Content uploaded by Elizabeth Meisinger
Author content
All content in this area was uploaded by Elizabeth Meisinger on Mar 18, 2016
Content may be subject to copyright.
This article was downloaded by: [University of Memphis Libraries]
On: 13 April 2015, At: 13:17
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK
Click for updates
Reading Psychology
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/urpy20
The Contributions of Oral and
Silent Reading Fluency to
Reading Comprehension
Katherine W. Pricea, Elizabeth B. Meisingerb, Max M.
Louwersec & Sidney D’Mellod
a Tangipahoa Parish Public School System,
Hammond, Louisiana
b Department of PsychologyUniversity of Memphis,
Memphis, Tennessee
c Department of Communication and Information
SciencesTilburg University, Tilburg, The Netherlands
d Department of PsychologyUniversity of Notre
Dame, Notre Dame, Indiana
Published online: 11 Apr 2015.
To cite this article: Katherine W. Price, Elizabeth B. Meisinger, Max M. Louwerse
& Sidney D’Mello (2015): The Contributions of Oral and Silent Reading Fluency to
Reading Comprehension, Reading Psychology, DOI: 10.1080/02702711.2015.1025118
To link to this article: http://dx.doi.org/10.1080/02702711.2015.1025118
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the
information (the “Content”) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any
losses, actions, claims, proceedings, demands, costs, expenses, damages,
and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.
This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Reading Psychology, 00:1–35, 2015
Copyright C
Taylor & Francis Group, LLC
ISSN: 0270-2711 print / 1521-0685 online
DOI: 10.1080/02702711.2015.1025118
THE CONTRIBUTIONS OF ORAL AND SILENT READING
FLUENCY TO READING COMPREHENSION
KATHERINE W. PRICE
Tangipahoa Parish Public School System, Hammond, Louisiana
ELIZABETH B. MEISINGER
Department of Psychology, University of Memphis, Memphis, Tennessee
MAX M. LOUWERSE
Department of Communication and Information Sciences, Tilburg University,
Tilburg, The Netherlands
SIDNEY D’MELLO
Department of Psychology, University of Notre Dame, Notre Dame, Indiana
Silent reading fluency has received limited attention in the school-based litera-
tures across the past decade. We fill this gap by examining both oral and silent
reading fluency and their relation to overall abilities in reading comprehension
in fourth-grade students. Lower-level reading skills (word reading, rapid auto-
matic naming) and vocabulary were included in structural equation models in
order to determine their impact on reading fluency and comprehension. Results
suggested that oral and silent reading fluency represent separate constructs, but
only oral reading fluency contributed to reading comprehension. Vocabulary was
found to contribute uniquely to comprehension even after controlling for reading
fluency.
Oral reading fluency has recently received increased attention
in the school-based literatures, even though it was once thought
of as a neglected area in reading research (Kuhn, Schwanen-
flugel, & Meisinger, 2010). The prominence of research on oral
reading fluency may be in part due to the information pre-
sented by the National Reading Panel (2000) outlining the im-
portance of fluency instruction and attainment. Although inter-
est in silent reading fluency has gradually increased over the past
several years, it has not garnered the level of attention that oral
Address correspondence to Elizabeth B. Meisinger, Department of Psychology, Uni-
versity of Memphis, 202 Psychology, Memphis TN 38152. E-mail: bmsinger@memphis.edu
1
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
2K. W. Price et al.
reading fluency has secured, with few studies having examined
the similarities and differences between oral and silent reading
fluency.
Oral reading provides several benefits for young readers.
Children are likely to first be exposed to literacy by adults read-
ing aloud to them. Later, as children are cementing their emerg-
ing literacy skills, they are likely to practice by reading aloud
with the support of a more proficient reader (Chall, 1996). Oral
reading also provides benefits to beginning or struggling read-
ers, as it allows for self-monitoring of progress (Hiebert, Samuels,
& Rasinski, 2012; Kuhn & Schwanenflugel, 2007), reinforcement
of letter–sound correspondence (Kuhn & Schwanenflugel, 2007),
and the use of both reading and listening comprehension skills to
facilitate understanding (Hoover & Gough, 1990; Kuhn & Schwa-
nenflugel, 2007). Additionally, oral reading results in longer time
on-task, as children generally read more slowly when they read
aloud (Rayner & Pollatsek, 1989).
However, as children reach the fourth grade, they are ex-
pected to effectively transition to silent reading. Students should
increasingly be able to read faster and with equivalent compre-
hension silently, no longer requiring the added support of oral
reading (Hiebert et al., 2012). Share (2008) pointed out, “silent
understanding rather than oral reading is the literacy benchmark
in knowledge-based societies” (p. 594).
Fluent oral reading skills have been shown to emerge be-
tween the first and third grade (Chall, 1996; Kuhn & Stahl, 2003).
However, far less research has been conducted on silent reading
fluency. In fact, some researchers (Share, 2008) have suggested
that the overarching dependence on oral reading provides an in-
complete picture of both reading fluency and reading develop-
ment. For instance, the emphasis on oral fluency might have led
to an overestimation of the importance of phonological variables
as well as overstated conclusions with regards to the cognitive pro-
cesses underlying oral and silent reading. Further, eye-tracking re-
search has demonstrated that, in skilled readers, the eye tends to
be ahead of the voice (Radach, Schmitten, Glover, & Huestegge,
2009; Rayner & Pollatsek, 1989). In short, even though oral read-
ing fluency is an important skill in young readers, emphasizing
oral reading fluency in research and its conclusions at the expense
of silent reading fluency is undesirable.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 3
Implicit in the current literature on reading fluency is the
idea that oral reading fluency and silent reading fluency involve
essentially the same processes (Share, 2008). English-language
studies of reading within the school-based literatures have mostly
been conducted using oral reading measures, and the results of
these studies have been largely assumed to apply to silent read-
ing. Few theoretical definitions of reading fluency make distinc-
tions between the two modalities, and rarely have studies explic-
itly examined their differences. Although discussions of this issue
within the literature are starting to appear (Hiebert et al., 2012)
and some theoretical definitions of reading fluency specify that
oral and silent reading involve separate skills (e.g., Kuhn et al.,
2010), there has yet to be an empirical study that thoroughly ex-
amines these issues in late elementary students.
The present study seeks to model the relations between both
silent and oral reading fluency in relation to comprehension.
First, however, a more thorough examination of the literature is
warranted. In the remainder of this section, the relation between
reading fluency and reading comprehension will be discussed, the
differences between oral and silent reading will be examined, and
finally the reading subcomponents modeled in the present study
will be presented.
The Link Between Reading Fluency and Reading Comprehension
Comprehension represents the sine qua non of the reading pro-
cess. As children transition from viewing reading as a word decod-
ing exercise to a meaning construction endeavor (Sweet & Snow,
2003), they are increasingly required to utilize their developing
comprehension skills to obtain knowledge both in and out of the
classroom. Oral reading fluency has been shown to be essential for
effective comprehension, although the directionality of this rela-
tionship is somewhat debated (e.g., Klauda & Guthrie, 2008; Kuhn
et al., 2010). Overall, studies examining oral reading fluency and
comprehension have found moderate to strong positive correla-
tions between the two in diverse samples comprised of students
from primary to secondary grades (Daane et al., 2005; Fuchs,
Fuchs, & Maxewell, 1988; Jenkins, Fuchs, van den Broek, Espin,
& Deno, 2003; Pinnell et al., 1995; Yovanoff, Duesbery, Alonzo, &
Tindal, 2005).
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
4K. W. Price et al.
Although researchers have been interested in the difference
between oral and silent reading for decades (Jones, 1932), rela-
tively few studies have examined comprehension differences be-
tween the two reading modes. Findings across such studies involv-
ing elementary students are somewhat varied, yet several themes
have emerged from the literature. First, oral reading may sup-
port comprehension in younger (Elgart, 1978; Fletcher & Pum-
frey, 1988) or low-ability readers (Burge, 1983; Fuchs & Maxwell,
1988; Miller & Smith, 1985). Second, at some point, children
may become equally proficient at comprehending across reading
modes. By adulthood, readers typically have equivalent compre-
hension after either reading mode (Holmes, 1985; Salasso, 1986).
Given the limited existent literature, further examination of oral
and silent reading fluency with regard to reading comprehension
seems warranted.
Modeling Reading Fluency
Several attempts have been made to parse apart literacy develop-
ment in children in order to determine the importance of var-
ious subcomponent skills of reading fluency and their relative
importance to reading comprehension (Berninger et al., 2010;
Kendeou, van den Broek, White, & Lynch, 2009; Ouellette &
Beers, 2010; Schwanenflugel et al., 2006; Vellutino, Tunmer, Jac-
card, & Chen, 2007). The majority of these studies have exam-
ined oral reading fluency in early elementary school students.
Kim, Wagner, and Foster (2011), however, examined the relations
among oral reading fluency, silent reading fluency, and reading
comprehension in a sample of first-grade students. Importantly,
oral and silent fluency were found to represent distinct constructs,
each predicting reading comprehension skill. The studies that
have looked at older students (Barth, Catts, & Anthony, 2009)
have solely examined oral reading fluency, largely overlooking the
importance of silent reading fluency within this age group.
Several subcomponents of the reading process have been
identified as being important for reading fluency and reading
comprehension, including phonological awareness, word reading
accuracy, naming speed, and vocabulary. Each of these subcom-
ponents will be discussed along with available evidence regarding
the relation of each to oral and silent reading fluency. It should be
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 5
noted that given the limited number of studies on silent reading
fluency, relatively little is known about the associations between
these reading subcomponents and silent reading fluency.
PHONOLOGICAL DECODING
Phonological skills, such as phoneme segmentation and
phonological (letter-sound) decoding, are essential for emergent
readers with small sight word vocabularies who rely heavily on de-
coding during reading (National Reading Panel, 2000). It is likely
that these skills are less important for more-skilled readers (Vel-
lutino et al., 2007), especially those who are able to utilize other
strategies for word reading, such as sight word recognition, analo-
gizing, prediction, and the use of context (Kuhn et al., 2010).
Nonword decoding is often used as a way to measure student’s
facility with letter knowledge, letter string, rime units, and speech
sounds (Schwanenflugel et al., 2006) and can be used as an in-
dicator of readers’ phonological processing skill (Siegel, 1993).
It is likely that skill in phonological decoding is more necessary
for the oral rendering of text than for silent reading because
silent understanding does not necessarily require the ability to
fully pronounce words. However, to our knowledge, no study to
date has examined the association between silent reading fluency
and phonological decoding.
WORD READING
Theories specifically relate proficient word reading to gen-
eral reading fluency development (e.g., La Berge & Samuels,
1974) and provide suggestions for how reading fluency mediates
the relation between word reading skill and comprehension. For
example, automaticity at the word level is thought to facilitate flu-
ent reading by freeing cognitive resources for use in comprehen-
sion. Indeed, word reading ability is moderately to strongly related
to fluency and comprehension within the literature (Barth et al.,
2009, Berninger et al., 2010; Ouellette & Beers, 2010; Vellutino,
Fletcher, Snowling, & Scanlon, 2004; Vellutino et al., 2007). Al-
though basic reading competencies such as phonological decod-
ing and word reading are essential to comprehension for emer-
gent readers, their contributions diminish across development
and are less predictive of comprehension as children gain in
proficiency and begin to encounter more complex texts (Floyd,
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
6K. W. Price et al.
Meisinger, Gregg, & Keith, 2012; Jenkins & Jewell, 1993; Shinn,
Good, Knutson, Tilly, & Collins 1992; Vellutino et al., 2007).
NAMING SPEED
Students’ rapid automatic naming skill (RAN), or the abil-
ity to provide rapid, fluent verbal responses to visual information
(e.g., naming letters or numbers), is related to reading develop-
ment in general, and more specifically to oral reading fluency
(Norton & Wolf, 2012; Wolf, Bowers, & Biddle, 2000) and word
reading fluency (Schwanenflugel et al., 2006), although its unique
contribution is less than that of word reading (Barth et al., 2009).
The relation between RAN and oral reading fluency is not surpris-
ing because children’s’ ability to produce oral language fluently in
response to visual stimuli probably underlies their ability to read
connected text aloud with appropriate fluency. Given that silent
reading fluency does not require verbal output, skills in RAN may
not be as central to its development. Overall, it seems probable
that RAN is more closely related to oral reading fluency than silent
reading fluency.
VOCABULARY
Vocabulary has been shown to be a strong predictor of
reading comprehension, even after controlling for word read-
ing, phonemic awareness, and letter knowledge (Muster, Hulme,
Snowling, & Stevenson, 2004). Further, it has been shown to con-
tribute unique variance to reading comprehension in studies ex-
amining various age groups, including children in the early el-
ementary (Ouellette & Beers, 2010), mid-elementary (Senechal,
2006), and late-elementary grades (Ouellette & Beers, 2010), as
well as young adults (Braze, Tabor, Shankweiler, & Mencl, 2007).
The relations between vocabulary and oral and silent reading flu-
ency have yet to be fully examined.
The Present Study
Few studies have carefully examined the relation between these
subcomponents in oral and silent reading fluency and compre-
hension in the late elementary years. Although several stud-
ies have modeled oral reading fluency (Berninger et al., 2010;
Kendeou et al., 2009; Ouellette & Beers, 2010; Schwanenflugel
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 7
FIGURE 1 The singular model, which represents reading fluency as a single
construct.
et al., 2006; Vellutino et al., 2007), especially in relation to read-
ing comprehension, the role of oral and silent reading fluency in
the reading process is unclear. The present study addressed this
important gap in the literature by examining both oral and silent
reading fluency and their relation to overall abilities in reading
comprehension in a normative sample of fourth-grade students.
Specifically, the current study utilized structural equation model-
ing (SEM; Kline, 2010) to test the viability of various theoretical
models depicting how oral reading fluency, silent reading fluency,
and reading comprehension relate to one another and to vari-
ous reading subcomponents. Two main research questions were
addressed in the present study: (a) Do oral reading fluency and
silent reading fluency represent distinct constructs? and (b) What
is the relation between the reading subcomponents (i.e. word
reading, nonword reading, rapid automatic naming, vocabulary)
and oral and silent reading fluency and reading comprehension?
In order to investigate whether or not oral and silent read-
ing fluency represent distinct constructs, two SEM models were
compared. Consistent with the broader literature that conflates
oral and silent reading fluency into a single construct, the sin-
gular model represented oral and silent reading fluency as a sin-
gle latent variable contributing to reading comprehension (see
Figure 1). In contrast, the split model represented both types of
reading fluency as separate skills each contributing directly to
reading comprehension (see Figure 2). As oral reading fluency
has been shown to develop prior to improvements in silent read-
ing fluency, in this split model, oral reading fluency was also spec-
ified to contribute directly to silent reading fluency.
Next we investigated the relation between the reading sub-
components and oral and silent reading fluency and reading com-
prehension by adding the reading subcomponent skills (i.e. word
reading, nonword reading, rapid automatic naming, vocabulary)
to the structural equation models. Each of the subcomponent
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
8K. W. Price et al.
FIGURE 2 The split model specifies two separate reading fluency constructs:
oral reading fluency and silent reading fluency.
skills was specified to contribute directly to the reading fluency
factors, and vocabulary was also specified to contribute directly to
reading comprehension.
We predicted that (a) those subcomponents that require the
oral rendering of text would be more closely related to oral rather
than silent reading fluency (i.e. word reading, rapid automatic
naming, and nonword reading), (b) that vocabulary would con-
tribute directly to comprehension above and beyond the contri-
butions of the other subcomponent skills and reading fluency,
and (c) that the reading subcomponent skills would contribute a
smaller proportion of the variance to comprehension than would
reading fluency.
Method
Participants
Participants were 106 fourth-grade students attending a public in-
termediate school located in the mid-south region of the United
States. All students attended general education classes; none were
excluded on the basis of special education status except for those
students in self-contained special education classes. The sample
was composed of 52% girls, and race/ethnicity was 56.2% Cau-
casian, 40.8% African American, 12.2% Hispanic, 2% multira-
cial, and 1% Asian or Pacific Islander. Approximately 52% of the
school population was eligible for free or reduced lunch.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 9
Measures
READING PASSAGE SELECTION
Reading passages for the oral and silent reading assessments
were drawn from the Qualitative Reading Inventory, Fourth Edi-
tion (QRI-4; Leslie & Caldwell, 2006). The QRI-4 is a criterion-
referenced, individually administered test of reading ability.
There are six fourth-grade expository passages available. Three
passages are biographies of famous Americans and three passages
are descriptive science and social studies materials about various
topics. The six1selected passages were counterbalanced across the
conditions using a Latin square procedure.
SILENT READING FLUENCY
Children’s silent reading fluency was assessed using underlin-
ing and slasher techniques.
Underlining. In order to monitor online silent reading flu-
ency, we used a measure that has been demonstrated to be a sub-
stitute for alternative measures of self-paced reading time (Price,
Meisinger, D’Mello, & Louwerse, 2012). The underlining proce-
dure was previously validated for use with late elementary-aged
students (Price et al., 2012). As students read the passage, they un-
derlined the text word-by-word in a smooth motion using a stylus
while their underlining time (approximating their reading time)
was recorded. Students were instructed to continue to underline
online with their reading (e.g., if students regressed, the regres-
sion was to be underlined, if they paused during reading, the un-
derlining was to pause). During the underlining of each passage,
the location of the stylus position on the screen was be recorded
at the rate of 10 Hz (i.e., 10 times per second) in order to track
various characteristics of students’ reading (e.g., rate, regressions,
pauses, etc.). The software was specifically designed for the un-
derlining task and was programmed in C# to run on the Windows
Operating System and is available to researchers upon request.
Alternate-form reliability estimates of 0.86 (mean word reading
time) were previously obtained for the underlining task (Price
et al., 2012). Validity estimates ranged from 0.56 to 0.73 with other
validated measures of silent reading fluency (Price et al., 2012).
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
10 K. W. Price et al.
The reading passages for the underlining procedure were
presented in size 12 Andale Mono (fixed width) font on a Dell
Latitude XT tablet personal computer (PC). Students read one
brief practice trial to familiarize themselves with the underlin-
ing procedure. After the practice passage, students completed
two trials, each with a different QRI-4 passage. After the student
finished reading each passage, eight comprehension questions re-
lated to the passage content were orally presented, and the stu-
dent provided oral responses which were scored by the examiner
as correct or incorrect based on provided criteria. These questions
were used to gauge reading comprehension and are described in
greater detail in a subsequent section. A single raw score from
the underlining measure was the mean number of words read per
minute across both passages.
Test of Silent Contextual Reading Fluency. The Test of Silent
Contextual Reading Fluency (TOSCRF; Hammill, Wiederholt, &
Allen, 2006) was group-administered to students in order to ob-
tain a second, timed measure of students’ silent reading fluency.
The TOSCRF is a standardized, norm-referenced group adminis-
tered test that yields standard scores and percentile ranks. It mea-
sures how quickly students can determine individual words within
a series of passages that increase in difficulty, from the preprimer
up through the adult reading level. Within each passage, words
are printed in uppercase, but spaces and punctuation are omitted.
Students were provided three minutes to draw lines or slashes be-
tween as many words as possible. The total score was derived from
the number of correctly marked words. Reported test–retest relia-
bility ranged from 0.84 to 0.92, whereas alternate form-delayed re-
liability ranged from 0.81 to 0.87. Validity estimates ranged from
0.67 to 0.85 with other validated measures of reading (Hammill
et al., 2006).
ORAL READING FLUENCY
Reading passages were individually administered to each stu-
dent in order to assess proficiency in the oral reading of con-
nected text. Students were provided a passage and asked to read
aloud while an administrator recorded any oral reading errors.
Modeling a common procedure (DIBELS Oral Reading Fluency
subtest; Good, Kaminski, & Dill, 2002), the following were con-
sidered word-reading errors: mispronunciations or substitutions,
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 11
omissions, and hesitations of more than three seconds. Although
standard administration of R-CBM oral reading fluency probes re-
quires the administration of three passages, for the purpose of this
study, only two passages were administered in order to mirror the
procedure for the underlining measure.2The time required for
the student to read the passage in its entirety was recorded by
the examiner. As in the underlining procedure, after the student
finished reading the oral reading passage, eight comprehension
questions tied to the passage were orally presented one at a time,
and the student provided oral responses which were scored by
the examiner as correct or incorrect based on provided criteria.
These questions were used to gauge reading comprehension and
are described in detail in a subsequent section. A single raw score
from the oral reading fluency passages was the mean number of
words read correctly per minute across both passages. Reliability
estimates for DIBELS Oral Reading Fluency ranged from 0.92 to
0.97; correlations with other measures of oral reading and read-
ing comprehension have ranged from 0.52 to 0.91 (Shaw & Shaw,
2002).
WORD READING
In order to provide a measure of students’ ability to rec-
ognize individual words in isolation, the Word Reading subtest
from the Wechsler Individual Achievement Test, Third Edition
(WIAT-3; Psychological Corporation, 2009) was individually ad-
ministered. The WIAT-3 is a standardized, norm-referenced test of
academic achievement that yields standard scores and percentile
ranks. Students were asked to read aloud from a provided list
of words. The subtest yielded two scores: accuracy and speed.
Only the accuracy score was included for the purposes of this
research. The Word Reading total score reflects the number of
words read aloud correctly in untimed conditions. The WIAT-
3 provides standard scores, and these were used in subsequent
analyses. The split-half reliability coefficient for the Word Read-
ing test was 0.98 for Grade 4; validity estimates with the WIAT-
2 (Wechsler, 2001) Word Reading subtest were 0.85 (Breaux,
2009).
NONWORD READING
The Pseudoword Decoding subtest of the WIAT-3 was in-
dividually administered as a measure of students’ nonword
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
12 K. W. Price et al.
reading. Nonword reading has been used as a gauge of student’s
phonological awareness in previous studies (Schwanenflugel
et al., 2008). Students were asked to read from a provided list of
phonetically regular, pronounceable non-words (e.g., vonk). The
Pseudoword Decoding total score reflects the number of words
read aloud correctly in untimed conditions. The WIAT-3 provides
standard scores and these were used in subsequent analyses. The
split-half reliability coefficient for the Pseudoword decoding sub-
test was 0.97 for Grade 4; the validity estimate with the WIAT-2
(Wechsler, 2001) Pseudoword Decoding subtest was 0.84 (Breaux,
2009).
VOCABULARY
In order to assess students’ expressive vocabulary, the Picture
Vocabulary subtest from the Woodcock Johnson Tests of Achieve-
ment, Third Edition (WJ III ACH; Woodcock, McGrew, & Mather,
2001) was individually administered. Students were shown a se-
ries of pictures and asked to orally provide a one-word name for
the picture. The score from Picture Vocabulary reflects the to-
tal number of correct responses provided. The WJ III ACH pro-
vides standard scores, and these were used in subsequent analyses.
Scores were derived using the WJ III ACH 2007 Normative Update
(Woodcock, McGrew, Schrank, & Mather, 2007). Test-retest relia-
bility estimates for students aged 9 years to 12 years ranged from
0.77 to 0.80.
RAPID AUTOMATIC NAMING
The Rapid Picture Naming subtest from the Woodcock John-
son Tests of Cognitive Abilities, Third Edition (WJ III COG)
was individually administered in order to gauge rapid automatic
naming. Students are asked to name as many pictures as pos-
sible within a three-minute time limit. The WJ III COG pro-
vides standard scores and these were used in subsequent anal-
yses. Scores will be derived using the WJ III COG 2007 Norma-
tive Update (Woodcock et al., 2007). Standard test-retest reliabil-
ity estimates for students aged 9 to 12 years ranged from 0.96 to
0.97, and analyses of the WJ III speeded tests indicate one-day
test-retest reliability of 0.78 for this age group (Woodcock et al.,
2007).
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 13
READING COMPREHENSION
Gates–MacGinitie Reading Test, Fourth Edition. Comprehension
was assessed using the Gates–MacGinitie Reading Test, Fourth
Edition (GMRT-4; MacGinitie, MacGinitie, Maria, & Dreyer, 2007)
Comprehension subtest. The GMRT-4 is a standardized, norm-
referenced group-administered test that yields normal curve
equivalent scores. Students were asked to silently read a se-
quence of 10 passages. Each passage was accompanied by a se-
ries of multiple-choice questions. Students were allowed 35 min-
utes to complete the subtest. Test–retest reliability estimates for
the GMRT-4 ranged from 0.83 to 0.85, internal consistency co-
efficients ranged from 0.96 to 0.97, and validity estimates with
other tests of reading comprehension ranged from 0.60 to 0.62
(MacGinitie, MacGinitie, Maria, & Dreyer, 2008).
Reading Maze. Comprehension was also assessed by a read-
ing maze task from the AIMSweb progress monitoring system
(Shinn & Shinn, 2002). The maze task is a standardized, group-
administered, multiple-choice cloze silent reading task. Students
read a narrative fiction passage in which the first sentence is left
intact, after which every seventh word is replaced by three word
choices in parentheses. One of the three word choices is cor-
rect, one is a near distracter (same word type but does not pre-
serve meaning), and one is a far distracter (not the same word
type, does not preserve meaning). Each student completed a short
practice passage and then had three minutes to read a grade-level
passage and complete the task. No students finished the passage
in less than three minutes. Test–retest reliability estimates of 0.90
have been reported for maze tasks similar to the ones used in this
study, and validity estimates have ranged from 0.77 to 0.85 for stu-
dents in Grades 3–5 (Fuchs & Fuchs, 1992).
Qualitative Reading Inventory, Fourth Edition. Each of the pre-
viously described comprehension measures provides a holistic es-
timate of students’ ability to comprehend written discourse. Ad-
ditional comprehension questions corresponding to the QRI-4
passages (Leslie & Caldwell, 2006) were also administered fol-
lowing the silent and oral reading procedures described above
in order to quantify students’ comprehension skills. The use
of these additional questions allows for a direct measure of
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
14 K. W. Price et al.
students’ comprehension following both oral and silent reading
of passages. The QRI-4 provides eight comprehension questions
for each passage: four explicit questions and four implicit ques-
tions. Explicit questions could be answered from material stated
directly in the text. Implicit questions required information that
must be inferred from the text and are based on the interac-
tion between information provided in the text and students’ prior
knowledge. Correct answers to implicit questions, however, must
be tied to the text and cannot be provided simply from prior
knowledge. Reliability estimates for the QRI-4 ranged from 0.80
to 0.99 and validity estimates with other tests ranged from 0.44
to 0.72 (Leslie & Caldwell, 2006). Further, evidence suggests that
questions from the QRI-4 are less reliant on participants’ decod-
ing skills to comprehend the passage text than other compara-
ble measures of reading comprehension (Keenan, Betjemann, &
Olson, 2008).
Procedure
Written parental consent and child assent were required for par-
ticipation in the study. The underlining, oral reading fluency, vo-
cabulary, RAN, word reading, and phonological decoding mea-
sures were individually administered in a quiet location in the
school. Administration of all individual measures was counterbal-
anced using a Latin square in order to address order effects. The
TOSCRF, GMRT-4, and maze tasks were group-administered fol-
lowing the completion of the individual measures. Order of the
tasks was counterbalanced across classes. All measures were ad-
ministered by graduate students in school psychology who were
trained by the lead investigator. Administrators were required to
reach 95% interrater agreement on all individually administered
measures prior to the beginning of data collection. The first day
of data collection for each administrator was then observed by
the lead investigator as a secondary fidelity check. Children re-
ceived a small gift (i.e. a pencil following the individual assess-
ments and candy following the group-administrated assessments)
as a token of appreciation for participating in the study. Teachers
received a gift card in appreciation of their participation in the
study.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 15
Analytic Technique
First, descriptive statistics and correlations were calculated in or-
der to determine the data’s suitability for further analyses. Struc-
tural equation modeling (SEM) using AMOS 18 was utilized in
order to explore the relations between reading comprehension,
oral and silent reading fluency, and the various subcomponents
of the reading process. This technique allowed us to build models
based on appropriate theory and then assess how the model fits
the observed correlations among the variables. Parameters were
estimated using maximum likelihood estimation, which is a com-
monly used and accepted approach. It is also considered to be the
most accurate approach when using normally distributed data and
is most appropriate with sample sizes smaller than approximately
N=250 (Kline, 2010). Several fit indices were examined for each
fitted model. First, the model χ2statistic associated with the p
value is reported, followed by the comparative fit index (CFI), the
Tucker Lewis Index (TLI), the Root Mean Square Error of Ap-
proximation (RMSEA), and the standardized root mean square
residual (SRMR). A non-significant χ2represents good model fit,
as do CFI and TLI values above 0.95, RMSEA values that are less
than 0.05, and SRMR values less than 0.08 (Kline, 2010). Addition-
ally, the Akaike Information Criterion (AIC) was examined for the
structural models in order to compare model fits; the AIC is a CFI
that is meaningful only when two models are estimated such that
the model with the lowest AIC value provides the best fit to the
data.
Results
Data Processing and Screening
Eight participants missed group-administered measures due to ab-
sence. Little’s Missing Completely at Random test (Little, 1988)
was used to determine these data were missing completely at ran-
dom and these participants were consequently dropped from the
dataset. These removals resulted in a final sample size of 98.3The
remaining data were screened in order to examine for missing
data points, outliers, and normalcy. No out of range data points
were found, but three additional missing data points were missing
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
16 K. W. Price et al.
at random (attributed to examiner error; Tabachnick & Fidell,
2007). These single points were imputed using the multiple im-
putation technique available in PASW Statistics 18.
Several univariate outliers across measures were identified
and were decreased to the level of the second highest score within
that same measure following the procedures outlined in Tabach-
nick and Fidell (2007). Subsequent examinations of the recalcu-
lated z-scores for each of the measures indicated that there were
no longer univariate outliers present (z≤3.29). Mahalanobis Dis-
tance was utilized in order to screen for multivariate outliers (us-
ing p<0.001 as the criterion), and none were found. Skewness
and kurtosis were found to be within acceptable limits (value di-
vided by standard error in order to convert to zscore; all z<3.29;
Tabachnick & Fidell, 2007) after correcting for univariate outliers.
No problems with curvilinear relationships were found based on
visual examination of bivariate scatterplots. There were no prob-
lems noted with multicollinearity or singularity (r<0.80; Kline,
2010).
Descriptive Statistics
Descriptive statistics and intercorrelations between variables are
presented in Table 1. Based on the performance for those mea-
sures for which the population mean was available (e.g., stan-
dard score or normal curve equivalent), the sample had some-
what weak reading comprehension (GMRT-4) and silent reading
fluency (TOSCRF) skills. The sample mean scores from the read-
ing subcomponent skills (word reading, nonword reading, rapid
naming, vocabulary) were also slightly lower than available pop-
ulation means. Overall, the patterns among the reading variables
were largely what would be predicted with the exception of the
underlining task. Although the underlining task was strongly cor-
related with the oral reading fluency measure, its relation with
the TOSCRF, the GMRT-4, and the maze task was weaker. In par-
ticular, the statistically significant but weak (r=0.20) relationship
between the two silent reading fluency measures was surprising.
In SEM analyses, variances are considered ill-scaled if they dif-
fer by greater than a factor of about 10, and ill-scaled covariance
matrices can result in problems due to the iterative nature of SEM
estimation (Kline, 2010). The current data set was considered
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
TABLE 1 Correlations, Covariances, and Descriptive Statistics
1 2 345 6 7 891011MSD
1 UL — 371.254.7 115.352.5−8.218.850.058.93.537.2 140.532.8
2ORF 0.53 ∗∗ —52.1 155.254.8.46 18.5 104.1 138.332.264.2105.421.5
3TOSCRF 0.20∗0.30 ∗∗ —36.716.01.23.337.929.213.622.394.28.2
4GMRT 0.24∗0.49 ∗∗ 0.31 ∗∗ —47.414.024.451.465.148.842.545.214.7
5Maze 0.25∗0.39 ∗∗ 0.30 ∗∗ 0.50 ∗∗ —1.75.815.424.415.612.719.26.5
6ULQ −0.08 0.01 0.05 0.29 ∗∗ 0.08 — 3.5.36 −0.88.63.96.33.2
7ORFQ 0.11 0.25∗0.11 0.47 ∗∗ 0.26∗0.31 ∗∗ —3.48.89.82.57.83.5
8WR 0.13 0.50 ∗∗ 0.48 ∗∗ 0.36 ∗∗ 0.24∗0.01 0.10 — 95.127.119..997.69.7
9PWD 0.01 0.48 ∗∗ 0.27 ∗∗ 0.33 ∗∗ 0.28 ∗∗ −0.02 0.19 0.73 ∗∗ —30.717.096.013.4
10 Vocab 0.01 0.17 0.19 0.37 ∗∗ 0.27 ∗∗ 0.30 ∗∗ 0.31 ∗∗ 0.31 ∗∗ 0.26∗—29.494.19.0
11 RPN 0.10 0.27 ∗∗ 0.25∗0.26 ∗∗ 0.18 0.11 0.07 0.19 0.12 0.30 ∗∗ —97.911.1
Original S21075.8 462.367.2 216.142.310.212.394.1 179.681.0123.2
Constant 1 2 4 2 5 10 10 4 3 4 3
Rescaled s21075.8 1849.2 1075.2 864.4 1057.5 1020.0 1230.01505.6 1616.4 1296.0 1108.8
Rescaled SD 32.843.032.829.432.531.935.138.840.236.033.3
Note. N = 98. 1. UL = underlining; 2. ORF = oral reading fluency; 3. TOSCRF = Test of Silent Contextualized Reading Fluency; 4. GMRT = Gates-
MacGinitie Reading Test, Fourth Edition Comprehension; 5. Maze = AIMSweb Maze; 6. UL Q = QRI-4 comprehension questions for underlining; 7.
ORF Q = QRI-4 comprehension questions for oral reading; 8. WR = WIAT-3 Word Reading; 9. PWD = WIAT-3 Pseudoword Decoding; 10. Vocab. =
WJ III ACH Picture Vocabulary; 11. RPN = WJ III COG Rapid Picture Naming. Correlations are presented below the diagonal, and covariances are
presented above the diagonal.
**p<0.01, *p<0.05.
17
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
18 K. W. Price et al.
ill-scaled according to this criterion. Using a method endorsed in
the literature (Kline, 2010), raw scores were each multiplied by a
constant, which serves to maintain correlations amongst the vari-
ables while modifying the variable means and variances, thus re-
sulting in a properly scaled covariance matrix. Specifically, within
each measure, raw scores were multiplied by the same constant.
Rescaling constants for each measure were selected in order to
result in an appropriately scaled matrix (i.e. all covariance values
within a factor of 10). Information about the constants used to
re-scale the covariance matrix is presented at the bottom of Ta-
ble 1. The means and standard deviations presented in Table 1
are those of the original, non re-scaled raw scores; however, the
re-scaled values were utilized for all further SEM analyses.
Structural Equation Modeling
The results were analyzed in three phases. First, measurement
models were analyzed to determine the feasibility of the various la-
tent variables under investigation. Second, structural components
for the reading fluency factor(s) were added in order to deter-
mine the fit of the hypothesized singular and split models. Finally,
subcomponent skills (i.e. word reading, nonword reading, rapid
automatic naming, and vocabulary) were added to the models.
Following this third phase, two alternative models were tested in
order to provide further evidence of the superiority of the final
model.
EVALUATION OF THE MEASUREMENT MODEL
Table 2 presents the fit indices for the singular fluency mea-
surement model with two latent variables: (a) reading compre-
hension with four indicators (GMRT-4, Maze, oral QRI-4 ques-
tions, underlining QRI-4 questions) and (b) reading fluency with
three indicators (ORF, underlining, TOSCRF). The singular mea-
surement model demonstrated adequate fit, suggesting that it in-
cluded viable latent factors to which structural components could
be added. A second measurement model was tested that exam-
ined a latent silent reading fluency factor with two indicators, the
TOSCRF and the underlining measure scores. Although model
fit was adequate, high standard errors for regression weights
indicated disturbance that can probably be attributed to the
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
TABLE 2 FitIndicesforEachoftheFittedModels
X2df pX2 CFI TLI RMSEA (90% CI) SRMR AIC
Singular Measurement
Model
0.18 13 0.18 0.97 0.94 0.06 (0.00–0.12) 0.07 —
Singular Structural Model 17.4 12 0.18 0.97 0.94 0.06 (0.00–0.12) 0.07 47.4
Split Structural Model 14.1 12 0.29 0.98 0.97 0.04 (0.00–0.12) 0.07 46.1
Singular Model with
Subcomponents
53.4 36 0.03 0.94 0.90 0.07 (0.02–0.11) 0.08 113.4
Split Model with
Subcomponents
25.7 27 0.44 1.00 1.00 0.01 (0.00–0.08) 0.06 105.5
Note. CI = confidence interval.
19
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
20 K. W. Price et al.
significant but weak correlation between the TOSCRF and the
underlining measure (see Table 1). In other words, the two silent
reading measures seemed to represent somewhat different aspects
of silent reading. Because of this issue, it was determined that it
would be more statistically sound to specify each silent reading
fluency measure as a manifest variable within subsequent models
as opposed to indicators of a single latent silent reading fluency
factor.
EVALUATION OF THE SEM MODELS
With data to indicate that the reading comprehension latent
factor and the singular reading fluency factor provide ample fit
to the data, structural components were added to both the sim-
ple singular model and the split model in order to address the
initial research question regarding whether oral reading fluency
and silent reading fluency represent distinct constructs. Fit in-
dices (see Table 2) for both models indicated adequate fit. How-
ever, a comparison of the fit indices for both structural models
indicated that in almost every instance, the fit indices for the split
model, depicting oral and silent reading fluency as separate con-
structs, demonstrated a stronger—albeit slightly stronger—fit to
the data.
EVALUATION OF THE STRUCTURAL MODELS WITH SUBCOMPONENTS
SKILLS
Next, we turned our attention to addressing the second re-
search question regarding the relations between the reading sub-
components (i.e. word reading, nonword reading, rapid auto-
matic naming, vocabulary) and oral and silent reading fluency.
Because the fit indices were rather similar and both the singu-
lar and split structural models indicated adequate fit, subcompo-
nent skills were added to both structural models. This ensured
that over reliance on a specific model based on previous decisions
did not obfuscate stronger, more complex structural models in
the third phase of the analyses. Fit for both the singular and split
structural models with subcomponent skills was deemed adequate
(see Table 2). However, a comparison of the fit indices for both
structural models indicated that in every case, the split model rep-
resented stronger fit to the data. Notably, a comparison of the
AIC values provided further evidence for the assertion that the
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 21
split model provided better fit to the data; the AIC value for the
split model (105.5) was lower than the AIC value for the singu-
lar model (113.4). Together these findings suggest that the split
model, in which oral and silent reading fluency were represented
as separate constructs, provided the best fit for our data and was
therefore accepted as the most viable model.
Standardized direct, indirect, and total effects from the split
model with subcomponents skills are presented in Table 3. These
coefficients, similar to beta weights from regression analyses, indi-
cate the proportion of standard deviation units that the endoge-
nous factor changes as a function of a one standard deviation
change in the exogenous factor. Standardized coefficient effect
sizes above 0.05 are considered small, effect sizes above 0.15 are
considered moderate, and effect sizes above 0.25 are considered
large (Kline, 2010). In addition to Table 2, the split model is pre-
sented in Figure 3, and significant and nonsignificant paths are
demarcated.
The amount of variance (i.e. squared multiple correlation)
explained for comprehension within the split structural model
was R2=0.47. Examining the standardized parameter estimates
in the split structural model indicates that oral reading fluency
(0.44) and vocabulary (0.46) both directly contributed signifi-
cantly to reading comprehension; indeed, both parameters fell
within the large range. In contrast, neither silent reading flu-
ency variable (underlining 0.02; TOSCRF 0.15) contributed sig-
nificantly to reading comprehension after controlling for the
other factors. Additionally, the oral reading fluency measure con-
tributed significantly to the underlining measure (0.62) to a large
degree, but it did not contribute significantly to the TOSCRF
(0.07).
An examination of the remaining subcomponent skills sug-
gested that, although word reading (0.29), nonword reading
(0.26), and rapid automatic naming (0.20) each contributed sig-
nificantly to the oral reading fluency measure, only word read-
ing contributed significantly to the TOSCRF (0.55), and none
of the subcomponent skills contributed significantly to the un-
derlining measure (all ≤|0.11|). The fact that word reading sig-
nificantly contributed to the TOSCRF measure can perhaps be
accounted for by the task demands of the TOSCRF. That is,
although the passages in the TOSCRF represented connected
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
22 K. W. Price et al.
TABLE 3 Standardized Direct, Indirect, and Total Effects on Reading
Comprehension for the Split Structural Model with Subcomponent Skills
Exogenous Variables Endogenous Variables within the Model
To To To To Comp-
ORF TOSCRF Underlining rehension
From Word Reading
direct
0.29∗0.55 ∗∗ −0.05 —
indirect — 0.02 0.18 0.22
total 0.29 0.57 0.13 0.22
From Nonword Reading
direct
0.26∗−0.19 −0.11 —
indirect — 0.02 0.16 0.09
total 0.26 −0.17 0.05 0.09
From Rapid Automatic
Naming direct
0.20∗0.14 −0.03 —
indirect — 0.01 0.12 0.11
total 0.20 0.16 0.09 0.11
From Picture Vocabulary
direct
−0.05 0.01 −0.04 0.36∗
indirect — 0.00 −0.03 −0.02
total −0.05 0.01 −0.07 0.34
From Underlining direct — — — 0.02
indirect — — — —
total — — — 0.02
From TOSCRF direct — — — 0.15
indirect — — — —
Total — — — 0 .15
From Oral Reading
Fluency direct
—0.07 0.62 ∗∗ 0.44 ∗∗
indirect — — — 0.02
total — 0.07 0.62 0.46
Note. ORF = Oral reading fluency; TOSCRF = Test of Contextualized Silent Reading
Fluency; Statistical significance is notated on direct effects only.
** p<0.01, *p<0.05.
text, students were required to identify and demarcate words
within the passage, which is perhaps more similar to a word
reading measure than the passage reading required in the un-
derlining task. Also, the indirect effects for the subcomponent
skills to reading comprehension (mediated by the reading flu-
ency variables) were not found to be statistically significant (see
Tabl e 3 ).
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 23
FIGURE 3 Split structural model with subcomponent relations. Statistically sig-
nificant path (p<0.05) coefficients are indicated by solid lines whereas non-
significant path coefficients are indicated by dotted lines.
EVALUATION OF ALTERNATIVE MODELS
Due to some surprising relations between the model variables
(i.e. the weak relation between underlining and the TOSCRF) as
well as the closeness of the fit of the singular and split models,
two alternative models were tested based on results from the pre-
viously run models. The fit indices for each of these alternative
models are presented in Table 4, and the fit indices for the orig-
inal split model with subcomponent skills are included as a base-
line for comparison.
First, the TOSCRF was removed from the model and the un-
derlining measure was utilized as the only silent reading fluency
variable. This model was tested in order to determine whether a
model with only a single silent text reading fluency measure would
better fit the current data set. Underlining was retained because
of the significant contribution of oral reading fluency to this vari-
able. In general, although the values for the fit indices indicated
excellent fit to the data, in each case, the fit indices were strongest
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
TABLE 4 Fit Indices for Each of the Fitted Alternative Models
X2df pX2CFI TLI RMSEA (90% CI) SRMR AIC
Split Model with Subcomponents 25.7 27 0.44 1.00 1.00 0.01 (0.00–0.08) 0.06 105.5
Split Model without TOSCRF 23.8 23 0.42 0.98 0.99 0.02 (0.00–0.09) 0.06 –
TOSCRF as a Subcomponent 28.5 27 0.38 0.99 0.99 0.02 (0.00–0.08) 0.06 106.5
Note. CI = confidence interval.
24
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 25
for the original split model with the TOSCRF and the subcompo-
nent skills. Because the AIC value is partially tied to the number
of manifest variables present in the sample, the removal of the
TOSCRF as a manifest variable reduces its appropriateness as a
method of comparing models. Therefore, the AIC was not pre-
sented for that model in Table 4.
Second, a model was tested in which the TOSCRF was con-
ceptualized as subcomponent skill contributing directly to the un-
derlining and the oral reading fluency measure. This is in contrast
to previous models where the TOSCRF served as a mediator be-
tween the subcomponent skills and reading comprehension. This
model tested the hypothesis that the TOSCRF is more akin to a
silent word reading fluency measure as opposed to a silent text
reading measure. Again, although the fit indices indicated ade-
quate fit to the data (see Table 4) across all indices, the values for
the original split model with subcomponent skills were stronger.
In this comparison, the split model and the model with the TO-
SCRF as a subcomponent had the same number of manifest vari-
ables, and therefore, the AIC was used an appropriate compara-
tive index. Notably, the AIC scores indicated that the split model
(AIC =105.5) had stronger fit to the data than did the TOSCRF
subcomponent model (106.5). Across all SEM models tested, con-
sistently stronger fit indices suggested that the split model pro-
vided a more viable explanation for the data than did the singular
or alternative models.
Discussion
Although oral reading fluency has received considerable atten-
tion in the school-based literatures across the past decade, silent
reading fluency has remained largely overlooked. The present
study contributed to our understanding of reading processes by
utilizing SEM to examine both types of reading fluency and their
relation to overall reading comprehension abilities in fourth-
grade students. Further, several subcomponents of the reading
process were included in the examined models in order to de-
termine how these additional skill sets support fluency and com-
prehension.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
26 K. W. Price et al.
Do Oral and Silent Reading Fluency Represent Separate Constructs?
The initial research question addressed whether oral and silent
reading represent distinct constructs in a normative sample of
fourth-grade students. It is largely assumed that oral and silent
reading involve essentially the same processes, but recently re-
searchers have advocated for the need to understand the unique
role of silent reading fluency (Hiebert et al., 2012; Share, 2008).
Consistent across analyses, models that split oral and silent read-
ing into separate constructs provided a better fit to our data.
These fluency-specific results are consistent with the limited ex-
tent literature on silent reading fluency and support the depic-
tion of oral and silent reading as separate constructs (Kim et al.,
2011).
Our results also indicated that oral reading fluency con-
tributed significantly to comprehension, which is consistent with
findings across a variety of diverse samples from students rang-
ing from the primary to the secondary grades (Daane et al., 2005;
Fuchs et al., 1988; Jenkins et al., 2003; Pinnell et al., 1995; Yavanoff
et al., 2005). This finding was supported at the holistic level, as
oral reading was more predictive of students’ reading comprehen-
sion within the SEMs. Further, students participating in this study
answered on average 1.5 more comprehension questions correctly
following oral as opposed to silent reading, even as the passages
were held constant (see Table 1).
Importantly, the silent reading fluency measures did not con-
tribute significantly to reading comprehension. This finding is
consistent with some previous literature, which suggests that prior
to fifth grade, students comprehend better after oral reading than
after silent reading (Elgart, 1978; Fletcher & Pumphrey, 1988;
Prior & Welling, 2001). An important caveat should be noted,
however. Interestingly, the two silent reading fluency measures
used in this study did not result in a stable, latent silent reading
fluency factor. This result, although somewhat surprising, perhaps
stemmed from the statistically significant but weak correlation be-
tween the two silent reading fluency measures. Although both
measures have been validated against other measures of silent
reading, it is notable that the two measures assess silent read-
ing fluency using very different methods. The underlining task
was a more ecologically valid passage-reading task, whereas the
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 27
TOSCRF was less similar to everyday silent reading tasks as it re-
quired students to identify and demarcate individual words within
text.
The Relation of Subcomponent Skills to Reading Fluency and Reading
Comprehension
The second aim of this study was to examine the contributions
of various subcomponent skills to oral and silent reading fluency
and reading comprehension in late elementary readers. The in-
clusion of these variables provided a more comprehensive model
of the reading process. Although several previous studies have ex-
amined these subcomponent skills and how they contribute to
oral reading fluency, reading comprehension, or both, few studies
have incorporated silent reading fluency in such investigations.
Specifically, we predicted that (a) those subcomponents that re-
quire the oral rendering of text would be more closely related to
oral rather than silent reading fluency (i.e. word reading, rapid
automatic naming, and nonword reading), (b) that vocabulary
would contribute directly to comprehension above and beyond
the contributions of the other subcomponent skills and reading
fluency, and (c) that the reading subcomponent skills would con-
tribute a smaller proportion of the variance to comprehension
than would reading fluency. An examination of the model pa-
rameters suggested that RAN, nonword reading, and word read-
ing (measures that required oral production of text) each con-
tributed significantly to the oral reading fluency measure. Also
consistent with our hypotheses, different patterns were observed
with regard to silent reading fluency. None of the subcomponent
skills contributed significantly to the underlining measure, and
only the word reading task contributed significantly to the TO-
SCRF. However, oral reading fluency did contribute significantly
to the underlining measure. This may suggest that oral reading
fluency, rather than the other early emerging reading subcom-
ponent skills, is supporting the development of silent reading
fluency.
Importantly, vocabulary stood out among the subcompo-
nents as an important contributor to reading comprehension. Al-
though the vocabulary factor did not contribute in any meaning-
ful way to the oral or silent reading fluency measures, it is notable
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
28 K. W. Price et al.
that it contributed strongly to the comprehension measure, even
when controlling for the fluency factors. These findings are con-
sistent with an emerging literature that suggests that vocabulary
is an important component of the reading process and should
not be overlooked in the examination of reading development
(Berninger et al., 2010; Ouellette & Beers, 2010).
With the exception of vocabulary, the impact of the subcom-
ponent skills on comprehension was lower than their effect on
reading fluency. Indeed, after taking into account oral and silent
reading fluency, the RAN, nonword reading, and word reading
variables did not contribute significantly to reading comprehen-
sion. These findings are consistent with previous studies which
have suggested that, although basic reading competencies are
essential to comprehension for emergent readers, their contri-
butions diminish across development and are less predictive of
comprehension as children gain in proficiency and begin to en-
counter more complex texts (Floyd et al., 2012; Jenkins & Jew-
ell, 1993; Shinn, Good, Knutson, Tilly, & Collins, 1992; Vellutino
et al., 2007). The particularly small total effect from nonword
reading is consistent with the literature, suggesting that the largest
gains in phonemic and phonological awareness occur in the first
year of reading instruction, and that by late elementary school,
non-word reading skills should be less predictive of students over-
all reading abilities (Share, 2008; Vellutino et al., 2007).
Limitations and Future Directions
The current study contributes to the literature regarding stu-
dents’ reading fluency, vocabulary, and comprehension develop-
ment in late elementary school. There are, however, limitations
that are worth pointing out. First, although the present sample
was adequate and the use of complex statistical analyses provided
the opportunity to examine relationships between the examined
factors, studies utilizing a longitudinal sample of older readers
(e.g., fifth and sixth grade) would provide important information
about developmental changes in this age group. Our results pro-
vide additional evidence to support the conceptualization of oral
and silent reading fluency as separate constructs, but the data fur-
ther indicate that students in the current sample had not yet fully
transitioned to silent reading for comprehension.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 29
A rather extensive battery of reading assessments was em-
ployed in this study, but some factors which may provide addi-
tional information about reading development were not exam-
ined. For example, inclusion of a non-verbal processing speed
measure would provide interesting information about how this
subcomponent is differentially related to oral and silent reading
fluency in late elementary readers. Such information would be
especially interesting given the findings in the present study that
indicated that RAN contributes significantly to oral reading but
not to either silent reading measure. It may be that non-verbal
processing speed shows the opposite pattern of relations. Two
measures of students’ silent reading fluency were included in the
present study, but these measures did not provide a stable factor.
As discussed previously, they might have been measuring differ-
ent aspects of silent reading fluency (i.e. word vs. text fluency).
Future studies should examine whether another text silent read-
ing fluency assessment such as a traditional paper-and-pencil mea-
sure (see Price et al., 2012) would shed light on this issue. Finally,
some studies in the literature have suggested that the relationship
between comprehension and fluency is bidirectional (Klauda &
Guthrie, 2008). Future studies should determine how the inclu-
sion of such a parameter would impact the present findings.
Theoretical and Practical Implications
Silent reading fluency and vocabulary are often overlooked in
studies modeling reading processes in elementary students, yet
our results suggest they are important variables to include. Fur-
thermore, our findings suggest the importance of differentiating
between oral and silent reading fluency interventions, especially
as students reach the late elementary grades and the curriculum
shifts from a focus on oral reading fluency and learning to read
to silent reading fluency and reading to learn. Although our data
suggest that proficient oral reading may support the development
of proficient silent reading, oral reading should not be allowed
to subsume its silent counterpart. Additionally, it may be nec-
essary to provide additional support or interventions specific to
silent reading (Scaffolded Silent Reading; Hiebert et al., 2012) to
fourth-grade students if they are expected to independently read
“in their head” and gain content area knowledge from texts.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
30 K. W. Price et al.
Our results also suggest that vocabulary provides important
information about the overall picture of reading development
and should be considered when modeling the reading process.
This finding is consistent with previous works, suggesting that chil-
dren’s comprehension is negatively impacted when texts contain
hard or unfamiliar words (Rayner & Pollatsek, 1989), and that ef-
fective vocabulary instruction focused on target words increases
children’s comprehension of text (Pullen, Tuckwiller, Konold,
Maynard, & Coyne, 2010). In sum, these results suggest that vo-
cabulary should remain an important component of reading cur-
ricula, and models overlooking vocabulary may provide an incom-
plete picture of the reading process.
Conclusion
The emphasis on oral reading fluency in the school-based lit-
eratures could have been explained by the fact that both types
of reading fluency should be considered as a single construct.
However, the current study provides evidence that oral and silent
reading fluency represent different constructs in late-elementary
readers, and that each type of fluency is differentially related
to comprehension. Interestingly, for the fourth-grade students
who participated in this study, oral rather than silent reading
fluency was supporting effective comprehension. Additional re-
search is needed to elucidate the developmental relation be-
tween silent reading fluency and reading comprehension to
ensure curricular goals are consistent with students’ reading
development.
Notes
1. The procedures originally included a third, group-administered silent read-
ing fluency measure that utilized the QRI-4 passages. Passages were counter-
balanced, taking in account the need to counterbalance including this third
measure. Due to time constraints during the group assessment, this third mea-
sure was dropped from the study, and therefore each student read only four of
the six possible QRI-4 passages. Students were equally likely to read any com-
bination of the four passages across the remaining oral reading fluency and
underlining tasks. Permission was obtained from the test publisher to present
the QRI-4 passages on a Tablet PC during the underlining task.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 31
2. Two passages were administered for each measure in order for the six fourth-
grade QRI-4 passages to be utilized. Additionally, a third passage was not ad-
ministered due to concerns about the length of testing time.
3. The final models were run with and without the missing participants and re-
sults were comparable, which suggests the removal of these eight participants
did not have a large impact on the final results.
Acknowledgments
We would like to express our sincere appreciation to Tera Bradley,
Sarah Irby, Rachel Peterman, and Colby Taylor for their assistance
in collecting the data, to Drs. Randy Floyd and Gilbert Parra who
served as dissertation committee members and provided valuable
feedback regarding this project, and to the students and teachers
who participated in this research project.
References
Barth, A. E., Catts, H. W., & Anthony, J. L., (2009). The component skills under-
ling reading fluency in adolescent readers: A latent variable analysis. Reading
and Writing,22, 567–590.
Berninger, V. W., Abbott, R. D., Trivedi, P., Olson, E., Gould, L., Hiramatsu, S.,
& Westhaggen, S. Y., (2010). Applying the multiple dimensions of reading
fluency to assessment and instruction. Journal of Psychoeducational Assessment,
28, 3–18.
Braze, D., Tabor, W., Shankweiler, D. P., & Mencl, W. E. (2007). Speaking up
for vocabulary: Reading skill difference in young adults. Journal of Learning
Disabilities,40, 226–243.
Breaux, K. C. (2009) WIAT-3 technical manual. San Antonio, TX: Pearson.
Burge, P. E. (1983). Comprehension and rate: Oral vs. silent reading for low
achievers. Reading Horizons, 23, 201–206.
Chall, J. S. (1996). Learning to read: The great debate (Revised). New York, NY:
McGraw-Hill.
Daane, M. C., Campbell, J. R., Grigg, W. S., Goodman, M. J., & Oranje, A. (2005).
Fourth-grade students reading aloud: NAEP 2002 Special Study of Oral Reading. U.S.
Department of Education. National Center for Education Statistics. Washing-
ton, DC: Government Printing Office.
Elgart, D. B. (1978). Oral reading, silent reading, and listening comprehension:
A comparative study. Journal of Reading Behavior,10, 203–207.
Fletcher, J., & Pumfrey, P. D. (1988). Differences in text comprehension amongst
7–8 year old children. School Psychology International,9, 133–145.
Floyd, R. G., Meisinger, E. B., Gregg, N., & Keith, T. Z. (2012). An expla-
nation of reading comprehension across development using models from
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
32 K. W. Price et al.
tell–Horn–Carroll theory: Support for integrative models of reading. Psychol-
ogy in the Schools,48, 725–743.
Fuchs, L. S., & Fuchs, D. (1992). Identifying a measure for monitoring student
reading progress. School Psychology Review,21, 45–58.
Fuchs, L. S., Fuchs, D., & Maxwell, L. (1988). The validity of informal reading
comprehension measures. Remedial & Special Education,9, 20–28.
Fuchs, L. S., & Maxwell, L. (1988). Interactive effects of reading mode, produc-
tion format, and structural importance of text among LD pupils. Learning
Disability Quarterly,11, 97–105.
Good, R. H., & Kaminski, R. A., & Dill, S. (2002). DIBELS Oral Reading Fluency.
In R. H. Good & R. A. Kaminski (Eds.), Dynamic Indicators of Basic Early Literacy
Skills (6th ed.). Eugene, OR: Institute for the Development of Educational
Achievement. Available: http://dibels.uoregon.edu/.
Hammill, D. D., Wiederholt, J. L., & Allen, E. A. (2006). Test of Silent Word Reading
Fluency. Austin, TX: Pro-ed.
Hiebert, E. H., Samuels, S. J., & Rasinski, T. (2012). Comprehension-based silent
reading rates: What do we know? What do we need to know? Literacy Research
and Instruction,51, 110–124.
Holmes, B. C. (1985). The effect of four different modes of reading on compre-
hension. Reading Research Quarterly,20, 575–585.
Hoover, W. A., & Gough, P. B. (1990). The simple view of reading. Reading and
Writing,2, 127–160.
Jenkins, J. R., Fuchs, L. S., van den Broek, P., Espin, C., & Deno, S. L. (2003).
Sources of individual differences in reading comprehension and reading flu-
ency. Journal of Educational Psychology,95, 719–729.
Jenkins, J. R., & Jewell, M. (1993). Examining the validity of two measures
for formative teaching: Reading aloud and maze. Exceptional Children,59,
421–432.
Jones, E. B. (1932). A comparison of comprehension results in oral and silent
reading. Peabody Journal of Education,9, 292–296.
Keenan, J. M., Betjemann, R. S., & Olson, R. K. (2008). Reading comprehension
tests vary in the skills they assess: Differential dependence on decoding and
oral comprehension. Scientific Studies of Reading,12, 281–300.
Kendeou, P., van den Broek, P., White, M. J., & Lynch, J. S. (2009). Predicting
reading comprehension in early elementary school: The independent contri-
butions of oral language and decoding skills. Journal of Educational Psychology,
101, 765–778.
Kim, Y., Wagner, R. K., & Foster, E. (2011). Relations among oral reading fluency,
silent reading fluency, and reading comprehension: A latent variable study of
first-grade readers. Scientific Studies of Reading,15, 338–362.
Klauda, S. L., & Guthrie, J. T. (2008). Relationships of three components of read-
ing fluency to reading comprehension. Journal of Educational Psychology,100,
310–321.
Kline, R. B. (2010). Principles and practice of structural equation modeling (3rd ed.).
New York, NY: Guilford Press.
Kuhn, M. R., & Schwanenflugel, P. J. (2007). Fluency in the classroom. New York,
NY: Guilford Press.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 33
Kuhn, M. R., Schwanenflugel, P. J., & Meisinger, E. B. (2010). Aligning theory
and assessment of reading fluency: Automaticity, prosody, and definitions of
fluency. Reading Research Quarterly,45, 232–253.
Kuhn, M. R., & Stahl, S. S. (2003). Fluency: A review of developmental and re-
medial practices. Journal of Educational Psychology,95, 3–21.
LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information
processing in reading. Cognitive Psychology,6, 293–323.
Leslie, L., & Caldwell, J. S. (2006). Qualitative Reading Inventory (4th ed.). Boston,
MA: Allyn and Bacon.
Little, J. A. (1988). A test of missing completely at random for multivariate data
with missing values. Journal of the American Statistical Association,83, 1198–1202.
MacGinitie, W. H., MacGinitie, R. K., Maria, K., & Dreyer, L. G. (2007). Gates-
MacGinitie Reading Tests, 4th edition technical report supplement forms S and T.
Rolling Meadows, IL: Riverside Press and Publishing.
MacGinitie, W. H., MacGinitie, R. K., Maria, K., Dreyer, L. G., & Hughes, K. E.
(2007). Gates-MacGinitie Reading Tests, (4th ed.). Rolling Meadows, IL: River-
side Press and Publishing.
Miller, S. D., & Smith, D. E. (1985). Differences in literal and inferential com-
prehension after reading orally and silently. Journal of Educational Psychology,
77, 341–348.
Muster, V., Hulme, C., Snowling, M. J., & Stevenson, J. (2004). Phonemes,
rimes, vocabulary, and grammatical skills as foundations of early reading de-
velopment: Evidence from a longitudinal study. Developmental Psychology,40,
665–681.
National Reading Panel. (2000). Report of the national reading panel. Washington
DC: Author.
Norton, E. S., & Wolf, M. (2012). Rapid automatized naming (RAN) and reading
fluency: Implications for understanding and treatment of reading disabilities.
Annual Review of Psychology,63, 427–452.
Ouellette, G., & Beers, A. (2010). A not-so-simple view of reading: How oral vo-
cabulary and visual-word recognition complicate the story. Reading and Writ-
ing,23, 189–208.
Pinnell, G. S., Pikulski, J. J., Wixon, K. K., Campbell, J. R., Gough, P. B., & Beatty,
A. S. (1995). Listening to children read aloud: Data from NAEP’s integrated reading
performance record (IRPR) at grade 4. The nation’s report card. Report No. 23-FR-
04. Washington, DC: Office of Educational Research and Improvement, U.S.
Department of Education.
Price, K. W., Meisinger, E. B., D’Mello, S. K., & Louwerse, M. M. (2012). Silent
reading fluency using underlining: Evidence for an alternative method of as-
sessment. Psychology in the Schools,49, 606–618.
Prior, S. M., & Welling, K. A., (2001). “Read in your head”: A Vygotskian analysis
of the transition for oral to silent reading. Reading Psychology,22, 1–15.
Psychological Corporation. (2009). Wechsler Individual Achievement Test (3rd ed.).
San Antonio, TX: Author.
Pullen, P. C., Tuckwiller, E. D., Konold, T. R., Maynard, K. L., & Coyne, M. D.
(2010). A tiered intervention model for early vocabulary instruction: The
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
34 K. W. Price et al.
effects of tiered instruction for young students at risk for reading disability.
Learning Disabilities Research & Practice,25, 110–123.
Radach, R., Schmitten, C., Glover, L., & Huestegge, L. (2009). How children
read for comprehension: Eye movements in developing readers. In R. K. Wag-
ner, C. Schatschneider, & C. Phythian-Sence (Eds.) Beyond decoding: The behav-
ioral and biological foundations of reading comprehension (pp. 75–106). New York,
NY: Guilford.
Rayner, K. & Pollatsek, A. (1989). Reading psychology. Hillsdale, NJ: Lawrence Erl-
baum.
Salasso, A. (1986) Cognitive processing in oral and silent reading comprehen-
sion. Reading Research Quarterly,21, 59–69.
Schwanenflugel, P. J., Meisinger, E. B., Wisenbaker, J. M., Kuhn, M. R., Strauss,
G. P., & Morris, R. D. (2006). Becoming a fluent and automatic reader in the
early elementary school years. Reading Research Quarterly,41, 496–522.
Schwanenflugel, P. J., Morris, R. K., Kuhn, M. R., Strauss, G. P., & Sieczko, J.
M. (2008). The influence of word unit size on the development of Stroop
interference in early word decoding. Reading and Writing: An Interdisciplinary
Journal,2, 177–203.
Senechal, M. (2006). Testing the home literacy model: Parent involvement in
kindergarten is differentially related to Grade 4 reading comprehension,
spelling, and reading for pleasure. Scientific Studies of Reading,10, 59–87.
Share, D. L. (2008). On the anglocentricities of current reading research and
practice: The perils of overreliance on an “outlier” orthography. Psychological
Bulletin,134, 584–615.
Shaw, R., & Shaw, D. (2002). DIBELS oral reading fluency-based indicators of third
grade reading skills for Colorado State Assessment Program (CSAP). (Technical Re-
port). Eugene, OR: University of Oregon.
Shinn, M. R., Good, R. H., Knutson, N., Tilly, W. D., & Collins, V. L. (1992).
Curriculum-based measurement: A confirmatory analysis of its relationship
to reading. School Psychology Review,21, 459–479.
Shinn, M. R., & Shinn, M. M. (2002). AIMSweb training workbook: Administration
and scoring of reading maze for use in general outcome measurement. Eden Prairie,
MN: Edformation.
Siegel, L. S. (1993). Phonological processing deficits as the basis of a reading
disability. Developmental Review,13, 246–257.
Sweet, A. P., & Snow, C. E. (2003). Rethinking reading comprehension: Solving prob-
lems in the teaching of literacy. New York, NY: Guilford Press.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (5th ed.),
Boston, MA: Pearson.
Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Specific
reading disability (dyslexia): What have we learned in the past four decades?
Journal of Child Psychology and Psychiatry,45, 2–40.
Vellutino, F. R., Tunmer, W. E., Jaccard, J. J., & Chen, R. (2007). Components of
reading ability: Multivariate evidence of a convergent skills model of reading
development. Scientific Studies of Reading,11, 3–21.
Wechsler, D. (2001). Wechsler Individual Achievement Test (2nd ed.). San Antonio,
TX: Pearson.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
Oral and Silent Reading Fluency 35
Wolf, M., Bowers, P., & Biddle, K. R. (2000). Naming-speed processes, timing,
and reading: A conceptual review. Journal of Learning Disabilities,33, 387–407.
Woodcock, R. W., McGrew, K. S., & Mather, N. (2001). Woodcock-Johnson III.
Rolling Meadows, IL: Riverside Publishing.
Woodcock, R. W., McGrew, K., Schrank, F. A., & Mather, N. (2007). The Woodcock-
Johnson III Normative Update (WJ III NU). Itasch, IL: Riverside Publishing.
Yovanoff, P., Duesbery, L., Alonzo, J., & Tindal, G. (2005). Grade-level invariance
of a theoretical causal structure predicting reading comprehension with vo-
cabulary and oral reading fluency. Educational Measurement: Issues and Practice,
24, 4–12.
Downloaded by [University of Memphis Libraries] at 13:17 13 April 2015
