meta-analyses (e.g., Blok, 1999; Bus, children ranging in age from 31 to 90 months. The overall
van IJzendoorn, & Pellegrini, 1995; Mol, Bus, & de Jong, 2009; effects of a variety of read-aloud interventions were medium
National Early Literacy Panel, 2009) and syntheses (Karweit in size for oral language (effect size [ES] = 0.63) and read-
& Wasik, 1996; Scarborough & Dobrich, 1994) have provided ing (ES = 0.41) outcomes. Blok (1999) identified study
important information to inform the use of read alouds with characteristics that acted as moderators for oral language and
young children, including (a) support for reading aloud as an wide reading outcomes, citing that instruction with younger
effective practice for improving child outcomes, (b) initial work children and smaller instructional group sizes were associated
to identify study features associated with stronger effects, (c) with larger effects. In addition, the use of untrained adult
some support of read alouds as a practice for children at risk for readers was associated with greater effects in wide reading.
reading difficulties, and (d) identification of methodological It is unclear though whether the provision of training for the
concern for further study. readers was a constant variable across studies in which this
Scarborough and Dobrich (1994) synthesized 31 studies was reported. An important, general conclusion of the meta-
conducted over a period of 30 years that explored the influence analysis was that although some studies revealed positive
of parentchild read-aloud experiences on language and lit- effects, the authors also reported some poor-quality research
eracy development. Effect sizes for some studies were reported and highlighted the need for additional empirical work on
and were moderately positive. Experimental groups outper- the use of reading aloud in educational settings as a means
formed control groups in interventions that involved parent to improve literacy outcomes.
training for shared book reading strategies. However, the An increase in high-quality research on story read alouds
strength of the association was inconclusive based on vari- has since occurred. A recent meta-analysis of early literacy
ability of results and methodology across the studies reviewed. practices for preschool and kindergarten children examined
While the extensively cited Scarborough and Dobrich experimental and quasi-experimental literacy literature
(1994) review was replete with evidence to support reading through 2003 (National Early Literacy Panel, 2009). The
to children, the authors also drew attention to the potential panel reported that shared reading interventions demonstrated
overconfidence that educators, parents, and researchers place moderate effects on childrens print knowledge and oral lan-
on the contributions of parentchild reading to the promotion guage skills. Shared reading was examined across settings
of literacy development. They cited an ongoing debate about (e.g., schools, home, pediatricians provided books to parents,
the bidirectional effect between child interest in reading and etc.) and across adult readers (parents, teachers). While the
being read to. In other words, does reading to young children overall effect size for children who were not at risk seemed
promote interest and skill development, or do inherent interest to be higher than for children who were at risk, the difference
and skill elicit more reading, thus impacting further skill in effects was not statistically significant. Information on
development? Scarborough and Dobrich concluded that read- features of instruction and the effects on specific literacy
ing aloud to children is a complex practice, worthy of addi- outcomes for the at-risk population were not analyzed.
tional research, and thus, subsequent studies and reviews on Our synthesis and meta-analysis differs from previous
the practice have since flourished. reviews in several ways. First, we focus on only teacher-
Karweit and Wasik (1996) complemented Scarborough delivered interventions, excluding parent interventions (e.g.,
and Dobrichs (1994) work by examining studies that focused Bus et al., 1995; National Early Literacy Panel, 2009). Second,
on the effects of storybook reading on 4- and 5-year-olds in we limit our review to studies that focus on students at risk
school settings. In their narrative review of studies that for reading difficulty. Third, because previous meta-analytic
included children from disadvantaged settings conducted results indicate no differences in effect size due to age differ-
between 1988 and 1994, the authors identified the following ences (National Early Literacy Panel, 2009), we have included
practices associated with positive outcomes: small group size preschool through third graders. Finally, previous syntheses
in favor of whole-group or individual story reading, repeated have focused on limited interventions (e.g., Bus et al., 1995,
readings to facilitate child involvement and comprehension, focused on frequency of book reading) or limited outcomes
vocabulary instruction using synonyms or role-playing, and (e.g., Mol et al., 2009, focused on vocabulary and print knowl-
teacher interaction and questioning strategies to support edge). Our synthesis considers all early reading and language
vocabulary learning and comprehension. Although the authors outcomes. We address the following questions:
identified disadvantaged children as a focus, the data were
not disaggregated to describe differential effects for children Research Question 1: What outcomes result from read-
with varying abilities or risk status. aloud practices in educational settings for children
Blok (1999) conducted a meta-analysis of 10 studies, considered at risk for reading difficulty?
published between 1968 and 1994, in which read alouds took Research Question 2: What features of read-aloud
place in educational settings. The studies included aggregated instruction are associated with improved outcomes
data for disadvantaged (not defined) and nondisadvantaged for children at risk for reading difficulty?
attrition between intervention and comparison group, sample or control groups and administered more than one measure
size of the intervention group at the beginning and end of the to assess performance on various outcomes. As such, to
study) and nine items for quasi-experimental designs (e.g., account for treatment and control differences specific to the
procedures used to equate groups, evidence of differential different types of outcomes, individual meta-analyses were
attrition between groups). Additional items allowed coders run for language, phonological awareness, word recognition,
to describe assessment measures. Finally, precision of out- print concepts, vocabulary, and reading comprehension. Esti-
comes for both effect size estimation and statistical reporting mation of the mean effect size, variability across studies, and
was coded using a series of eight forced choice or yes/no moderator effects were assessed using a multilevel model
questions (e.g., evidence of assumptions of independence, approach using HLM 6.0 software (Raudenbush & Bryk,
normality, and equal variance). Effect sizes were calculated 2002). Multilevel models are aptly applied to meta-analytic
using reported performance on outcome measures and sam- data for a number of reasons: (a) Multilevel modeling can
ple sizes for each treatment and comparison condition. account for subjects nested within studies by calculating the
Six researchers participated in a series of training sessions variance at both the subject and study level; (b) multilevel
(approximately 8 hours total) on interpreting the code sheet. modeling provides flexibility in mixed model estimation; and
Following training, researchers coded one article to determine (c) the researcher can simultaneously estimate the average
interrater agreement (agreements divided by agreements plus effect size (fixed effect), variance of the effect (random effect),
disagreements). Interrater reliability of .95 for study charac- and residual variance modeled by moderator variables.
teristic coding, and .91 for effect size coding was achieved. Multilevel modeling has the benefit of estimating results
Teams of two researchers coded each article, resolved any based on several methods, including restricted maximum
differences in coding, and reached final decisions by consen- likelihood (REML; Raudenbush & Bryk, 1986). Tradition-
sus. After coding was completed, information was compiled ally, ordinary least squares or weighted least squares are
in table format. Table 1 includes information on study design, used in a regression analysis, with both methods designed
number and age of participants, type and duration of inter- to reduce the squared differences between predicted and
vention, and person who implemented the intervention. In observed values of the dependent variables. In the context
Table 2, interventions are described and effect sizes are of a multilevel design however, both methods are less effi-
reported when appropriate data were available. cient and require more participants to achieve results com-
Article categorization. A team of two researchers indepen- parable to those estimated by REML (Kreft & de Leeuw,
dently categorized articles by intervention type. A total of 1998). Although multilevel analyses are typically performed
six read-aloud intervention types were identified. Descriptors with the presumption that the variances are previously
for each intervention type were developed (see Table 3) and unknown, the meta-analysis must be performed under the
used by a third researcher to verify the categorization. known variance models. In these models, the variances must
Effect size calculation. The primary statistic used to analyze be calculated and entered into the multivariate data matrix
the relationship between treatment and outcomes across the along with the study ID, effect size, and moderator variables.
outcome measures was Cohens d (Cohen, 1988). This statistic The estimation of multilevel meta-analytic models is per-
indicates, in standard deviation units, the extent to which the formed in a two-stage process. Initially, the unconditional
treatment groups outperformed the control groups. For studies model is run to estimate the grand mean effect size and the
in this synthesis that employed a treatment-comparison design, variance of the true effects. The Level-1 model (within-
effect sizes can be interpreted as d = 0.2 as small, d = 0.5 as studies) is summarized as
medium, and d = 0.8 as a large effect (Cohen, 1988). The
authors calculated Cohens d by taking the difference of the dj = j + ej, (1)
treatment and control group means and dividing by the pooled
standard deviation. In instances where F test statistics were where dj = the effect size from a study; j = the true effect
provided instead of means and standard deviations, the effect size of dj across all studies, or the corresponding population
size was calculated using the F statistic and degrees of free- parameter; and ej = the sampling error associated with dj as
dom. Nine studies did not report sufficient data to allow effect the estimate of j.
size calculation. Moreover, it is known that effect sizes from Should statistically significant variability be observed, the
small studies tend to be biased; thus, a sample weighted esti- analyst may perform a conditional analysis, which includes
mate of the standardized difference was computed. the Level-2 moderator variables. The purpose of this second
Meta-analysis plan. We conducted a meta-analysis of a sub- model is to generate an effect size for prediction and estimate
set of 18 studies that employed a treatment-comparison design remaining residual variance after controlling for the entered
and included information about outcomes and sample sizes Level-2 moderators, summarized as
that allowed for meta-analytic analysis (studies are denoted
in tables). Many of the studies included multiple treatment j = 0 + 1W1j +...sjWsj + uj, (2)
where 0 = grand mean effect size, s = regression coefficients, whether the unconditional or conditional model is selected
W1j,...,Wsj = moderator variables predicting j, and uj = is based on the estimation of (i.e., observed score variance),
Level-2 random error. whereby a value of zero indicates that no variance in study
When combining the two equations, two sources of error effect sizes remains after partitioning sampling errors, thus
are contained; one for the unconditional model (uj) and one the Level-2 model would not be needed. Should be signifi-
for the conditional Level-2 model (ej). Determination of cantly different from zero, the conditional model would be
Table 2. (continued)
Design/Intervention Measures Results/Findings
Coyne, Simmons, Dameenui, and Vocabulary: Taught words T > C; ES = 0.74
Stoolmiller(2004)b Vocabulary: Untaught words T > C; ES = 0.20
T: Storybook
C: Control
Hargrave and Snchal (2000)b PPVT-Ra T < C; ES = -0.12
T: Dialogic reading EOWPVT-Ra T > C; ES = 0.17
C:Typical instruction comparison Book vocabulary T > C; ES = 0.07
Justice and Ezell (2002) Print concepts T > C; ES = 0.41
T: Print-focused Print recognition T > C; ES = 3.28
C: Picture-focused Words in print T > C; ES = 1.45
letter orientation, and T > C; ES = 0.72
discrimination T > C; ES = 0.48
Alphabet knowledge T > C; ES = 0.67
Literacy terms
Justice, Meier, and Walpole (2005)b Vocabulary: Elaborated words T > C; ES = 1.59
T: Word elaboration Vocabulary: Nonelaborated words T > C; ES = 2.06
C: Comparison
Korat and Shamir (2007)b Vocabulary T1 > C; ES = 1.00
T1: E-book T2 > C; ES = 0.89
T2: Adult reader Word recognition T1 > C; ES = 0.10
C: Typical practice comparison T2 < C; ES = -0.10
Phonological awareness T1 > C; ES = 0.18
T2 < C; ES = -0.09
Lonigan, Anthony, Bloomfield, Dyer, and WJ-LCa T1 < C; ES = -0.23
Samwel (1999)b PA: Alliteration oddity T2 < C; ES = -0.07
T1: Shared reading PA: Blending T1 < T2; ES = -0.15
T2: Dialogic reading PA: Elision T1 > C; ES = 0.98
C: Typical curriculum comparison PA: Rhyme T2 > C; ES = 1.43
EOWPVT-Ra T1 < T2; ES = -0.32
ITPA-VEa T1 > C; ES = 0.03
PPVT-Ra T2 > C; ES = 0.04
T1 < T2; ES = -0.01
T1 < C; ES = -0.30
T2 > C; ES = 0.11
T1 < T2; ES = -0.37
T1 < C; ES = -0.04
T2 > C; ES = 0.56
T1 < T2; ES = -0.40
T1 > C; ES = 0.07
T2 < C; ES = -0.03
T1 > T2; ES = 0.11
T1 < C; ES = -0.01
T2 < C;ES = -0.01
T1 = T2; ES = 0.00
T1 > C; ES = 0.04
T2 < C; ES = -0.49
T1 > T2; ES = 0.49
Lonigan and Whitehurst (1998)b Semantic complexity: T1 > C1; ES = 0.96
T1: High compliance dialogic reading Number of different wordsa T2 > C2; ES = 0.97
T2: Low compliance dialogic reading Semantic complexity: T1 > C1; ES = 0.18
C1: Control at high compliance center Number of modifiersa T2 > C2; ES = 1.04
C2: Control at low compliance center Semantic complexity: T1 < C1; ES = -0.10
Number of nouns T2 > C2; ES = 1.38
Semantic complexity: T1 > C1; ES = 0.73
Number of verbs T2 > C2; ES = 0.13
Language production: T1 > C1; ES = 1.17
Number of words T2 > C2; ES = 1.31
(continued)
Table 2. (continued)
Table 2. (continued)
employed to test if the residual variance could be explained moderators in the analyses. In this equation qq (UC) repre-
by the selected moderator variables. sents the variance component for the unconditional model,
Furthermore, at both the unconditional and conditional and qq (C) is the variance component for the conditional
levels, empirical Bayes (EB) estimates may be derived to model (Raudenbush & Bryk, 2002).
compare with the maximum likelihood estimates. This pro- Prior to meta-analysis, estimation of the file-drawer prob-
cess allows the researcher to supply the grand mean effect lem was conducted to determine the extent to which unpub-
size and create a normal distribution of effect sizes, shrinking lished, unsuccessful studies would need to be incorporated
the extreme estimates closer toward the mean. An advantage to make results nonsignificant. Since most published studies
of this procedure is that the analyst may observe the degree report statistically significant findings, it is supposed that
of unreliability that exists within a particular study effect greater nonpublished research exists for the topic. Results
size. Whereas the REML approach considers all values as indicated that the file drawer problem did not pose a threat to
likely, the EB estimates generate a normal distribution around the current study (1.72 < 1.96).
the grand mean for the individual cases, with those showing
the most shrinkage as evidence that the units have little data,
or may considered to be outliers (Raudenbush & Bryk, 2002). Results
Moreover, from both models, the variance components for A total of 27 articles, comprising 29 studies (Beck & McKeown,
the grand means can be used to estimate the proportion of 2007, and Coyne et al., 2007, each contain two studies) are
reduction in variance accounted when using the selected included in this synthesis and reflect a variety of study designs.
Table 4. Quality of Treatment-Comparison Studies includes a description of each studys treatment and com-
parison conditions.
Element Number of Studies
In all, 17 studies reported number of intervention sessions
Random assignment to conditions 11 with a range from 3 to 155, and a mean of 30 sessions. The
Fidelity of treatment reported 12 number of sessions was not reported in 12 studies. In addition,
Standardized dependent measures 14 7 studies reported the length of sessions, with a range of 6 to
Random assignment, treatment 2
40 minutes and a mean of 24 minutes. Among treatment-
fidelity, and standardized measures
comparison studies, the number of intervention sessions aver-
aged 29, while the length of sessions averaged 20 minutes.
Available information about frequency and length of sessions
Therefore, we conducted several types of analyses to fully is reported in Table 1.
explain the results of these studies. First, we analyzed study
features (e.g., study design and use of random assignment).
Second, we conducted a meta-analysis of all treatment- Meta-Analysis
comparison design studies to determine the effect of read- In this section, we report results of the meta-analysis by
aloud interventions on several early reading outcomes, along outcome (see Table 6) coupled with report of our investiga-
with a follow-up moderator analysis to examine differences tion of moderating effects of intervention type on each out-
between studies using criterion measures (hard criteria; i.e., come (see Table 7).
below average achievement) or background characteristics
(soft criteria; i.e., low socioeconomic status [SES], low
school quality, history of family risk) and differences using Language Outcomes
standardized or researcher-developed measures. Finally, we The range of the 23 weighted effect sizes for language outcomes
synthesized all single-group, single-subject, multiple treat- was -0.48 to 1.79, with an overall weighted mean from the
ment, and three treatment-comparison studies not included unconditional model of 0.29, t(23) = 3.16, p = .005. This sug-
in the meta-analysis by outcome and intervention type. gests that children who received read-aloud interventions sig-
nificantly outperformed children in the comparison group on
measures of language. The variance component from the mul-
Study Features tilevel model of 0.003 was not statistically significant (p .500;
In the following section, we summarize information about Table 6), indicating that while intervention children outper-
study characteristics and design elements. Detailed informa- formed comparison children, there was not significant variability
tion about each study is included in Tables 1 and 2. among studies. Although the range suggested larger deflections
Sample characteristics. Sample sizes ranged from 1 to from the mean in terms of magnitude, the amount of error in
254 children, with a mean of 56.6 and a median of 58 partici- effect sizes precluded a rejection of the null hypothesis.
pants. The majority of studies targeted preschool through kin-
dergarten children (n = 22). However, three studies included
first graders and two studies included second or third graders. Phonological Awareness Outcomes
Study design. This corpus of studies included 21 treatment- A large grand mean difference was observed (d = 0.78) for the
comparison, 5 multiple-treatment, 2 single-group, and 1 single- phonological awareness effect sizes, t(21) = 5.23, p < .001. The
subject study. Several design elements strengthen the reliability range of effect sizes was -0.22 to 3.01, and the error around the
and lend credibility to findings from treatment-comparison estimates was relatively small. This resulted in a significant vari-
studies. These include use of random assignment, fidelity of ance component (0.42, p < .001), suggesting that studies varied
treatment procedures, and the use of standardized dependent considerably in the magnitude of difference in performance on
measures (Raudenbush, 2005; Shadish, 2002; U.S. Department phonological awareness outcomes between treatment and control
of Education, 2003). The number of treatment-comparison children. The second stage of analysis included the modeling of
studies that possesses these design elements is reported in variance with selected moderators, namely, types of interventions
Table 4. Of note, 2 treatment-comparison studies (Lonigan, used. Within the conditional model analysis, four types of treat-
Anthony, Bloomfield, Dyer, & Samwel, 1999; Whitehurst ment were identified, including dialogic reading, computer
et al., 1994) reported all three elements. assisted, limited questioning, and repeated reading. Across the
Interventions. Studies were placed into one of six categories 14 effect sizes, dialogic reading was the most frequently observed
(computer assisted, dialogic reading, limited questioning, intervention and was set as the referent condition. The mean
repeated reading, vocabulary, and other) based on the authors effect size from the conditional model was 0.84, indicating that
description of the intervention. The distribution of intervention children who were exposed to dialogic reading outperformed
type by design is reported in Table 5. In addition, Table 2 control children. Computer-assisted interventions were weaker
Study Design
Intervention Type Treatment-Comparison Multiple Treatment Single Group Single Subject Marginal Totals
Dialogic 10 2 1 0 13
Extended vocabulary 3 1 1 0 5
Repeated reading 1 1 0 0 2
Limited questioning 2 0 0 0 2
Computer assisted 2 0 0 0 2
Other 2 1 0 1 4
Marginal totals 20 5 2 1 28a
a
Korat and Shamir (2007) included a computer-assisted treatment and a limited questioning treatment in their multiple treatment design study and are
not included in the study totals.
than dialogic reading by -0.39 units, with a fitted mean effect size of -0.23. Neither computer-assisted nor limited questioning
size of 0.45. Limited questioning interventions were also weaker mean effect sizes were statistically different from dialogic read-
than dialogic reading by -1.07, producing a fitted mean effect ing, likely due to the relatively few number of effect sizes
represented by the two conditions. Conversely, children in the this outcome in the unconditional model was 0.70 and was
repeated reading condition demonstrated a significantly larger statistically different from 0, t(20) = 4.94, p < .001. Significant
mean effect size (i.e., 2.59) compared to dialogic reading, t(19) variability existed in the mean and was modeled by dialogic
= 2.27, p < .05. A follow-up moderator analysis tested the extent reading, computer-assisted interventions, and a group of
to which differences in the definition of risk criteria used in the interventions labeled here as other. The mean effect size for
study related to residual unexplained variance in the effect sizes. dialogic reading was 0.60, t(18) = 3.81, p < .01. Computer-
Testing differences between studies using standardized or assisted interventions showed stronger differentiation between
researcher-developed measures was not warranted as all studies treatment and control groups, with a fitted mean of 1.27. This
using phonological awareness used researcher-developed assess- value was more than 0.50 standard deviation units larger than
ments. Results from this test indicated that no significant differ- the effect size for dialogic reading and was significantly
ences were observed between studies using soft or hard criteria different. Conversely, the group of effect sizes other dem-
for risk (p = .64). onstrated a small negative deflection from dialogic reading
By modeling treatment differences, the amount of variance (-0.18); however, this was not considered to be a significant
was reduced from 0.42 in the unconditional model to 0.41. deviation. The secondary moderator analysis tested only the
Using the estimates from both models, the variance reduction differences between researcher and standardized measures,
was 0.20, indicating that 2% of the variance in effect sizes was as all study samples met soft criteria for risk status. The test
explained by the type of intervention that was administered. of differences between researcher and standardized measures
indicated that no statistically significant effects were observed
(p = .07). While the amount of variance in effect sizes decreased
Print Concepts Outcomes as a function of intervention types, only 4% was explained
A total of 10 effect sizes were calculated for print concepts using these covariates, and a significant portion remained that
outcomes, with a range of -0.77 to 2.35. The grand mean could be explained by other variables.
from the unconditional model was 0.86 and was significantly
different from a mean of 0, t(10) = 3.17, p = .01, with signifi-
cant variance existing among these 10 effect sizes (p < .001). Vocabulary Outcomes
Across studies with print concept outcomes, three different Vocabulary outcomes comprised the largest number of effect
interventions were codeddialogic reading, repeated reading, sizes (n = 51), with a range across the weighted estimates of
and computer assisted. Similar to the phonological awareness -1.19 to 4.73. From the unconditional model, a large mean
analysis, dialogic reading was set as the referent condition, coefficient of 1.02 was observed, which was significantly
with a mean effect size of 0.70 observed. Given the large different from 0, t(50) = 5.83, p < .001, with a variance
standard error associated with this mean (0.43), a rejection component that was relatively large (1.68) when compared
of the null hypothesis was not warranted. Similarly, the fitted to values from the other outcomes. A total of five intervention
means of both repeated reading (1.30) and computer assisted categories measured vocabulary outcomes, including com-
(0.18) were associated with large standard errors and were puter assisted, limited questioning, vocabulary, other inter-
not significantly different from dialogic reading. The second- ventions, and dialogic reading again set as the referent. The
ary moderator analysis tested only the differences between results from the conditional model indicated that the mean
soft and hard risk criteria, as all print concepts measures were effect size for the referent was moderate in nature (0.57) and
researcher developed. Results indicated that the mean effect was also significantly different from 0, t(46) = 2.97, p < .01.
size for studies using soft criteria was significantly less (i.e., Vocabulary interventions showed to be weaker in magnitude
-2.11) than studies using hard criteria, t(7) = -4.84, p = .001. than dialogic reading by -0.42 units, but due to the large
The variance components between the unconditional and standard error, this was not a statistically meaningful differ-
conditional models did decrease as a function of including ence. Similarly, while the mean effect size for limited ques-
intervention type; however, this only resulted in a 6% reduc- tioning was larger in magnitude than the referent (i.e., 1.43),
tion. When accounting for differences due to risk criteria, the a large standard error precluded a meaningful difference from
variance components decreased by a total of 82%. Residual dialogic reading. Interventions identified as other produced
variance after covarying interventions suggested that signifi- a substantially larger effect size compared to the referent, with
cant variability still existed that could be explained by other a fitted value of 2.18, and limited questioning interventions
noncoded predictors. were significantly stronger as well with a fitted mean of 2.00.
These fitted means are representative of the studies based on
inclusion criteria; however, the number of studies is relatively
Comprehension Outcomes few. The supplementary test of both risk criteria and measures
Comprehension outcomes were represented by 22 effect sizes, differences indicated no significant relationships existed
with a range of -0.52 to 1.47. The grand mean effect size for between these moderators (p = .51 and .16, respectively) and
and two vocabulary studies (Beck & McKeown, 2007, received the dialogic intervention alone on the Iowa Test of
Study 2; Coyne et al., 2007, Study 2) reported at least one Basic Skills (Hoover, Hieronymus, Frisbie, & Dunbar, 1993)
vocabulary outcome measure. Second graders who received Word Analysis subtest (ES = -1.36) and oral reading error
dialogic reading intervention (Reutzel et al., 1994) slightly rate (ES = -1.01). However, the dialogic intervention only
outperformed children in a more fluency focused intervention group read a greater number of words per minute at posttest
on a standardized measure of vocabulary (ES = 0.19). However, (ES = 0.63)
Bygrave (1994) reported no significant differences between
groups at posttest. Pemberton and Watkins (1987) examined
the effect of language recastings on vocabulary outcomes. Discussion
They provided one group of prekindergarteners with multiple In this synthesis, we sought to examine the effects of read-aloud
readings of a text written with language recastings and the practices on the literacy outcomes of children at risk for read-
comparison group with repeated readings of a story with ing difficulties. Read-aloud instruction has been examined
the same content, but no language recastings. Teachers read in several instructional formats, including dialogic reading;
the book aloud to children six to seven times. Authors reported repeated reading of stories; story reading with limited ques-
no statistically significant differences between the treatment tioning before, during, and/or after reading; computer-assisted
and comparison groups on a standardized test of vocabulary story reading; and story reading with extended vocabulary
(ES = 0.05). However, both groups demonstrated statisti- activities. Dialogic reading has received the most examina-
cally significant vocabulary gains from pre- to posttest. tion in the literature. A few additional read-aloud instructional
Two authors designed vocabulary-based interventions techniques have been examined in single studies, reviewed
delivered during read-aloud sessions (Beck & McKeown, in this synthesis as other read-aloud interventions.
2007, Study 2; Coyne et al., 2007, Study 2). Both studies Previous syntheses of storybook reading have indicated
provided one treatment group with rich, extended vocabulary a lack of high-quality research, qualifying the findings and
instruction that included a variety of interactive activities decreasing the ability to make robust statements regarding
before, during, and following read alouds that focused on the effects of read alouds on literacy outcomes for children
recognizing, discussing, and answering questions about target (Blok, 1999; Scarborough & Dobrich, 1994). Our synthesis
vocabulary words. However, the second treatment condition of the current literature indicates the amount of high-quality
differed between these two studies. When Coyne and col- research has increased. We found 20 treatment-comparison
leagues (2007, Study 2) compared an extended vocabulary studies of read-aloud interventions specifically for children
treatment with read alouds containing simple definitions of with risk factors for reading difficulties. In addition, 14 studies
target words provided during reading, kindergarteners in the included standardized assessments of childrens literacy out-
extended vocabulary treatment scored significantly higher comes, allowing for reliable and valid examination of broad
on researcher-developed measures of expressive (ES = 1.70), literacy outcomes following interventions. The larger and
receptive (ES = 0.99), and context vocabulary (ES = 1.12). higher-quality corpus of studies allowed us to examine more
In addition, children with higher Peabody Picture Vocabulary differentiated categories of independent and dependent
Test (PPVT) scores prior to treatment were more likely to variables than has been previously reported.
learn word meanings through extended instruction than chil- The meta-analysis revealed significant, positive effects for
dren with lower initial PPVT scores. read-aloud interventions on childrens language, phonological
Beck and McKeown (2007, Study 2) compared an extended awareness, print concepts, comprehension, and vocabulary
vocabulary treatment (rich instruction) with a more inten- outcomes. These results suggest that read-aloud interventions
sive intervention where children received more frequent provide children at risk for reading difficulties with higher
extended vocabulary instruction for a longer duration (more literacy outcomes than children who do not participate in
rich instruction). While effect sizes could not be calculated, these interventions. Although the effect on language outcomes
authors reported that among both kindergarten and first grad- was small and not moderated by intervention type, all other
ers, the pre- to posttest gain in words learned among children mean effects were large with intervention type explaining
who received more rich instruction was higher than children some of the variance in outcomes. These results are in line
in the rich instruction condition. with the recent National Early Literacy Panel (2008) review
of the effects of shared-reading interventions. The panel
reported significant effects for oral language (including
Word Recognition listening comprehension and vocabulary) and print knowl-
One dialogic reading study (Reutzel et al., 1994) reported a edge but did not disaggregate these individual outcomes for
word recognition outcome. Reutzel and colleagues (1994) children at risk for reading difficulties. Encouragingly, our
reported that second graders who received the dialogic inter- findings suggest children at risk for reading difficulties do
vention plus fluency instruction outperformed children who benefit from read-aloud interventions in these areas. In
contrast to the National Early Literacy Panel (2008) results, examination of the vocabulary studies indicates children are
we also found significant effects for phonological awareness making gains in the words that are taught in the intervention
outcomes, suggesting a possible additional benefit for children with the storybook, but demonstrating less gain in uninstructed
who are at risk for reading difficulties when read-aloud inter- vocabulary words. Two studies in this synthesis also noted
ventions are implemented in educational settings. differential effects on vocabulary outcomes, reporting children
Dialogic reading has the most causal evidence to support with the lowest vocabulary at pretest made the most gains
its effects on childrens literacy outcomes with eight experi- in vocabulary (Coyne, Simmons, Dameenui, & Stoolmiller,
mental studies. The meta-analysis indicated moderate to large 2004; Justice, Meier, & Walpole, 2005). Importantly, our
mean effect sizes for dialogic reading interventions on child synthesis reveals that participation in the different types of
outcomes of phonological awareness, print concepts, reading read-aloud interventions seems to stimulate vocabulary devel-
comprehension, and vocabulary. Thus, the evidence suggests opment for children at risk for reading difficulties, even on
that extended dialogue around the read aloud with transfer standardized measures of vocabulary.
of storytelling to the child can improve literacy outcomes for Despite the positive effects for read-aloud interventions,
children at risk for reading difficulties. only a small amount of variance was accounted for by inter-
The meta-analysis also revealed that computer-assisted vention type in the outcomes. The meta-analysis indicates
interventions demonstrated small to large mean effects on that other unknown factors beyond the provided intervention
literacy outcomes while limited questioning interventions explain significant amounts of variance in child outcomes on
revealed a small, negative mean effect size on phonological each of the measures. Similarly, several studies examining
awareness yet a large effect size related to vocabulary out- multiple read-aloud treatments reported no significant dif-
comes. In both cases, there were few effect sizes contributing ferences in outcomes for the different treatment conditions,
to these mean effects, and the mean effects for limited ques- though significant pre- to posttest gains were realized in each
tioning were not significantly different than the mean effect treatment. These findings may suggest that while read-aloud
for dialogic reading. Computer-assisted interventions dem- interventions are valuable for childrens literacy outcomes,
onstrated significantly higher effects on reading comprehen- some of the specific features related to improvement have not
sion and vocabulary outcomes for children than the dialogic been fully realized in the literature. Currently, the strongest
reading interventions. The higher effects on vocabulary may evidence comes from dialogic reading interventions, suggest-
be the result of the outcome measures. The computer-assisted ing that incorporating extended childadult dialogue and
studies used researcher-developed measures assessing student questioning around storybooks is a valuable practice in edu-
knowledge of the meaning of the words in the book read cational settings. Computer-assisted read alouds demonstrate
aloud during the intervention. In contrast, the dialogic reading promise for improving childrens literacy outcomes as well.
studies largely measured broad vocabulary outcomes (beyond The studies examining computer-assisted read alouds gener-
words in the intervention storybooks) on standardized mea- ally included interactive features allowing the child to engage
sures of vocabulary. in key aspects of the story line or character manipulation.
The practice of repeatedly reading stories to children at The literature provides little information regarding the
risk for reading difficulties has received limited attention, long-term effects of read-aloud interventions. Only three
with only two studies examining the intervention effects studies examined delayed or long-term outcomes of the inter-
on only two literacy outcomes, language and print concepts. ventions (Coyne et al., 2007, Study 1 and Study 2; Hindson
A large mean effect was noted for print concepts; however, et al., 2005). In both of their studies, Coyne et al. (2007)
these effects are reported from only one study and were not reported children in the read-aloud intervention continued to
significantly different from the print concepts outcomes for outperform children in the control group 6 weeks after inter-
dialogic reading. A second study found no differences in vention on words taught in the intervention. Hindson et al.
language outcomes between two versions of a repeated read- (2005) reported 1-year follow-up results indicating children
ing treatment, but reported significant differences between in the read-aloud intervention continued to perform at grade-
pretest and posttest scores for children involved in the treat- level expectations, but children not at risk for difficulties made
ments. The evidence suggests that repeated reading interven- more gains in the year, thus widening the gap. Future research
tions may have potential for positively effecting student on read-aloud interventions should provide additional infor-
outcomes, but additional research is needed. mation on the long-term outcomes of participation in the
Intervention type accounted for 28% of the variance in interventions to better inform practice.
vocabulary effect sizes. Although seemingly counterintuitive,
the meta-analysis indicates that read alouds with extended
vocabulary had the lowest mean effect size for vocabulary Limitations
outcomes. However, this small effect was not significantly We have the most confidence in reporting findings related to
different than the moderate effect for dialogic reading. A closer dialogic reading given the relatively large number of studies
examining this type of intervention. Only two to three studies Bus, A. G., van IJzendoorn, J. H., & Pellegrini, A. D. (1995). Joint
each have examined computer-assisted read alouds, repeated book reading makes for success in learning to read: A meta-
reading of read alouds, or limited questioning. Thus, often there analysis on intergenerational transmission of literacy. Review
was a lot of error around the estimates. The low number of of Educational Research, 65, 121.
studies with common intervention features precluded us from *Bygrave, P. L. (1994). Development of listening skills in students
examining more specific features of intervention implementation in special education settings. International Journal of Disability,
such as group size and amount of time in intervention that may Development, and Education, 41, 5160.
moderate student outcomes. These factors may explain some Catts, H., & Hogan, T. (2003). Language basis of reading disabili-
of the additional variance in child literacy outcomes and warrant ties and implications for early identification and remediation.
continued research to inform early intervention efforts for chil- Reading Psychology, 24, 223246.
dren at risk for reading difficulties. Finally, one should interpret Cohen, J. (1988). Statistical power analysis for the behavioral
the findings of the supplementary tests of differing risk criteria sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
and standardized or researcher-based assessments with caution. Collins-Stanley, T., & Gan, S. (1996). Choice of romantic or vio-
The very small number of effect sizes contributing to some of lent or and scary fairy-tale books by preschool girls and boys.
these supplementary tests was underpowered to accurately detect Child Study Journal, 26, 279303.
statistically significant differences, meaning that we cannot *Combs, M. (1987). Modeling the reading process with enlarged
completely rule out the possibility of Type I error. texts. The Reading Teacher, 40, 422426.
Cooper, H. (1998). Synthesizing research: A guide for literature
Declaration of Conflicting Interests reviews. Thousand Oaks, CA: Sage.
The author(s) declared no conflicts of interest with respect to the *Coyne, M. D., McCoach, D. B., & Kapp, S. (2007). Vocabulary
authorship and/or publication of this article. intervention for kindergarten students: Comparing extended
instruction with embedded instruction and incidental exposure.
Funding Learning Disabilities Quarterly, 30, 7488.
The author(s) disclosed receipt of the following financial support *Coyne, M. D., Simmons, D. C., Dameenui, E. J., & Stoolmiller, M.
for the research and/or authorship of this article: This work was (2004). Teaching vocabulary during shared storybook read-
supported by The Meadows Foundation. ings: An examination of differential effects. Exceptionality,
12, 145162.
References Daly, E., & Martens, B. (1994). A comparison of three interven-
*References marked with an asterisk indicate studies included in tions for increasing oral reading performance: Application of the
the meta-analysis. instructional hierarchy. Journal of Applied Behavior Analysis,
*Aram, D., & Biron, S. (2004). Joint storybook reading and joint 27, 459469.
writing interventions among low SES preschoolers: Differen- *Davies, P., Shanks, B., & Davies, K. (2004) Improving narrative
tial contributions to early literacy. Early Childhood Research skills in young children with delayed language development.
Quarterly, 19, 588610. Educational Review, 56, 271286.
*Beck, I. L., & McKeown, M. G. (2007). Different ways for dif- Dickinson, D. K., & Smith, M. W. (1994). Long-term effects of
ferent goals, but keep your eye on the higher verbal goals. In preschool teachers book readings on low-income childrens
R. K. Wagner, A. Muse, & K. Tannenbaum (Eds.), Vocabulary vocabulary and story comprehension. Reading Research
acquisition: Implications for reading comprehension (pp. 182204). Quarterly, 29, 105122.
New York: Guilford. Ezell, H. K., & Justice, L. M. (2005). Shared storybook reading:
*Bellon-Harn, M. L., Hoffman, P. R., & Harn, W. E. (2004). Use Building young childrens language and emergent literacy skills.
of cloze and contrast word procedures in repeated storybook Baltimore, MD: Brookes.
reading: Targeting multiple domains. Journal of Communication Gambrell, L., & Marinak, B. (2009). Reading motivation: What
Disorders, 37, 5375. the research says. Retrieved from http://www.readingrockets
Blok, H. (1999). Reading to young children in educational settings: .org/article/29624
A meta-analysis of recent research. Language Learning, 49, *Hargrave, A. C., & Snchal, M. (2000). A book reading interven-
343371. tion with preschool children who have limited vocabularies: The
*Boykin, A. W., & Cunningham, R. (2001). The effects of move- benefits of regular reading and dialogic reading. Early Child-
ment expressiveness in story content and learning context on the hood Research Quarterly, 15, 7590.
analogical reasoning performance of African American children. Hindson, B., Byrne, B., Fielding-Barnsley, R., Newman, C.,
Journal of Negro Education, 70, 7283. Hine, D. W., & Shankweiler, D. (2005). Assessment and early
Bus, A. G. (2001). Joint caregiver-child storybook reading: A route instruction of preschool children at risk for reading disability.
to literacy development. In S. B. Neuman, & D. K. Dickinson Journal of Educational Psychology, 97, 687704.
(Eds.), Handbook of early literacy research (Vol. 1, pp. 179191). Hoover, H., Hieronymus, A., Frisbie, D., & Dunbar, S. (1993).
New York, NY: Guilford. Iowa tests of basic skills. Chicago, IL: Riverside.
Institute of Education Sciences. (2003). What Works Clearing- National Center for Educational Statistics. (2006). The condition of
house study review standards. Retrieved http://ies.ed.gov/ncee/ education 2006. Washington, DC: U.S. Department of Education.
wwc/help/idocviewer/doc.aspx?docId=19&tocId=4 National Early Literacy Panel. (2008). Developing early literacy:
*Justice, L. M., & Ezell, H. K. (2002). Use of storybook reading to Report of the National Early Literacy Panel. Washington, DC:
increase print awareness in at-risk children. American Journal National Institute for Literacy.
of Speech-Language Pathology, 11, 1729. National Early Literacy Panel. (2009). Developing early literacy:
*Justice, L. M., Meier, J., & Walpole, S. (2005). Learning new Report of the National Early Literacy Panel. Washington, DC:
words from storybooks: Findings from an intervention with at- National Institute for Literacy.
risk kindergarteners. Language, Speech, and Hearing Services National Research Council. (1998). Preventing reading difficulties
in Schools, 36, 1732. in young children. Washington, DC: National Academy Press.
Karweit, N., & Wasik, B. A. (1996). The effects of story reading Neuman, S. B., & Dickinson, D. K. (Eds.). (2001). Handbook of
programs on literacy and language development of disadvan- early literacy research. New York, NY: Guilford.
taged preschoolers. Journal of Education for Students Placed at *Pemberton, E. F., & Watkins, R. V. (1987). Language facilitation
Risk, 1, 319348. through stories: Recasting and modeling. First Language, 7,
Kim, A., Vaughn, S., Wanzek, J., & Wei, S. (2004). Graphic orga- 7989.
nizers and their effects on the reading comprehension of stu- *Peskin, J., & Astington, J. W. (2004). The effects of adding meta-
dents with learning disabilities. Journal of Learning Disabilities, cognitive language to story texts. Cognitive Development, 19,
37, 105118. 253273.
*Korat, O., & Shamir, A. (2007). Electronic books versus adult Purcell-Gates, V. P., McIntyre, E., & Freppon, P. A. (1995). Learn-
readers: Effects on children emergent literacy as a function ing written storybook language in school: A comparison of low-
of social class. Journal of Computer Assisted Learning, 23, SES children in skills-based and whole language classrooms.
248259. American Education Research Journal, 32, 659685.
Kreft, I., & de Leeuw, J. (1998). Introducing multilevel modeling. Raudenbush, S. W. (2005). Learning from attempts to improve
Thousand Oaks, CA: Sage. schooling: The contribution of methodological diversity.
Lonigan, C. J. (1994). Reading to preschoolers exposed: Is the Educational Researcher, 34(5), 2531.
emperor really naked? Developmental Review, 14, 303323. Raudenbush, S. W., & Bryk, A. S. (1986). A hierarchical model
*Lonigan, C. J., Anthony, J. L., Bloomfield, B., Dyer, S. M., & for studying school effects. Sociology of Education, 59, 117.
Samwel, C. (1999). Effects of two preschool shared read- Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear mod-
ing interventions on the emergent literacy skills of children els: Applications and data analysis methods. Thousand Oaks,
from low-income families. Journal of Early Intervention, CA: Sage.
22, 306322. *Reutzel, D. R., Hollingsworth, P. M., & Eldredge, J. L. (1994).
*Lonigan, C. J., & Whitehurst, G. J. (1998). Relative efficacy of Oral reading instruction: The impact on student reading
parent and teacher involvement in a shared-reading intervention development. Reading Research Quarterly, 29, 4062.
for preschool children from low-income backgrounds. Early Scarborough, H. S., & Dobrich, W. (1994). On the efficacy of
Childhood Research Quarterly, 17, 265292. reading to preschoolers. Developmental Review, 14, 245302.
Lyon, G. R. (1998). Why reading is not a natural process. Educa- Shadish, W. R. (2002). Revisiting field experimentation: Field notes
tional Leadership, 55(6), 1418. for the future. Psychological Methods, 7, 318.
McCoach, D. B., OConnell, A. A., Reis, S. M., & Levitt, H. *Talley, S., Lancy, D. F., & Lee, T. R. (1997). Children, story-
(2006). Growing readers: A hierarchical linear model of chil- books, and computers. Reading Horizons, 38, 116128.
drens reading growth over the first two years of school. Journal Ukrainetz, T. A., Cooney, M. H., & Dyer, S. K. (2000). An investiga-
of Educational Psychology, 98, 1428. tion into teaching phonemic awareness through shared reading
Mol, S. E., Bus, A. G., & de Jong, M. T. (2009). Interactive and writing. Early Childhood Research Quarterly, 15, 331355.
book reading in early education: A tool to stimulate print *Ulanoff, S. H., & Pucci, S. L. (1999). Learning words from books:
knowledge as well as oral language. Review of Educational The effects of read aloud on second language vocabulary acqui-
Research, 79, 9791007. sition. Bilingual Research Journal, 23, 319332.
*Morrow, L. M. (1984). Reading stories to young children: Effects U.S. Department of Education. (2003). Random assignment in
of story structure and traditional questioning strategies on com- program evaluation and intervention research: Questions and
prehension. Journal of Reading Behavior, 16, 273288. answers. Retrieved from: http://www.ed.gov/rschstat/eval/
*Morrow, L. M. (1988). Young childrens responses to one-to-one resources/randomqa.pdf
story readings in school settings. Reading Research Quarterly, van Kleeck, A. (2004). Joint caregiver-child storybook reading: A
23, 89107. route to literacy development. In C. A. Stone, E. R. Silliman, B.
National Center of Educational Statistics. (2004). The condi- J. Ehren, & K. Apel (Eds.), Handbook of language and lit-
tion of education 2004. Washington, DC: U.S. Department of eracy: Development and disorders (pp. 175208). New York,
Education. NY: Guilford.
Vaughn, S., Kim, A., Sloan, C. V. M., Hughes, M. T., Elbaum, B., & disabilities, reading, effective instruction, and response to
Sridhar, D. (2003). Social skills interventions for young children intervention.
with disabilities: A synthesis of group design studies. Remedial
and Special Education, 24, 215. Yaacov Petscher, PhD, is the Director of Research at the Florida
*Verhallen, M. J. A. J., Bus, A. G., & de Jong, M. T. (2006). The Center for Reading Research, and an associate in research at Florida
promise of multimedia stories for kindergarten children at-risk. State University.
Journal of Educational Psychology, 98, 410419.
*Wasik, B. A., & Bond, M. A. (2001). Beyond the pages of a Sharon Vaughn, PhD, holds the H. E. Hartfelder/Southland Corp.
book: Interactive book reading and language development in Regents Chair in Human Development and is the Executive Director
preschool classrooms. Journal of Educational Psychology, 93, of the Meadows Center for Preventing Educational Risk an orga-
243250. nized research unit at the University of Texas. Her current interests
*Wasik, B., Bond, M., & Hindman, A. (2006). The effects of a are investigating effective interventions for students with reading
language and literacy intervention on Head Start children and difficulties and students who are English language learners
teachers. Journal of Educational Psychology, 98, 6374.
Wasik, B. H., & Hendrickson, J. S. (2004). Family literacy prac- Christie Cavanaugh, PhD, is an educational consultant. Her
tices. In C. A. Stone, E. R. Silliman, B., Ehren, J., & Apel, K. research interests are language, early literacy, and RTI.
(Eds.), Handbook of language and literacy development and
disorders (pp. 154174). New York, NY: Guilford. Jennifer Heckert, PhD, is an Assistant Professor of Special Educa-
*Whitehurst, G. J., Arnold, D. S., Epstein, J. N., Angell, A. L., tion in the Inclusive Education Department at Kennesaw State
Smith, M., & Fischel, J. E. (1994). A picture book reading inter- University. Her research interests include collaboration between
vention in day care and home for children from low-income general and special education, preparing general educators to effec-
families. Developmental Psychology, 30, 679689. tively instruct students with LD, and principal instructional leader-
ship for students with LD.
About the Authors
Elizabeth Swanson, PhD, is a senior research associate at The Guliz Kraft, PhD, received her doctoral degree from University of
Meadows Center for Preventing Educational Risk at The University Texas at Austin in Learning Disabilities and Behavior Disorders. She
of Texas at Austin. Research interests include effective reading is a World Bank Family Network, Margaret McNamara Fund 2005
instruction for students who struggle with reading, with a particular fellow and is a member of WINPEACE (Women Initiation of Peace).
focus on students with learning disabilities.
Kathryn Tackett, PhD, is a research analyst at the American Insti-
Jeanne Wanzek, PhD, is an assistant professor in special education tutes for Research in Washington, DC. She also serves as an adjunct
at Florida State University and on the research faculty at the Florida professor for the Department of Special Education at George Mason
Center for Reading Research. Her research interests include learning University.