You are on page 1of 28

The Effects of a Supplementary Computerized Fluency

Intervention on the Generalization of the Oral Reading


Fluency and Comprehension of First-Grade Students

Lenwood Gibson Jr., Gwendolyn Cartledge, Starr E. Keyes, Christopher D. Yawn

Education and Treatment of Children, Volume 37, Number 1, February 2014,


pp. 25-51 (Article)

Published by West Virginia University Press


DOI: https://doi.org/10.1353/etc.2014.0003

For additional information about this article


https://muse.jhu.edu/article/535928

Access provided by University of Athens (or National and Kapodistrian Univ. of Athens) (18 May 2018 19:56 GMT)
EDUCATION AND TREATMENT OF CHILDREN Vol. 37, No. 1, 2014

The Effects of a Supplementary Computerized


Fluency Intervention on the Generalization of
the Oral Reading Fluency and Comprehension of
First-Grade Students
Lenwood Gibson, Jr.
The City College of New York
Gwendolyn Cartledge
The Ohio State University
Starr E. Keyes
Bowling Green University
Christopher D. Yawn
The City College of New York

Abstract
The current study investigated the effects of a repeated reading interven-
tion on the oral reading fluency (ORF) and comprehension on generalization
passages for eight, first-grade students with reading risk. The intervention
involved a commercial computerized program (Read Naturally Software
Edition [RNSE], 2009) and a generalization principle requiring a “greater
magnitude” of responding. A multiple probe experimental design with two
treatment phases was used to determine the effects of the intervention. Phase
I used a standard end-of-year benchmark score (i.e., 40 CWMP) as fluency
criteria for all participants. During Phase II, fluency criteria were changed
and individualized for each participant based on performance during Phase
I. Data were collected on ORF and word retell fluency (WRF) across treat-
ment and generalization probes. Results showed ORF and comprehension
increases in both phases; however, satisfactory generalization did not occur
for most of the participants until the second phase was implemented. These
results are discussed relative to classroom implications and directions for fu-
ture research.

R eading is probably the most important academic skill because


it is the basis on which many other academic activities are built
(Flanagan, West, & Walston, 2004; Lyon, 1998; 2003). There is evidence

Correspondence to Lenwood Gibson, The City College of New York School of Educa-
tion, NAC Building, Room 3/227F 160 Convent Ave & 138th St. New York, NY. 10031;
e-mail: lgibson@ccny.cuny.edu.

Pages 25–51
26 GIBSON, et al.

that students who fail to acquire adequate reading skills during early
primary grades are much more likely to experience chronic difficulties
in school (Alexander, Entwisle, & Horsey, 1997; McIntosh, Flannery,
Sugai, Braun, & Cockrane, 2008; Simner & Barnes, 1991). Considering
what is at stake, it is particularly important to help young readers ac-
quire critical reading skills such as reading fluency. Reading fluency is
defined as the ability to read connected text with speed, accuracy, and
proper expression (Bursuck & Damer, 2011). This skill is of critical im-
portance because it is positively correlated with reading comprehen-
sion (Fuchs, Fuchs, Hosp, & Jenkins, 2001; Martens et al., 2007; Reis
et al., 2007), and is based on the idea that students who read fluently
are much more likely to concentrate on the meaning of the text rather
than on the pronunciation of specific words.
A way to assess a student’s proficiency in reading fluency is by
measuring the rate at which he/she reads connected text passages out
loud. This measure is referred to as oral reading fluency (ORF) and
rates are typically measured using novel or unfamiliar text passag-
es during one-minute timed tests. Interventions for increasing oral
reading fluency appear to favor repeated reading instruction (Na-
tional Institute of Child Health and Human Development [NICHD],
2000), which entails repeatedly reading the same passage smoothly
and accurately to meet a specified criterion (e.g., 40 correct words per
minute).
Recent research examining ORF instruction focused on the ben-
eficial effects of multi-component strategies (Lo, Cooke, & Starling,
2011; Martens et al., 2007). These strategies allow researchers to imple-
ment and study the different instructional strategies used to increase
reading fluency. For example, Lo et al. (2011) effectively used an
eight-component strategy to increase the ORF of three second-grade
students who struggled in reading. Each component was designed
to address and strengthen deficits in reading fluency. The current
study also used a multi-component strategy to increase the ORF of
first grade students.
In addition the type of strategies used to increase ORF an-
other important focus has been on generalization (Ardoin, Eckert,
& Cole, 2008; Ardoin, McCall, & Klubnik, 2007; Begeny, Daly III &
Valleley, 2006; Martens, et al., 2007; Silber & Martens, 2010). There
are a number of ways to promote generalization but use of multiple
exemplars as described by Stokes and Baer (1977) has gained atten-
tion when studying ORF. In this approach learners are exposed to
multiple examples of correct responding to increase the likelihood of
this responding transferring to novel or unlearned stimulus condi-
tions. Ardoin and colleagues (2008) compared repeated reading to the
COMPUTERIZED FLUENCY INTERVENTION 27

multiple exemplar approach which required the participants to read


three variations of the same story and then read either a high word
or medium word overlap generalization passage. The results dem-
onstrated that both phases produced higher rates of ORF on training
passages but generalization occurred only with passages containing
medium word overlap during the multiple exemplar condition. The
inconclusive findings for the multiple exemplar approach pointed to
the need for more research.
In an extension of the Ardoin et al. (2008) study, Silber and
Martens (2010) also investigated the use of a multiple exemplar ap-
proach and compared it to a listening passage preview/repeated read-
ing strategy. Both conditions produced higher ORF rates on training
passages compared to the control group; however, only the multiple
exemplar intervention produced generalization gains for oral reading
fluency compared to the control group but no differences were noted
between the two experimental conditions. Albeit encouraging, these
modest generalization gains warrant further investigations of gener-
alization principles that might produce even greater effects (Silber &
Martens, 2010).
Another option for promoting generalization involves teaching
behavior to levels required by natural contingencies of reinforcement
(Cooper, Heron, & Heward, 2007). Cooper et al. (2007) suggest when
planning for generalization, practitioners should identify and rein-
force levels of performance that are desired in the absence of training
conditions. In the case of promoting generalization of ORF, students
should read practice passages at a rate that is expected on untrained
material. Many repeated reading studies used set performance lev-
els as criteria to indicate “fluent reading” (Kostewicz & Kubina, 2010;
Musti-Rao, Hawkins, & Barkley, 2009) but it is possible that only
training students to reach a predetermined performance level is not
sufficient to produce generalization. Cooper et al. (2007) suggest the
following:
The learners should be taught to emit the target behavior at a rate
commensurate with the naturally occurring contingency, with
more accuracy, within a shorter latency, and/or at a greater magni-
tude. Generalization planning should include identification of the
levels of performance necessary to access existing criteria for rein-
forcement. (p. 635-636)
Taking this recommendation into consideration, the current
study focused on manipulating the performance expectations dur-
ing the practice passages to determine which level of performance
would be better at promoting generalization on untrained passages.
28 GIBSON, et al.

The performance goals were changed to assess if generalization was


more effective under higher or standard criteria. Thus, the goal cri-
terion was based on the best individual performance of each par-
ticipant in the first phrase. Furthermore, the reinforcer was the
completion of the practice story and moving to the next step in the
RNSE program sequence, which is more in line with the “naturally
occurring contingencies” of reinforcement suggested by Cooper et
al. (2007).
In addition to investigating effective ways to increase gener-
alization, the use of supplemental instructional programs is also a
concern. With the recent emphasis on response to intervention (RtI;
Fuchs & Fuchs, 2006) there is good evidence that supplemental in-
terventions can be effective for students showing academic risk (e.g.,
Cartledge, Yurick, Singh, Keyes, & Kourea, 2011). An important is-
sue (and potential problem) with the use of supplemental interven-
tions is the time and resources needed to implement them properly.
This is especially true of supplemental instructional strategies that
use multi-component interventions, where schools may lack the per-
sonnel and time needed for small group/individualized implemen-
tation. One option is to deliver the intervention through computer
software, allowing more for independent pupil delivery rather than
relying solely on the teacher. The current study implements a com-
puter program explicitly as supplementary instruction to determine
its effectiveness.
There were three purposes to this study. First, the researchers
questioned to what extent would increasing the ORF criteria on prac-
tice passages promote generalization of ORF to untrained grade level
passages? A second related question focused on method of delivery.
That is, could a multi-component, computer software program (Read
Naturally Software Edition, [RNSE], 2009) be used effectively to in-
crease the ORF of the first grade participants on both training and
generalization passages? A final question pertained to comprehen-
sion. An ostensible purpose for helping learners become more fluent
in their reading is to keep reading from being so laborious to facilitate
comprehension and the appeal of reading. There is some evidence
of a positive relationship between fluency and reading comprehen-
sion (Kim, Wagner, & Lopez, 2012; Neddenriep, Fritz, & Carrie, 2011).
Therefore, we wanted to further verify those findings and questioned
to what extent would this intervention improve comprehension, that
is, the word retell fluency of the first-grade participants on both train-
ing and generalization passages?
COMPUTERIZED FLUENCY INTERVENTION 29

Method
Participants and Settings
This study included eight students enrolled in two separate
schools (i.e., four students in each school). All the students were in
first grade, ranging in age from 6 to 8 years and were selected because
of their difficulties in reading fluency and comprehension. Prior to
inclusion in the study potential participants were administered two
Dynamic Indicator of Basic Early Literacy Skills (DIBELS; Good &
Kaminski, 2002) winter benchmark assessments (i.e., Nonsense Word
Fluency and Oral Reading Fluency). Each participant needed to meet
the following selection criteria: score in either the emerging (30-49) or
established (50 or higher) range on Nonsense Word Fluency (NWF).
The NWF measure was used to ensure that participants were able to
decode grade level words without too much difficulty. Once partici-
pants exhibited sufficient decoding skills, they needed to score either
at risk (0-7 cwpm) or some risk (8-19 cwpm) on the ORF assessment.
These are the cutoff scores established by DIBELS to indicate the level
of reading risk for the middle of first grade. Table 1 presents the de-
mographic information for each participant and his/her pre-interven-
tion DIBELS scores.
School One was a charter school located in a predominately Af-
rican American section of the city. The residents were considered low
to low-middle income with nearly 70% of the students eligible for free
or reduced lunch. The school listed the ethnic make-up of its student
body as 96% African American, 1% American Indian, 1% White, and
2% unknown. At the time of this study the enrollment of 663 students
and 28 teachers, the student-teacher ratio was 23.6:1. Thirteen percent
of its students were considered to have a disability compared to the
14% average for the state. Reading Mastery was the core instructional
program for reading, emphasizing phonemic awareness skills begin-
ning at the kindergarten level and the school used a response to inter-
vention (RtI) model to place students in the reading program.
School Two was an elementary charter school located within a
predominately African American community and enrolled approxi-
mately 220 students. African Americans made up 97% of these stu-
dents and 94% of the student body were economically disadvantaged.
The reading curriculum at the primary grades was a Direct Instruction
program with a heavy emphasis on culturally relevant literature. Read-
ing instruction was delivered using a small group format, similar to the
RtI model. Although both schools were implementing the RtI model
from the beginning of the school year, the number of years the schools
used this approach is not known. All participants engaged in their dai-
ly classroom reading instruction in addition to the intervention.
30 GIBSON, et al.

Table 1
Demographic and DIBELS pre-assessment information for
participant in School One and Two
DIBELS winter benchmark
Name Age Race Gender
NWF ORF

Marcus 7-5 AA M 36 Emerging 7 At-risk

Cleveland 6-5 AA M 39 Emerging 9 Some

Michael 7-1 AA M 39 Emerging 14 Some

Tylon 6-10 AA M 31 Emerging 9 Some

Lawrence 8-0 AA M 30 Emerging 6 At-risk

Stewie 6-4 AA M 32 Emerging 7 At-risk

Sheltrie 6-3 AA F 32 Emerging 16 Some

Amber 8-0 AA F 42 Emerging 15 Some

Dependent Variables
The first dependent variable for this study was the number of
correct words read during a one-minute timed reading. At the end
of the training sequence, the participant read the same passage while
the experimenter (i.e., the first author) recorded the number of correct
words in a one-minute timing. A word was counted correct when a
participant pronounced it accurately within three seconds of the pre-
vious word being read. A word was counted incorrect if the partici-
pant mispronounced it or failed to read it within three seconds of the
previous word being read.
The second dependent variable was the number of correct words
read per minute on generalization passages. These passages were
taken from the first grade AimsWeb progress monitoring assessment
for oral reading fluency. The experimenter administered each gener-
alization probe directly following every third ORF and comprehen-
sion treatment probes. Data were collected on the number of correct
words read at the end of the one-minute timing. A word was counted
as correct when it was pronounced accurately within three seconds
of the previous word being read. A word was counted incorrect if the
student mispronounced it or failed to read the word within 3 seconds
of the previous word being read.
The third and fourth dependent variables were the number of
words retold on a one-minute timing following the ORF treatment
and generalization probes. The experimenter instructed participants
COMPUTERIZED FLUENCY INTERVENTION 31

to verbally recall everything they remembered about what they just


read, upon completing the ORF treatment or generalization probes. A
word was counted retold correctly when it met the following criteria:
(a) it was in or directly related to the passage read, (b) it was part of a
comprehensive statement, and (c) was not repeated. Words that were
repeated outside of the context of a new statement were not counted
as correct. Non-words such as “um” were not counted.
Treatment Probes
The experimenter administered treatment probes once the par-
ticipants reached their goal and they completed the reading compre-
hension test. Probes consisted of one-minute readings of the practiced
passage and were taken from the “pass timing” activity of the RNSE
sequence. Validity and reliability data for these passages are as fol-
lows: reliability coefficients for first grade passages were 0.874 and
validity coefficients were 0.779. The experimenter instructed partici-
pants to click start and begin reading the treatment story displayed
on the computer screen. If the participant did not correctly pronounce
a word or if three seconds elapsed without the participant reading
the word, the experimenter provided the correct pronunciation to the
participant and instructed him/her to go to the next word. The missed
word was scored as incorrect. At the end of one minute the experi-
menter instructed the participant to click on the last word that was
read. The program automatically calculated the number of correct
words the participant read.
Following the pass timing treatment probe, the experimenter ad-
ministered the word retell probe. The participants were instructed to
tell the experimenter everything they remembered about the story. A
timer was set for one minute and the experimenter said, “Go.” Every
word directly related to the treatment story was counted as correct.
Para-words or utterances such as “um” and repeats were not counted
as a retell.
Generalization Probes
Following the completion of three treatment probes, the experi-
menter administered a generalization probe. These probes consisted
of a one-minute timed reading selected from the AimsWeb ORF prog-
ress monitoring passages and the word retell probe of that passage.
Median reliability coefficients for first grade passages were 0.91 and
predictive validity was between 0.47 and 0.57. These probes were
identical to treatment probes with the exception of the use of Aim-
sWeb progress monitoring passages. That is, the same procedures
used to administer the treatment probes were also used to administer
32 GIBSON, et al.

the generalization probes. The only difference was instead of practice


passages being presented, a novel first grade Aimsweb passage was
presented. All of the stories from the generalization probes were pre-
sented on the computer screen.
Definition and Measurement of Independent Variables
Each participant engaged with the computerized reading pro-
gram, Read Naturally Software Edition. They were all trained on how
to use this program and received the computerized instruction three
to four times per week. The entire intervention was implemented be-
tween 14-16 weeks. The program consisted of the following instruc-
tional sequence for each selected treatment story: key words, one min-
ute cold reading, read along, practice reading, comprehension test,
and pass timing.
Key words. Prior to reading the treatment story, the program in-
troduced participants to the new key words for that story. Each les-
son began with approximately three to four key words that were in
the treatment story. At the beginning of the activity each key word
was presented to the participant on the computer screen. The com-
puter read each word to the participants. It then prompted the par-
ticipants to read the words aloud but quietly, and then click on the
word to hear the definition. The computer then read the definition,
and sometimes provided the word used in a sentence and an accom-
panying picture. If the participants needed to hear any of the key
words again, they were instructed to click the word, and the com-
puter would read it again. This practice continued until the partici-
pants thought they were able to read all of the key words correctly.
No feedback was provided regarding the accuracy of reading the key
words. Each participant decided if they needed additional practice
with each key word.
One-minute cold reading. This activity consisted of the comput-
er displaying the selected treatment story on the screen. All treat-
ment stories were selected according to the order they appeared on
the screen when entering the program. Participants followed the
sequence across the screen from left to right and at the end of the
screen they went to the next row. The same picture that the partici-
pants clicked on to enter the lesson continued to be displayed in this
screen and the screen for all of the other activities. The program in-
structed the participants to read the treatment story for one minute
and click on any words that they did not know. After clicking start,
the computer highlighted the lines as the participants moved their
cursor over the words, tracking as they read. The participants read
aloud but in a quiet voice and were able to click any words they did
COMPUTERIZED FLUENCY INTERVENTION 33

not know. Because the computer had no way of detecting reading


errors, participants were responsible for indicating if they did or did
not know a specific word. At the end of one minute, the participants
were prompted by the computer to click on the last word that was
read. The computer prompted the students to click on any words
they missed to hear them read again. The computer calculated and
displayed the number of words read and also displayed the “goal”
for the story. The goal for Phase I of the intervention was set at the
benchmark level of end of first grade ORF (i.e., 40 words correct per
minute). The goal for Phase II was based on the individual perfor-
mance of each participant. The experimenter did not intervene or
prompt the participants at any time during the cold reading. Things
such as articulation errors could not be taken into account. Addition-
ally, each story was accompanied by one picture that appeared on
the screen during the cold read, the read along, the practice reading,
and the timed reading checkout. This picture remained consistent
through each story sequence but was not present during the compre-
hension activity.
Read along. Following the completion of the cold reading, the
computer introduced the treatment story. The computer read the sto-
ry to the participants and they were prompted to read quietly along
with the computer. As the computer read the treatment story, the sen-
tences were highlighted in blue. During this phase participants were
required to listen and read along three consecutive times for each
treatment story.
Practice reading. During the practice reading activity, the com-
puter instructed the participants to read the treatment story. While
reading, the participants tried to reach their goal of the set criterion
within a one-minute time period. The computer instructed the par-
ticipants to click a clock icon at the top of the page (directly above
the treatment story) and begin reading. If a student did not know a
word in the story, the student could click on the word to hear it read
and the word was counted as incorrect. At the end of one minute, the
bell rang and the computer instructed the participants to click on the
last word they read. The total number of words read correctly was
automatically calculated and displayed in a text box to the right of
the treatment story. If the participant did not reach his/her goal, he/
she was instructed to practice the treatment story again. Every time
the participant did not reach his/her goal, the computer instructed the
participants to continue practicing. Once the student reached his/her
goal, the student was instructed to click the next button to proceed
to the next activity. Although specific data were not collected on the
number of times it took each participant to reach their goal, anecdotal
34 GIBSON, et al.

observations indicated that participants generally met their goals after


1 or 2 attempts in Phase I and 3 or 4 attempts in Phase II.
Comprehension tests. Once each participant reached his/her goal
the student was introduced to the comprehension activity. This activ-
ity consisted of a series of multiple choice and short answer questions
about the story. The participants were instructed to answer all of the
questions by either clicking the response they thought was correct or
by dictating their answer to the experimenter (this accommodation
was made to account for the age and lack of typing experience of the
participants). If an answer was incorrect it was presented again until
it was correctly answered. Once the participant answered all of the
questions, the program continued to re-present any of the questions
that were answered incorrectly until the participant provided the cor-
rect answer. This comprehension test was included as a part of the
program sequence but the scores were not used in the analysis of the
dependent variables.
Pass timing. The final component of the software sequence was
the pass timing. After the participants completed the comprehension
quiz they were instructed to find a teacher to “pass them.” The ex-
perimenter served as the teacher for the pass timing. This timing was
identical to the cold timing and the scores (i.e., CWPM) were used as
treatment probe data.
Experimental Design and Conditions
A multiple probe across students design was used for this study.
This design consisted of three tiers with two students in the first tier
and one student in both the second and third tier. There were four
students for each school and each school served as its own separate
experiment. The experimenter made treatment decisions for each
school independently regardless of what was occurring with the other
school’s participants. It should be noted that throughout this study all
participants received their daily reading instruction in their classroom
(i.e., Reading Mastery or direct reading instruction).
Baseline. Baseline consisted of one-minute timed readings of a
selected connected text passage. These passages were novel to the par-
ticipants and each baseline probe was a cold reading. The experiment-
er presented the selected story in a Word document on the computer
screen. The selected treatment story was presented with similar font
type and size compared to the RNSE stories. One important difference
that should be noted is that the Word document was not equipped to
provide the highlighted sentence tracking. Participants were instruct-
ed to read as many words as they could for one minute, which mir-
rored the directions provided by the RNSE program. The treatment
COMPUTERIZED FLUENCY INTERVENTION 35

probe passages were selected from the Read Naturally SE 1.0 grade
level and generalization passages were selected from first grade Aim-
sweb progress monitoring passages. The experimenter set a timer for
one minute and started it once the first word was read or 3 seconds
had elapsed. If the participants failed to read the first word or any of
the subsequent words within 3 seconds, the experimenter provided
the correct pronunciation and it was scored as incorrect. At the end of
the one-minute timing, the experimenter counted the total number of
correct words and recorded this number on the data collection sheet.
Training. Following the baseline phase, the experimenter intro-
duced the RNSE intervention and a training sequence that familiar-
ized them with the computer program. This training consisted of the
experimenter guiding the participants through all of the activities for
one of the stories. The experimenter supervised the participants as
they navigated through one story of the program. Prior to and follow-
ing each activity, the experimenter provided explicit instructions and
a review of the activity. The participants were given the opportunity
to ask questions. The participants needed to exhibit a thorough under-
standing of how to use the program before the training was complete.
The experimenter used a training checklist to ensure that the partici-
pants were able to adequately engage with the software. This checklist
consisted of 12 items that corresponded to specific steps in engaging
with the program. For example, it checked if the participants clicked
on the appropriate places in the program or listened to the “modeled
reading” three times before the independent practice phase. All par-
ticipants exhibited 100% correct on the training checklist before train-
ing was considered completed.
Phase I. Once the participants completed baseline and were
trained on how to use the Read Naturally software program, Phase
I of the intervention was implemented. During this phase, the ex-
perimenter placed the participants in front of the computer and in-
structed them to click on the prescribed treatment story. The pre-
determined oral reading fluency goal for the first phase was set at
40 words read correctly by the end of the one-minute timing. This
goal was selected because it is the benchmark for the end of first
grade; meaning, by the end of first grade students are considered
to be at low reading risk, if they independently read 40 words of a
novel connected text passage within one minute. During treatment
sessions, participants engaged in all of the activities in the software
program for the selected treatment story. Once the participants met
their goal on the practice activity and completed the comprehension
quiz, they were presented with the pass timing, which was the treat-
ment probe. If the participants did not complete the entire sequence
36 GIBSON, et al.

by the end of the session, they continued where they left off during
the next session.
Phase II. Phase II was identical to phase I with the exception of an
increase in the oral reading fluency goal for each participant. During
this phase, the experimenter increased the participants’ goals based
on their individual performance in Phase I. The two highest treat-
ment probes in Phase I were averaged and used to set the new goal
for Phase II.
Interobserver Agreement (IOA)
A second observer independently recorded data for 40% of
the baseline, treatment, and generalization probes. In addition, IOA
was conducted for word retell fluency across 33% of baseline, treat-
ment, and generalization probes. The experimenter recorded all of the
probes using an audio/video recorder and provided these recordings
along with a copy of the passage to the second observer. The experi-
menter calculated inter-observer agreement using the exact agree-
ment method. In this method, agreement was scored for each interval
in which observer one and observer two scored identical frequencies
of behavior. Each observer recorded either the number of CWPM or
number of words retold in a minute for treatment and generalization
passages. IOA was calculated using the following formula: Agreement
Frequency/(Agreement Frequency + Disagreement Frequency) x 100
= ___%. The experimenter calculated inter-observer agreement sepa-
rately for each passage and then averaged all of the passages across all
of the participants for each dependent variable. Inter-observer agree-
ment for CWPM on treatment probes was 99% (range 95%-100%) and
was 97% (range 94% – 100%) for generalization probes. The IOA for
word re-tell fluency on treatment probes was 98% (range 95% -100%)
and on generalization probes it was 96% (93%-100%).
Procedural Integrity
An independent observer collected procedural integrity data in
35% of baseline, treatment, and generalization sessions. Several types
of procedural integrity data were collected during this study. The first
type of data collected was used to assess the implementation of the
independent variable. This assessment used a checklist to ensure the
Read Naturally software functioned properly throughout treatment
sessions. The checklist consisted of 11 items to measure things such as,
presentation of the selected story or if the computer presented the en-
tire sequence without crashing. The percentage of steps in which the
program operated correctly was tallied and calculated across all ses-
sions and all participants. The total percentage of correct steps ranged
COMPUTERIZED FLUENCY INTERVENTION 37

from 82% to 100% and averaged 97.6%. There were several occasions
on which the computer program froze and needed to be rebooted. This
accounted the lower percentage of steps being completed correctly.
The second type of procedural integrity involved the collection
of probe data. The experimenter needed to provide specific directions
to the participants during probes and also needed to engage in specific
procedures during treatment sessions. A second observer used treat-
ment recordings to collect data using a checklist. The total percentage
of steps correctly implemented was averaged across all participants.
The procedural integrity data ranged from 71% to 100% with an aver-
age of 89%. Discrepancies occurred mainly when the experimenter
failed to provide specific directions to participants during probe ses-
sions.
The third and final type of procedural integrity data involved
assessing how the participants used the software program. A check-
list was used to collect data on the number of steps each participant
performed correctly. An independent observer completed this check-
list for each participant in at least 33% of all treatment sessions. The
total percentage of steps completed correctly was averaged across
all participants. The percentage ranged from 66% to 100% and aver-
aged 85%. The main reasons for lower percentages on this checklist
involved participants not clicking on unknown words and/or not fol-
lowing along during the read along.
Social Validity
Social validity measures were used to assess the participants’
and teachers’ satisfaction with the treatment procedures. Following
the completion of the study, the experimenter administered question-
naires to the participants. This questionnaire was designed to solicit
information regarding what each participant thought about the pro-
gram and its effects on their reading skills. Teachers also completed
questionnaires after the participants completed the study. Teacher
questionnaires consisted of five multiple choice questions that were
designed to solicit opinions about the effects of the intervention on the
reading fluency of their students. There was an additional space for
the instructor to provide comments.
Results
The results for this study are presented separately for each
school (i.e., for each set of students). Figures 1 and 2 present the re-
sults of the multiple probe data for the participants at School One. Fig-
ure 1 presents the oral reading fluency (ORF) rates for correct words
per minute (CWPM) for both the treatment and generalization probes.
38 GIBSON, et al.

100 CWPM GEN CWPM


80 BL TX phase I TX phase II
60

40

20
Marcus
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

100 Individualized criteria


80

60

40

20
Cleveland
per Minute

0
Minute

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Wordsper

100
CorrectWords

80

60 1st grade benchmark


Correct

40

20
Michael
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

100

80

60

40

20
Tylon
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Probe
ProbeSessions
Sessions

Figure 1: Oral reading fluency on treatment and generalization probes for par-
ticipants attending School One
COMPUTERIZED FLUENCY INTERVENTION 39

60
Retell-TX Retell-Gen
50 BL TX phase I TX phase II
40
30
20
10 Marcus
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

60
50
40
30
20
10 Cleveland
Minute
Re-Tell per Minute

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
WordRe-Tell

60
50
40
Words

30
20
10 Michael
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

60
50
40
30
20
10 Tylon
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

Probe Sessions
Probe Sessions

Figure 2: Word retell fluency on treatment and generalization probes for par-
ticipants attending School One
40 GIBSON, et al.

120 CWPM GEN CWPM


100 BL TX- Phase I TX- Phase II
80
60
40
20
Lawrence
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

120
individualized criteria
100
80
60
40
20 Stewie
Minute

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Minute
perper

120
Words

100
Words

80
1st grade benchmark
60
Correct
Correct

40
20
Sheltrie
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

120
100
80
60
40
20 Amber
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Probe Sessions
Probe Sessions

Figure 3: Oral reading fluency on treatment and generalization probes for par-
ticipants attending School Two
COMPUTERIZED FLUENCY INTERVENTION 41

Retell Gen Retell


70
60 BL TX- Phase I TX- Phase II
50
40
30
20
10 Lawrence
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

70
60
50
40
Minute

30
per Minute

20
10 Stewie
0
Re-Tell per

0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
WordRe-Tell

70
Word

60
50
40
30
20
10 Sheltrie
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

70
60
50
40
30
20
10 Amber
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Probe
Probe Sessions
Sessions

Figure 4: Word retell fluency on treatment and generalization probes for par-
ticipants attending School Two
42 GIBSON, et al.

Figure 2 presents the data for the number of words retold per minute
on both treatment and generalization probes. Figures 3 and 4 present
the results of the multiple probe data for the participants at School
Two. Figure 3 presents the ORF rates for treatment and generalization
probes and Figure 4 presents the word retell fluency (WRF) data for
treatment and generalization probes.
Participants in School One
The overall results of the intervention show the participants in
School One increased ORF and WRF across treatment phases. A func-
tional relationship was observed between the dependent variable (i.e.,
ORF) and the intervention on treatment stories using a multiple probe
design across participants. Table 2 presents intervention and general-
ization data for individual participants’ ORF scores. Table 3 presents
intervention and generalization averages for WRF scores. The follow-
ing participants attended School One: Marcus, Cleveland, Michael,
and Tylon.
During baseline, low stable rates of responding occurred for
all four participants with the average ORF being16 CWPM. Once the
treatment phase was implemented for each participant there was an
immediate increase in level of responding on treatment probes. Dur-
ing treatment Phase I fluency rates increased to an average of 54.5
CWPM across all four participants. Data in Phase I were somewhat
variable but there were no overlapping points when compared to
baseline data. Once Phase II was implemented additional increases in
responding were noted and participants averaged 65.2 CWPM. These
data were more stable with some degree of overlapping data points
with Phase I but not with baseline. Similar results were demonstrated
for the WRF on treatment story probes. During baseline participants
averaged 7.2 words retold per minute. Following the introduction of
the treatment, the number of words retold increased to an average to
29 words per minute in Phase I and 31.2 in Phase II.
Results for generalization probes indicated that the participants
were able to increase their ORF on generalization passages. During
baseline participants averaged 23.1 CWPM with some variability in
responding across sessions. Following the introduction of the RNSE,
participants increased their responding to 24.8 CWPM in Phase I.
There was no change in level of responding and a flat trend for 3 or
4 participants. Once Phase II was implemented a change in level and
increasing trends were noted for 3 of the 4 participants. The average
CWPM for Phase II was 39 CWPM.
The results of the word retell fluency for generalization passages
followed a similar trend. During baseline the participants averaged
COMPUTERIZED FLUENCY INTERVENTION 43

Table 2
Individual ORF scores across baseline, Phase I, Phase II
for intervention and generalization conditions

Intervention Generalization
Name
BL PhI PhII BL PhI PhII
Marcus 12.6 41.7 45.5 19 17 30.6

Cleveland 16.3 51.6 69.9 22 20 40.5

Michael 15.3 63.7 74.3 24 28 47

Tylon 19.8 61.1 70.8 27.3 34.2 38

Lawrence 12.7 47.3 58 11 17.3 26.8

Stewie 10.6 40 70.1 8.0 13.6 24.5

Sheltrie 20.3 78.4 100 24.5 22.3 43.4

Amber 27.7 70.4 80.8 21.6 35.6 49.6

Table 3
Individual WRF scores across baseline, Phase I, Phase II
for intervention and generalization conditions

Intervention Generalization
Name
BL PhI PhII BL PhI PhII
Marcus 6.3 30 29.8 6.0 9.7 18.8

Cleveland 1.6 27.4 36 3.0 8.6 22

Michael 9.0 32 32.6 12.5 21.3 26.3

Tylon 12 26.6 26.4 9.3 17.6 18.6

Lawrence 7.3 30 29.9 11 12 18.8

Stewie 3.3 20.7 30 3.0 9.0 14.5

Sheltrie 18.6 35 36.7 12.5 15.3 25

Amber 16 38.8 43.7 13 19 28

7.7 words retold per minute. In Phase I of the intervention this aver-
age nearly doubled to 14.3 words retold per minute. In Phase II the av-
erage increased further to 21.5 words retold per minute, which nearly
tripled the baseline average.
44 GIBSON, et al.

Participants in School Two


The overall results of the intervention show the participants in
School Two also increased ORF and WRF across phases. Table 2 and
Table 3 present intervention and generalization data for individual
participants’ ORF and WRF scores, respectively. The following partic-
ipants attended School Two: Lawrence, Stewie, Sheltrie, and Amber.
During baseline, the average ORF of the four participants was
17.8 CWPM with stable rates of responding. Once Phase I was imple-
mented there was an immediate increase in responding for all par-
ticipants and they averaged 59 CWPM. There was a slight increasing
tread for all of the participants. During Phase II, additional increases
in responding were noted with participants averaging 77.2 CWPM.
There was some degree of overlap between Phase I and II. Similar
results were demonstrated for the WRF on treatment story probes.
During baseline participants averaged 11.3 words retold per minute.
Following the introduction of the treatment the number of words re-
told increased to an average to 31.1 in Phase I and 35 in Phase II.
Results for generalization probes showed that the participants
were also able to increase their ORF on generalization passages. Dur-
ing baseline, participants averaged 16.3 CWPM with only slight in-
creases once Phase I was implemented. During Phase I participants
averaged 22.2 CWPM but 3 of 4 either exhibited flat or decreasing
trends and there was considerable overlap between baseline and
Phase I. During Phase II the average ORF increased to 36.1 CWPM
with all data showing an increasing trend.
The results of the word retell fluency for generalization passages
followed a similar trend. During baseline the participants averaged
9.9 words retold per minute. During Phase I of the intervention this
average increased to 13.8 words retold per minute, and in Phase II the
average increased further to 21.6 words retold per minute.
Social Validity
Data on social validity were collected using questionnaires com-
pleted by school personnel and the participants of this study. Overall
the results of these questionnaires indicated that the school person-
nel believed the intervention was important for their students and it
helped improve the reading performance of their students. Three of
the six teachers surveyed indicated that the fluency of their student
either improved “a lot” or “somewhat’ with only one teacher indi-
cating that there was only “a little” improvement. Additionally, all
of the teachers responded that they believed their students enjoyed
the program and would allow their students to participate in similar
instructional interventions in the future. It should be noted that none
COMPUTERIZED FLUENCY INTERVENTION 45

of the teachers actually ever observed their students engaging with


the program. Similar results were noted when the students completed
their questionnaires, with all of the participants indicating that they
enjoyed the program, believed they became stronger readers, and
would want to continue to use the program in the future.
Discussion
This study examined the effects of a multicomponent repeated
reading strategy on ORF, generalization, and comprehension. All of
the participants in this study increased their ORF and WRF on both
training and generalization passages. A multiple probe experimental
design with two phases was used to investigate the intervention, with
results showing increases in ORF and WRF scores.
Methods designed to increase generalization of ORF to un-
trained passages have become increasingly important in recent re-
search (Adroin et al., 2008; Ardoin et al., 2007; Lo et al., 2011; Silber
& Martens, 2010). This is because students will benefit far greater if
the fluency evidenced with practice passages transfers to untrained,
typical materials in the classroom and larger environment. Previous
studies using multiple exemplar procedures (Ardoin et al., 2007; 2008;
Silber & Martens, 2010) reported some increases in ORF generaliza-
tion, but collective results were either minimal or not much greater
than typical repeated reading procedures. The current study exam-
ined a different generalization principle to train to levels expected in
the natural environment. Accordingly, the researcher structured the
intervention so that criteria for the practice passages were greater
than the end of first grade benchmark criterion of 40 CWPM (Bur-
suck & Damer, 2011; Good & Kaminski, 2002). For some participants
this increase was slight (e.g., 46 CWPM) but for others it was substan-
tial (e.g., 97 CWPM). The 40 CWPM criteria produced only marginal
increases in generalization. In fact, the average ORF score for all of
the participants on generalization passages during Phase I was 23.5
CWPM. This represented an increase of only 3.8 words per minute
over the baseline average of 19.7 CWPM. Three of the eight partici-
pants actually decreased in ORF on generalization passages during
Phase I of the intervention and only two of the participants reached
or approached the end of the year benchmark of 40 CWPM by the end
of Phase I.
These findings are consistent with other studies showing only
marginal gains in generalization to untrained passages (Adroin et al.,
2007, 2008; Silber & Martens, 2010). During Phase I of this interven-
tion, it was noted that all of the participants were reading the training
passages at or above the selected criterion during treatment probes
46 GIBSON, et al.

but there were only minimal increases in generalization for most par-
ticipants. Furthermore, a close analysis of the students’ performances
showed, for example, that during baseline, five of the eight students
(Marcus, Cleveland, Michael, Tylon, and Sheltrie) did as well or bet-
ter on generalization probes as they did on treatment probes. Four of
these students (Marcus, Cleveland, Michael, and Sheltrie) continued
to perform at the same level in Phase I as in baseline. Tylon differed in
that his generalization data continued the slightly upward trend be-
gan in baseline. These observations could legitimately raise questions
regarding the power of the training, especially because the two par-
ticipants who did achieve criteria in Phase I (Tylon and Amber) were
trending upward across conditions and their performance could be
viewed simply as a function of good classroom instruction and read-
ing maturation.
For Phase II, the researcher theorized that increasing the train-
ing criteria according to individual performances in Phase I would
result in greater generalization gains. Increasing fluency criteria
is consistent with Baer’s (1999) suggestion of increasing the rate of
performance during training sessions to promote generalization. In
this study, the average criterion went from 40 CWPM in Phase I to
70 CWPM in Phase II, across all participants. Phase II criteria were
obtained by averaging the two highest data points from Phase I for
each participant. Following this increase, the average ORF on gener-
alization probes increased from 23.5 CWPM in Phase I to 36.9 CWPM
in Phase II. Particularly important in these data is the analysis of re-
sponses for four participants (Marcus, Cleveland, Michael, and Shel-
trie), who at the onset of Phase II, compared to Phase I, showed de-
cisive, upward goal-based trends on the generalization probes. The
generalization probes for these participants in the previous phase
were either flat or moving downward and the immediate upward
trend in the subsequent phase indicates the beneficial training effects
in Phase II. These mixed results contribute to information on the pro-
cess of promoting generalization of ORF. The fact that the instruction
had been substantially intensified and that only three of the eight stu-
dents met criteria during Phase II points to the difficulty in achieving
40 CWPM for some students. Nevertheless, the critical importance of
early reading warrants the investment.
Even with the mixed results the current study contributes to
the research literature in several ways. First, it investigated the use
of increased criteria to produce generalization effects. Cooper et al.
(2007) recommended training behaviors to “natural contingencies of
reinforcement,” which suggests using criteria that require students
to perform as they would in the typical classroom environment. In
COMPUTERIZED FLUENCY INTERVENTION 47

the current study in order for students to continue progressing in


the training sequence, they needed to reach their set criterion on the
practice passage. Theoretically, the naturally occurring reinforcer was
reaching or surpassing the set criterion and passing to the next step
in the training sequence. During the first phase of the intervention all
of the participants routinely surpassed the set criterion of 40 CWPM;
however, during this phase increases in ORF on generalization pas-
sages occurred for only two participants. Once the criterion for each
participant was increased in phase two, participants had to exert more
effort to reach both their individual goals and the naturally occurring
contingency of reinforcement. We contend that this manipulation was
most likely responsible for the increases in ORF on generalization pas-
sages for three participants. The data show that 5 of 8 participants
reached benchmark criteria by the end of the study.
Relative to ORF increases on both practice and generalization
passages, all of the participants also made corresponding gains in
comprehension measures. The current study used a word retell flu-
ency (WRF) procedure to assess comprehension. The participants in
both schools were not able to retell many words during the baseline
probes; however, once the intervention was implemented there were
substantial increases in the number of retold words on treatment
probes. These increases in word retell reflect the ORF gains found
on both treatment and generalization passages, and are in line with
the literature on the positive correlation between reading fluency and
comprehension (Kim, Petscher, Schatschneider, & Foorman, 2010;
Pearce & Gayle, 2009)
The use of the RNSE in this study was a way to bridge the gap
between the need for effective supplemental instruction and the re-
source realities of the classroom. Although this program was imple-
mented under experimental conditions, once students learned the
program, the demands on the experimenter were minimal. The only
time adult supervision was required was at the beginning of the
program for training and to listen to the pass timings at the end of
each treatment story. The study outcomes underscore the potential
of computer-based programs to serve as delivery mechanisms for ef-
ficient and effective supplemental instruction for students at risk for
academic failure, who would otherwise go lacking.
A major limitation of this study is the fact that generalization data
appear to be trending upward for some of the participants through-
out this study. This is more evident for participants Michael, Tylon,
and Amber but inconsistent patterns for generalization probes exist
for all participants during baseline and treatment Phase I. Given that
treatment probes were a measure of the repeated reading strategy,
48 GIBSON, et al.

generalization probe data are very important in establishing treat-


ment effects. It could be argued that participants’ performance on
generalization probes may have continued to improve regardless
of the increased criteria requirements. This may not hold true for
Michael, Sheltrie, and Cleveland because their final generalization
probe in Phase I was equal to or lower than baseline probes and their
first generalization probe in Phase II was higher than all their probes
in baseline and Phase I. Future research should establish procedures
to ensure stable rates of responding on generalization passages be-
fore any criteria changes are implemented.
Another limitation of this study was the conditions in which the
intervention was implemented. Although the experimenter did not
need to interact with the participants because of the computerized in-
structional format, the intervention was implemented outside of the
typical classroom environment with someone other than the class-
room teacher or regular school personnel. It is unclear if the results
would be the same if the program was implemented under typical
classroom and school conditions. Nevertheless, the findings of inter-
ventions such as this study are even more convincing when treatment
is carried out by regular school staff. The generalization passages,
which were taken from the AIMSWeb progress monitoring probes,
point to another study limitation. Even though these passages were
valid generalization measures, it was not determined if participants’
ORF would generalize to their classroom reading material. Future re-
search should test similar procedures under more naturalistic condi-
tions such as conducting a second generalization probe in the class-
room with typical first grade reading material.
A third limitation for this study was the way IOA was collect-
ed and calculated. Each observer counted the total number of cor-
rect words per minute or words retold per minute with each story
counting as an interval. This method was used because all treatment
sessions were recorded but the second observer recorded her score
independently. There was no way to determine where disagree-
ments would have occurred on specific words. This method was used
because a second observer was not available during live treatment
probes, so sessions were recorded. Considering that IOA was very
high across all measures, it is possible that it was inflated because of
the way it was calculated.
The procedure of increasing the ORF criteria based on student
performance during training sessions has only been studied on a lim-
ited basis. With the mixed results of this study further evaluation of
these procedures is warranted. Future research might seek to replicate
these results by implementing similar procedures; however, the use of
COMPUTERIZED FLUENCY INTERVENTION 49

an increasing criterion design may shed light on the optimal increases


needed to promote generalization in ORF.
The classroom implications of this study are considerable. Teach-
ers and students can greatly benefit from these interventions. The re-
sults of this study provide support for the use of computer-based pro-
grams as supplemental instruction, as well as provide limited support
for the generalization principle of teaching behavior to levels required
by natural contingencies. In this case, researchers chose to exceed ex-
pected levels on practice passages in anticipation that generalization
performances would fall within the desired grade range. Systemati-
cally raising expectations based on student performance is probably a
useful instructional strategy that deserves more attention.

References
Alexander, K. L., Entwisle, D. R., & Horsey, C. S. (1997). From first
grade forward: Early foundations of high school dropout. So-
ciology of Education, 70, 87-107.
Ardoin, S. P., Eckert, T. L., & Cole, C. A. S. (2008). Promoting gen-
eralization of reading: A comparison of two fluency-based
interventions for improving general education Student’s oral
reading rate. Journal of Behavioral Education, 17, 237-252.
Ardoin, S., McCall, M., & Klubnik, C. (2007). Promoting generaliza-
tion of oral reading fluency: Providing drill versus practice
opportunities. Journal of Behavioral Education, 16, 54-69.
Begeny, J. C., Daly, E., & Valleley, R. J. (2006). Improving oral read-
ing fluency through response opportunities: A comparison of
phrase drill error correction with repeated readings. Journal of
Behavioral Education, 15, 229-235.
Bursuck, W. D., & Damer, M. (2011). Reading instruction for students
who are at risk or have disabilities. Boston, state: Pearson/Allyn
& Bacon.
Cartledge, G., Yurick, A., Signh, A., Keyes , S., & Kourea, L. (2011).
Follow-up study of the effects of a supplemental early read-
ing intervention of the reading/disability risk of urban prima-
ry learners. Expectionality, 19(1), 140-159.
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior
analysis (2nd ed.). Upper Saddle River, NJ: Pearson.
Flanagan, K. D., West, J., & Walston, J. (2004). National study shows
variations in early learning environments as they relate to
children’s reading development. ERS Spectrum, 22(4), 15-24.
50 GIBSON, et al.

Fuchs, D., & Fuchs, L. S. (2006) Introduction to response to interven-


tion: What, why, and how valid is it? Reading Research Quar-
terly, 4(1), 93-99.
Fuchs, L. S., Fuchs, D., Hosp, M. K., & Jenkins, J. R. (2001). Oral read-
ing fluency as an indicator of reading competence: A theo-
retical, empirical, and historical analysis. Scientific Studies of
Reading, 5, 239-256.
Good, R. H., & Kaminski, R. A. (Eds.). (2002). Dynamic indicators of
basic early literacy skills (6th ed.). Eugene, OR: Institute for
the Development of Education Achievement. Retrieved July
13, 2012, from http://dibels.uoregon.edu
Kim, Y., Petscher, Y., Schatschneider, C., & Foorman, B. (2010). Does
growth rate in oral reading fluency matter in predicting read-
ing comprehension achievement? Journal of Educational Psy-
chology, 102, 652-667.
Kim, Y., Wagner, R., & Lopez, D. (2012). Developmental relations be-
tween reading fluency and reading comprehension: A longi-
tudinal study from Grade 1 to Grade 2. Journal of Experimental
Child Psychology, 113, 93-111.
Kostewicz, D., E., & Kubina Jr., R. M. (2010). A comparison of two
reading fluency methods: Repeated readings to a fluency cri-
terion and interval sprinting. Reading Improvement, 47 (1), 43-
63.
Lo, Y., Cooke, N. L., & Starling, A. L. (2011). Using a repeated reading
program to improve generalization of oral reading fluency.
Education & Treatment of Children, 34, 115-140.
Lyon, R. G. (2003). What principals need to know about reading. Prin-
cipal, 83(2), 14-18.
Lyon, R. G., & National Center for Learning Disabilities, Inc., New York,
NY. (1998). The NICHD research program in reading development,
reading disorders and reading instruction: A summary of research
findings, keys to successful learning: A national summit on research
in learning disabilities. Retrieved from http://www.eric.ed.gov/
contentdelivery/servlet/ERICServlet?accno=ED430366
Martens, B., Eckert, T., Begeny, J., Lewandowski, L., DiGennaro, F.,
Monarello, S., …, Fiese, B. (2007). Effects of a fluency-building
program on the reading performance of low-achieving sec-
ond and third grade students. Journal of Behavioral Educa-
tion, 16, 38-53.
McIntosh, K., Flannery, K. B., Sugai, G., Braun, D. H., & Cochrane,
K. L. (2008). Relationships between academics and problem
COMPUTERIZED FLUENCY INTERVENTION 51

behavior in the transition from middle school to high school.


Journal of Positive Behavior Interventions, 10, 243-255.
Musti-Rao, S., Hawkins, R. O., & Barkley, E. A. (2009). Effects of re-
peated readings on the oral reading fluency of urban fourth-
grade students: Implications for practice. Preventing School
Failure, 54 (1), 12-23.
National Institute of Child Health and Human Development. (2000).
Report of the National Reading Panel. Teaching children to read:
An evidence-based assessment of the scientific research literature on
reading and its implications for reading instruction (NIH Publica-
tion No. 00-4769). Washington, DC: U.S. Government Print-
ing Office.
Neddenriep, C. E., Fritz, A. M., & Carrier, M. E. (2011). Assessing
for generalized improvements in reading comprehension
by intervening to improve reading fluency. Psychology in the
Schools, 48, 14-27.
Pearce, L. R., & Gayle, R. (2009). Oral reading fluency as a predictor
of reading comprehension with American Indian and White
elementary students. School Psychology Review, 38, 419-427.
Read Naturally, Inc. (2009). Read Naturally Software Edition [RNSE].
St. Paul, MN. Retrieved from www.ReadNaturally.com
Reis, S. M., McCoach, D. B., Coyne, M., Schreiber, F. J., Eckert, R. D.,
& Gubbins, E. J. (2007). Using planned enrichment strategies
with direct instruction to improve reading fluency, compre-
hension, and attitude toward reading: An evidence-based
study. Elementary School Journal, 108(1), 3-23.
Silber, J., & Martens, B. (2010). Programming for the generalization of
oral reading fluency: Repeated readings of entire text versus
multiple exemplars. Journal of Behavioral Education, 19, 30-46.
Doi:10.1007/s10864-010-9099-0
Simner, M. L., & Barnes, M. J. (1991). Relationship between first-grade
marks and the high school dropout problem. Journal of School
Psychology, 29, 331-335.
Stokes, T. F., & Baer, D. M. (1977). An implicit technology of gener-
alization. Journal of Applied Behavior Analysis, 10, 349-367.
Doi:10.1901/jaba.1977.10-349

You might also like