Excerpted from Chapter 13 of Early Childhood Literacy, edited by Timothy Shanahan, Ph.D., & Christopher J. Lonigan, Ph.D.
Copyright© 2013 by Paul H. Brookes Publishing Co. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Beyond the National Early Literacy Panel
The chapters of this book build on the findings of NELP. They highlight
emerging research in topics summarized by NELP, offer additional perspective
on the findings reported by NELP, and suggest additional areas
in which the knowledge base of early literacy and professional practice
must be increased. In this section, we summarize some of these key
points and identify some crosscutting issues raised by the authors of
the chapters.
NELP restricted its review and synthesis to empirical studies published
in peer-reviewed journals. In part, this was done to limit the
scope of searches conducted (i.e., alternative search strategies could
have included conference reports, dissertations, and other unpublished
work), to ensure that searches within a literature were comprehensive
(i.e., many alternative literatures are not abstracted and cannot be
systematically searched), and to provide some level of quality control
over the studies that were included (i.e., peer review ensures that work
included meets some quality standard, however slight). In Chapter 2,
Griffin and McCardle point out several large-scale projects supported
by the U.S. government that have relevant data concerning the development
and promotion of early literacy skills. Many of these data sets can
be used to answer questions based on the analyses of large and often
nationally representative samples of children and families. Regarding
some of these data sets, one can question the degree to which depth was
sacrificed for breadth in the design of assessment batteries or can be
concerned that the measurement operations below the construct labels
are, by design, black boxes (e.g., specific item content is not available for
most measures used in the data sets for the Early Childhood Longitudinal
Study). However, there is no question that these data sets and the
reports derived from them to date provide useful information concerning
young children's development. Many of the studies noted by Griffin
and McCardle are embedded in the findings reported in other chapters
of this book (e.g., Chapter 8).
In Chapter 3, Spencer, Spencer, Goldstein, and Schneider provide
a review of assessments that can be used for different purposes. They
note that the NELP report identifies skill domains in which assessment
would provide useful information for purposes of screening, identification,
and progress monitoring children's literacy levels and proficiency.
These skill domains also are relevant to child and classroom assessment
systems that could be used to identify teachers who might need support
and professional development to achieve the goal of supporting children's
development. Although the chapter authors identify a host of measures
for these different purposes, data on the appropriateness and validity of
many of these measures for these purposes are, in most cases, weak or
nonexistent (Lonigan, Allan, & Lerner, 2011; Wilson & Lonigan, 2009,
2010). A number of standardized diagnostic measures are validated for
use with preschool-age children. However, the cost, training, and administration
time required for most of these measures make them impractical
for widespread use in most early childhood environments. The best use of
these measures is to provide in-depth assessments of children's relative
strengths and weaknesses in specific skill domains.
Additional work is needed on the development and validation of
measures that can be used for screening and progress monitoring of a
subset of skills identified by NELP. As an example of such work, the State
of Florida, in 2009–2010, developed a set of assessments that teachers
in its state-funded preschool program can use to screen and monitor
children's progress related to oral language, phonological awareness,
print knowledge, and mathematics (see Florida Department of Education,
2008), in line with Florida's Early Learning Standards. These
assessments are part of an effort by the state to create assessment systems
that can be used by teachers to inform instruction. For example,
the state previously developed the Florida Assessment for Instruction
in Reading, which is a system of screening, progress monitoring, and
diagnostic measuring of children in kindergarten through Grade 12. A
large-scale field trial of the preschool measures indicated that they can
be administered reliably by children's preschool teachers and are valid
for identifying children at risk of not meeting Florida's school-readiness
classification in the fall of kindergarten.
In Chapter 4, Spinrad, Valiente, and Eisenberg highlight the growing
number of research findings of a connection between children's
self-regulation and their academic skills. As noted by Spinrad and colleagues,
studies have shown that a variety of socioemotional constructs
are related to children's language, reading, and math skills, including
effortful control, executive function, social competence, peer relationships,
teacher relationships, motivation, emotionality, emotion understanding,
and emotion regulation. These constructs were not a part
of the NELP report's set of predictor variables because 1) many of the
studies of these variables involve children in elementary school or 2) the
studies were not published until after the cutoff date for NELP's literature
review (e.g., Blair & Razza, 2007; McClelland et al., 2007). This
area of research raises a number of interesting questions concerning
the dynamic interaction of academic skills and behavioral and cognitive
development. At present, there are many questions concerning definitions
of constructs (e.g., Allan & Lonigan, 2011; Willoughby, Blair, Wirth,
& Greenberg, 2010) and the degree to which these self-regulation constructs
independently influence academic skills, including growth in
academic skills (see Duncan et al., 2007; Ponitz, McClelland, Matthews,
& Morrison, 2009).
As previously noted, a frequent criticism of the NELP report is that
it emphasized evidence concerning code-related skills and instruction;
however, this was a function of the amount of higher-quality published
evidence concerning these skills and instructional activities and not a
decision on the part of the panel to focus on one area and not another.
In Chapter 5, Phillips and Piasta expand the evidence summarized by
NELP on code-focused outcomes and interventions to include more
recent studies. They confirm the NELP findings regarding the predictive
nature of phonological awareness and alphabet knowledge. In addition,
they summarize some more recent work on teaching phonological
awareness and print knowledge. By the cutoff date of NELP's literature
review, there were few studies identified that examined the effect of
teaching children about the alphabet only—in contrast to studies that
taught the alphabet in conjunction with phonological awareness instruction.
Having looked at the results of more recent studies on the effects of
teaching alphabet knowledge, Phillips and Piasta conclude that teaching
both letter names and letter sounds is important. They also discuss
issues surrounding whole-class versus small-group instructional methods
and note that current data suggest a need to increase attention to
phonological awareness instruction in preschool classrooms.
The NELP synthesis of studies concerning interventions involving
shared reading is just one of a number of such syntheses that have
been reported. In Chapter 6, Pentimonti, Justice, and Piasta review
some of these other summaries of the effects of shared reading, and
they summarize the effects of interventions involving shared reading
that also involve word elaborations and print referencing. They make a
point of noting that the effects of these shared-reading strategies may
be substantially reduced when used by teachers instead of parents or
researchers and that the effects seem to vary as a function of child age
and risk status. They note that despite longstanding calls for shared
book reading in early childhood education settings, there is considerable
variability in how often book reading takes place in preschools and
child care centers, with a not small number of preschool and child care
classrooms with low levels of book-reading activity. Pentimonti and colleagues
also discuss a relatively underresearched area, that is, the types
of books that are used in shared reading with young children (e.g., narrative
versus expository), and they make suggestions for types of books
to be used. They also highlight the need for more research on the relative
benefits of small-group versus whole-class shared reading, on types
of books used in shared reading, and on optimal ways to promote language
skills beyond vocabulary.
In Chapter 7, Wasik and Feldman summarize results from additional
studies of parent and home programs. They provide a detailed
summary of the history of evaluations of the now-defunct Even Start
program, which was the U.S. government's two-generation ("family literacy")
program. Several randomized evaluations of Even Start failed
to demonstrate positive benefits of the program on most child or parent
outcomes. The last randomized evaluation of Even Start, the Classroom
Literacy Interventions and Outcomes [CLIO] study (Judkins et al.,
2008), evaluated the impact of deploying evidence-based classroom and
parent interventions in Even Start programs; however, the CLIO study
yielded few positive findings. Wasik and Feldman discuss one possible
reason for the lack of consistent positive effects from family literacy
programs—fidelity. That is, for an intervention to be effective, those
it is intended to affect must receive it in some minimally potent dose.
Teachers in classrooms may not implement evidence-based practices
as intended, and parents may not participate in the available program
components with sufficient frequency for those components to make a
difference. Wasik and Feldman also discuss findings from research with
older children in which parents are provided with explicit instructional
strategies to teach reading skills, and they suggest that it may be useful
to develop and evaluate explicit instructional practices for parents of
preschool children to use—beyond shared reading—that will help their
children acquire early literacy skills.
Few preschool curriculum or professional development studies
were included in the NELP synthesis because there are few empirical
evaluations of preschool curricula that are published in peer-reviewed
journals or that meet standards of quality that allow causal conclusions.
In Chapter 8, Lonigan and Cunningham provide a summary of current
evidence concerning preschool curricula and professional development
programs. Beyond studies of programs that suggest the value of high-quality
early childhood education, there is a relative dearth of well-designed
studies that evaluate specific preschool early literacy curricula.
Most commercially available curricula have no available studies, or
they have studies from which valid conclusions about the effects of the
curricula cannot be drawn, and there are relatively few positive results
from the few well-designed studies available. Such findings indicate that
substantially more work is needed to develop and evaluate preschool
curricula. At present, most of the curricula adopted by preschools either
have no evidence of effectiveness or no evidence that a particular curriculum
is better than a generic alternative. As for professional development
programs, the available evidence indicates that professional development
for teachers seems unlikely to offer an easy solution, with most
studies of professional development failing to show better outcomes for
children whose teachers received the professional development than for
children whose teachers did not receive the professional development;
the one study demonstrating a significant advantage of professional
development involved a scope and intensity of professional development
well beyond that which programs typically adopt or can afford. Commonalities
among curricula and professional development with evidence
of effectiveness include explicit instruction that is at odds with the historical
philosophy of early childhood education.
In Chapter 9, Dickinson and Darrow highlight the significance of
language development for becoming a skilled reader and note the relative
paucity of interventions shown to be effective in promoting language
development skills. They note that most studies showing positive effects
on language development involve relatively narrow and short-term interventions
and have not involved language-focused curricula. They discuss
a number of impediments to deploying effective, literacy-focused curricula
and also discuss aspects of fidelity of implementation that may be
associated with more versus less positive outcomes for children, including
the fact that in many classrooms, initial levels of language teaching
are low—requiring large changes to move into effective levels of
Instruction—adoption of effective strategies may be low and variable in
application, and use of strategies may be highly variable throughout the
day.
In Chapter 10, Hogan, Cain, and Bridges provide an expanded
description on the ways in which oral language skills contribute to
skilled reading, particularly reading comprehension. Based upon
the simple view of reading model (Gough, Hoover, & Peterson, 1996;
Hoover & Gough, 1990), they describe a changing situation for skilled
reading that initially draws heavily on code-related skills and simple
language processes in children's early years but that increasingly relies
on a variety of more complex language processes to yield good reading
comprehension skills. Given the changing nature of the reading task
as children progress through school, difficulties may not be apparent
until the more complex language processes are required. Hogan and
colleagues discuss the need to promote these more complex language
skills that allow children to integrate multiple elements within a text,
develop a mental model of the text that can be interrogated and modified,
and identify causal relations in text to provide children with the
language skills that will ultimately be needed to have skilled reading
once the reading tasks make increasing demands on more complex
language processes.
One area that NELP noted it was unable to provide significant evidence
for concerned special populations, including children with disabilities.
This was because few group-design studies reported results
in ways that allowed effects to be isolated within a part of the study
sample and because studies with single-subject designs were excluded
from the NELP synthesis because they do not yield effect sizes that can
be combined with effect sizes of group-design studies. In Chapter 11,
Carta and Driscoll summarize the evidence of the effectiveness of interventions
with special needs populations. As with the NELP summary,
the largest number of interventions evaluated with special needs populations
concerned code-focused instruction. These interventions, which
typically include a large component of explicit instruction, are effective
at promoting the skills they target (i.e., phonological awareness or print
knowledge). However, most of the studies of these interventions have
involved children who have speech and language impairments rather
than more general developmental delays or other cognitive/sensory
impairments. Carta and Driscoll note the need for additional studies
in this area—studies that involve a broader inclusion of children with
special needs, instruction that targets more than code-focused skills,
instruction that takes place in typical educational settings, and inclusion
of longitudinal outcomes so that the longer term effects of these
interventions on children's conventional literacy skills can be evaluated.
Finally, in Chapter 12, Kaefer, Neuman, and Pinkham highlight
some of the strengths and weaknesses of using meta-analysis to identify
important correlates of an outcome and effective instructional practices
for promoting development of an outcome. They correctly note, as we
have elsewhere (Lonigan & Shanahan, 2010), that any meta-analysis is
limited by the research that has been completed and reported by the
time the analysis is done. Clearly, a meta-analysis cannot provide a
synthesis of research not yet conducted or a synthesis of good ideas.
Moreover, different sets of inclusion and exclusion criteria can affect
the conclusions of a synthesis; Kaefer and colleagues use the example
of the effects of shared-reading interventions in which different meta-analyses
have supported stronger or weaker effects, in part because of
the studies included. They highlight a "mixed methods" approach to
meta-analysis in which qualitative aspects of studies are evaluated in an
attempt to understand patterns. Such an approach is likely what happens
in most meta-analyses, however, and it certainly describes most of
the meta-analyses reported by NELP. That is, once the primary meta-analysis
is completed, features of individual studies (e.g., methods,
population, components of intervention) are analyzed to try to explain
reliable and replicable features associated with variations in effect sizes.
Although some of these features typically are envisioned at the outset
of a meta-analysis, it is also usually the case that other, unanticipated
features are evaluated after the studies have been examined first. The
essential elements here, however, are reliability and replicability. Any
study may produce larger or smaller effects and have unique features;
however, those unique features may not be responsible for the larger
or smaller effects. To link any feature to an outcome, it is necessary
that the feature be evident in multiple studies, and then its presence or
absence has to correlate with the outcome variations.