Content uploaded by Heidi Anne E. Mesmer
Author content
All content in this area was uploaded by Heidi Anne E. Mesmer on Jun 17, 2022
Content may be subject to copyright.
TEXT DECODABILITY AND THE FIRST-GRADE
READER
Heidi Anne E. Mesmer
College of Education, Oklahoma State University,
Stillwater, OK, USA
This study investigated the effects of highly decodable text and coordinated
phonics instruction on first graders’ word recognition strategies. The quanti-
tative study sought to examine the validity of a major claim about highly
decodable text—that it enables readers to apply phonics instruction to a greater
extent than less decodable text. All participants received the same fourteen-day
phonics instruction. However, participants read either highly decodable or less
decodable text following the instruction. Treatment participants reading highly
decodable text were found to apply letter=sound knowledge to a greater extent
than control participants. They also were more accurate and relied on exam-
iners less for assistance. Treatment and control participants did not differ in
self-correction rates. As a replication of an earlier study, this work suggested
that readers with knowledge of the alphabetic principle, given the same phonics
instruction, will apply it more in a highly decodable context.
Recently, researchers, educators, and policymakers have expressed a
renewed interest in the instructional materials used with beginning read-
ers. Several articles exploring text issues have been written (Brown, 1999;
Cole, 1998; Hiebert, 1999; Johnston, 1998; Mesmer, 1999; Watson, 1997).
The topic ‘‘readers and texts’’ has become an inquiry focus of the Center
for the Improvement of Early Reading Achievement (CIERA), and a great
variety of instructional materials has flooded the textbook market. Due
in part to policies in many states, highly decodable text has become inc-
reasingly visible (California Department of Education, 1996; Texas
Department of Education, 1997; Virginia General Assembly, 2001).
The topic of text decodability is all the more heated within the context
of struggling readers. A number of programs that are typically used with
struggling readers, including those in special education programs, rely on
Address correspondence to Heidi Anne E. Mesmer, Oklahoma State University, College of
Education, 249 Willard Hall, Stillwater, OK 74074. E-mail: Haemesmer@hotmail.com
Reading & Writing Quarterly, 21: 61786, 2005
Copyright #Taylor & Francis Inc.
ISSN: 1057-3569 print
DOI: 10.1080/10573560590523667
61
highly decodable materials (Wilson, 1996). Nonetheless, opinions vary
regarding its use with struggling readers. Some researchers question any
use of highly decodable text (Allington, 2001; Allington & Woodside=Jiron,
1998; Hicks & Villaume, 2001; Routman, 1997), while others insist it to be
foundational to reading acquisition (Fletcher, Francis, & Foorman, 1997;
Groff, 1999; Grossen, 1997; Kameenui & Simmons, 1997; Moats & Hall,
1999; Stein, Johnson, & Gutlohn, 1999). Still others believe that decod-
ability should be considered in choosing instructional materials but should
not serve as the only textual scaffold available to readers (Brown, 1999;
Cole, 1998; Hoffman, 2001; Mesmer, 1999, Hiebert, 1998).
Unfortunately, opinions and hypotheses dominate discussions because only
a small amount of research informs current knowledge. The National Reading
Panel (2000), Snow, Burns, & Griffin (1998), and others (Allington, 1997;
Pikulski, 1998) have noted this gap. In response to calls for study, researchers
have analyzed reading materials (Hoffman, Patterson, Assaff, Sailors, Mast, &
McCoy, 1993; Menon & Hiebert, 1999; Stein et al., 1999) or investigated how
readers respond to texts of varying levels of decodability (Hoffman, Roser,
Paterson, Salas, & Pennington, 2000; Mesmer, 2001b). In 1998, Pikulski pin-
pointed the need for an examination of readers’ responses to text by asking,
‘‘How many studies have been published in which the nature of beginning
reading texts is systematically varied and the effects measured?’’ (p. 30). The
present study responds to this need.
Decodability, like other textual scaffolds, may have some benefits.
However, it may also produce side effects. Ultimately, classroom teachers
need to understand when they might use texts with some level of decod-
ability and when these texts are not useful. This study sought a sophisti-
cated understanding of the use of decodability as a textual scaffold.
Decodability is not a long-term solution: like the training wheels on a
bicycle, it is designed to facilitate future independence.
BACKGROUND
This section will briefly review the current literature. The first portion
examines definitions of highly decodable text, and the following section
describes its theoretical purposes. The third section analyzes studies that
examined the effects of decodable text on readers, and the final section
presents a developmental model for the present study.
What is Decodable Text?
Most define decodable text by the presence of the following two features: a
proportion of words with phonically regular relationships between letters
62 H. A. E. Mesmer
and sounds, and a degree of match between the relationships represented
in text and those that the reader has been taught (Allington & Woodside-
Jiron, 1998; Groff, 1999; Hiebert, 1998; Stein et al., 1999; Willows, Borwick,
& Hayvren, 1981).
Phonic Regularity
In most studies, phonic regularity refers to letter=sound patterns that
are structurally simple, follow common letter=sound associations, and
recur in many different words (Ehri, 1998; Hiebert, 1998; Hoffman, 2001;
Hoffman et al., 1993; Juel & Roper-Schneider, 1985; Menon & Hiebert,
1999). Phonic regularity has been assessed by examining texts using many
methods, such as rating scales, rimes, and bigram versatility.
Rating scales have been used extensively to determine phonic regularity.
The following represent some tools:
1. a three-level rating scale based on Venezky’s (1970) work (Juel & Roper-
Schneider, 1985)
2. an eight-level system (Menon & Hiebert, 1999)
3. The Scale for Text Accessibility and Support-Grade 1 or STAS 1
(Hoffman et al., 1993, 2000).
(See Mesmer, 2001a, for a more detailed discussion of these systems).
Given the many ways of accounting for phonic regularity, several uni-
fying conclusions can be made about what regularity is and how it relates
to decodability. First, regular patterns, be they rimes or bigrams, recur in
many different words. Second, the most regular patterns have one-to-one
letter=sound correspondences (cat, trap, stump) or predictable two-to-one
patterns (seat, boat, blue). Third, highly decodable text is usually not
distinguished by a great number of regular words but instead by the
clustering of words with similar word parts together within specific pas-
sages (Hiebert, 1999; Juel & Roper-Schneider, 1985). Thus, in highly
decodable text, specific words may not be repeated more often, but certain
universal word parts are (rimes, bigrams, and specific patterns).
Lesson-to-Text Match
Currently decodable text is characterized by a degree of match between
letter=sound relationships represented in text and those that the reader
has been taught. Here, this feature is called lesson-to-text match (LTTM),
but in other works it has been called ‘‘instructional consistency,’’ ‘‘wholly
decodable,’’ and ‘‘text related,’’ (Hoffman, 2001; Reutzel & Daines, 1987;
Stein et al., 1999). In most studies, LTTM is expressed as a percentage of
words matching phonics lessons specified in teacher’s manuals. In seven
studies, researchers specifically evaluated this variable (Barr, Dreeben,
Text Decodability and the First-grade Reader 63
& Wiratchai, 1983; Beck, 1981; Beck & Block, 1979; Meyer, Greer, &
Crummey, 1987; Reutzel & Daines, 1987; Stein et al., 1993, 1999). Con-
sistently, they found highly decodable text to contain more words matching
phonics lessons. The purpose of the LTTM feature is to specify what has
been taught, because it can be independently measured much more easily
than learned information.
There are three main limitations of the LTTM feature. First, in previous
studies conceptualizations of LTTM took into account lessons specified in
teacher’s editions despite the fact that actual phonics lessons will differ
(Ball & Cohen, 1996; Pressley, Rankin, & Yokoi, 1996). Like Memer
(2001b), the present research examines actual phonics instruction as it
coordinates with words in text. Second, in previous studies, the nature of
the lessons is not specified even though the type of phonics instruction, use
of modeling, opportunity for practice, and quality of feedback all influence
how ‘‘decodable’’ words are to the reader. For this reason, this study’s
methodology specifies exactly the nature of lessons so that judgments can
be made about the instruction’s potential for rendering words decodable. A
final limitation of the LTTM feature is that no optimal percentage of LTTM
has been established by research (Allington, 2001). A major purpose of this
study is to report the percentage of words coordinated with lessons and to
also report the effects on readers.
Both features of the decodability definition, phonic regularity and LTTM,
contain restrictions. Using the analogy of readability, the limitations of
decodability may be better understood. Readability is established by
examining word- and sentence-level features that are independent of the
reader. However, there is clearly a reader=text interaction that also influ-
ences the readability of a book. This reader=text interaction is based on
individual differences that are difficult, if not impossible, to measure. Like
readability, decodability can be independently but not comprehensively
measured by calculating regularity and LTTM. Just as educators have
accepted readability as a gross measure of text difficulty, so also can they
accept decodability as a gross measure of word level difficulty at specific
literacy stages. As a whole, the decodability construct is functional and
draws on over twenty years of research.
What Are Theoretical Purposes of Decodable Text?
To become proficient, beginners must read great amounts of text, yet they
are significantly challenged by the task of identifying words in print.
Textbook authors have employed a variety of textual scaffolds to assist
readers as they identify words in text. All textual scaffolds, be they con-
nected to decodability or not, represent a manipulation of the reading
material at some level. Materials that do not contain textual scaffolds are
64 H. A. E. Mesmer
written with aesthetic purposes as the primary authorial force. Thus, the
purpose of decodability, like other scaffolds, is to assist the literacy
acquisition process.
Decodability, however, prompts readers to use specific strategies to
identify words. As a textual scaffold, decodability has two purposes. First,
it offers a context through which readers can use letter=sound strategies
(Adams, 1990; Gough, 1997; Stahl, Duffy-Hester, & Stahl, 1998). Theore-
tically, students reading highly decodable text should be able to apply
phonics instruction due to the coordination between letter=sound infor-
mation and words in print (Adams, 1990; Barr, 1972; Bryne, 1991; Gough &
Juel, 1991). In contrast, when students learn isolated phonics rules or
complete worksheets, they are not likely to generalize decoding strategies
to reading. Second, the regularity in decodable text along with the lesson-
to-text match focuses attention on letter=sound strategies. Readers who
apply letters and sounds during connected reading must be attending
consciously to them (Adams, 1990). Decodable text encourages this
conscious attention.
The Effects of Decodable Texts and Instruction on Word
Recognition Strategies
Four examinations of readers’ responses to decodable text can be found in
the literature (Foorman, Frances, Fletcher, Schatschneider, & Mehta, 1998;
Hoffman et al., 2000; Juel & Roper-Schneider, 1985; Mesmer, 2001b). A
quasi-experimental study by Juel and Roper-Schneider (1985) compared
the abilities of two first grade groups (N¼93), one reading in a highly
decodable text and the other in a high-frequency text. Both groups
received the same supplemental, scripted phonics program. The groups did
not differ in their abilities to read words in their own basals, but the de-
codable group performed better in reading pseudo-words at November=
December and February assessments. For this reason, the current study
was timed during the first two-thirds of the first grade year.
In a larger study, Foorman et al. (1998) compared three groups of at-risk
first and second graders (N¼285), one of which received direct phonics
instruction coordinated with decodable text. The other two groups rec-
eived either an embedded or implicit phonics program. Researchers mea-
sured decoding, receptive vocabulary, phonological processing, spelling,
and comprehension. The readers in the highly decodable text group had
superior performance on both the decoding and comprehension measures,
with the differences being most robust in decoding.
A smaller study contrasted readers’ abilities to apply phonics instruction
in a decodable and control context (Mesmer, 2001b). This study eschewed
pseudo-word assessments in favor of decoding abilities used during
Text Decodability and the First-grade Reader 65
connected reading. For fifteen days, all participants received the same
phonics instruction but were assigned to read either highly decodable or
control texts. Students reread texts, and data were collected using running
records. Oral reading errors were rated for the match between the pro-
nounced and written word. The treatment group had more errors in which
two or more phonemes matched the print than the control group.
Hoffman et al. (2000) also examined student behaviors as related to
predictability, decodability, and text-leveling systems. These researchers
rated little ‘‘books’’ for difficulty as well as decodability and predictability.
Participants were introduced to sets of books via three treatments, and
their accuracy, fluency, and reading rates were measured. Participants’
accuracy was positively and significantly correlated with more highly
decodable texts, indicating that decodability may have assisted students in
recognizing words. Participants’ fluency was negatively correlated with less
decodable text. Thus, as words became less decodable, participants were
less fluent.
According to these studies, young readers may be influenced in very
specific ways and at very specific times by text decodability. Readers
seemed to be most influenced during the first portion of first grade, a
period that in fact may be more precisely connected to word recognition
phases. Readers using decodable text also appear to pay greater attention
to letter=sound information, both in a global sense as well during con-
textualized reading. Furthermore, even when readers have been given the
same phonics instruction, those who read highly decodable text appeared
to apply this instruction more than control groups. Interestingly, three of
these studies included significant numbers of English Language Learners
(ELLs). A contribution of the present study is that the sample is different
demographically.
A Model Based on Word Recognition Phases
Ultimately, decodability is a tool, one that can be used correctly or incor-
rectly, appropriately or inappropriately. Having discussed how decodability
may work in the previous section, this section examines when it may work.
As mentioned earlier, the theoretical purpose of decodability is to assist
readers in applying letter=sound strategies. With this purpose in mind, the
question remains: ‘‘When, developmentally, do readers need the most
assistance to apply letter=sound knowledge?’’
The word recognition literature specifies a series of developmental
phases through which readers pass as they become automatic. In these
stages, readers are distinguished by the extent to which they apply the
alphabetic principle. Prealphabetic readers do not use letter=sound cor-
respondences to recognize words (Ehri, 1991, 1994, 1995, 1998; Ehri
66 H. A. E. Mesmer
& McCormick, 1998). Partial alphabetic readers understand the alphabetic
principle and gradually apply their knowledge by using beginning or final
consonants. As readers move into full alphabetic reading, they connect all
letters in a word with the phonemes that the letters represent. At the
consolidated alphabetic phase, readers internalize multiletter units and
patterns. Finally, in the automatic phase, readers retrieve most of their
words by sight and are free to focus most of their attention on meaning.
Once readers understand and begin to apply their alphabetic knowledge
at the partial alphabetic stage, they often struggle to fully apply it in the
next phase. However, moving students into the full alphabetic phase is a
crucial goal for first-grade teachers (Ehri, 1998). Highly decodable text
may be particularly useful in helping young readers transition from the
partial alphabetic phase into full alphabetic reading. At this transition,
readers know enough about letter=sound correspondences to begin app-
lying them but not enough to handle the full range of complexities in
English. In addition, they need help to decode vowels, the most complex
letter=sound relationships in English.
PURPOSE AND QUESTIONS
The study had the following four major purposes:
1. to examine the theoretical claims about text decodability
2. to identify students who would benefit from highly decodable text
3. to contrast reading behaviors of groups (accuracy, self-corrections, need
for assistance, repetitions, and insertions)
4. to make improvements to an earlier study.
The study’s design chiefly investigated whether or not use of highly
decodable text resulted in a greater application of letter=sound informa-
tion. Both treatment and control groups received the same phonics
instruction, but the treatment group read highly decodable text and the
control group read less decodable text. The essential question was, ‘‘When
treatment and control groups were given the same phonics instruction,
did treatment readers apply this instruction more as they read a highly
decodable text than control readers who read less decodable text?’’
This major query was examined through an outcome variable in which
readers’ oral reading errors were rated for their match to the printed sti-
muli. Each error was assigned a point for each phoneme matching the print
in the correct sequence. For example, the word ‘‘ship’’ has three phonemes.
The error ‘‘lip’’ has two corresponding phonemes and would receive a two
rating; the error ‘‘show,’’ a rating of one; and the error ‘‘got,’’ zero. Errors
Text Decodability and the First-grade Reader 67
with two or more phonemes matching the print were called graphically
similar. This division corresponds with the developmental juncture of the
model. A reader who correctly pronounces zero or one phonemes is most
likely a pre- or partial-alphabetic reader using either no alphabetic infor-
mation or only initial sounds. A reader pronouncing at least two or more
phonemes correctly is more likely to move into the full alphabetic phase by
using more than simply initial or final sounds. The treatment group was
expected to have a higher average number of errors with two or more
phonemes matching the print.
The second purpose of the study was directly related to the type of
student who might benefit from highly decodable text. This purpose was
exploratory in nature but deemed vital given state mandates for the use
of highly decodable text in first grade and the preponderance of highly
decodable text in programs for struggling readers. Using a statewide screen-
ing measure, the study selected students who were beginning readers
demonstrating an understanding of the alphabetic principle and some
knowledge of sight words. Students who did not demonstrate these skills
at a minimal level or who had clearly surpassed beginning reading stages
were not selected.
The third purpose was to build on earlier work by contrasting reading
behaviors (Hoffman et al., 2000). Through data collected on running
records, the researcher examined word accuracy, the number of self-
corrections, repetitions, and the number of words provided by the assessor.
This query was related both to participants’ autonomy as they read highly
decodable text as well as the possible side effects of decodability. Treat-
ment participants were expected to be more autonomous as evidenced by
higher percentages of word accuracy, fewer appeals to the examiner to
pronounce words, and higher self-correction rates. As a side effect, they
were expected to repeat more often due to their increased focus on
decoding.
Finally, in three crucial ways, this research improved upon a previous
study. First, it was timed during the first one-third of the first grade year,
when previous researchers had found highly decodable text to have its
most pronounced effects. Second, the decodable and control texts were
more uniform because the same publishers produced both. Third, a new
scale, containing eight levels instead of three, was used to measure phonic
regularity. The following questions guided the study:
1. When compared to control participants receiving the same phonics
instruction, did treatment participants have more errors in which two or
more phonemes matched the printed word?
2. When compared to control participants receiving the same phonics
instruction, did treatment participants have higher percentages of word
68 H. A. E. Mesmer
accuracy, fewer occasions in which they relied upon examiner assistance
to pronounce words (tolds), more self-corrections, and more repeated
words or phrases?
METHODS
Sample
One school in a metropolitan area in the southeast participated in the
study. The school belonged to an urban district with approximately fifty
elementary schools. The school was labeled a ‘‘model’’ school by the
system, and many parents petitioned to have their children attend the
school. This fact made the population of students at this school different
than other schools, with higher poverty rates. The population at this school
consisted of 47% African-Americans, 51% Caucasian, and 2% other; 26% of
the students received free or reduced priced lunches.
Students from three first-grade classrooms participated in the study. In
these three classrooms, teachers shared similar practices. Teachers used
the morning message, shared reading, journaling, writer’s workshop, and
independent reading. Teachers also used a variety of reading materials,
including literature, big books, and ‘‘little books,’’ but did not make great
use of highly decodable text. Prior to the intervention, the researcher met
with the teachers and discussed the purpose of the study. Teachers agreed
to the intervention and were eager for the additional small group instruc-
tion that their students would receive.
Fifty-one participants were screened for this study and 23 participated
in the study. Participants had summed scores between 22 and 44 on the
statewide screening measure. Seventeen students had scores that were too
high to qualify for the study, five students had scores that were too low, one
student discontinued due to excessive absences, and five students were
eliminated due to the need to maintain equal numbers across classrooms.
Those who participated in the study understood the alphabetic principle,
could apply this to initial and final consonants, and had acquired some sight
word knowledge. They had neither begun to fully analyze words nor had
spellings that included many vowels.
The sample consisted of ten females and thirteen males. Eleven parti-
cipants were Caucasian and twelve were African-American. An equal
number of children from each classroom participated in the study (eight
per classroom), with the exception of one classroom, from which one
participant was lost due to excessive absences. In each classroom, the eight
participants were divided into two small groups, one receiving the treat-
ment and one the control. The intervention took place outside of the
regular classroom, with the researcher meeting separately with treatment
Text Decodability and the First-grade Reader 69
and control groups. A total of six small groups existed across the three
classrooms.
Materials
The Phonics Readers Levels 274 by Sundance Publishers were chosen
as the treatment text because they were highly decodable on three
measures—LLTM, regularity, and number of syllables. This pronounced
decodability made them especially viable for a short-term intervention. The
control books included the Alpha Kids Readers Levels 275, also published
by Sundance publishers. These texts could be best described as ‘‘little
books.’’ They contained short sentences, many high-content words, close
print7picture match, natural language, and some predictable elements.
Analysis of Decodability
During the total intervention, each group read fourteen books. The
content words in the treatment and control text were compared on LLTM,
phonic regularity, total number of running words, number of syllables, and
repetitions. Content words are nouns, verbs, adjectives, and adverbs,
whereas structure words are pronouns, determiners, quantifiers, preposi-
tions, intensifiers, coordinate and adverbial conjunctions, conjunctive
adverbs, relative pronouns, and auxiliary and linking verbs (Clark & Clark,
1977). Content words were counted as opposed to structure words because
of the significant differences between these word-types, both in purpose
and in frequency. To understand the essence of a text, readers must
recognize these content words. For each feature, words were analyzed and
book set means compared using t-tests (a¼.05).
In the areas of LTTM, specific phonic regularity categories, and number
of syllables, the highly decodable and control texts differed significantly
(see Table 1 for specific decodability statistics). Note that the LTTM takes
into account only those sounds taught during the research intervention.
Procedures for measuring the LLTM followed those used in earlier
studies (Mesmer, 2001b). Each content word was analyzed for its match
to the letter=sound focus of lessons delivered by the researcher. For each
book, an average LTTM was obtained by comparing the number of content
words matching lessons to the total number of content words in the
book. For example, every short =e=word that appeared in the text when
the short =e=was the focus of the lessons was counted as matching. Later,
when the sound focus shifted to short =a=,bothshort=a=and short =e=
words were counted as matching. The treatment text contained about
40% of words matching the lessons, whereas the control text contained
only about 8%.
70 H. A. E. Mesmer
Phonic regularity was evaluated using the scale developed by Menon &
Hiebert (1999). This scale included eight levels and was chosen because of
its specificity in categorizing words. Level 1 words included short one- and
two-letter words, and Level 2 words contained short vowel c-v-c and v-c
patterns. Level 3 words contained clusters (tr, str, sl) and digraphs (th, sh,
ch) but not words ending in r, l, or gh. Level 4 words contained the silent
e pattern, and Level 5 words vowel digraphs (ea, ee, oa). Level 6 words
contained r-controlled words and words ending in 11. Finally, Level 7 words
included diphthongs (oi), and Level 8 words were multisyllabic.
The mean percent of words per book was reported for each word type.
The treatment and control books did not differ significantly at all levels of
phonic regularity. The most notable differences occurred at Level 2 (short
vowels): as might be expected, the decodable text contained many more
Level 2 words in comparison to the control text. The decodable text also
contained fewer silent e, vowel digraph, and diphthong words.
The words for the two sets of books were analyzed for the number of
syllables each contained. For each book, an average number of syllables per
content word was measured. As was found in previous studies, all book sets
TABLE 1 Decodability Statistics for Treatment and Control Books
Treatment Control
Feature MSDMSDT p
Percent words matching lessons=book
(LLTM)
40.3 (16.3) 7.7 (2.11) 6.5 .0000
Number of content words=book 87.9 (34.3) 52.5 (29.5) 2.8 .009
Number of total words=book 122.5 (45.8) 92.5 (42.9) 1.7 .07
Number of repetitions=word=
book
4.4 (0.9) 3.1 (0.7) 3.8 .0009
Number of syllables=word 1.1 (0.3) 1.3 (0.0) 2.5 .01
Percent of content words=book in following categories:
Regularity 1 (1 & 2 letter words) 5.1% (5.9) 2.5% (3.9) 1.3 .19
Regularity 2 (short vowel o-v-c) 54.2% (18.5) 15.4% (12.7) 6.3 .00
Regularity 3 (blends &
diagraphs)
18.1% (14.0) 10.4% (2.9) 1.6*.13
Regularity 4 (silent e) 2.1% (2.8) 12.1% (13.3) 2.6*.02
Regularity 5 (vowel digraphs) 0.0% (0.3) 12.8% (10.2) 4.3*.00
Regularity 6 (r- and l-controlled) 8.7% (8.0) 4.5% (3.7) 1.7 .09
Regularity 7 (diphthongs) 5.0% (4.1) 12.9% (10.7) 1.6*.03
Rather than reporting Level 8, the percent of multisyllable words, the table reports the aver-
age number of syllables per word. Thus, Level 8 is not reported, and the percentages for Levels
177 do not collapse to 100.
*indicates that a test for unequal variances was conducted.
Text Decodability and the First-grade Reader 71
differed significantly on the number of syllables per word, with the decod-
able books containing fewer syllables per word than the control books.
Number of Content Words, Running Words,
and Repetitions
The number of words contained in a text can place demands on the
reader, and for this reason affect the difficulty level of the text. Therefore,
each set was compared on the number of content words per book set.
When the number of content words was compared, the text sets differed
with the control books, on average containing about 52 words compared to
the 87 words in the decodable text. Depending on one’s perspective, this
difference might favor either the control or decodable readers. Assuming
that the reading of predictable, less decodable texts is usually driven by
memory, the depressed number of content words in the control set would
theoretically have enabled the control group to read more accurately
because they would have had to remember fewer words. From this per-
spective, the difference favored the control group. However, the fewer
number of content words could have disadvantaged the control group due
to the less amount of time spent reading and the fewer opportunities to
practice reading words containing the patterns taught.
For this reason, a second analysis was conducted comparing the total
number of running words per book as opposed to only content words. In
this analysis, the two book sets did not differ significantly. Collectively,
fourteen highly decodable texts contained exactly 1724 words, and the
control books exactly 1298 words. Within-set books varied widely in num-
bers of words, as indicated by the high standard deviations for both total
number of content words and total number of running words.
Finally, each word within each book was analyzed for the number of
times that it was repeated. The book means were than converted to group
means for both decodable and control sets, yielding an average number of
repetitions per word per book. On average, words in the highly decodable
set were repeated one time more than those in the control set. Although
repetition can characterize both highly decodable and less decodable text,
repetition in highly decodable text appears to be higher.
Design and Procedures
Due to the numerous questions about highly decodable materials as well as
calls for the manipulation of text, this study had an experimental design.
The independent variable was text, either a highly decodable text or a less
decodable text. The dependent variables included the degree of application
of letter=sound information, word accuracy, percentage of errors self-
corrected, number of repetitions, and number of words provided by the
72 H. A. E. Mesmer
researcher. Both treatment and control participants received this phonics
instruction. The phonics instruction intervention lasted fourteen days, and
each participant spent twenty minutes per day in this instruction. Although
all participants received the same phonics instruction, this instruction took
place within their prospective treatment groups, outside of the regular
classroom. Following each lesson, participants read either highly decodable
or less decodable text, depending on group. Maintaining internal validity
was a primary concern in the study, and the following procedures increased
control:
1. selecting students from three different classrooms to control for teacher
effects.
2. balancing the number of treatment and control participants within each
classroom.
3. counterbalancing the times during the day that groups met for lessons.
4. selecting participants within a specific score range to decrease the
sample’s heterogeneity.
5. controlling the phonics lessons.
Preintervention Procedures
Students were assessed by their classroom teachers using a statewide
literacy screening measure. Administered to first graders in fall and spring
of each year, this screening gives teachers diagnostic information about
their students and identifies students who are at risk for reading failure. All
students in the first grade are administered a twenty item, preprimer word
list and a spelling test. The spelling assessment approximated students’
applied knowledge of letter=sound information. Students are assigned
points for correct letter=sound representation in words based on various
features of the English language (e.g., initial consonants, final consonants,
blends, digraphs, short vowels, and long vowels). In addition, a bonus point
is assign if the word is spelled correctly. The total number of points that a
student can earn for the beginning of the year spelling is 57. The raw
scores for both the word list and spelling list are summed. If the summed
score is below 28, then other diagnostic measures are administered,
including phonological awareness, concepts about print, and letter
knowledge. Readers with scores at 28 or below qualify for additional
assistance.
Within classrooms, participants were randomly assigned to treatments.
Four participants from each classroom were in the treatment group and
four were in the control group. When the total number of qualifying par-
ticipants within a classroom was not equal to eight, participants were
ordered by their summed scores alternately eliminated from the top and
then the bottom of the score range until only eight participants remained.
Text Decodability and the First-grade Reader 73
Intervention Procedures
The researcher delivered the daily twenty-minute lesson. The lesson
components were ordered to scaffold participants’ alphabetic learning. The
lesson was based on a model called ‘‘Hear-it, Read-it, Spell-it,’’ with which
students first practiced hearing a target vowel sound with phonemic seg-
mentation activities, practiced reading that sound in words, spelled the
sound in a dictated word, and then read books. For the treatment group,
the words in the books contained the target sound, and for the control
group, there was no control on words in text (See Appendix A, Lesson
Matrices).
Instruction within each activity progressed from teacher modeling to
group guided practice to independent practice. During independent prac-
tice, each participant individually completed one to two items within a
given activity (e.g., reading words, segmenting words, or reading pages).
The researcher gave participants specific, corrective feedback for inde-
pendent practice items. When items were not correct, the researcher asked
a scaffolding question. If the student was not successful after the scaf-
folding question, then the researcher modeled the correct procedure and
asked the student to repeat. If the student was still not successful, then the
researcher completed the item with the student.
Phonemic Segmentation
The phonemic awareness activities specifically focused on segmenting
words because this skill is crucial for later decoding (Muter & Snowling,
1998; Nation & Hulme, 1997; Snow et al., 1998). In addition, phonemic
segmentation corresponds with the skills that a full-alphabetic reader must
have in order to be successful (Ehri, 1998). The researcher first demon-
strated how words could be stretched out and then modeled a segmenting
procedure for tapping the table each time a new sound was heard. After the
word was segmented, it was pronounced completely. Toward the end of the
lesson sequence, when students were learning both short e and short a,
they were asked to differentiate between two words containing these
sounds using the methods described above. The Lesson Matrices detail
the specific words used for modeling, guided practice, and independent
practice.
Reading Words
During this part of the lesson, a strategy for blending sounds together in
words was modeled. The process for reading the words involved three
parts. Covering the last letter in the word, participants articulated the first
and second letter=sounds and then blended these together. After blending
these two sounds, they added the ending sound to the word and finally
articulate the blended word.
74 H. A. E. Mesmer
Dictated Writing
At the end of each lesson, participants wrote one word containing focus
sounds. During this segment of the lesson, the researcher said the word in
isolation and then used the word in a sentence to ensure participants’
familiarity with it. After this point, the researcher stretched the word out,
slowly enunciating each sound, and then asked the participants to write the
word. While the participants wrote the word, the researcher alternately
said the word in a standard fashion and stretched the word. All of the
writing was recorded in little journals.
Reading a New Book
The last segment of the lesson included introducing and reading a new
book. Prior to reading the book, each participant was asked to look
through the book at the pictures, thinking about the content of the story.
After this, a conversation followed, posing possible predictions about the
story. During the first reading of the book, the entire group was asked to
read the book chorally. This independent reading gave participants prac-
tice in applying the skills that they had just learned. Depending on the
group’s success, the researcher read along with the participants or stayed
silent. At the second reading, each child was asked to volunteer to read a
page of the book.
Approximately 90% of the time, all participants volunteered to read a
page. When a child did not volunteer, the researcher asked the participants
to read a page and volunteered to read along with him=her. At points when
a participant could not supply the correct word, the researcher waited
three seconds and then supplied the word.
Rereading
Before each lesson, one participant from each group individually
reread the book from the previous day with a graduate assistant. Each
participant orally reread a familiar book every fourth day. During this
time, oral reading errors were collectedusingrunningrecordproce-
dures (Clay, 1993). To consistently collect types of oral reading errors,
standard procedures were established. Because ‘‘Tolds’’ were counted
as errors, participants were given ample opportunity to figure a word
out and encouraged to use strategies to help them. Participants were
told a word if after two-second wait, a generic prompt, and then a five-
second wait they could not correctly pronounce it. The running
records and the lessons were not performed by the same person to
avoid any bias in providing words to students. Repetitions, insertions,
omissions, and substitutions were denoted using standard markings
(Clay, 1993).
Text Decodability and the First-grade Reader 75
Post Intervention Procedures
After the intervention was completed, researchers coded the data. First,
the average word accuracy rate for each running record was calculated.
Insertions, omissions, substitutions, and tolds were all counted as errors.
Repetitions were not counted as errors because, in many cases, readers
repeated words or phrases to monitor themselves. Second, researchers
rated each substitution for its graphic similarity as described earlier. After
rating each substitution, a count was tabulated of the number or errors
with ratings of 071 and those greater than or equal to 2. The number of
self-corrections also was tabulated. Finally, any running record with 85%
word accuracy was not used in this analysis. This resulted in two pieces of
data, one from each group, being omitted.
FINDINGS
Analysis
Due to the design and two-sample comparison, Student’s t appeared to be
the most appropriate path for data analysis. However, the assumptions of
this test were examined, and an inspection of the data indicated that t-tests
would not be sufficient for several reasons. The distributions of many of the
sampling means were not normal and often moderately skewed. The t-test
is relatively robust to a violation of the normal distribution assumption if
the total sample is over 40, which it was not. In addition, the sample means
did not have equal variances. The t-test is also robust to this violation when
sample sizes are equal, which they were not.
For these reasons, the non-parametric equivalent to the t-test was the
Wilcoxon Rank Sum Test. In a recent comparison of Student’s t-test and
this test, the Wilcoxon better maintained Type 1 errors at the nominal
rate for non-normal distributions (MacDonald, 1999; Zimmerman, 1994). In
addition, with non-normal distributions, the Wilcoxon gave better power,
an advantage that was increased when sample sizes were unequal
(MacDonald, 1999). The Wilcoxon places the scores of each participant in
rank numerical order while retaining the group identity (treatment of
control). Then the ranks are summed by group. Higher rank sums indicated
higher overall scored. As specified in APA, the sum of ranks (W)are
reported along with related pvalues. In addition, zscores are reported.
The baseline equivalence of participants was established using the sta-
tewide assessment described earlier. Table 2 shows the mean scores and
standard deviations on the spelling assessment, word list, and the summed
scored. This information is reported to make the findings more under-
standable. However, the means were not analyzed parametrically: neither
the word list scores (treatment W¼111.5, control W¼119.0, p¼.93) nor
76 H. A. E. Mesmer
the spelling scores (treatment W¼107.5, control W¼123.5, p¼.87) nor
the summed scores (treatment W¼104.5, control W¼126.5, p¼.71)
differed significantly. On average, participants had made some inroads into
reading and knew some sight words but did not appear to be full alphabetic
readers.
The major question of the study related to the degree to which alpha-
betic information was applied during text reading. This was measured by
the GSE variables. The GSE 1 represented substitution errors in which
zero or one phoneme matched the printed word, and the GSE 2 category
represented substitution errors in which two or more phonemes matched
the printed word. Total GSEs or Total Substitutions represented the combi-
nation of these categories. Participants had an average of 3.5 running records.
Table 3 shows the data for the aforementioned categories.
TABLE 3 Sum of the ranks, z scores, and p values for GSE < 1, GSE > 2, and Total
GSEs categories
Treatment Control
N¼11 N¼12
Mean errors with 071 rating (GSE <1) 4.4 (5.4) 6.2 (5.9)
W(sums of the ranks)*94.5 136.5
z1.0
Mean errors with 2þrating (GSE 52) 5.2 (2.3) 3.1 (1.9)
W(sums of the ranks)
{
139.1 91.5
z2.0
Mean total substitutions (total GSEs) 9.6 (6.9) 9.3 (6.8)
W(sums of the ranks)
{
112.5 118.5
z.14
Mean errors represents the mean number of errors in a given category over 3.5 points of
measurement per participant.
*pvalue ¼.26.
{
pvalue ¼.03.
{
pvalue ¼.88.
TABLE 2 Participant’s Mean Scores on the Word List and Spelling Assessments
Word List
Max ¼20
Spelling
Max ¼57
Summed score
Benchmark
score ¼28
MSDMSDMSD
Treatment 12.9 (5.12) 19.7 (5.3) 32.6 (6.28)
Control 12.5 (4.17) 20.4 (4.6) 33.0 (6.46)
Text Decodability and the First-grade Reader 77
Essentially, the data showed that the groups had the same total number
of substitution errors as indicated by Mean Total Substitutions. However,
the treatment group had significantly more errors in the GSE 2 category,
with a rating of 2 or greater. The reverse was expected for the GSE 1
category, with ratings of 071, but this was not the case. Therefore, the
answer to the first question, ‘‘Do treatment participants have more errors in
which two or more phonemes match the printed word?’’ is yes. By this
outcome variable, treatment participants applied letter=sound information
to a greater extent than control participants.
How did readers in the treatment group compare to control readers in other
behaviors? Were they more autonomous as they read? Table 4 reports the
means and standard deviations on word accuracy, tolds, self-corrections,
repetitions, and insertions along with the Wilcoxon Scores (sum of the ranks),
the z scores, and the pvalues. The parametric data not only give the reader an
understanding of the general tendencies of the sample but also support the
decision to use non-parametrics. Groups differed in three reading behaviors:
word accuracy, number of tolds, and repetitions. The treatment group was
slightly more accurate than the control group, relied on the examiner less or
pronounce unknown words, and had more repetitions. Thus, the treatment
TABLE 4 Sum of the ranks, Z scores, and p values for Reading Behaviors
Treatment Control
N¼11 N¼12
Mean percent of words read
accurately (SD)
95.6 (2.3) 92.6 (4.2)
W(sums of the ranks)*136.0 95.0
zscore 1.7
Mean number of tolds (SD) 0.2 (0.2) 1.7 (1.8)
W(sums of the ranks)
{
79.5 151.5
zscore 2.2
Mean percent of errors self-
corrected (SD)
20.0 (14.2) 19.5 (23.1)
W(sums of the ranks)
{
119.5 111.5
zscore .6
Mean number of repetitions
(SD)
2.3 (1.0) 1.5 (1.2)
W(sums of the ranks)
§
132.5 98.5
zscore 1.5
Mean number of insertions
(SD)
0.5 (0.3) 0.6 (0.8)
W(sums of the ranks)
}
114.7 117.0
zscore .23
*p¼.03;
{
p¼.02;
{
p¼.26;
§
p¼.05;
}
p¼.79.
78 H. A. E. Mesmer
participants were more accurate, less likely to appeal to the examiner to
pronounce words, but also more likely to repeat words or phrases. However, in
other reading behaviors the treatment group did not respond as predicted.
Note that the treatment group did not differ with respect to numbers of self-
corrections. Thus, the treatment group was more autonomous in some ways
but not in self-correcting.
This study is limited in several respects. First, the intervention phase was
fourteen days; clearly, a minimal amount of time using any type of highly
contrived texts is desirable. However, it is likely that the transition from the
partial to full-alphabetic phase scaffolded by decodable text lasts longer than
fifteen days. The need to control for actual instruction, a unique study feature,
partially justified this limitation. Future studies may extend the intervention
phase of the study. In addition, the small number of participants limited the
study. This study does not establish the long-term influences of decodability.
IMPLICATIONS
This work enhances current knowledge about the benefits and side effects
of highly decodable text in three ways. First, the findings support one of the
major theoretical claims about decodable text—that it results in a greater
application of letter=sound information during text reading. These findings
replicate similar data from a previous study (Mesmer, 2001b). Second, the
results here document the actual behaviors of reading highly decodable
text. Although highly decodable text lived up to one of the theoretical
assertions made about it, it did not result in readers who were necessarily
more autonomous in all reading behaviors. Finally, the screening and
selection of participants for this study suggested that readers must have
some knowledge of the alphabetic system in order to demonstrate the
behaviors found in this study. In addition, many readers in the first one-third
of first grade did not need the intervention’s short vowel focus and highly
decodable text, at least according to the screening instrument used here.
Living up to Theoretical Claims
The major finding suggests that highly decodable text does live up to
theoretical claims made about it. The phonic regularity and lesson-to-text
match in highly decodable text resulted in greater application of let-
ter=sound knowledge to recognize words. In this study, readers of highly
decodable text applied letter=sound knowledge during text reading to a
greater degree than treatment readers did. Although both readers of hi-
ghly decodable text and less decodable text had similar numbers of
substitutions, the quality of treatment readers’ substitutions showed more
attention to letter=sound information. Because the purpose of phonics
Text Decodability and the First-grade Reader 79
instruction is application during text reading, the finding indicates that
decodability facilitates this application.
As the primary finding of the study, this result intersected with other work.
This finding was supported by a very similar study (Mesmer, 2001b) as well as
by the Juel & Roper-Schneider (1985) study. In the latter study, readers of
more highly decodable text outperformed a control group on a generic
decoding measure at the beginning and middle of the year, whereas readers in
the Mesmer study applied letter=sound knowledge more during connected
reading. Thus, when readers received the same phonics instruction, as they
did in both of these studies, they were likely to internalize and apply that
instruction if they had exposure to more highly decodable text.
To a teacher working with readers, be they struggling or normally-
developing, the implication is clear—if you want readers to generalize
letter=sound information taught in a synthetic phonics lesson to connected
reading, then you must give them text that has some phonic regularity and
coordination with phonics instruction. In this study, the coordination with
phonics instruction (LTTM) was 40% over the fifteen-day intervention. One
contribution of the present study is that it examined the effects of
decodable text when it matched with actual, as opposed to scripted,
instruction. Thus, educators employing decodability scaffolds must make
certain that the text coordinates with phonics lessons that are being
delivered in the classroom. It is possible for a stand-alone phonics program
and a series of highly decodable texts to operate separately and thus to
have no benefit for the students. In this case, the phonics instruction may
focus on one sound while the text words emphasize another.
The specific phonic regularity features that resulted in this increased
application of phonics instruction are discussed in greater detail below. The
highly decodable text in this study was characterized by very specific pat-
terns. On average, about 54% of the words in this study’s highly decodable
texts were short-vowel c-v-c words (e.g., pin, bag, mat). This base of words
coordinated with the phonics lessons. The fact that the books featured short
vowel words made coordinating instruction easier. For a teacher, the easily
identifiable sound focus in highly decodable text is an added advantage. One
can retrieve a set of more highly decodable books, quickly identify the
sound focus, and then concentrate on teaching that sound. This focused
instruction is especially important for readers who are not breaking the code
and need the benefits of very consistent teaching to generalize behavior.
Words Accuracy, Tolds, Repetitions, and Highly
Decodable Text
In this study, the treatment group displayed reading behaviors that indicate
both benefits and potential side effects of using highly decodable text. The
80 H. A. E. Mesmer
major benefit appeared to be that the treatment group was more autono-
mous in its reading behaviors in two ways: it was slightly more accurate
than the control group, and it relied less on the examiner to provide words.
Although both groups read within acceptable word accuracy limits for
instructional level text (90% or above), the treatment group was more
accurate. This slight edge in word accuracy may have been the result of the
treatment group using the decodability scaffolds to help it identify words.
Therfore, one effect of using highly decodable text was a slight increase in
accuracy; this increased accuracy occurred with readers who understood
and could apply the alphabetic principle.
The second piece of evidence supporting increased self-confidence was
the lower number of tolds for the treatment group. This difference was
more pronounced than the word accuracy differences and seemed to
indicate that the treatment group again felt more empowered by the match
between the instruction that they had received and the textual scaffolds in
place. The control group, on the other hand, may have been frustrated by
the plethora of patterns that they confronted and more likely to give up on
applying phonics instruction. Although the situation clearly seems irra-
tional, this is the exact position we place readers in when they are trying to
make sense of the alphabetic code and have little contextual opportunity to
do so. In this day of heavy emphasis on phonics instruction, the study
affirms the need to consider text along with phonics instruction.
If a teacher has the goal of increasing student accuracy, attention to
letter=sound information, and independence, then using highly decodable
text and coordinating phonics instruction may be effective for readers with
basic alphabetic knowledge. The implication may be even more important
for struggling readers, some of whom reportedly have this knowledge yet
reach an obstacle when they need to more fully apply it (Brown, 1999;
Hicks & Villaume, 2001; Mesmer, 1999). However, the children in this study
did not read highly decodable text exclusively; their classroom libraries and
instruction provided them with a variety of other materials. Thus, a balance
of texts existed throughout the study. This implies that the use of highly
decodable text need not be exclusive—in fact, exclusive use may not be
desired at all.
In two ways, highly decodable text did not have the intended effects.
First, the treatment group did not self-correct at greater rates. Certainly a
characteristic of skilled reading is the ability to self-correct, and better
decoders usually do this at higher rate (Ehri, 1998). In fact, self-correcting
is perhaps the pinnacle of autonomous reading. Although the treatment
group showed some independence in not relying on the examiner to pro-
nounce words, its independence did not extend to greater self-monitoring.
Both groups self-corrected about 20% of their errors, or about 1 out of 5.
While this is not the optimal 1:3 ratio, it is respectable. The self-correcting
Text Decodability and the First-grade Reader 81
behaviors may not have differed because both treatment and control
groups may have made use of the scaffolds present within each of the
different texts. Thus, the control group may have relied on the print=
picture match, predictability, and other features to assist them as they self-
corrected. On the other hand, the treatment group depended on phonic
regularity and coordination with their phonics lessons.
The second effect of using highly decodable text was an increased level
of repetition for the treatment group. Although neutral, this finding can be
analyzed as potentially negative or positive. Repetition can be viewed as a
self-monitoring behavior, one in which skilled readers participate in order
to self-monitor. However, excessive repetition can accompany disfluency.
When analyzed from this perspective, the finding implies the possible
propensity for readers of highly decodable text to be less fluent.
As a behavior accompanying disfluency, the increased repeating beha-
vior may be accounted for developmentally. Several researchers have
documented a decreased fluency as readers initially learn to decode (Chall,
1983). Therefore, the increased repetition may have been a predictable
developmental result of increased attention to decoding. The key to using
highly decodable text lies in identifying the development of the reader and
the teaching goals. Some developing readers actually compromise accuracy
for fluency and may need to focus on more the latter temporarily at the
expense of the former. If the teaching goal of greater application of
letter=sound information temporarily outweighs the goal of fluency then
the teaching compromise may be acceptable.
Sampling and Timing
The study’s developmental model was not validated by this research;
nonetheless, it was useful to the study in three ways. First, it assisted in
setting screening standards for participants. The participants in this study
had a basic understanding of the alphabetic principle, could apply this to
spelling, and could recognize some preprimer words. Although the
screening did not specifically verify the participants’ use of initial con-
sonants to decode, previous research has established that use of initial
consonants in spelling and use of initial consonants in decoding appear
simultaneously (Stahl, McKenna, Kovach, & Eakle, 2001). These partici-
pants were at least partial alphabetic readers but not operating effortlessly
in consolidated alphabetic phases. Whether or not these participants were
transitioning from partial alphabetic to full alphabetic reading as specified
in the model is not clear. Future research should fully test and validate this
model.
The second benefit of the developmental model was that it allowed for
an application of study results to struggling readers regardless of grade
82 H. A. E. Mesmer
level. Although this study took place with first graders, the participants
could be located along a developmental continuum upon which readers of
any age might be located. The extension to struggling readers is that given
the same abilities, they may also demonstrate behaviors similar to those
observed in this study.
Finally, the model has a pivotal policy implication: the key to using
highly decodable text is determining students’ development along a con-
tinuum of reading behaviors. If highly decodable text were used with
readers who did not have a firm understanding of the alphabetic principle,
then it is possible that the same results would not have been found. Policies
stipulating the use of highly decodable text in the first grade assume that
readers have the prerequisite skills previously mentioned. But in this study,
five of those screened did not; in addition, seventeen would not have
benefited from the intervention because their skills seemed to surpass the
requisite alphabetic knowledge. Therefore, policies about decodable text
should specify that assessment data be used to determine readers’ devel-
opment and that such data may be used in deciding whether or not
decodability is an appropriate textual scaffold. Educators should keep in
mind that a major goal of early reading instruction is increasingly sophis-
ticated application of the alphabetic principle until readers are accessing
words automatically. Part of this progression is reading full alphabetic
reading in the first grade (Ehri, 1998). In addition, a variety of texts,
including highly decodable text, should be available to teachers.
Given its short intervention length and other limitations, this work adds
only a piece to the puzzle over a highly disputed material. It does suggest
how readers might actually interact with decodability scaffolds. No one
material, be it controlled, predictable, or decodable, will work with all
readers all the time. Future research should focus on whether or not
decodability results in readers who are more able to handle a range of
materials containing no textual scaffolds as they mature. In addition,
investigations of the use of highly decodable text with struggling readers
should be undertaken, as well as examinations validating the study’s
developmental model.
REFERENCES
Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge,
Mass: MIT Press.
Allington, R. (1997). Overselling phonics: Five unscientific assertions about reading instruc-
tion. Reading Today, 15,1715.
Allington, R. (2001). What really matters for struggling readers. New York: Addison Wesley
Longman.
Allington, R. & Woodside Jiron, H. (1998). Decodable text in beginning reading: Are mandates
and policy based on research? ERS Spectrum, 16,3713.
Text Decodability and the First-grade Reader 83
Ball, D. L. & Cohen, D. K. (1996). Reform by the book: What is—or might be—the role of
curriculum materials in teacher learning and instructional reform? Educational
Researcher, 25,678.
Barr, R. (1972). The influence of instructional conditions on word recognition errors. Reading
Research Quarterly, 7, 5097529.
Barr, R., Dreeben, R., & Wiratchai, N. (1983). How schools work. Chicago: The University of
Chicago Press.
Beck, I. L. (1981). Reading problems and instructional practices. In G. A. MacKinnon and T. G.
Waller (Eds.), Reading research: Advances in theory and practice, (Vol. 2, pp. 53793).
Newark, DE: Academic Press.
Beck, I. L. & Block, K. K. (1979). An analysis of the dimensions that affect the development
of code-breaking ability in eight beginning reading programs. Pittsburgh: University of
Pittsburgh, Learning Research and Development Center.
Brown, K. (1999). What kind of text—for whom and when? Textual scaffolding for beginning
readers. The Reading Teacher, 53, 2927307.
Bryne, B. (1991). Experimental analysis of the child’s discovery of the alphabetic principle. In
L. Rieben and C. A. Perfetti (Eds.), Learning to read: Basic research and its implica-
tions (pp. 75784). Hillsdale, NJ: Lawrence Erlbaum Associates Publishers.
California Department of Education. (1996). Guide to the California Reading Initiative of
1996. Sacramento: Author.
Chall, J. S. (1983). Stages of reading development. New York: McGraw Hill.
Clark, H. C. & Clark, E. V. (1977). Psychology and language. New York: HBJ.
Clay, M. M. (1993). An observation survey of early literacy achievement. Portsmouth, NH:
Heinemann.
Cole, A. (1998). Beginner-oriented texts in literature-based classrooms: The segue for a few
struggling readers. The Reading Teacher, 51, 4887501.
Ehri, L. C. (1991). Learning to read and spell words. In L. Rieben and C. A. Perfetti (Eds.),
Learning to read: Basic research and its implications (pp. 57774). Hillsdale, NJ:
Lawrence Erbaum Associates Publishers.
Ehri, L. C. (1994). Development of the ability to read words, In R. Barr, M. Kamil,
P. Mosenthall, and P. Pearson (Eds.), Handbook of reading research, (Vol. 2,
pp. 3837417). New York: Longman.
Ehri, L. C. (1995). Phases of development in learning to read words by sight. Journal of
Research in Reading,18, 1167125.
Ehri, L. C. (1998). Grapheme7phoneme knowledge is essential for learning to read words
in English. In J. L. Metsala and L. C. Ehri (Eds.). Word recognition in beginning literacy
(pp. 3740). Mahwah, NJ: Lawrence Erlbaum Associates Publishers.
Ehri, L. C. & McCormick, S. (1998). Phases of word learning: Implications for instruction with
delayed and disabled readers. Reading and Writing Quarterly: Overcoming Learning
Difficulties,14, 1357163.
Fletcher, J. M., Francis, D. J., & Foorman, B. (1997). Only 15720 percent. Reading Today,5,
2718.
Foorman, B., Francis, D., Fletcher, J., Schatschneider, C., & Metha, P. (1998). The role of
instruction in learning to read: Preventing reading failure in at-risk children. The Journal
of Educational Psychology,90,37755.
Gough, P. (1997). Critical balances: Early instruction for lifelong reading. Reading online
[On-line serial]. Available at www:==readingonline.org=critical=houston=gough.
Gough, P. B. & Juel, C. (1991). The first stages of word recognition. In L. Rieben and C. A.
Perfetti (Eds.), Learning to read: Basic research and its implications (pp. 47756).
Hillsdale, NJ: Lawrence Erlbaum Associates Publishers.
84 H. A. E. Mesmer
Groff, P. (1999). Decodable words in reading textbooks: Why they are imperative. The National
Right to Read Foundation. Available at http:==www.nrrf.org=27_decode_txtbooks.htm.
Accessed October 20, 2004.
Grossen, (1997). Thirty years of research: What we know about how children learn to
read. Santa Cruz, Calif.: The Center for the Future of Teaching and Learning.
Hicks, C. P. & Villaume, S. K. (2001). Finding our own way: Critical reflections on the literacy
development of two Reading Recovery Children. The Reading Teacher,54, 3987412.
Hiebert, E. (1999). Text matters in learning to read. CIERA Report 1-001. Ann Arbor, Mich.:
Center for the Improvement of Early Reading Achievement.
Hoffman, J. V. (2001). Decodable texts for beginning reading instruction: Leadership or pol-
itics in California and Texas? The California Reader,34,278.
Hoffman, J. V., McCarthey S.J., Abbott, J., Christian, C., Corman, L., Curry, C., Dressman, M.,
Elliott, B., Maherne, D., & Stahle, D. (1993). So what’s new in the new basals? A focus on
first grade. Journal of Reading Behavior,26,47773.
Hoffman, J. V., Roser, N. L., Patterson, E., Salas, R., & Pennington, J. (2000). Text leveling and
‘‘little books’’ in first grade reading. Journal of Literacy Research,33(3), 5077528.
Johnston, F. (1998). The reader, the text, and the task: Learning words in first grade. Reading
Teacher,51, 6667675.
Juel, C. & Roper-Schneider, D. (1985). The influence of basal readers on first grade reading.
Reading Research Quarterly,20, 1347152.
Kameenui, E. J. & Simmons, D.C. (1997). Decodable texts and language of dichotomy: A
response to Allington. Reading Today,15(2), 18.
MacDonald, P. (1999). Power, Type I, and Type III error rates of parametric and nonparametric
statistical tests. The Journal of Experimental Education,67, 3677379.
Menon, S. & Hiebert, E. H. (1999). Literature anthologies: The task for first grade readers.
Ciera report# 1-009. Ann Arbor, Mich.: Center for the Improvement of Early Reading
Achievement
Mesmer, H. A. (1999). Scaffolding a crucial transition using text with some decodability. The
Reading Teacher,53, 1307142.
Mesmer, H. A. (2001a). Decodable text: A review of what we know. Reading Research and
Instruction,40, 1217142.
Mesmer, H. A. (2001b). In J. Hoffman, D. Schalbert, C. Fairbanks, J. Worthy, and B. Maloch.
(Eds). Examining the theoretical claims about decodable text: Does text decodability lead
to greater application of letter=sound knowledge in first grade readers? 50th Yearbook of
the National Reading Conference, 4447459. National Reading Conference: Oak Creek, WI.
Meyer, L. A., Greer, E. A., & Crummey, L. (1987). An analysis of decoding, comprehension, and
story text in four first-grade reading programs. Journal of Reading Behavior, 6,69798.
Moats, L. & Hall, (1999). Straight talk about reading: How parents can make a difference
during the early years. Chicago, IL: Contemporary Books.
Muter, V. & Snowling, C. (1998). Concurrent and longitudinal predictors of reading: The role of
metalinguistic and short-term memory skills. Reading Research Quarterly,33, 3207337.
Nation, K. & Hulme, C. (1997). Phonemic segmentation, not onset7rime segmentation, pre-
dicts early reading and spelling skills. Reading Research Quarterly,32, 1547167.
National Reading Panel, (2000). Report of the National Reading Panel: Teaching Children
to Read. Washington, DC: National Institute of Child Health and Human Development.
Pikulski, J. J. (1998). Business we should finish. Reading Today, 15(5), 30.
Pressley, M., Rankin, J., & Yokoi, L. (1996). A survey of instructional practices of outstanding
primary-level literacy teachers. Elementary School Journal,96, 3637384.
Reutzel, D. R., & Daines, D. (1987). The text-relatedness of reading lessons in seven basal
reading series. Reading Research and Instruction,27,26735.
Routman, R. (1997). Routman thanks Allington. Reading Today,15(5), 30.
Text Decodability and the First-grade Reader 85
Snow, C., Burns, M. S., & Griffin, P. (1998). Preventing reading difficulties in young chil-
dren. Washington, DC: National Academy of Sciences Press.
Stahl, S., Duffy-Hester, A., & Stahl, K. A. D. (1998). Everything you wanted to know about
phonics (but were afraid to ask). Reading Research Quarterly,33, 3387335.
Stahl, S., McKenna, M., Kovach, J. J., & Eakle, J. (2001). Phonological awareness, word
recognition, and spelling: Rethinking the relation. Paper presented at the Annual Meeting
of the National Reading Conference, December, San Antonio, Texas.
Stein, M., Johnson, B., & Gutlohn, L. (1999). Analyzing beginning reading programs: The
relationship between decoding instruction and text. Remedial and Special Education,
20, 2577287.
Texas Department of Education. (1997). 1997 Proclamation of the State Board of Educa-
tion Advertising for Bids on Instructional Materials. Austin, TX: Author.
Venezky, R. L. (1970). The structure of English orthography. The Hague: Mouton.
Virginia General Assembly. (2001). House Joint Resolution No. 794. Richmond, VA: Author.
Watson, D. (1997). Beyond decodable texts—supportive and workable literature. Language
Arts,74, 6357643.
Willows, D. M., Borwick, D., & Hayvren, M. (1981). The content of school readers. In G. E.
MacKinnon and T. G. Waller (Eds.), Reading research: Advances in theory and practice
(vol. 2, pp. 957175). New York: Academic Press.
Wilson, B. A. (1996). The Wilson Reading System. Millbury, Mass.: Wilson Language Training
Corp.
Zimmerman, D. W. (1994). Not on the influence of distribution shape on parametric and
nonparametric tests. The Journal of General Psychology,121, 3917401.
86 H. A. E. Mesmer
