Abstract
This study investigated the effects of a reading intervention that
integrated vowel pattern analysis and children's literature on the
word decoding performance of second graders with reading disabilities.
The intervention evaluated students' abilities to decode a set of
training words using 3 common vowel patterns (syllable types), in
isolation and in context. Additional sets of novel words and nonsense
words using the same vowel patterns were presented to evaluate
generalization of the reading intervention. A multiple-baseline design
across vowel patterns was used to analyze the experimental effects of
the decoding performance of 5 students. All 5 second graders
demonstrated substantial gains in their ability to decode words with the
3 vowel patterns. The students increased their accuracy reading words in
isolation, as well as in context. In addition, all students increased
their decoding accuracy of both novel a nd nonsense generalization
words. These increases in decoding accuracy were maintained during
follow-up observations. The effects suggest the value of integrating
code-based and meaning-based teaching strategies when designing
interventions for students with reading disabilities.
Learning to read is among the most critical academic
accomplishments of childhood, and teaching a child to read is one of the
most important tasks of a teacher. Over the years, the approaches to
reading instruction have varied greatly. Advocacy of different reading
methods has been as much a function of ideology as research (Adams,
1994; Chall, 1989; NRP, 2000; Vellutino, 1991). In recent years, most
educators have come to agree that effective reading instruction should
include elements that teach five critical areas of literacy: phonemic
awareness, phonics, fluency, vocabulary, and text comprehension (NRP,
2000). To deliver effective reading instruction that promotes learning
in these areas, instructional approaches have been designed that
incorporate both meaning-based and code-based instruction (Blachman et
al., 2004; Pressley, 1998; Rasinski & Padak, 2008; Simmons,
Kuykendall, King, Cornachione, & Kame'enui, 2000).
While the logic and creativity of a more balanced approach is
encouraging, two problems to the development of "balanced"
reading approaches exist. First, there is no universally agreed upon
model of balanced reading instruction. Each model of balanced reading
instruction places a different emphasis on the scope and sequence of
skill instruction; indeed there remains tremendous variability in the
visibility and implementation of skill instruction in some models (Lapp
& Flood, 1997; Rasinski & Padak, 2008). In an early review of
balanced instruction, Freppon and Dahl (1998) pointed out that although
the importance of code-based instruction within meaningful text is
accepted universally across reading philosophies, huge differences
existed in the implementation of teaching approaches. Some reading
researchers advocate integrating skills in context and de-emphasizing
explicit instruction of the code (Lapp & Flood, 1997), while others
promote separate and explicit skill instruction (Pullen, Lane, Lloyd,
Nowak & Ryals, 2005).
A second challenge to delivering balanced reading instruction
involves the limited empirical support for the various models of
combined instruction (IDEA, 2002-2004; Pikulski & Chard, 2005;
Pressley, Gaskins, & Fingeret, 2006; Snow, Burns, & Griffin,
1998). Indeed, the very nature of academic research often results in
de-constructing instructional packages to assess the relative
contributions of each component of instruction. While this effort
provides critical information about the value of the individual methods,
it hinders investigations of combined or integrated instructional
packages. This focus on combined methods has increasingly been
recommended by many reading researchers in recent years (Blachman et
al., 2004; NRP, 2000; Rasinski & Padak, 2008).
To date, much of the empirical support for integrated approaches to
teaching reading has been applied by logical extension of sound
investigations of single approaches. For example, there is unequivocal
support for (a) reading aloud to children (Bus, van Ijzendoorn, &
Pellegrini, 1995), (b) enhanced exposure to print materials (Clay, 1991;
Prior & Gerard, 2004), (c) attention to the alphabetic principle
(Blachman et al., 2004; DiLorenzo, Rody, Bucholz, & Brady 2011;
McGuinness, McCuinness, & Donohue, 1995; Ritchey, 2004), (d)
development of phonemic awareness (Foorman et al., 2003; Shaywitz,
2003), (e) development of fluency and accuracy (Good, Simmons, &
Kame'enui, 2001; Katzir et al., 2006; Ming & Dukes, 2008;
Samuels, 2002), (f) the linkage of reading to writing and vocabulary
development (Campbell, Brady, & Linehan, 1991; Nelson & Stage,
2007), and (g) the use of repeated readings (Nelson, Alber & Gordy,
2004; Yurick, Robinson, Cartledge, Lo & Evans, 2006). What has yet
to be studied systematically is how combinations of different
instructional strategies might be integrated to promote reading.
The purpose of this study was to explore a reading intervention
that integrated code-based strategies (vowel pattern analysis) along
with reading-for-meaning elements (children's literature) on the
decoding performance of second graders with reading disabilities.
Consistent with the recommendations of reading researchers, this
investigation incorporated several principles (IRA/NAEYC, 1998; NRP,
2000; Rasinski & Padak, 2008) of balanced instruction, including
designing instruction:
* in a "whole-to-part-to-whole" sequence (i.e., reading
instruction begins with connected text, then proceeds to skill
development, then concludes with skill application in meaningful text);
* with systematic and explicit instruction; and
* where decoding skills are taught within the context of
children's literature.
In this study, the integrated intervention included both
meaning-based and code-based elements, where skills were taught
explicitly and in the context of children's literature.
Research Questions
Two questions were addressed:
1. Will students who receive a reading intervention integrating
meaning-based and code-based strategies increase their reading accuracy
on training words that contain three common vowel patterns, in isolation
and in context?
2. Will students who receive this intervention increase their
reading accuracy of novel (untrained) words and nonsense words that
contain the same vowel patterns as the training words?
Method
Participants and Setting
Five students (three boys; two girls) with reading disabilities
diagnosed by school psychologists were selected from a class in a
private elementary school for students with learning disabilities. All
children were from upper middle-class families, and spoke English as
their primary language. Two students had at least one Hispanic parent,
and all were Caucasian. All students read below grade level; on Woodcock
Reading Mastery Tests-Revised (WRMT-R) (Woodcock, 1987), students ranged
from 9 months to 1 year-8 months below their chronological ages on Word
Identification scores, 8 months to 2 years-1 month on Word Attack
scores, and 8 months to 1 year-7 months on Comprehension scores. Two
students (Laurie and Mike) participated in the special education program
the previous year; none of the other students had been enrolled
previously in special education programs. None of the participants
received medication. A summary of participant characteristics is found
in Table 1.
Six criteria were used to select students for the study. This
included (a) second-grade placement, (b) willingness of parent and child
to participate, (c) regular attendance, and (d) full-scale IQ > 90.
The fifth criterion involved students' phonics abilities. The
phonics criteria included the ability to (a) produce consonant sounds
when shown letter symbols, with at least 80% accuracy, (b) match
consonant sounds to letter symbols, with at least 80% accuracy, and (c)
score less than 50% accuracy reading words that contained two or more of
the three vowel patterns (i.e., Magic e, Double Vowels, and Closed). To
demonstrate the first two phonics criteria, a teacher-made consonant
symbol-sound test was administered individually to each prospective
participant. The student was asked (a) to say the appropriate consonant
sound when presented with the letter symbol, and (b) to point to the
letter symbol when given the consonant sound. To assess accuracy with
the three vowel patterns, the monosyllabic real and nonsense words
subtest of the Decoding Skills Test (Richardson & DiBenedetto, 1985)
was administered. Students were asked to read monosyllabic real and
nonsense words (i.e., words containing short vowels, long vowels with
silent e, and vowel digraphs) that conform to the three common vowel
patterns. The performance of the five students matched against the
phonics criteria is presented in Table 2.
Finally, the sixth criterion required a substantial discrepancy
between students' IQ scores and their word identification scores on
the WRMT-R. Any discrepancy of one or more standard deviations can be
considered evidence of a significant reading disability. All five
participants showed at least a full standard deviation discrepancy
between (a) either their full scale IQs or their verbal IQs, and (b)
their WRMT-R word identification scores (see Table 1).
Dependent Measures
Pre- and post-intervention measures. The WRMT-R was selected as the
pre-test/post-test measure. Form G was used as the pre-test and Form H
was used as the post-test. Three subtests were used for this study: word
identification, word attack, and passage comprehension.
Reading accuracy: Training words. Two procedures were used to
measure accuracy in reading the training words. First, accuracy of
reading words in isolation was measured. To do this, a master list of
150 training words containing 50 of each of the three vowel patterns was
developed. This was called the "training set." Each day, 15
words (5 words representing each vowel pattern) were selected randomly
from the master list by picking individual word cards from large
envelopes containing all possible words. Each student was asked to read
these words as the words were presented individually on 3" x
5" white index cards in mixed order.
Second, reading accuracy in context was evaluated. To do this,
sentence strips containing the "training set" words were
developed. These sentences were taken directly from the books used for
the training set. The target word(s) were written on the back of each
card. Each day, 5 sentences that included training words from each vowel
pattern were selected randomly from large envelopes containing all
possible sentences (a total of 15 sentences). Each student was asked to
read each sentence as it was presented individually on 3" x
11" sentence strips in mixed order.
Reading generalization: Novel words and nonsense words. Two
measures of potential generalization were assessed: accuracy of novel
words and accuracy of nonsense words. First, reading accuracy of novel
words (different than "training set") was measured. To do
this, a master list of 150 novel words containing 50 of each of the
three vowel patterns was developed. Only novel words that did not rhyme
with the "training set" words were selected. Each day, 5 words
from each of the three vowel patterns were selected randomly from the
master list (a total of 15 generalization novel words) from large
envelopes containing all possible words. Each student was asked to read
these words as they were individually presented on index cards in mixed
order.
Second, reading accuracy of nonsense words was evaluated. A list of
150 nonsense words containing 50 of each of the three vowel patterns was
developed. Each day, 5 nonsense words from each of the vowel patterns
were selected randomly from the master list (a total of 15
generalization nonsense words). Each student was asked to read these
words as they were individually presented on index cards, also in mixed
order. The purpose of using novel and nonsense words was to evaluate
transfer and generalization of the three vowel patterns.
To determine the order of presentation, a coin was tossed. Words in
isolation (training set) and novel words (generalization set) were
presented first when "heads" occurred; conversely, words in
context (training set) and nonsense words (generalization set) were
presented first when "tails" appeared. However, training words
were always presented before generalization words.
All the word cards and sentence strips used for daily measurement
were computer generated and then laminated. They were printed with black
ink, and the size, shape, and format were consistent. The index cards
selected for individual words (training words in isolation, novel words,
and nonsense words) were 3" x 5" and white in color. The
training words in context were printed on sentence strips, approximately
3" x 11".
Data Collection
For potential pre- and post-test changes, the word identification,
word attack, and passage comprehension subtests of the WRMT-R were
administered individually to each participant. Form G was used prior to
baseline (pre-test) and H was used during the last week of the study
(post-test). All assessments were administered by a doctoral level
special educator who worked as a reading diagnostician and teacher at
the school.
For the daily reading accuracy measures, a data collector recorded
each word and sentence presented to each student as "correct"
or "incorrect." These data were collected during a morning
homeroom period, approximately 23 hours after the previous day's
reading intervention (i.e., data were not collected during or
immediately after the intervention). Data were collected separately for
training words in isolation, training words in context, novel words, and
nonsense words.
Three individuals served as data collectors including a doctoral
student in special education, a master's student participating in a
student teaching internship, and a paraprofessional employed by the
school. In training sessions conducted prior to the study, each observer
demonstrated a minimum of 80% agreement on each coding category for
training and generalization words, against both other observers.
Intervention
Each day, approximately 40 minutes of group instruction was
provided to all students in the class, including the 5 participants
selected for the study. The instruction consisted of two components:
shared reading of children's literature (i.e., meaning-based
activities with Big Books), and explicit phonics instruction using vowel
patterns (i.e., code-based instruction of syllable types). The time
allocated for instruction was divided evenly between these two
components (i.e., 20 minutes for meaning-based activities, followed by
20 minutes for the code-based activities). The intervention lasted 15
days for the "Magic e" vowel pattern, 11 days for "Double
Vowels", and 5 days for "Closed Vowels." The teacher
remained the same throughout the study, and all of the reading
instruction occurred only during this time. On 6 occasions an additional
investigator observed to assure the fidelity of the intervention. During
these observations the investigator compared the delivery of the
intervention to the sequence of planned activities. On each occasion
there was complete agreement between the planned sequence and the
delivered intervention.
Meaning-based instructional activities. Each day the group
instruction began with a shared reading of a specific story from one of
10 books from the Big Books series (Wright Group, 1998; 1996). Each book
was selected because it contained at least 5 examples of different words
of the vowel pattern targeted for instruction. For example, instruction
on the "Magic e" vowel pattern incorporated the books The
Jigaree and My Sloppy Tiger Goes to School; instruction on the
"Double Vowels" incorporated the books My Boat and Meanies;
instruction on "Closed Vowel" patterns used the books
Grandpa-Grandpa and Mrs. Wishy Washy. Each time a Big Book was
introduced, the teacher followed a three-step activity sequence. First,
the teacher read the story. Second, students were asked to join the
teacher as she read the story again (choral reading). Finally,
individual students who volunteered read short segments of the story
aloud.
Each Big Book was used for a minimum of two consecutive days of
instruction. The first day a Big Book was introduced, the teacher drew a
semantic map of the story on the board. The semantic map served as a
graphic organizer and summarized story grammar (e.g., characters,
settings, problems, and solutions). The teacher used the semantic map as
a medium for guiding students' discussion. The mapping was followed
by guided questioning, which targeted both literal and inferential
comprehension of the story. The second day a Big Book was used, the
teacher prompted students to illustrate parts of the story by drawing
their favorite characters or events from the story. Following the
illustrations, students wrote words, phrases, or sentences about their
illustrations. When a Big Book was used for more than two days, the
teacher drew from the full range of instructional activities.
Code-based instructional activities. Following each day's
shared reading activities, explicit instruction was provided on one of
the three targeted vowel patterns in this study ("Magic e",
"Double Vowels," and "Closed Vowels"). The teacher
provided instruction on a single vowel pattern at a time: "Magic
e" was the first pattern introduced. When the intervention on
"Magic e" was withdrawn, instruction on the "Double
Vowel" pattern began. The "Closed Vowel" pattern was
taught after the "Double Vowel" intervention was withdrawn.
To deliver the code-based instruction, the teacher selected words
from the story in the Big Books, and wrote them on the whiteboard. The
words were written using a black marker for the consonants and the
vowels were written in red. (For example, in the story The Jigaree,
three of the "Magic e" words are take, ride, and home; for the
word take, the a and e were written in red.) Next, the teacher described
the rule for the vowel pattern, applied it to the words on the
whiteboard, and emphasized the position of the vowels within the word.
(For example, "in the word take, the 'Magic e' at the end
gives all its power to the a in the middle, so it can say its long vowel
sound - /[bar.a]/. Since the e gave all its power to the a, the final e
is silent.")
Following the whiteboard activity, the teacher provided each
student with word cards, each of which contained a single word that
incorporated the targeted vowel pattern. These cards included words from
the Big Book used that day, as well as other words from the same word
family or rime (for example, the word take from The Jigaree was one of
the words on the card, and other cards included cake, bake, and snake).
These word cards became the training words in this study, and were the
medium for subsequent instructional activities. To use the word cards,
each student first read the word on the card, and then copied it into
the correct position on a Vowel Pattern Chart, a graphic organizer that
contained the six common vowel patterns or syllable types (Cheyney &
Cohen, 1999); for this study, only three patterns were used. Next,
students traced the vowels in the word on the chart using a red marker,
and made the American Sign Language hand sign for the letter
representing the vowel. On alternate days, students constructed words
related to the Big Book stories (training words selected by the teacher)
using poker chips, where each chip had a single letter, with black
letters for consonants, and red letters for vowels. Whenever the poker
chip activity was used, students followed that with a guided spelling
activity. During guided spelling the teacher drew blank lines on the
whiteboard (black lines for consonants, red lines for vowels), and asked
students to spell the selected word by inserting a letter onto each
blank line. Students who did not participate in the activity at the
whiteboard did this activity on response boards at their desks.
Summary. The intervention included multiple components, and no
effort was made to isolate or evaluate any individual elements of it. To
the contrary, the meaning-based instructional activities were intended
to provide a story context to learning, while the code-based activities
provided explicit instruction of the rules and conventions of regular
vowel use. This integration provided a purpose for learning particular
decoding strategies.
Booster sessions. Additional instruction was provided for two of
the five students (Eddie and Christie) whose progress appeared
"stalled" after several days of the group intervention. To
provide the additional instruction the students met with the teacher for
5-10 minutes of practice on a portion of the group lesson that was
provided earlier in the day. During the additional practice sessions,
the students received individual guidance on the vowel pattern presented
earlier in the story, identified the training words, and explained the
rule that applied to the vowel pattern. This additional instruction was
not different from the group instruction provided earlier in the day,
but provided an opportunity for more practice and individual feedback.
The additional booster sessions occurred between Days 15 and 37 and were
conducted approximately 4 hours after the group instruction. Christie
received 9 booster sessions; Eddie received 15.
Research Design and Data Analysis
A multiple-baseline design across vowel patterns was implemented
for each of the five students in this study. Replication of the
experimental effects within each student was observed for each of the
three vowel patterns. Replication across the five students was observed
when similar effects were obtained only after the intervention was
applied to them. Data were collected daily to measure students'
accuracy when reading words with the targeted vowel patterns in (a)
training words in isolation, (b) training words in context, (c) novel
words, and (d) nonsense words.
Each student's multiple baseline design consisted of three
phases. During Baseline, students participated in group reading
activities that did not include any instruction on vowel patterns. These
activities included small group reading games and independent worksheets
with all participants. This phase included repeated measures of
performance concurrently on the three dependent measures (vowel
patterns). Baseline on the Magic-e words for all participants lasted
from Day 1-6. Baseline for the Double Vowel words lasted from Day 1-22,
and for the Closed Vowel words, baseline lasted from Day 1-33.
During Intervention, the group instruction with the meaning-based
and code-based activities was implemented. The intervention was applied
to one vowel pattern at a time so that the Magic-e intervention was
implemented on Days 7-22. The intervention for the Double Vowel pattern
was implemented from Day 23-33; intervention with the Closed Vowels was
delivered from Day 34-38.
Follow-up probes were conducted to determine whether potential
increases in decoding were maintained. This phase also was implemented
one vowel pattern at a time, after the intervention for each vowel
pattern ended. During the follow-up, there was no further instruction on
the previous vowel pattern. In other words, when the Magic-e
intervention ended, the follow-up observations of Magic-e began, and the
intervention on Double Vowels started. The first follow-up observations
were made during the students' daily reading lessons. Following
these daily sessions, additional follow-up probes were made across a
total of seven weeks, including probes made after a two week winter
vacation. Follow-up observations started on Day 23 for Magic-e words,
Day 34 for Double Vowels, and Day 39 for Closed Vowels.
This study was implemented in a classroom (group) context, with the
results of each student examined individually. Although the traditional,
single subject research standards for implementing phase changes were in
place (i.e., trend, directionality, and stability of the data), not all
of the students' data met those standards simultaneously for both
their words in context and their isolation words. To maintain a group
instruction format, phase changes were implemented based on a majority
trend, so that proceeding from baseline to the intervention for from
intervention to the follow-up) met the traditional standards for at
least 3 of the 5 participating students.
The pre- and post-test scores were analyzed by raw score, standard
score, and age-equivalents for each child. Age-based norms were used to
obtain standard scores. To control for maturation, a derived
age-equivalent gain was calculated by subtracting four months (length of
the study) from the actual age-equivalent gain.
Results
Interobserver Agreement
To establish agreement on the accuracy of observers'
recording, the following procedure was used. While one observer scored
each word, a second observer was positioned between and to the side of
the primary observer and the student, such that she could view each word
card, but could not view the data collection sheet held by the primary
observer. Interobserver agreement checks were calculated for each
category of vowel patterns for both the training and generalization
words. To determine the level of agreement, we used the exact agreement
calculation for each vowel pattern (A / A+D x 100 = %).
On 48% of the study days, at least one student participated in an
agreement check. Agreement checks on individual students ranged from 9%
to 11 % of their observation sessions. Agreement on all vowel patterns
combined averaged 98%, and ranged from 85% for the lowest vowel pattern
(Closed Vowels in novel words for one student) to 100% for 45 of 60
categories (vowel patterns x types of training and generalization words
x students).
Effects on Training Words
The effects of the intervention on training words are presented for
each student in Figures 1-5. Each vowel pattern was analyzed separately
for words presented in isolation and in context. Group mean scores for
each vowel pattern, during each phase of the study, are presented in
Table 3.
The Magic e vowel pattern results indicate a low level of accuracy
during baseline for all five students. Accuracy for words presented in
isolation was consistently lower than for words in context. During
intervention, the overall mean score for words in isolation was 56%
(compared to 13% during baseline), and 70% for words in context
(compared to 38% during baseline). When instruction on the Magic e vowel
pattern was removed, all five students continued to recognize this Magic
e pattern. The mean score for words in isolation was 96%, and 98% for
words in context.
The Double Vowel pattern showed higher but more variable scores
during baseline. When the intervention was applied, the overall mean
score for words in isolation increased from 33% to 81%, and from 54% to
88% for words in context. The change was fairly rapid; gains were
noticeable in most students in three to four days. During the follow-up
condition, all five students continued to recognize the Double Vowel
pattern. The mean score for words in isolation was 95%, and 99% for
words in context.
Accuracy of the Closed Vowel pattern was higher than the other
patterns but inconsistent during baseline. When the intervention was
implemented, the change was rapid, and gains were noticeable almost
immediately. The overall mean score for words in isolation was 98%
(compared to 81% during baseline), and 97% for words in context
(compared to 88% during baseline). Follow-up results continued to be
positive, with mean score averages of 98% accuracy for both types of
training words.
Effects on Generalization Words
The effects on generalization words are presented in Figures 6-10.
Group mean scores for the novel and nonsense (generalization) words each
vowel pattern, during each phase of the study, are presented in Table 4.
During baseline, accuracy of the Magic e vowel patterns for the
novel and nonsense words (generalization words) was low and stable for
all five students. When the students received the intervention on the
training words, the overall mean score for novel words increased to 41%
(compared to 4% during baseline), and 31% for nonsense words (compared
to 2% during baseline). During the follow-up, all five students
continued to recognize the Magic e pattern on the generalization words.
The mean score for novel words was 87%, and 83% for nonsense words.
The Double Vowel pattern was low for most students, and variable
for all of them during baseline. When the intervention was implemented
with the training words, the mean accuracy for generalization words
(novel words and nonsense words) increased to 55%. During follow-up, the
mean score for novel and nonsense words increased to 89% and 88%
respectively.
The Closed Vowel pattern was variable and inconsistent during
baseline. When the intervention was implemented with the training words,
the mean score for generalization novel words increased to 88% (compared
to 55% during baseline), and 82% for nonsense words (compared to 47%
during baseline). Follow-up results continued to be positive, with mean
scores of 93% and 91%.
Effects on Standardized Test Measures
Effects on the three subtests of the WRMT - R (i.e., word
identification, word attack, and passage comprehension) are provided in
Table 5. For the word identification subtest, all five students made
positive gains in all measures (i.e., raw scores, standard scores, and
actual age equivalent). Overall, the mean score gain for the five
students was 12.8 words correct (raw score), 6.4 points (standard
score), 5.6 months (actual age equivalent), and 1.6 months (corrected
age-equivalent). Raw score means increased from 21.4 (pre-test) to 34.2
words correct (post-test), while actual raw scores ranged from 5 to 29
(pre-test) and from 26 to 42 (post-test). Standard score means increased
from 76 (pre-test) to 82.4 (post-test), while actual scores ranged from
61 to 84 (pre-test) and from 77 to 92 (post-test).
Difference scores also were calculated for each student. For raw
scores and standard scores, this figure represents posttest minus
pretest scores. However, to correct for the length of the study, 4
months was subtracted from each student's age-equivalent gain,
resulting in a corrected age-equivalent (age-equivalent difference
figure on Table 5). Differences in raw scores (range of 6 to 21),
standard scores (range of 1 to 16), and actual age-equivalents (range of
2 to 8 months) were notable. Even the corrected age-equivalents (range
of -2 to +4 months) deserve mention considering norms are based on an
average population, and this study included five students with reading
disabilities. Mike demonstrated the most obvious difference score (21
for raw score, 16 for standard score, 8 for actual age-equivalent, and 4
months for corrected age-equivalent), while Eddie displayed the smallest
difference scores (6 for raw score, 1 for standard score, 2 for actual
age-equivalent, and -2 for corrected age-equivalent). Overall, these
gains were substantial given the fact that the words used in the
intervention were monosyllabic and followed regular vowel patterns,
whereas many of the words in this subtest were multisyllabic and
irregular.
For the word attack subtest, all five students made substantial
gains. In fact, the greatest gains were observed in this subtest. The
overall mean score gain was 7.4 words correct (raw score), 7.8 points
(standard score), 8.2 months (actual age-equivalent), and 4.2 months
(corrected age-equivalent). Raw score means increased from 6 (pre-test)
to 13.4 words correct (post-test), while actual raw scores ranged from 2
to 13 (pre-test) and from 7 to 17 (post-test). Standard score means
increased from 79.4 (pre-test) to 87.2 (post-test), while actual scores
ranged from 65 to 93 (pre-test) and from 80 to 94 (post-test).
Differences in raw scores (range of 5 to 14), standard scores
(range of 1 to 15), and actual age-equivalents (range of 5 to 14 months)
were notable. The corrected age-equivalents (range of 1 to 10 months)
are notable considering the reading difficulties of these children.
Individually, Mike demonstrated the most obvious difference scores (13
words correct for raw score, 14 points for standard score, 14 months for
actual age-equivalent, and 10 months for corrected age-equivalent).
Christie showed similar difference gains (9 for raw score, 15 for
standard score, 10 months for actual age-equivalent, and 6 months for
corrected age-equivalent). Laurie displayed the smallest difference
scores (4 for raw score, 1 for standard score, 5 months for actual
age-equivalent, and 1 month for corrected age-equivalent). The word
attack subtest is usually considered the most difficult because it
contains mono- and multisyllabic nonsense words, and students must rely
on their phonetic ability alone to decode them.
For the passage comprehension subtest, overall increases also were
demonstrated. The mean score gain was 6 responses correct (raw score),
3.4 points (standard score), 4.8 months (actual age-equivalent), and .8
months (corrected age-equivalent). Raw score means increased from 11.6
(pre-test) to 17.6 (post-test), while actual raw scores ranged from 8 to
15 (pre-test) and from 15 to 21 (post-test). Standard score means
increased from 78.4 (pre-test) to 81.8 (post-test), while actual scores
ranged from 71 to 87 (pre-test) and from 79 to 85 (post-test).
Differences ranged from 3 to 9 (raw score), from -3 to +8 (standard
scores), and from 2 to 7 months (actual age-equivalents). The corrected
age-equivalent differences ranged from -2 to +3 months.
Discussion
In recent years some reading researchers have called for an
increased focus on combining meaning-based and code-based instructional
methods when teaching reading (Blachman et al., 2004; NRP, 2000;
Rasinski & Padak, 2008). This study combined the elements most
commonly cited as a balanced literacy (i.e., meaning-based and
code-based) intervention (Campbell et al., 1991; Eldredge, 1991; Mather,
1992; Rasinski & Padak, 2008; Ritchey, 2004; Simmons et al., 2000).
Meaning-based reading approaches serve as important motivators to
learning for children, but integrating code-based strategies (e.g.,
vowel pattern analysis) is critical for improving reading acquisition.
The results of this study demonstrate that students with reading
disabilities benefited from a reading intervention that integrated
teaching strategies based on vowel pattern analysis and children's
literature. All five students substantially increased their word reading
accuracy after the intervention was implemented, and their accuracy
generally became much more stable. Positive changes were seen on all
three vowel patterns that were targeted during the instruction for the
training words in isolation and in context. Positive changes also were
observed for the novel and nonsense generalization words. In addition,
the gains were maintained throughout the follow-up observations. In
fact, performance during the follow-up condition actually improved in
some instances, surpassing the accuracy obtained during intervention
(see Tables 3 and 4). This might be attributed to time and practice;
once the intervention for each vowel pattern was withdrawn, the students
continued to read books that included the previously taught patterns.
This additional practice during the follow-up condition might have
allowed students to develop automaticity on words with the targeted
vowel patterns. Overall, the changes in this study occurred within a
relatively short period of time, and because the intervention was
delivered within the context of whole group, class-wide instruction,
these results were encouraging.
One positive result of the intervention was the increase in
decoding accuracy for training words, both in isolation and in context.
This result is understandable since the training words were the
foundation of the intervention. Among the training effects, words in
context (story sentences from the books) were more easily read than the
training words in isolation. We believe that connecting the vowel
pattern instruction to literature provided, in part, a
"purpose" for students to learn these graphophonic skills. For
the words in isolation, the students had to depend on the graphophonic
cueing system alone; in the absence of context cues, acquisition of
training words in isolation was less impressive. Although some research
suggests that students might rely on context cues at the expense of
code-based cues (Nicholson, 2004), other work shows that context can
serve to strengthen reading by providing a purpose for learning the code
(Blachman et al., 2004; DiLorenzo et al., 2011; Freppon & Dahl,
1998; Mudre & McCormick, 1989; Rasinski & Padak, 2008).
In addition to the improvements on the training words, all students
showed increases in decoding generalization (i.e., novel and nonsense)
words, although novel words were recognized more easily than nonsense
words. This too might be due, in part, to the meaning attached to the
words; the novel words were "real" and had real referents.
They were likely encountered in children's oral, if not written,
language experiences and might have been the source of the meaning that
made the difference. The nonsense words, of course, had no meaning.
Applying the vowel pattern strategy to the novel generalization words
might have been aided by the meaning and familiarity of the actual
words. Without word meaning, students were less able to decode the
nonsense words with similar ease. For these words, students had to rely
on the decoding strategies alone. These results support Lovett et
al.'s (1994) findings that strategy training is an effective means
of increasing decoding generalization. Whether taught to use
phonological analysis, blending skills, letter-sound correspondence or
various metacognitive decoding strategies, students in the Lovett study
showed substantial generalization. In the current study, teaching
students to decode patterns of vowels (rather than individual sounds)
formed the basis of our decoding strategy, and was used by the students
to decode the novel and nonsense words. These results are important
given the Lovett et al. (1994) observation that, "generalization
... is necessarily the true test of efficacy for any intervention"
(p.820).
Limitations and Future Research
Despite these encouraging findings, the study has several
limitations. Our effort to deliver the instruction as a whole-group
intervention precluded making experimental changes solely on individual
student performance. As a result, phase changes were made when at least
3 of the 5 participating students met the traditional standards for
behavior analytic studies. Ideally, experimental alterations would have
called for delaying implementation of the intervention, for example, in
instances where a student's baseline data were highly variable. As
a result, the experimental control across all 5 students, 3 vowel
patterns, and 3 experimental conditions was sometimes compromised, and
the results must be considered with this limitation in mind. In
addition, a convention of behavior analytic research is that 20% of all
observations typically have a secondary assessment. Although we
conducted an agreement check on 48% of the study days, the absence of a
secondary observer to conduct an agreement check on at least 20% of each
student's sessions is a clear limitation. Fortunately, this
limitation is mitigated somewhat by high degree of agreement across the
different coding categories.
Another limitation involves the choice of training and
generalization words. Throughout the study, some decoding errors did not
reflect vowel miscues, but rather errors with consonant blends,
digraphs, reversals, and transpositions. For example, Eddie read
"tain" for train, "wish" for with, and
"flo" for foal. This limitation could have been avoided had we
developed a coding system that discriminated vowel pattern errors from
errors related to consonant blends, digraphs, and soft c and g sounds.
Alternately, we could have explicitly taught these graphemes prior to
implementing the study, thus decreasing the probability that errors
would include the non-vowel types.
Because the intervention integrated a number of different
strategies that were either code-based or meaning-based, it is
impossible to cull out the most and least effective elements. This
limits any conclusions about which components of the intervention might
be unnecessary in the delivery of decoding instruction. Although reading
researchers increasingly recommend interventions that bring more balance
to instruction (IRA/NAEYC, 1998; Rasinski & Padak, 2008), there is
much to be learned about the relative contributions of different
instructional approaches. This study did not isolate these
contributions; rather, this study presented a full package of strategies
in a whole-class setting. Given the encouraging results, we hope to
further investigate different elements of the intervention, and urge
others to pursue a component analysis as well.
There are several opportunities for future research. A logical
question involves comprehension. In our study, there was no repeated
measure of story comprehension, although the post-test measure of
comprehension from the WRMT-R suggested some gains. It is unclear
whether interventions that integrate vowel pattern analysis and
children's literature might improve comprehension of the target
stories (a training effect) or other more global measures of reading
comprehension (a generalization effect). Future research should examine
potential co-varying relationships that occur between decoding and
comprehension as a result of these integrated interventions. In
addition, future investigations might incorporate decodable books (books
that contain phonetically controlled text) rather than children's
literature that frequently incorporates irregular words that do not
follow the phonetic code. Also, investigations are needed on ways to
teach other vowel patterns that were not used in this study, apply
syllabication rules to multisyllabic words, and to use vowel pattern
analysis to improve reading fluency. Finally, many reading researchers
continue to search for the most and least effective components of
intervention packages, and this will require multiple component analysis
studies.
In conclusion, this study demonstrated that a reading intervention
that integrated vowel pattern analysis and children's literature
was effective in increasing the decoding performance of young children
with reading disabilities. The increases were observed in training words
as well as generalization words. In addition, these increases were
maintained well after the intervention ended. This intervention offers a
practical and effective approach to reading instruction, that can help
students with reading disabilities conquer the code and master the
meaning, thereby linking the "romance, precision, and
generalization" (Whitehead, 1929) of reading.
References
Adams, M. J. (1994). The progress of the whole-language debate.
Educational Psychologist, 29, 217-222.
Blachman, B. A., Schatschneider, C., Fletcher, J. M., Francis, D.
J., Clonan, S. M., Shaywitz, B. A., & Shaywitz, S. E. (2004).
Effects of intensive reading remediation for second and third graders
and a 1-year follow-up. Journal of Educational Psychology, 96, 444-461.
Bus, A., van Ijzendoorn, M., & Pellegrini, A. (1995). Joint
book reading makes for success in learning to read: A meta-analysis on
intergenerational transmission of literacy. Review of Educational
Research, 65, 1-21.
Campbell, B., Brady, M. P., & Linehan, S. (1991). Effects of
peer mediated instruction on the acquisition and generalization of
written capitalization skills. Journal of Learning Disabilities, 24,
6-14.
Chall, J. S. (1989). Learning to read: The great debate twenty
years later. A response to "Debunking the great phonics myth."
Phi Delta Kappan, 71, 521-538.
Cheyney, W., & Cohen, E.J. (1999). Focus on phonics: Assessment
and instruction. Bothell, WA: Wright Group.
Clay, M. (1991). Becoming literate. Portsmouth NH: Heinemann.
DiLorenzo, K.E., Rody, C.A., Bucholz, J.L., & Brady, M.P.
(2011). Teaching letter-sound connections with picture mnemonics:
Itchy's Alphabet and early decoding. Preventing School Failure
55(1), 28-34.
Eldredge, J. L. (1991). An experiment with a modified whole
language approach in first-grade classrooms. Reading Research and
Instruction, 30, 21-38.
Foorman, B.R., Chen, D., Carlson, C., Moats, L., Francis, D. J.,
& Fletcher, J.M. (2003). The necessity of the alphabetic principle
to phonemic awareness instruction. Reading and Writing: An
Interdisciplinary Journal. 16, 289-324.
Freppon, P.A., & Dahl, K.L. (1998). Balanced instruction:
Insights and considerations. Reading Research Quarterly, 33, 240-251.
Good, R. H., Simmons, D.C., & Kame'enui, E. J. (2001). The
importance and decision-making utility of a continuum of fluency-based
indicators of foundational reading skills for third-grade high-stakes
outcomes. Scientific Studies of Reading, 5, 257-288.
Institute for the Development of Educational Achievement.
(2002-2004). Big ideas in reading. Retrieved June 8, 2007, from
http://reading.uoregon.edu/big_ideas/trial_bi_index.php.
International Reading Association & the National Association
for the Education of Young Children (1998). Learning to read and write:
Developmentally appropriate practices for young children. The Reading
Teacher, 52, 193-216.
Katzir, T., Kim, Y., Wolf, M., O'Brien, B., Kennedy, B.,
Lovett, M., & Morris, R. (2006). Reading fluency: The whole is more
than the parts. Annals of Dyslexia, 56, 51-82.
Lapp, D., & Flood, J. (1997). Where's the phonics? Making
a case (again) for integrated code instruction. The Reading Teacher, 50,
696-700.
Lovett, M. W., Borden, S. L., DeLuca, T., Lacerenza, L., Benson, N.
J., & Brackstone, D. (1994). Treating the core deficits of
developmental dyslexia: Evidence of transfer of learning after
phonologically- and strategy-based reading training programs.
Developmental Psychology, 30, 805-822.
Mather, N. (1992). Whole language reading instruction for students
with learning disabilities: Caught in the cross fire. Learning
Disabilities Research & Practice, 7, 87-95.
McGuinness, D., McGuinness, C., & Donohue, J. (1995).
Phonological training and the alphabet principle: Evidence for
reciprocal causality. Reading Research Quarterly, 30, 830-852.
Ming, K., & Dukes, C. (2008). Fluency: A necessary ingredient
in comprehensive reading instruction in inclusive classrooms. TEACHING
Exceptional Children Plus, 4(4) Article 6. Retrieved [Nov.12, 2008] from
http://escholarship.bc.edu/education/tecplus/vol4/iss4/art6.
Mudre, L. H., & McCormick, S. (1989). Effects of
meaning-focused cues on underachieving readers' context use,
self-corrections, and literal comprehension. Reading Research Quarterly,
24, 89-113.
National Reading Panel (2000). Report of the National Reading
Panel: An evidence-based assessment of the scientific research
literature on reading and its implications for reading instruction.
Washington, DC: National Institute of Child Health and Human
Development.
Nelson, J.,S., Alber, S.R., & Gordy, A. (2004). Effects of
systematic error correction and repeated readings on the reading
accuracy and proficiency of second graders with disabilities. Education
and Treatment of Children, 27, 186-198.
Nelson, R.J., & Stage, S.A. (2007). Fostering the development
of vocabulary knowledge and reading comprehension though
contextually-based multiple meaning vocabulary instruction. Education
and Treatment of Children, 30, 1-22.
Nicholson, T. (2004). Do children read words better in context or
in lists? A classic study revisited. In D. Wray (Ed.), Literacy: Major
themes in education. Reading processes and teaching (Vol 2, pp. 29-44).
Abingdon, Oxfordshire: Routledge.
Pikulski, J. J. & Chard, D.J. (2005). Fluency: Bridge between
decoding and reading comprehension. The Reading Teacher, 58, 510-519.
Prior, J., & Gerard, M.R. (2004). Environmental print in the
classroom: Meaningful connections for learning to read. Newark, DE:
International Reading Association.
Pressley, M. (1998). Reading instruction that works: The case for
balanced teaching. New York: Guilford Press.
Pressley, M., Gaskins, I., & Fingeret, L. (2006). Instruction
and development of reading fluency in struggling readers. In S.J.
Samuels & A.E. Farstrup (Eds.), What research has to say about
fluency instruction (pp.47-69). Newark, DE: International Reading
Association.
Pullen, P.C., Lane, H.B., Lloyd, J.W., Nowak, R., & Ryals, J.
(2005). Effects of explicit instruction on decoding of struggling first
grade students: A data-based case study. Education and Treatment of
Children, 28, 63-76.
Rasinski, T. V., & Padak, N. D. (2008). From phonics to
fluency: Effective teaching of decoding and reading fluency in the
elementary school. Boston: Pearson, Allyn & Bacon.
Richardson, E., & DiBenedetto, B. (1985). Decoding skills test.
Parkton, MD: York Press.
Ritchey, K.D. (2004). From letter names to word reading: The
development of reading in kindergarten. Reading Research Quarterly, 39,
374-376.
Samuels, S. J. (2002). Reading fluency: Its development and
assessment. In A. E. Farstrup & S. J. Samuels (Eds.), What research
has to say about reading instruction (3rd ed., pp. 166-183). Newark, DE:
International Reading Association.
Shaywitz, S. E. (2003). Overcoming dyslexia: A new and complete
science-based program for reading problems at any level. New York:
Alfred A. Knopf.
Simmons, D.C., Kuykendall, K., King, K,. Cornachione, C., &
Kame'enui, E.J. (2000). Implementation of a school-wide reading
improvement model: "No one ever told us it would be this
hard!" Learning Disabilities Research & Practice, 15, 92-100.
Snow, C., Burns. M.S., & Griffin, P. (1998). Preventing reading
difficulties in young children. Washington DC: National Academy Press.
Vellutino, F. R. (1991). Introduction to three studies on reading
acquisition: Convergent findings on theoretical foundations of
code-oriented versus whole-language approaches to reading instruction.
Journal of Educational Psychology, 83, 437-443.
Whitehead, A. N. (1929). The aims of education and other essays.
New York: The Free Press.
Woodcock, R. W. (1987). Woodcock reading mastery tests-revised.
Circle Pines, MN: American Guidance Service.
Wright Group (1998). The story box books [Literature collection].
San Diego, CA: Author.
Wright Group (1996). Sunshine books [Literature collection]. San
Diego, CA: Author.
Yurick, A.L., Robinson, P.D., Cartledge, G., Lo, Y., & Evans,
T.L. (2006). Using peer-mediated repeated readings as a fluency-building
activity for urban learners. Education and Treatment of Children, 29,
469-506.
Author Note
The authors wish to thank Dr. Kyle Bennett of the Center for Autism
and Related Disabilities at Florida Atlantic University for his graphing
skill and effort.
E. Judith Cohen
Florida International University
Michael P. Brady
Florida Atlantic University
Correspondence to Michael P. Brady, Dept. of Exceptional Student
Education, Florida Atlantic University, 777 Glades Rd., Boca Raton, FL
33431; e-mail: mbrady@fau.edu.
Table 1
Participant Characteristics
Student Age Gender Full Verbal WRMT * WID Gates-MacGinitie
(yr-mo) scale IQ **
IQ
Vocab Comp
Eddie 7-8 M 104 121 80 1.2 1.4
Christie 8-2 F 103 104 74 1.5 1.4
Laurie 7-8 F 102 95 84 1.3 1.4
Mike 7-10 M 109 95 61 1.0 K
Ricky 7-11 M 94 102 81 1.6 1.5
* Woodcock Reading Mastery Tests - Revised, Word Identification
Subtest, standard score.
** Cates-MacGinitie Reading Test scores are grade equivalents based on
administration one month prior to baseline.Table 2
Phonics Criteria
Student Produce Match Accuracy Accuracy Accuracy
consonant sounds to magic e double vowels closed words
sounds letters words
Eddie 81% 100% 0% 0% 10%
Christie 95% 100% 0% 0% 30%
Laurie 90% 100% 0% 0% 50%
Mike 86% 100% 0% 0% 30%
Ricky 95% 100% 0% 0% 80%Table 3
Training Word Accuracy: Group Means
Vowel Pattern Baseline Intervention Follow-up
Words in Isolation
Magic e 13% 56% 96%
Double vowels 33% 81% 95%
Closed 81% 98% 98%
Words in Context
Magic e 38% 70% 98%
Double vowels 54% 88% 99%
Closed 88% 97% 98%Table 4
Generalization Word Accuracy: Group Means
Vowel Pattern Baseline Intervention Follow-up
Novel Words
Magic e 4% 41% 87%
Double vowels 22% 55% 89%
Closed 55% 88% 93%
Nonsense Words
Magic e 2% 31% 83%
Double vowels 17% 55% 88%
Closed 47% 82% 91%Table 5
Effects on Standardized Test Measures: Woodcock Reading Mastery
Tests-Revised
Student Word Identification Word Attack Passage Comp
RS SS AE RS SS AE RS SS AE
Eddie
Pretest 22 80 6-9 2 77 6-1 12 83 6-9
Posttest 28 81 6-11 7 83 6-7 16 82 7-0
Difference 6 1 -2 5 6 +2 4 -1 -1
Christie
Pretest 23 74 6-9 2 65 6-1 9 71 6-7
Posttest 33 75 7-1 11 80 6-11 18 79 7-2
Difference 10 1 0 9 15 +6 9 8 +3
Laurie
Pretest 28 84 6-11 13 93 7-0 15 87 7-0
Posttest 42 92 7- 17 94 7-5 18 84 7-2
Difference 14 8 +3 4 1 +1 3 -3 -2
Mike
Pretest 5 61 6-2 3 76 6-2 8 72 6-6
Posttest 26 77 6-10 16 90 7-4 15 79 7-0
Difference 21 16 +4 13 14 +10 7 7 +2
Ricky
Pretest 29 81 6-11 10 86 6-10 14 79 6-11
Posttest 42 87 7-6 16 89 7-4 21 85 7-5
Difference 13 6 +3 6 3 +2 7 6 +2
Note. Time between pre- and posttest = 16 weeks
Standard scores derived from age-based norms.
Difference for RS (raw score) and SS (standard score) = posttest minus
pretest scores.
Difference for AE (age equivalent) = AE gain minus 4 months (length of
study).