Content uploaded by Stephanie Al Otaiba
Author content
All content in this area was uploaded by Stephanie Al Otaiba on Jan 01, 2015
Content may be subject to copyright.
Journal of Educational Psychology
2001,
Vol. 93, No. 2, 251-267Copyright 2001 by the American Psychological Association, Inc.
O022-O663/Ol/$5.O0 DOI: 10.1037//0022-0663.93.2.251
Is Reading Important in Reading-Readiness Programs?
A Randomized Field Trial With Teachers as Program Implemented
Douglas Fuchs, Lynn S. Fuchs, Anneke Thompson,
Stephanie Al Otaiba, Loulee Yen,
Nancy J. Yang, and Mary Braun
Peabody College of Vanderbilt University
Rollanda E. O'Connor
University of Pittsburgh
The purpose of the present study was to examine the effectiveness and feasibility of phonological
awareness training, with and without a beginning decoding component. Thirty-three teachers in 8 urban
schools were assigned randomly within their schools to 3 groups: control, phonological awareness
training, and phonological awareness training with beginning decoding instruction and practice. Follow-
ing training, teachers in the 2 treatment groups conducted the treatments for about 20 weeks. In each
teacher's class, pre- and posttreatment data were collected on 12-14 children (N = 404); 312 children
were tested again the following fall. At the end of kindergarten, the 2 treatment groups performed
comparably and outperformed controls on the phonological awareness measures. On alphabetic (reading
and spelling) tasks, however, the group participating in phonological awareness training with beginning
decoding instruction did better than the other 2 groups. In the fall of the next year, many of these
between-group differences remained but were less impressive. Implications are discussed for bridging
research and practice.
Phonological awareness is "very hot." So say 25 experts from
the United States, Australia, and Canada who were asked to rate 27
reading-related topics for Reading Today, a publication of the
International Reading Association (Cassidy & Wenrich, 1998). In
support of the experts' rating, one should consider the following:
1.
The National Research Council's Committee on the Preven-
tion of Reading Difficulties in Young Children recommended that
"activities that direct young children's attention to the phonolog-
ical structure of spoken words .. . and that highlight the relations
between print and speech" be included in early childhood pro-
grams (Snow, Burns, & Griffin, 1998, p. 321).
2.
The Houston public schools recently spent millions of dollars
to provide all K-3 teachers with materials and 30 hr of training so
Douglas Fuchs, Lynn S. Fuchs, Anneke Thompson, Stephanie Al
Otaiba, Loulee Yen, Nancy J. Yang, and Mary Braun, Department of
Special Education, Peabody College of Vanderbilt University; Rollanda E.
O'Connor, Department of Instruction and Learning, University of Pitts-
burgh.
This research was supported in part by Grants H324V980001 and
H023C0251 from the Office of Special Education Programs in the U.S.
Department of Education and by Core Grant HD 15052 from the National
Institute of Child Health and Development. The article does not necessarily
reflect the positions or policies of these funding agencies, and no official
endorsement by them should be inferred. Portions of this article were
presented at the annual meeting of the American Educational Research
Association in Montreal, Canada, the Biennial Conference on Evidence-
Based Policies and Indicator Systems in Durham, England, and the Pacific
Coast Research Conference in La Jolla, California (all in 1999).
We thank Steve Graham, Karen Harris, Joe Jenkins, Ingrid Oxaal, and
Joanna Williams for their help in the planning stage of the study. We also
thank our study teachers from whom we learned much about kindergarten.
Correspondence concerning this article should be addressed to Douglas
Fuchs, Box 328 Peabody, Vanderbilt University, Nashville, Tennessee
37203.
Electronic mail may be sent to doug.fuchs@vanderbilt.edu.
they can conduct phonemic awareness lessons and other activities
thought to boost reading development (see Bradley, 1998).
3.
The Reading Excellence Act, a $260 million program target-
ing 500,000 children in Grades K-3, says that reading covers six
vital areas, one of which is phonemic awareness (International
Reading Association, 1999).
4.
In 1998, the National Institute for Child Health and Human
Development (NICHD) and the Office of Special Education Pro-
grams in the U.S. Department of Education cosponsored the Na-
tional Summit on Research in Learning Disabilities to help dis-
seminate research showing causal connections between
phonological awareness and beginning reading (Lyon, n.d.).
5.
The seven-page cover story in a November 22, 1999, issue of
Newsweek was on dyslexia. It highlighted NICHD researchers'
work on phonological awareness and concluded by recommending
in part that parents wishing to help their children learn to read
should "lobby for phonemic-awareness testing and .. . instruction"
(Kantrowitz & Underwood, 1999, p. 78).
Explaining "Very Hot"
Why has phonological awareness grabbed our attention so? We
suspect there are many explanations; we offer four, moving from
the general to the specific. First, everyone agrees reading is pivotal
to virtually all school learning. Second, many children are not
learning to read. On the National Assessment of Educational
Progress, 40% of fourth graders read below an established standard
(see Snow et al., 1998). Approximately 4.5 million African Amer-
ican students and 3.3 million Hispanic children "were reading very
poorly in fourth grade" (Snow et al., 1998. p. 97).
Third, phonological awareness is often equated with early in-
tervention. It is widely acknowledged that reading failure begins
early. When young children fail to learn to read, they begin to
dislike reading, they read less than their more skilled classmates,
251
252FUCHS
ET AL.
and
as a
consequence, they lose
an
important means
of
gaining
vocabulary, background knowledge,
and
information about text
structure (e.g.. Juel, 1996).
The
(word-) poor
get
poorer
as the
(word-) rich
get
richer
(see
Stanovich, 1986). Juel (1988) esti-
mated that
the
probability
of a
poor reader
at the end of
Grade
1
remaining
a
poor reader
at the end of
Grade
4 is .88.
Hence,
reading failure
can set in
quickly
and
thereafter
can
become
difficult
to
remediate. This would appear
to
argue
for
intervening
before children experience sustained failure
and
lose their
self-
confidence
and
enthusiasm
for
learning (e.g., Juel, 1996). Many
American policymakers
and
researchers
are
bullish
on
early inter-
vention programs like Head Start because they believe they work
and,
in the
long
run,
save taxpayers money (e.g., Snow
et
al.,
1998;
The Consortium
on
Renewing Education, 1998).
Fourth,
an
impressive amount
of
correlational
and
experimental
evidence links phonological awareness
to
reading. Typical
(non-
disabled) kindergartners with relatively strong phonological
awareness read better in subsequent grades than
do
classmates
who
show comparatively weak phonological awareness
in
kindergarten
(e.g., Bradley
&
Bryant,
1983;
Bryant, MacLean, Bradley,
&
Crossland, 1990; Juel, Griffith,
&
Gough, 1986; Lundberg, Olofs-
son.
&
Wall, 1980; Stanovich, Cunningham, & Cramer, 1984),
and
the strength
of
this correlation endures after controlling
for
intel-
ligence, vocabulary, letter knowledge, memory,
or
social class
(e.g., Bradley
&
Bryant,
1985;
Goswani
&
Bryant,
1990;
Share,
Jorm, MacLean,
&
Matthews,
1984;
Torgesen, Wagner,
& Ra-
shotte,
1994;
Wagner
&
Torgesen,
1997;
Wagner, Torgesen,
Laughon, Simmons.
&
Rashotte, 1993).
In
other correlational
studies, children with severe reading problems have demonstrated
poor phonological awareness skills (e.g., Bradley
&
Bryant,
1983;
Vellutino
&
Scanlon, 1987).
Experimental work
on
phonological awareness
is
typically
re-
ferred
to as the
"training studies."
At
least
60 of
these studies
are
described
in the
literature; they usually involve preschool
and
kindergarten children,
and
they
can be
divided into five types:
Investigators either
(a)
train only phonological awareness
and
estimate program effectiveness exclusively
on
phonological
awareness measures (e.g., Content, Kolinsky, Morais, & Bertelson,
1986;
Olofsson
&
Lundberg,
1983;
Rosner, 1974);
(b)
train only
phonological awareness
and use
both phonological awareness
and
beginning reading measures
to
gauge treatment effects (e.g., Bren-
nan
&
Ireson.
1997;
Lundberg, Frost,
&
Peterson, 1988; Olofsson
& Lundberg,
1985;
Schneider, Kuspert, Roth,
&
Vise,
1997;
Torneus. 1984; Weiner. 1994);
(c) use
substitutes
for, or
analogues
of,
letters
in
letter-sound correspondence instruction together with
phonological awareness training (e.g., Cunningham, 1990;
Fox &
Routh,
1984;
Vellutino
&
Scanlon, 1987);
(d) use
real letters
for
letter-sound correspondence training
in
combination with phono-
logical awareness training (e.g., Ball
&
Blachman, 1991; Bradley
& Bryant, 1983);
or (e)
integrate phonological awareness training
with beginning reading instruction (e.g., Hatcher, Hulme,
&
Ellis,
1994;
Wallach
&
Wallach,
1976;
Williams, 1980).
Overall, these researchers' experimental work indicates that
phonological awareness
can be
trained;
the
training
can
produce
a
positive, albeit small, effect
on
reading development;
and its
influence
can be
enhanced when integrated with letter-sound
or
beginning reading instruction. Such findings have
led
many schol-
ars (e.g., Adams,
1990;
Juel,
1988;
Snow
et al., 1998) and
poli-
cymakers (e.g.. Lyon.
n.d.) to
recommend inclusion
of
phonolog-
ical awareness elements
in
reading-readiness instruction before
first grade.
Generalizability
of the
Training Studies
An important
and
obvious assumption
is
that
the
correlational
and experimental training studies represent
a
requisite database—a
critical mass
of
knowledge—to call
for
important modifications
in
preschool
and
kindergarten
(and
first-grade) reading-readiness
programs. Such changes would entail the redesigning
of
preservice
and inservice professional development
and the
redeployment
of
technical assistance money from governmental agencies. Despite
a
clear need
for
action
and a
desire
no
doubt felt
by
many research-
ers
to
provide help—not incidentally proving
the
real-world value
of research
to
those
who
think
it has
little value
(see
Kaestle,
1993)—we must
ask
whether
the
training studies have produced
reading-readiness programs that
are
indeed ready
for
export
to
classrooms. This concern
was
prompted
by an
impression,
ex-
pressed
by
others before
us
(e.g., Blachman, Ball, Black,
& Tan-
gel,
1994;
Byrne
&
Fielding-Barnsley,
1995;
O'Connor, Notari-
Syverson,
&
Vadasy, 1996; Troia, 1999), that most training studies
have been conducted outside classrooms
and
with individual chil-
dren
or
small groups
of
students
or
have been conducted inside
classrooms
but
with research
staff,
rather than teachers, conducting
the training.
How many training studies involved teachers
as
program imple-
menters?
To
answer this question,
we
read qualitative (Troia,
1999)
and
quantitative
(Bus & van
IJzendoorn,
1999)
reviews
of
the training studies
and all
articles referenced
in the
reviews that
(a) explored
the
effectiveness
of
phonological awareness training,
decoding training,
or
both;
(b)
involved preschool
or
kindergarten
children;
and (c)
were published
in
peer-review journals.
We
then
hand searched seven scholarly journals published
for the
years
1985
to 1999,
inclusive, looking
for
similar studies.
We
searched
Child Development, Exceptional Children, Journal of Educational
Psychology, Journal of Experimental Child Psychology, Journal of
Learning Disabilities, Reading Research Quarterly,
and
Scientific
Studies
of
Reading.
The
studies discovered through
the
hand
search, added
to
those found
in the
qualitative
and
quantitative
reviews, totaled
41.
Authors
of
only
4 of
these studies asked preschool teachers
to
train their children
and
conduct
the
experimental program (Byrne
& Fielding-Barnsley, 1995; Haddock, 1976; Rosner, 1974; White-
husrt
et al.,
1994); authors
of an
additional
9
studies, reported
in 8
articles,
did so at
kindergarten (Blachman
et al., 1994;
Brady,
Fowler, Stone,
&
Winbury,
1994; Bus, 1986;
Lundberg
et al.,
1988;
O'Connor
et
al., 1996; Olofsson & Lundberg, 1983; Schnei-
der
et al., 1997;
Tur-Kaspa, Mioduser,
&
Leitner,
in
press).
Fur-
ther,
1 of 4
preschool studies (Byrne
&
Fielding-Barnsley,
1995)
and
6 of 9
kindergarten studies
(Bus,
1986; Lundberg
et al., 1988;
Olofsson
&
Lundberg, 1983; Schneider
et al.,
1997; Tur-Kaspa
et
al.,
in
press) were conducted
in
Australia,
the
Netherlands,
Den-
mark, Sweden, Germany,
or
Israel; that
is,
only
6 of the 13
published phonological awareness training studies with teachers
as
trainers were conducted
in
English with American preschoolers
and kindergartners.
This
is
important
for at
least
two
reasons. First, American
children typically enter kindergarten
1
year earlier than
do
German
children
and
1
to 2
years before many Scandinavian children
(see
IS READING IMPORTANT?253
Lundberg et al., 1988; Schneider et al., 1997). The typical German
and Scandinavian kindergartner, therefore, has probably reached a
relatively advanced level of cognitive development (Lundberg et
al.,
1988). Second, English has a less regular orthography than do
several European languages such as Danish, Finnish, and German.
There are more concerns about the 13 studies. None of the
researchers provided data on the fidelity (or accuracy) with which
teachers implemented the training program; investigators of only 2
studies (Bus, 1986; Whitehurst et al., 1994) randomly assigned
teachers (and their classrooms) to experimental and control
groups; and authors of only 6 studies (Bus, 1986; Byrne &
Fielding-Bamsley, 1995; Haddock, 1976; Olofsson & Lundberg,
1983;
Rosner, 1974; Tur-Kaspa et al., in press) controlled for
Hawthorne effects. Without fidelity information, one cannot at-
tribute (positive or negative) findings to the treatment. Without use
of random assignment and a means of controlling for possible
Hawthorne effects, one cannot dismiss many nontreatment factors
as responsible for study outcomes.
In sharing these observations, we do not wish to be harsh. We
know the many difficulties involved in field-based research. We
are respectful of the effort reflected in these 13 studies, and we
understand that no field-based investigation is perfectly conceptu-
alized and executed. Still, without well-controlled experiments
showing that teacher-implemented programs strengthen children's
reading readiness, the field may not rightfully claim that phono-
logical awareness training has been validated as effective and
feasible for classroom use.
Purpose of the Study
As indicated, many researchers who have implemented training
studies have taught phonological awareness in isolation, rather
than in combination with letter-sound or reading instruction. In
some cases, such a strategy may reflect the cultural norms of study
participants (see Lundberg et al., 1988). However, this approach
also seems to reflect a belief that phonological awareness is a
prerequisite of reading and that phonological awareness training
should be conducted prior to, rather than concurrent with, formal
reading instruction (e.g., Brennan & Ireson, 1997; Lundberg et al.,
1988).
Nonetheless, there is considerable support for combining
the two (e.g., Ball & Blachman, 1991; Bryant & Bradley, 1985;
Byrne & Fielding-Bamsley, 1993; Hatcher et al., 1994; Williams,
1980).
Thus, further exploration of this issue is necessary, espe-
cially with teachers implementing treatments. Therefore, one of
two treatment conditions in the present kindergarten study repre-
sents an integration of phonological awareness training with be-
ginning decoding instruction and practice. We contrasted this
group with (a) a treatment group that participated in phonological
awareness training but not in decoding instruction and practice and
(b) a no-treatment control group.
To draw relatively clear connections between these three groups
and outcome data, we (a) randomly assigned kindergarten teachers
(and their intact classes) within schools to the three study condi-
tions;
(b) collected treatment fidelity data on teachers and students;
(c) observed teachers in all classrooms to describe their general
literacy programs; (d) used well-regarded phonological awareness,
beginning reading, and spelling measures; (e) required the phono-
logical awareness and beginning reading treatments to run 20
weeks and 16 weeks, respectively; and (f) tested participants again
in the fall of the next year. To strengthen generalizability of our
findings, we recruited a relatively large number of schools (TV =
8),
kindergarten teachers {N = 33), and children (N = 404). Half
the schools were Title I (high-poverty) schools, and half served
mostly middle-class children. In each classroom, we collected data
on low-achieving, average-achieving, and high-achieving stu-
dents,
25 of whom had disabilities.
Further, we explored the specificity of treatment effects, that is,
(a) whether children who participate in phonological awareness
training but not in decoding instruction and practice demonstrate
improved performance on phonological awareness measures but
not on reading or spelling measures and (b) whether students in
both phonological awareness training and decoding instruction and
practice show growth on phonological awareness, reading, and
spelling measures. Although such results may seem self-evident on
the basis of research showing that "you get what you teach,"
investigators have not frequently examined the specificity of treat-
ment effects with respect to phonological awareness training and
decoding instruction in kindergarten, especially when teachers
have conducted the interventions.
Finally, this study represents an initial evaluation of Peer-
Assisted Learning Strategies (PALS) in kindergarten. PALS is a
classwide, peer-mediated approach to instruction and practice that
accelerates reading achievement in Grades 2-5 (e.g., Fuchs, Fuchs,
Mathes, & Simmons, 1997; Simmons, Fuchs, Fuchs, Hodge, &
Mathes, 1994) and Grade 1 (Mathes, Howard, Allen, & Fuchs,
1998).
On the basis of this prior work, we wondered whether
PALS might strengthen kindergartners' early reading develop-
ment. Nevertheless, we were uncertain as to whether students as
young as 5 years can make meaningful use of peer-mediated
strategies. So we examined students' (and teachers') fidelity of
PALS implementation as well as their progress on reading-related
outcomes.
Method
Treatments
"Ladders" was our phonological awareness training treatment. "Lad-
ders + PALS" was our phonological awareness training and beginning
decoding approach. Before describing the nature of these treatments, we
provide some context by discussing the reading/language arts programs of
the teachers in the study.
Reading/language arts context. On the basis of one-to-one interviews
with all participating teachers, their completion of a questionnaire, and a
20-week presence in their classrooms, we concluded that multiformity, not
uniformity, characterized what and how they taught. Virtually all teachers
made use of whole-language activities. However, their focus and emphasis
varied (e.g., journal writing vs. shared reading), as did the quality of their
teaching. Nearly two thirds of our teacher sample used the school district's
formally adopted text: Harcourt-Brace Treasury of Literature: First Street
Collection for Kindergarten (Farr & Strickland, 1995). A majority of the
teachers said they made frequent use of First Street's "Big Books." About
half of the teachers reported using High Hat (Goldman & Lynch, 1986)
animal stories and picture cards emphasizing letter-sound correspondence
and blends. Only
1
(control) teacher systematically led her students through
a phonics-based program of beginning reading. We failed to discern
clear-cut differences in teachers' reading/language arts programs between
schools or study groups. The one exception to this conclusion is that a
majority of control teachers taught alphabet letter naming, whereas most
treatment teachers did not. We do not know whether our treatments were
254FUCHS ET AL.
responsible for this difference. Finally, there were no statistically signifi-
cant differences among the three study groups in the number of hours per
week teachers reported devoting to language arts, F(2, 30) = .00, ns, or to
reading, F(2, 30) = .81, ns. For language arts, the mean number of hours
per week reported by the Ladders + PALS, Ladders, and control teachers
was 11.46, 11.55, and 11.55, respectively; for reading, teachers re-
ported 7.00 hours, 5.55 hours, and 7.64 hours, respectively.
Ladders. With Rollanda O'Connor's help, we chose 15 Ladders activ-
ities from more than 80 lessons in her Ladder.s-To-Literacy workbook
(O'Connor, Notari-Syverson, & Vadasy, 1998b). These 15 activities were
chosen for two reasons: to help teachers promote phonological awareness
among their students and to differentiate Ladders from the beginning-
decoding PALS component. Of the 15 activities, 10 were designed to
stimulate word and syllable awareness, rhyming, first-sound isolation,
onset-rime-level blending, and sound segmentation. Six of
the
10 activities
were chosen to promote the blending or segmenting of sounds in
consonant-vowel-consonant words; only 1 activity required manipulations
of printed letters. Teachers conducted the 10 activities three or more times
during 2 nonconsecutive weeks. The remaining 5 Ladders activities were
journal writing, "letter sound of the week," "morning message," nursery
rhymes and poems, and shared storybook reading. Only journal writing and
morning message presented students with printed letters. Each of these 5
activities was conducted at least once per week for the entire implemen-
tation period. All 15 Ladders activities were teacher led, were directed to
the whole class, and required 5 to 15 min each day of implementation. The
maximum time teachers devoted to Ladders each week was 45 min (15
min X 3 days), or 10% of their reading/language arts program. Teachers
conducted Ladders for 20 weeks.
O'Connor, Notari-Syverson, and Vadasy (1996, 1998a) reported that
Ladders improved the early literacy skills of special-needs students and
nondisabled low-income children. However, O'Connor et al. used 25
activities, in contrast to our 15, and their activities deliberately placed
greater emphasis on three-phoneme segmentation and blending tasks.
Ladders + PALS. The PALS component of the more complex Lad-
ders + PALS treatment required children to work in dyads with same-age
peers on as many as 51 PALS lessons. With the Rapid Letter Naming
(RLN) Test, a predictor of future reading performance (e.g., Levy &
Stewart,
1991;
O'Connor, Jenkins, & Slocum, 1995; Torgesen, Wagner, &
Rashotte, 1997;
Wolf,
1991), we paired the highest scoring student in each
class with the lowest scoring student, the second-highest scoring student
with the next-to-last scoring student, and so on. If a teacher believed a pair
to be socially incompatible, the pair was reassigned. Each student in each
pair took a turn as reader (tutee) and coach (tutor). Pairs remained together
for 4 to 6 weeks, at which point the teacher named new pairs. Teachers
trained their students to work productively and cooperatively. Before each
PALS lesson, for about 5 min, the teachers modeled new letter sounds and
sight words, increasing the likelihood that their students would experience
success. Previous research indicates that PALS strengthens the word at-
tack, word identification, reading fluency, or reading comprehension of
first-grade students (Mathes et al., 1998) and second through fifth-grade
students (e.g., Fuchs et al., 1997; Simmons et al., 1994). Teachers con-
ducted PALS three times per week for 16 weeks. Sessions lasted about 20
min beyond the brief teacher-led instruction. This represented 10% to 15%
of teachers' reading/language arts time. Thus, the combined Ladders +
PALS treatment accounted for a maximum of 20% to 25% of teachers'
reading/language arts program. This was more than double the time of the
Ladders treatment. Nevertheless, as already documented, the total amount
of reading/language arts time for students in the two treatment groups was
virtually identical.
In kindergarten PALS, the What Sound? activity (WS) is the first of two
activities. Its purpose is to help students learn the correct sounds of all 26
letters except x. In this activity, the coach points to a printed letter and asks,
"What sound?" The reader responds, after which the coach provides praise
for an appropriate answer or a standard correction for an incorrect answer.
The coach's correction is as follows; "Stop. You missed that sound. That
sound is [letter sound]." The coach again asks, "What sound?" After a
correct response, the coach says, "Good. Read that line [of letters] again."
There are 51 WS lessons printed on separate, 9-in. X 11-in. sheets of paper.
Each sheet contains four lines of upper- and lowercase letters, with six
letters per line. Interspersed among the letters are prominent black stars.
When the pair gets to a star, the coach tells the reader, "Good job!" A new
letter sound is introduced in approximately every other lesson. The order in
which letters are introduced in the WS lessons, as well as the procedure and
layout, of the lessons, was influenced by Carnine, Silbert, and Kameenui
(1990).
The second PALS activity is called What Word? (WW). It requires
children to read aloud sight words, decodable words, and simple sentences.
Reading orally permits a child's partner and teacher to monitor word
identification skill and to present corrective feedback when errors are
made. The sight words are /, the, is, was, on, and has. Teachers introduce
the first of them in Lesson 29 and new ones at an approximate rate of one
every three lessons. On the same lesson sheet containing the WS activity,
the coach points to both newly presented and previously learned sight
words and asks, "What word?" The correction procedure for this activity is
similar to the one used in the WS activity.
Also on the same lesson sheet are decodable words (i.e., words that can
be sounded out with letter sounds practiced in earlier lessons) representing
as many as five word families: at, an, ap, ad, and am. Each letter of each
decodable word is placed in a "sound box" (Elkonin, 1973). The coach
says,
"Read the word slowly." The reader slowly says the letter sounds as
he or she touches the individual sound boxes. Then the coach asks, "What
word?"
Simple sentences (still on the same lesson sheet) are composed of
the sight words and decodable words learned in previous WW lessons.
When the reader misreads a sentence, the coach applies a correction
procedure similar to those used in connection with the reading of sight
words and decodable words in the WS activity. Teachers implementing the
Ladders + PALS treatment were trained and expected to conduct the
Ladders lessons with the same frequency and degree of integrity as did the
Ladders-only teachers.
Control. Prior to the study, control teachers were told they were
members of a control group, and we discussed with them (and with the
teachers in the Ladders and Ladders + PALS treatments) why such a group
was important in evaluating the effectiveness of Ladders and PALS. We
asked control teachers to continue their reading/language arts instruction
and not to conduct Ladders or PALS lessons for the study's duration. We
promised that, following completion of the study, we would give them
copies of Ladders and PALS manuals. Such promises and discussion were
deemed necessary because teachers within a given school had been as-
signed randomly to the three study groups, with the result that many control
classrooms were literally next door to Ladders and Ladders + PALS
classrooms.
School, Teacher, and Student Selection
Schools. We recruited four Title I and four non-Title I schools in the
Metro-Nashville Public Schools system. Douglas Fuchs and Anneke
Thompson discussed the study with principals and teachers of 10 schools;
only 1 school declined participation. Of the remaining 9, we chose the
first 8 that agreed to participate. Median percentages of African American
kindergarten children in the Title I and non-Title I schools were 52% and
21%,
respectively. For the study classrooms only, the median proportion of
children receiving free or reduced lunch was 81% in Title I schools and
29%
in non-Title I schools. The average price of a home in the commu-
nities surrounding the four Title I schools was $53,923, versus $136,309 in
the non-Title I school communities.
Teachers. Douglas Fuchs and Anneke Thompson explained to the
teachers the study's rationale and purpose, the general nature of the
Ladders and Ladders + PALS treatments, and that we wished to randomly
assign them to treatment conditions. Among the eight study schools, 38
IS READING IMPORTANT?
255
of
43
kindergarten teachers volunteered
to
participate.
We
chose
33,
ensuring that 18 taught
in
Title
I
schools and that 15 taught
in
non-Title
I
schools. Title
I
teachers were overselected (i.e., 1 extra teacher
for
each
of
the three study groups)
to
compensate
for an
expected greater student
turnover
in
those schools.
The
33
teachers were assigned
to
study groups
by
means
of
stratified
randomization. Specifically, teachers
in
each
of the
eight schools were
assigned randomly
to
the three study groups. Title
I
teachers
(n =
18)
and
non-Title
I
teachers
(n = 15)
were distributed equally among
the
three
groups such that
6 and 5,
respectively, were assigned
to
each, totaling
11
teachers
per
group.
We
conducted one-way analyses
of
variance
(ANOVAs; Ladders
+
PALS
vs.
Ladders
vs.
controls)
and
found
no
significant differences among groups
on
teachers' class size
or
years
of
teaching experience. We ran chi-square analyses and did not obtain reliable
relations between study groups
and
teachers' age, gender, highest degree
earned,
or
race
(see
Table
1).
Students.
Two
criteria guided
the
selection
of
students
as
study
par-
ticipants:
the RLN
test
and
teacher judgment. Project staff individually
administered the RLN test
to all
students
in
participating classrooms.
The
RLN test
is a
timed measure
of
alphabet letter names,
and as
mentioned
earlier, research indicates that
it is a
respectable predictor
of
future reading
performance. Children with
the
lowest
six
scores
in
each
of
the
33
class-
rooms were designated
by
staff as low achievers (LA). Four students with
scores
in the
middle
of the
distribution were assigned average-achiever
(AA) status.
The
four students with
the
highest scores were labeled high
achievers (HA).
LAs
were overselected because,
in a
pilot study, these
students proved most likely
to
leave before school
was out.
The names
of
the
6
LA,
4
AA, and
4 HA
students were shown
to
each
teacher
for
validation. When
a
teacher
did not
concur with
an
LA, AA,
or
HA rating,
the
child
in
question
was
replaced
by a
student with
the
next
closest RLN score,
and his or her
name was then presented
to the
teacher
for approval.
In
this manner, 379 children were assigned
to an
LA, AA,
or
HA category. Across
the 33
study classes,
an
additional
25
children were
identified
as
special-education students
on the
basis
of a
current individ-
ualized educational plan (IEP). They,
too,
were rated
LA, AA, or HA,
resulting
in a
final study sample
of
404 children.
Two-factor ANOVAs (treatment: Ladders
+
PALS
vs.
Ladders
vs.
controls; student type:
LA vs. AA vs. HA) on
number
of
school
absences,
age, and
report card grades indicated students were compa-
rable across study group. Further, there were
no
significant interactions
between study group
and
student type. However,
two
significant main
effects were identified
for
student type. Follow-up tests revealed that
(a)
HA
students
had
fewer school absences than
did LA and AA
students, whereas
LA and AA
students were similar
in
this regard;
and
(b)
LA
students
had
lower report card grades than
did AA or HA
students whose grades were comparable.
In
addition, chi-square tests
indicated
no
reliable relations between
the
study groups
and
students'
Title
I,
English-as-a-second-language (ESL), special-education,
or re-
tention status,
or
between study groups
and
students' race
or
gender
(see Table
2).
Staff and Classroom-Based Assistance
The four project staff were women.
One
worked
40 hr per
week,
and
three worked
20 hr per
week.
Two
were special-education graduate
stu-
dents,
a
third was
a
graduate student
in
developmental psychology, and the
fourth had
a
master's degree
in
special education. Three were experienced
classroom teachers.
The
number
of
study teachers assigned
to
each staff
person ranged from
5 to 11
(median
=
8.50).
Across the three study groups, staff administered pre- and posttreatment
tests
and
follow-up tests, collected demographic data
on
students
and
teachers, and conducted structured interviews with teachers regarding their
reading/language arts programs. Staff also gathered data
on how fre-
quently,
and
with what degree
of
fidelity, Ladders
and
Ladders
+
PALS
teachers implemented treatment activities.
In
addition, they answered
teachers' questions and occasionally offered suggestions about how best
to
implement the treatments.
In
Ladders
+
PALS classes, staff helped teach-
ers train their students
in
PALS. Staff visited each Ladders and Ladders
+
PALS classroom twice each week
for the
first
2
months
of
treatment.
Afterward, visits were reduced
to
once
a
week.
A
visit usually lasted
about
40 min.
Table
1
Teacher Demographic Data
by
Study Group
Variable
Age
21-30
31-40
41-50
51 +
Class size (No.
of
pupils)
Female teachers
Highest degree
earned
B.S./B.A.
M.EdVM.S.
Ed.S/Ph.D
Race
African American
Caucasian
Other
Teaching experience
(in years)
Ladders
+
PALS
(n
= 11)
M
SD n (%)
3(27)
1(9)
5(45)
2(18)
18.55
1.81
11(100)
5(45)
6(55)
0(0)
1(9)
9(82)
1(9)
15.82
9.96
Ladders
(n=
11)
M
SD
n (%)
1(9)
3(27)
5(45)
2(18)
19.00
1.61
11(100)
4(36)
7(64)
0(0)
1(9)
9(82)
1(9)
15.18
8.64
Control
(»
= ID
M
SD n (%) f
)f(dj)
3.47 (6)
2(18)
1(9)
7(64)
1(9)
19.18
2.36 .31
11(100)
.79(4)
3(27)
8(73)
0(0)
1.50(4)
2(18)
9(82)
0(0)
15.00
8.80 .02
Note. Because
of
rounding, percentages
may not sum to
100. PALS
=
Peer-Assisted Learning Strategies.
"df
=
2, 30.
256
FUCHS
ET AL.
Table
2
Student Demographic Data
by
Study Group
and
Student Type
Variable and student
type
Absences'1
LA
AA
HA
Chronological age
LA
AA
HA
ESLC
LA
AA
HA
IEP classification
Language/speech
LA
AA
HA
Other
LA
AA
HA
Kindergarten report
cardf
LA
AA
HA
Race
African American
LA
AA
HA
Caucasian
LA
AA
HA
Other
LA
AA
HA
Gender (female)
LA
AA
HA
Title I8
LA
AA
HA
Ladders +
M
1.96
2.59
1.28
5.65
5.66
5.61
36.50
38.72
37.75
PALS
(n = 133)
SD
2.31
2.58
1.80
.48
.35
.94
6.26
1.75
7.22
«(%)
4(3)
2(2)
2(2)
7(5)
0(0)
2(2)
0(0)
0(0)
0(0)
14(11)
19(14)
17(13)
31 (23)
16(12)
24(18)
6(5)
4(3)
2(2)
18(14)
22(17)
25(19)
28(21)
22(17)
24(18)
M
2.33
1.69
1.02
5.55
5.80
5.72
35.87
38.56
38.89
Ladders
(n = 136)
SD
2.47
2.08
1.46
.35
.45
.44
5.58
2.05
2.40
n(%)
3(2)
1(1)
0(0)
5(4)
0(0)
0(0)
0(0)
1(1)
0(0)
18(13)
19(14)
18(13)
33 (24)
17(13)
18(13)
6(5)
3(2)
4(3)
21(15)
15(11)
23(17)
32 (24)
20(15)
22(16)
(
M
2.88
2.05
1.45
5.61
5.60
5.68
35.44
37.21
38.18
Control
n = 135)
SD
3.88
3.83
2.19
.38
.35
.37
5.51
3.38
6.56
n(%)
3(2)
0(0)
1(1)
4(3)
3(2)
1(1)
1(1)
0(0)
1(1)
18(13)
15(11)
19(14)
35 (26)
20(15)
20(15)
5(4)
2(2)
1(1)
22 (16)
17(13)
24(18)
28(21)
19(14)
21 (16)
F(df)bF(df)c
.91 (2, 395)
6.66*
(2, 395) 0.99 (4, 395)
.50(2,379)
1.13(2,379) 1.14(4,379)
.94 (2, 349)
8.89**
(2, 349) .50 (4, 349)
.20 (2)
.33(1)
.33(1)
.88 (2)
.33(1)
.64(2)
.60 (2)
.11(2)
.24 (2)
.49 (2)
.90 (2)
.12(2)
.67 (2)
2.00 (2)
.43 (2)
.44(2)
.08 (2)
.36 (2)
.23 (2)
.21 (2)
Note. PALS = Peer-Assisted Learning Strategies; LA = low achievers; AA = average achievers; HA = high achievers.
a F values for treatment main effect.
b F values for student-type main effect.
c F values for Treatment X Student-Type interaction.
d Number of absences in the fourth 6-week period of the school year.
e Number of children who were in the English-as-a-second-language program.
f Kindergarten report card score during the first 6-week period of the school year. Scores range from 0 to 40.
8 Title I status refers to school, not to individual students. For example, the number 28 associated with LA students in the Ladders + PALS group indicates
that 28 such children attended a Title I school. We do not know whether each (or any) of these children received free or reduced lunch.
*p<.01.
**/>< .001.
Measures
RLN. We have discussed how RLN was used to identify student
participants as LA, AA, and HA. It was also used as a pretreatment
variable. RLN assesses the number of letter names a student identifies in 1
min. Students are given a 9-in. X 11-in. sheet displaying 52 letters. Each
letter of the alphabet is displayed in upper- and lowercase, on seven lines,
in random order. Uncovering one line at a time, the examiner points to a
IS READING IMPORTANT?257
letter and says, "Go across the page and tell me the names of as many
letters as you can. Try to name each letter. If you come to a letter you don't
know, I'll tell it to you. It's OK if you don't know the names of many of
these letters. The important thing is to try your best." If the student does not
respond in 3 s, the examiner says the name of the letter and moves to the
next. The number of letters named correctly in
1
min is the student's score.
If the student completes the test in less time, the score is prorated.
Ladders fidelity. To determine how often Ladders activities were used,
monthly calendars were distributed to the teachers. The calendars offered
a reasonable implementation sequence and timeline. The teachers recorded
the activities they implemented and when they did them.
To explore the quality of Ladders implementation, staff assigned teach-
ers a weekly global rating ranging from 1 {poor) to 3 (excellent). The
ratings were designed to reflect (a) lesson clarity; (b) how well the
teacher's presentation jibed with the intent of the lesson; and (c) the degree
to which the students—LA and special-education students in particular—
were engaged. Before these ratings were conducted, staff simultaneously
coded videotapes of three teachers conducting the same activity with
varying levels of fidelity. Complete (100%) agreement was obtained
among all staff on these tapes before the treatment fidelity ratings com-
menced in classrooms.
PALS fidelity. On two occasions, staff used checklists to evaluate the
accuracy with which teachers and students used PALS. The checklist for
the first occasion was developed for the WS activity. It consists of 12
teacher behaviors and 64 student behaviors. Teacher behaviors address
how they begin and end the activity and how they monitor the lesson.
Student behaviors address purposefulness, organization, and accuracy of
implementation.
The second PALS checklist evaluates implementation of both WS and
WW activities. It comprises 13 teacher behaviors and 214 student behav-
iors.
The staff observed three randomly chosen pairs during a PALS lesson.
Pairs with an ESL student, a special-education student, or a student new to
the PALS program were not observed. Observers rotated from one pair to
another after each partner had been observed in the roles of reader and
coach. On both PALS checklists, behavior is scored as "demonstrated,"
"not demonstrated," or "not applicable." Each observation yields an overall
teacher score, an overall (averaged) student score, and a combined teacher
and student classroom score. Scores are derived by dividing the number of
behaviors demonstrated by the number of behaviors demonstrated and not
demonstrated and multiplying by 100. For both the first and second fidelity
checks, two staff members simultaneously observed teachers and students
in four classes (2 in Title I schools and 2 in non-Title I schools) to
determine interrater agreement. For the first fidelity check, agreement was
100%,
99%, and 99% for teacher, student, and teacher-student combined,
respectively. For the second check, respective data were 100%, 100%,
and 98%.
Rapid Letter Sound (RLS). The RLS test, which is based on a measure
developed by Levy and Lysunchuk (1997), assesses the number of letter
sounds a student identifies in 1 min. The student is given a 9-in. X 11-in.
sheet that provides four practice letters and displays, in random order,
all 26 lowercase letters on five lines. The examiner says, "I'm going to
show you some letters. You tell me what sound the letters make. If you
don't know the sound a letter makes, don't worry. What's important is that
you try your best." The examiner then shows the student only the four
practice letters and says, "This letter says Ibl. Your turn. What sound does
it say?" After providing the student with practice, the examiner uncovers
the first line of the test proper and says, "You're doing a great
job.
Now it's
just going to be your turn. Go as quickly and carefully as you can.
Remember to tell me the sounds the letters make. Try your best. If you
don't know a letter sound, it's okay." If the student fails to give a response
in 3 s, the examiner moves to the next letter. After 1 min, the number of
sounds expressed correctly is the score. If a student completes the test prior
to 1 min, the score is prorated. The RLS test was administered at pre- and
posttreatment and at follow up.
Segmentation. This timed, 1-min measure closely resembles the Yopp-
Singer Test (cf. Yopp, 1988) and assesses children's ability to deconstruct
words into component sounds. It consists of 3 three-phoneme practice
words (e.g., dog) and 22 two- or three-phoneme words. The examiner uses
the practice words to help the student understand the task. If the student
cannot produce a single correct sound in at least one practice word, the test
is stopped. For the remaining 22 words, the examiner says, "Say the sounds
in [word]." If the student does not say one correct sound in 4 consecutive
words, the test is terminated. The student's score is the number of correct
phonemes expressed in 1 min. The segmentation test was given before and
after treatment and at follow up.
Word Attack Subtest of
the
Woodcock Reading Mastery Test—Revised,
Form G (Woodcock, 1987). This measure evaluates students' ability to
pronounce pseudowords. It is commonly used (e.g., Torgesen et al., 1997)
because it is widely regarded as a sensitive test of decoding skill and
reading progress (e.g., Felton & Wood, 1992; Juel et al., 1986; Tunmer &
Hoover, 1992). The Word Attack subtest contains 45 nonsense words,
progressing in presentation from most easy to most difficult. The test is
discontinued after six consecutive errors. Students earn 1 point for each
correctly pronounced word. Scores range from 0 to 45. Split-half and
test-retest reliabilities reported in the manual are .95 and .90, respectively,
for first grade. No such information was given for the kindergarten children
in the standardization population. This subtest was given at pretreatment, at
posttreatment, and at follow up.
Word Identification Subtest (Word ID) of the Woodcock Reading Mas-
tery Test—Revised, Form G (Woodcock, 1987). The Word ID subtest is
a frequently used measure of real-word reading ability (Vellutino & Scan-
Ion, 1987; Torgesen et al., 1997). It requires children to read single words
out of context. It consists of 100 words. For each child, we started with
Item 1 and discontinued testing after six consecutive errors. The subtest
was administered and scored in standard fashion. Students earn 1 point for
each correctly pronounced word. Scores range from 0 to 100. Split-half and
test-retest reliabilities are .99 and .94, respectively, for first grade; again,
no reliability or validity data were found for kindergarten children. The
subtest was administered both before and after treatment implementation
and at follow up.
Blending. We created a test consisting of 22 three-sound words (e.g.,
soap, mom, food). On four practice items preceding the test proper, the
examiner says, "I'm going to say some sounds. If you put the sounds
together, they make a word." Then the examiner says, "/C/ /a/ lil. What
word is that?" If the student commits four consecutive errors, the test is
stopped. Otherwise, the examiner records the number of words blended
correctly in
1
min and continues testing until all 22 test items are presented.
The number of words identified correctly in
1
min is the student's blending
score, which ranges from 0 to 22. The blending test was administered at
posttreatment and at follow-up.
Spelling Subtest of the Wechsler Individual Achievement Test. The
Spelling subtest consists of 50 words. To obtain a sufficient sample of
spelling behavior, we administered to every child the first 12 test items
(Items 1 to 6 address letter recognition; Items 7 to 12 ask children to spell
single words). Otherwise, the test was administered in standard fashion:
For each item, the examiner said the word, then said the word in a sentence,
then repeated the word. Students had 10 s to write the word. Scoring was
conducted in accordance with the developmental scoring rubric created by
Tangel and Blachman (1992). We used a developmental score because of
our study participants' age: Whereas many kindergartners are incapable of
traditional spelling, many can write the first letter of a word, or they may
spell a word phonetically. Accordingly, Items 1 to 6 were scored in
standard fashion; but on Items 7 to 50, students could receive a score of 0
to 6, resulting in a total Spelling score ranging from 0 to 270. We created
explicit scoring criteria for Items 7 to 50, reflecting Tangel and Blachman's
(1992) guidelines. Three staff members independently scored 20% of the
protocols, achieving 94% agreement. After scoring was completed by these
staff members, one member rescored 10% of the protocols scored by
258FUCHS
ET AL.
others; agreement
was 96%. The
Spelling subtest
was
administered
at
posttreatment
and at
follow-up.
PALS teacher questionnaire. PALS teachers completed
a
posttreat-
ment questionnaire, which asked them
to
evaluate
the
effectiveness
of
PALS
on a
5-point Likert-type scale
and to
explain
how
they regularly
found time
to
implement
the
activities. Although
the
teachers were asked
to identify themselves
on the
questionnaires,
we
told them repeatedly
(when we handed them the instrument and
on
the questionnaire itself) that
we were interested
in
their candid appraisals
and
opinions
and
that their
individual responses would
not be
shared with anyone other than project
staff.
Procedure
Training. Teachers attended
a
full-day workshop during which
we
discussed phonological awareness
in
terms
of
blending sounds into words,
segmenting words into sounds,
and
rhyming words
to
hear similarities
of
sound.
We
also discussed
the
connections between sounds
and
printed
letters
and
between letters
and
words,
and we
described
the 15
Ladders
activities teachers would
be
asked
to
implement.
We
explained that some
Ladders activities
are
relatively easy (e.g., Guess
My
Word; word sounds
are stretched
in
exaggerated fashion [mmaasskk]), several Ladders activi-
ties
are
more challenging (e.g.,
I'm
Thinking
Of;
requires awareness
of
onset-rime [m-ask]),
and a few
activities
are
more difficult still (e.g.,
segmenting words
at
the phonemic level
[m-a-s-k]).
We demonstrated how,
irrespective
of
difficulty,
all
Ladders activities
can be
presented with
varying degrees
of
teacher support, permitting teachers
to use
each with
virtually
all
students
in the
class.
We
also emphasized
the
importance
of
involving
LA
children
as
well
as HA
children, maintaining
a
brisk-paced
lesson,
and
cycling through
the
activities
to
promote
a
sense
of
freshness
and novelty. Finally, each teacher was given
a
Ladders manual,
and
staff
members
and
teachers carefully reviewed
its
content.
Ladders
+
PALS teachers attended
an
additional half-day workshop
to
prepare them
to
train their students
in
PALS. Teachers' logistical concerns
were addressed
by
providing suggestions
on how to
pair students, assign
seats,
and
schedule activities.
A
proposed time line
for
PALS implemen-
tation was discussed, and teachers were given detailed examples
of
typical
lessons.
To
promote familiarity with these lessons, teachers formed dyads
and, under the direction
of
the
staff,
role played each activity as both coach
and reader. Teachers were encouraged
to
frequently model
for
their
stu-
dents cooperative
and
supportive PALS behavior and, during PALS
les-
sons,
to
walk about the classroom, giving praise
to
student pairs who were
working collegially. Finally, teachers were given
a
comprehensive manual,
written expressly
for
them, that included scripted PALS lessons and acetate
overheads
to
facilitate student training. (We encouraged teachers
to
trans-
late
the
scripts into their own words.)
Fidelity. Fidelity-of-treatment observations occurred weekly
in
each
Ladders classroom between Week
6 and
Week 20, inclusive. PALS treat-
ment fidelity data were collected
in
each classroom
on two
occasions.
In
Week
9,
staff members observed teachers'
and
students'
use of the WS
activity;
in
Week
16,
staff members observed teachers'
and
students'
implementation
of
both WS and WW activities. Guiding these observations
was
a
checklist reflecting important PALS behaviors
for
teachers
and
students.
A
description
of
the checklist
and
related information
is in the
Measures section.
Testing
and
scoring.
All
tests were administered
to
students
on a
one-to-one basis. Testing
in
kindergarten
(pre- and
posttreatment)
was
accomplished
in two
sessions.
The RLN
test
was
administered first
at
pretreatment. Other measures were given
in
random order across children
and pre- and posttreatment testing. Following completion of the treatments,
during
the
second testing session,
the
blending test
was
always adminis-
tered before
the
spelling test.
The
staff was trained
to
administer the tests
according
to
instructions
in a
user's manual.
In the
absence
of
such
instructions, they practiced agreed-on procedures
to
ensure standard