ArticlePDF Available

Enhancing Maintenance and Generalization of Sight Words Taught With Incremental Rehearsal: Applying the Depth of Processing and Generalization Frameworks

Authors:

Abstract and Figures

Incremental rehearsal (IR) is a flashcard technique that has produced strong effects for a variety of outcomes including word recognition. We utilized theory-based modifications to IR to enhance maintenance and generalization of sight words. We utilized a within-subjects design in which 41 participants in 2nd and 3rd grade were taught seven unknown words in each of three IR variants—IR, IR with vocabulary (IR-V, which leveraged the depth of processing framework), and IR with context (IR-C, which leveraged Stokes and Baer’s, 1977, generalization framework). Auditory working memory and decoding skills were measured as potential moderators. Maintenance and generalization were high across conditions, but maintenance was greater in IR-V and IR-C than IR. IR was the most efficient intervention variant. A potential moderating effect of decoding skills, but not of auditory working memory, was noted. Researchers and practitioners should consider the extent to which theory-based modifications increase effectiveness of an intervention, while also considering impacts on efficiency.
Content may be subject to copyright.
School Psychology Quarterly
Enhancing Maintenance and Generalization of Sight
Words Taught With Incremental Rehearsal: Applying the
Depth of Processing and Generalization Frameworks
Shawna Petersen-Brown and Matthew K. Burns
Online First Publication, November 29, 2018. http://dx.doi.org/10.1037/spq0000294
CITATION
Petersen-Brown, S., & Burns, M. K. (2018, November 29). Enhancing Maintenance and
Generalization of Sight Words Taught With Incremental Rehearsal: Applying the Depth of
Processing and Generalization Frameworks. School Psychology Quarterly. Advance online
publication. http://dx.doi.org/10.1037/spq0000294
Enhancing Maintenance and Generalization of Sight Words Taught With
Incremental Rehearsal: Applying the Depth of Processing and
Generalization Frameworks
Shawna Petersen-Brown
Minnesota State University, Mankato
Matthew K. Burns
University of Missouri
Incremental rehearsal (IR) is a flashcard technique that has produced strong effects for a variety of
outcomes including word recognition. We utilized theory-based modifications to IR to enhance main-
tenance and generalization of sight words. We utilized a within-subjects design in which 41 participants
in 2nd and 3rd grade were taught seven unknown words in each of three IR variants—IR, IR with
vocabulary (IR-V, which leveraged the depth of processing framework), and IR with context (IR-C,
which leveraged Stokes and Baer’s, 1977, generalization framework). Auditory working memory and
decoding skills were measured as potential moderators. Maintenance and generalization were high across
conditions, but maintenance was greater in IR-V and IR-C than IR. IR was the most efficient intervention
variant. A potential moderating effect of decoding skills, but not of auditory working memory, was noted.
Researchers and practitioners should consider the extent to which theory-based modifications increase
effectiveness of an intervention, while also considering impacts on efficiency.
Impact and Implications
Maintenance and generalization were high across all conditions—incremental rehearsal (IR), IR with
vocabulary (IR-V), and IR with context (IR-C). Although there was an inconsistent and small benefit
to theory-based modifications in IR-V and IR-C, efficiency was much lower with these components
added. Practitioners should carefully monitor the impact of theory-based modifications on student
outcomes, and future research should investigate the use of theory-based modifications with specific
samples (e.g., nonresponders) and when implemented in ways where efficiency is maintained.
Keywords: incremental rehearsal, flashcard technique, maintenance, generalization, auditory working
memory
Flashcard interventions are often used to enhance reading out-
comes including word recognition, reading fluency, and compre-
hension (Burns & Boice, 2009; Burns, Dean, & Foley, 2004; Nist
& Joseph, 2008). Incremental rehearsal (IR; Tucker, 1989) is a
flashcard intervention that introduces unknown items one at a time
along with a high proportion of known items (approximately 80%
to 90% of items are known) to maintain a high rate of success. IR
also uses many opportunities to practice the unknown stimulus and
gradually increases the number of trials between exposures to the
unknown stimulus. Previous research has shown that IR led to high
retention with a variety of information types and student popula-
tions and with retention intervals ranging from 1 day to 30 days
after teaching (Burns & Boice, 2009; MacQuarrie, Tucker, Burns,
& Hartman, 2002; Nist & Joseph, 2008). IR has been used to teach
words (Burns, 2007; Nist & Joseph, 2008), letter sounds (Volpe,
Burns, DuBois, & Zaslofsky, 2011), and math facts (Codding,
Archer, & Connell, 2010). A recent meta-analysis showed a large
effect size for IR across information type, student population, and
age group (Burns, Zaslofsky, Kanive, & Parker, 2012).
Researchers have documented the effectiveness of IR but have
only begun to examine theoretical underpinnings, which can in-
form underlying mechanisms and moderators of intervention ef-
fects (e.g., Burns, 2011; Hughes, 2015), and efficiency, a critical
practical consideration. Within IR, there have been some attempts
to understand causal mechanisms, which may have implications
for improving efficiency and effectiveness. Szadokierski and
Burns (2008) found that the number of words retained was strongly
affected by the number of times students practiced the word being
learned, but there was only a small effect for percentage of known
items while rehearsing. Therefore, the high opportunities to respond
(OTR) was identified as a potential causal mechanism for the effects
of IR.
Shawna Petersen-Brown, Department of Psychology, Minnesota State
University, Mankato; Matthew K. Burns, Educational, School, & Coun-
seling Psychology, University of Missouri.
Correspondence concerning this article should be addressed to Shawna
Petersen-Brown, Department of Psychology, Minnesota State University,
103 Armstrong Hall, Mankato, MN 56001. E-mail: shawna.petersen-
brown@mnsu.edu
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
School Psychology Quarterly
© 2018 American Psychological Association 2018, Vol. 1, No. 999, 000
1045-3830/18/$12.00 http://dx.doi.org/10.1037/spq0000294
1
IR has questionable or inconsistent efficiency (Skinner, 2008)
when efficiency is defined as the number of items learned per
instructional minute (Skinner, Belfiore, & Watson, 2002). There-
fore, efforts to further investigate or enhance IR’s efficiencies are
important in future IR research. One additional weakness in the
current IR research is an inconsistent emphasis on maintenance
and generalization (Burns et al., 2012). Below we will discuss
potential moderators for IR, enhancing the efficiency of IR, and
using theory to modify IR to enhance maintenance and general-
ization.
Potential Moderators for Word Recognition With IR
A moderator is a variable that affects the strength and direction
of the relationship between the independent and dependent vari-
ables (Baron & Kenny, 1986). For instance, researchers have
investigated verbal ability as a potential moderator of IR’s effects,
but it was poorly correlated (r.06 to .16) with retention of words
taught with IR (MacQuarrie et al., 2002), even among students
with disabilities (r.02 to .10; Burns & Boice, 2009). Two
moderators were especially pertinent to this study. The modifica-
tions made to IR require one to manage additional verbal infor-
mation; thus, we selected working memory as a potential moder-
ator of the effects of IR modifications. Additionally, decoding
skills may facilitate word recognition as described below, render-
ing interventions like IR with and without modifications unneces-
sary. Below we discuss working memory and decoding skill as
potential moderators for the effects of IR.
Working Memory
Working memory is a system that both maintains newly ac-
quired information and retrieves stored information (Baddeley,
1992). According to Baddeley (1992), the phonological loop is a
component within the working memory system which processes
verbal information. Attention to this component is particularly
important in this investigation of the impact of IR modifications
which require managing additional verbal information. Research-
ers have demonstrated the relationship between verbal working
memory capacity and reading (e.g., Alloway & Alloway, 2010;
Dixon, LeFevre, & Twilley, 1988; Swanson & Howell, 2001) and
language (e.g., Archibald & Gathercole, 2006) outcomes. A con-
nection between verbal working memory capacity and IR effec-
tiveness has been found in previous research because student
learning through IR decreased when a student’s working memory
capacity was exceeded (Haegele & Burns, 2015).
Decoding
In addition to working memory, a student’s current level in more
basic skills is also likely to play a role in intervention outcomes.
Proficiency in decoding, or the ability to use letter-sound corre-
spondence rules to identify words (Gough & Tunmer, 1986), may
facilitate word recognition and successful sight word instruction
(Aaron et al., 1999; Pikulski & Chard, 2005; Samuels, 1988). The
connection between decoding skills and sight word recognition
suggests more skilled decoders may benefit more from IR for sight
words, although a relationship between the effectiveness of IR for
sight words and decoding has not previously been investigated.
Efficiency of IR
Beyond identifying potential moderators of IR, investigating the
efficiency of IR is an important area for future research as this has
previously been identified as an area of concern. Efficiency is
calculated as items learned per instructional minute (Skinner et al.,
2002) and is an important characteristic to consider when selecting
interventions (Cates et al., 2003). Researchers have questioned the
efficiency of IR (Skinner, 2008) because IR requires more time to
complete than other approaches (MacQuarrie et al., 2002). How-
ever, meta-analytic research indicated variability among five stud-
ies that compared efficiency of IR with other drill methods and an
overall effect size that approximated 0 (Burns et al., 2012). Nist
and Joseph (2008) compared IR with traditional flashcard ap-
proaches and determined that IR was the least efficient of the three
because it required the most time. However, students maintained
more through IR and learned more words over the same time
period. Burns and Sterling-Turner (2010) used number of words
maintained in the numerator and found IR to be as efficient as
traditional flashcard approaches. Therefore, the research regarding
the efficiency of IR is inconclusive.
Deeper Processing to Improve Maintenance of Sight
Word Recognition
Research has indicated that IR’s efficiency may be equal to that
of other approaches when maintenance, rather than words learned,
is considered. Maintenance is defined as correctly performing a
response over time without reteaching (Alberto & Troutman,
2003) and is critical if the intervention is to benefit the child over
the long term (Ardoin, 2006). The depth of processing framework
suggests mechanisms for increasing maintenance outcomes, thus it
served as the basis of an intervention modification for enhancing
maintenance in this study. In proposing this framework, Craik and
Lockhart (1972) suggested that memory was procedural and was a
byproduct of cognitive processes (e.g., defining or analyzing)
carried out on information rather than a byproduct of a specialized
encoding process.
According to the depth of processing framework, the type of
procedure carried out on information impacts its durability in
memory, where more meaningful processes lead to greater dura-
bility (Craik & Lockhart, 1972). Craik and Tulving (1975) iden-
tified and investigated three levels of processing that ranged from
most superficial to deepest, (a) structural (relating to the physical
properties of a stimulus), (b) phonemic (relating to the sound of the
stimulus), and (c) semantic or contextual (relating to the definition
or context of a stimulus). Semantic processing led to superior
same-day retention, time or effort spent processing did not account
for results, and other variations (such as intentional learning) did
not affect results (Craik & Tulving, 1975).
Initial research indicated that deeper processing may enhance IR
outcomes. Petersen-Brown and Burns (2011) used semantic pro-
cessing to enhance maintenance outcomes of IR with 61 second-
and third-grade students by having students state each unknown
word and rehearse a brief definition (IR with vocabulary [IR-V]).
The students maintained and generalized 84.1% and 82.1% of the
words, respectively, with IR, and 93.2% and 93.7%, respectively,
within IR-V. Additional research is needed to replicate these
findings and to examine additional variables such as efficiency.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2PETERSEN-BROWN AND BURNS
Contextual Reading Practice to Benefit Generalization
of Sight Word Recognition
Generalization is a critical intervention outcome, although it has
been investigated inconsistently in previous research (Burns et al.,
2012). Generalization is defined as the application of a learned
stimulus in a way that differs from original learning with regard to
setting, stimuli, or behavior exhibited (Alberto & Troutman,
2003). Stokes and Baer (1977) formulated the initial framework of
approaches to generalization programming based on the principles
of applied behavior analysis (e.g., sequential modification, intro-
duce to natural maintaining contingencies, train sufficient exem-
plars, train loosely, and program common stimuli). Attention to
generalization is lacking in the research (Burns et al., 2012),
Stokes and Baer’s (1977) framework remains the most well-
established generalization framework; we utilized this framework
as a basis for IR modification in the current study. Training
sufficient exemplars and programming common stimuli are most
pertinent to this study. Training sufficient exemplars involves
teaching examples or instances of a specific concept or skill until
an individual is able to correctly respond to all examples of the
skill or concept. Programming common stimuli refers to making
the training or learning situation as similar to situations when the
individual will need to apply the specific concept or skill. Within
the current investigation, contextual reading practice was used to
execute these two strategies. Contextual reading practice serves to
train sufficient exemplars by teaching multiple examples of the
skill (i.e., reading a word in a variety of contexts) to facilitate
correct responding to all examples of the skill (i.e., reading a word
in any context). Contextual reading practice also serves to program
common stimuli because it enhances the similarity between the
training or learning situation (i.e., the IR session with contextual
reading practice) and situations in which the individual will apply
the skill (i.e., reading within age-appropriate context) settings. The
strategies suggested in this framework are not consistently applied
in intervention research (Burns, 2004b) and have not been ex-
tended to IR research.
Purpose
The purpose of the current study was to compare the effects of
three IR variants on maintenance and generalization of unknown
words. Adding a vocabulary component to IR (IR-V) was hypoth-
esized to enhance maintenance because adding semantic informa-
tion facilitates processing the taught words at a greater depth.
Adding contextual reading practice to IR (IR-C) was hypothesized
to enhance generalization because it trains multiple exemplars and
maximizes similarities between the training and generalization
contexts.
Second, we investigated whether adding components to IR impacts
its efficiency. The current investigation did not compare efficiency
between IR and other flashcard methods; rather, it measured effi-
ciency and compared between IR variants. Considering efficiency of
an intervention modification is an important consideration when de-
ciding whether it should be used (Cates et al., 2003).
These IR modifications (adding vocabulary or contextual read-
ing practice) require processing additional information such as a
word’s definition and therefore may place greater demands on
working memory. Therefore, we hypothesized a moderating effect
in which the benefit of adding a vocabulary component or contex-
tual reading practice is greater for participants with greater work-
ing memory capacity. Additionally, more proficient readers may
be more likely to spontaneously decode items regardless of main-
tenance. Therefore, we hypothesized a moderating effect in which
the benefit of adding a vocabulary component or contextual read-
ing practice is less for participants with better decoding skills.
Method
We used a within-subjects group design in this study. There
were three repeated measures in this study, corresponding to the
three IR variants. We used a within-subjects design to promote
comparison among IR variants and to reduce the potential that
differences in participants would account for differences in out-
comes among IR variants.
Participants and Setting
Forty-one second and third grade students recruited from a
suburban elementary and middle school in the Upper Midwest
participated in the study, with 40 students being the targeted
sample size as a result of an a priori power analysis and accounting
for possible attrition. Using G
Power 3.1, a sample size of 31 was
needed assuming a large effect size based on previous related
research (Petersen-Brown & Burns, 2011), three predictors, and
two tested predictors in order to achieve power of .80 and a Type
I error probability of .05. Thirty-two participants (78.0%) were in
second grade, and 26 (63.4%) females participated. Thirty partic-
ipants (73.2%) were Caucasian, five (12.2%) were African Amer-
ican, five (12.2%) were Asian, and one (2.4%) was Hispanic. To
ensure sampling across levels of reading achievement, we col-
lected the reading achievement quartile based on the fall admin-
istration of the Measures of Academic Progress (Northwest Eval-
uation Association, 2011) for all participants whose parent(s) or
guardian(s) returned informed consent. Participants included those
for whom informed consent was obtained first while also ensuring
that each quartile was equally represented (10 participants from the
first quartile, 10 from the second, 11 from the third, and 10 from
the fourth).
Interventionists
Six interventionists conducted sessions. Five interventionists
were graduate students in a local educational psychology program
with previous experience implementing IR. These five interven-
tionists included two fourth-year doctoral candidates who imple-
mented 43% of sessions (n53) and 13% of sessions (n16),
two second-year doctoral students who implemented 27% of ses-
sions (n33) and 9% of sessions (n11), and one first-year
doctoral student who implemented 3% (n4) sessions. One
interventionist was an advanced undergraduate psychology student
without previous experience implementing IR and implemented
5% of sessions (n6). Each interventionist was trained in the
study procedures and delivered approximately the same number of
intervention sessions across IR variants to control for teacher
effects. Students did not work with the same interventionist across
intervention sessions.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
THEORY-BASED MODIFICATIONS TO IR
Materials
Locating known and unknown words. We selected 25
words as potential known words and 204 words as potential
unknown words to permit the individualized identification of both
known and unknown words. The words were selected from the
1,000 most common English words and elementary and interme-
diate content words in math, social studies, geography, and science
(Frye & Kress, 2006). The lists contain words that participants are
likely to encounter when reading a book or content area materials
and are therefore considered to be high utility. The 204 potential
unknown words were randomly selected from the 501st word
through the 1,000th word and from the elementary and interme-
diate content area lists. The 25 potential known words were ran-
domly selected from the first 100 words in the common word list
(Frye & Kress, 2006) because they are the most commonly used
words in written material, and thus, most likely to be known.
Word length has previously been connected to decoding diffi-
culty, particularly for struggling readers (Manis, 1985) and has
often been associated with word difficulty in readability indices
(e.g., Coleman & Liau, 1975). Words for the current study were
sorted into the following groups: three- and four-letter words (e.g.,
son, meet), five-letter words (e.g., equal), six-letter words (e.g.,
voyage), seven-letter words (e.g., polygon), eight-letter words
(e.g., boundary), and nine- to 13-letter words (e.g., tradition,
evaporation).
Intervention materials. We created two sets of flashcards
using the 204 potential unknown words. For the IR and IR-V
variants, we printed the unknown words on the front of 3 5
cards. A short definition (10 words or fewer) and a sentence
including the unknown word were written on the back of the card.
Sentences implied the word’s definition (e.g., “When the movie
commences, it is time to be quiet”).
For the IR-C condition, we printed the unknown words on the
front of 4 6 cards along with three sentences containing the
unknown words. Including multiple sentences with the unknown
word is an example of training multiple exemplars from Stokes
and Baer’s (1977) framework. Previous studies have suggested
three exemplars is effective to enhance generalization (Ardoin,
Eckert, & Cole, 2008; Hupp, 1986). Examples showed some
variability in usage. The three sentences featuring the word similar
included, “My mom and aunt look similar,” “The blue and yellow
balls are similar,” and “Is my car similar to yours?”
Intervention Conditions
Each student participated in three variants of IR (IR, IR-V, and
IR-C) with one variant presented in each of three one-on-one
instructional sessions. Instructional sessions were held 1 week
apart, and IR variants were presented in a random order for each
participant. In each session, interventionists taught seven different
unknown words and used eight known words. Previous research
with second-grade students found an average of five words learned
from IR with a standard deviation of 2.07 (Burns, 2004a). There-
fore, seven words were taught during each intervention session to
ensure a number that was reasonable for a majority of the sample.
Sessions were estimated to require from 10 (IR) to 15 (IR-V and
IR-C) minutes each.
IR. The interventionist began with one unknown word in front
of the eight known words. The interventionist presented the un-
known word by saying, “This word is _____. What word is this?”
When the participant said the word correctly, the interventionist
replied, “Good. Can you use it in a sentence?” The participant
provided a sentence that included the unknown word (used seman-
tically or syntactically correctly or incorrectly). Then, the partici-
pant read the unknown word and the first known word. Next, the
participant read the unknown word and the first two known words,
followed by the unknown word and the first three known words.
The procedure continued until the participant read the unknown
word and all eight known words. The last known word was
removed, and the second unknown word was added to the front.
The interventionist presented the newly added unknown word, and
the procedure continued until seven unknown words were taught.
If the participant made an error in reading a word at any time
during the procedure, the interventionist said, “This word is ____.
What word is this?” When the participant said the word correctly,
the interventionist said, “Good,” and continued with the procedure.
This procedure resulted in 9 42 opportunities to read each un-
known word (and 1 opportunity to use it correctly or incorrectly in
context) depending on when in the session the word was practiced,
not including any error correction.
IR-V. Interventionists conducted the IR-V procedure as de-
scribed above under the IR heading with the following modifica-
tions. The interventionist presented each unknown word by saying,
“This word is ____. _____ means ____. What is this word, and
what does it mean?” using definitions on the back of the card.
When the participant correctly stated the unknown word and its
meaning, the interventionist said, “Good. Please use it in a sen-
tence.” If the unknown word was used incorrectly or if the partic-
ipant could not provide a sentence, the interventionist provided the
sentence from the back of the card. Whenever the participant saw
the unknown word while it was being taught, he or she read the
word and provided the definition. If the participant made an error
in reading the unknown word or in providing the definition, the
interventionist corrected the participant by saying, “This word is
____. _____ means ____. What is this word, and what does it
mean?” After a word was taught and was no longer at the front of
the stack, the participant did not supply the definition, and the
correction procedure did not include the definition. This procedure
resulted in 9 42 opportunities to read each unknown word and
nine opportunities to practice its definition (and 1 opportunity to
use it correctly in context) depending on when in the session the
word was practiced, not including any error correction.
IR-C. Interventionists conducted the IR-C procedure as de-
scribed above under the IR heading, with the following modifica-
tions. The interventionist presented the IR-C card and said, “This
word is _____. What is this word?” When the participant said the
word correctly, the interventionist replied, “Good. Now read
_____ in the sentences on the card.” While that unknown word was
at the front of the stack, the participant read the word and one of
the sentences on the card. After a word was taught and was no
longer at the front of the stack, the participant did not read any
sentences. This variation was hypothesized to enhance generaliza-
tion by training multiple exemplars and programming common
stimuli. This procedure resulted in 20 –53 opportunities to read
each unknown word (11 of those times in context), not including
any needed error correction, depending on when in the session the
word was practiced.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4PETERSEN-BROWN AND BURNS
Measures
Dependent measures included maintenance and generalization
of taught words after 1 week. Efficiency in words maintained per
minute was also studied. Measures of working memory and de-
coding skills were administered and investigated as potential mod-
erators in the study.
Maintenance of taught words. Interventionists measured
maintenance of taught words by showing participants the words
they learned 1 week prior on flashcards showing only the taught
words. The word was maintained if the participant read the word
correctly within 2 s. The word was not maintained if the participant
did not say the word within2sorsaid the word incorrectly. The
data for this variable consisted of the number of words maintained
after 1 week.
Generalization of taught words. Interventionists measured
generalization of taught words by showing participants the words
they learned 1 week prior within novel sentences sequentially on a
single page, a method used in previous research to measure gen-
eralization (Nist & Joseph, 2008). If the participant read the target
word correctly within2sofapproaching the word, it was gener-
alized. The word was not generalized if the participant did not say
the word within2sorsaid the word incorrectly. The data for this
variable consisted of the number of words generalized after 1
week.
Efficiency. We calculated efficiency in each condition by
dividing words maintained at 1 week by the number of minutes
needed to complete the IR condition. The data were number of
words maintained per minute.
Verbal working memory. Each student completed the Audi-
tory Working Memory (AWM) subtest of the Woodcock-Johnson
III Tests of Cognitive Abilities (Woodcock, McGrew, & Mather,
2001) as a measure of verbal working memory. In the AWM
subtest, participants hear a sequence of common nouns and num-
bers (e.g., bread 9 tub 1 5) and must say the words first in the order
they heard them followed by the numbers in the order they heard
them (e.g., bread tub 9 1 5). Median internal consistency reliability
of the AWM subtest is .87 (McGrew & Woodcock, 2001). The
data consisted of the standard score (M100, SD 15) based on
grade level.
Decoding skills. Each student completed the Word Attack
(WA) subtest of the Woodcock-Johnson III Tests of Achievement
(Woodcock et al., 2001) as a measure of decoding skill. In the WA
subtest, participants read a series of nonsense words with a variety
of phonetic features. Median internal consistency reliability is .87
(McGrew & Woodcock, 2001). The data were the standard score
(M100, SD 15) based on grade level.
Procedure
Prior to the start of data collection, interventionists were trained
on the procedures for the current study. Next, unknown words
were determined for each participant and randomly distributed
across the three IR variants. The order of IR variants was then
randomly determined for each participant with all possible orders
of IR variants being equally likely. Four sessions were conducted
with each participant. All sessions were 1 week apart and included
implementation of each IR variant and assessment of maintenance
and generalization of taught words, decoding skill, and verbal
working memory as described below.
Training interventionists. Prior to the beginning of the study,
we conducted a 90-min training session with the interventionists in
which we reviewed the procedures for IR, IR-V, and IR-C and
assessment procedures and facilitated practice. Interventionists
were blind to the hypotheses and were told that different variations
on IR were being investigated. Following training, interventionists
implemented each IR variant while we observed and provided
feedback until each interventionist correctly implemented 100% of
steps listed on the fidelity checklists (described below).
Determining condition order and unknown words taught in
each session. The order of IR variant presentation, as well as the
unknown words to be taught in each session, was randomly deter-
mined in the first session. Unknown words were identified on the
day of the first session as described below. A form was utilized
that allowed each participant’s unknown words to be randomly
assigned to variant and IR variants to be randomly ordered.
First session and identifying unknown words. First, inter-
ventionists completed the WA subtest. Next, interventionists se-
lected unknown words. Words were identified as unknown each
time a participant read a word incorrectly or took longer than 2 s
to read the word. Interventionists began with the list of six-letter
words. If the participant read six words in a row correctly, the
interventionist progressed to the next list of longer words. If the
participant read six words in a row incorrectly, the interventionist
progressed to the next list of shorter words. This process continued
until the interventionist identified 21 unknown words, which were
randomly assigned to IR variant using the form referenced above.
Then, the interventionist located the flashcards corresponding to
the IR variant to be implemented on that day. Thus, we identified
individualized lists of 21 unknown words for each participant, and
seven unique words were taught during teach IR variant. Then, the
interventionist selected eight known words and ensured the par-
ticipant was able to quickly identify them. Finally, the interven-
tionist conducted the IR variant randomly selected to occur first.
Second session. During the second session 1 week later, the
interventionist presented the participant with a set of seven sen-
tences; each sentence included one of the seven words taught the
previous week. The participant read the sentences. Taught words
were defined as generalized or not generalized based on proce-
dures described in the Measures section. Next, the interventionist
presented the flashcards used the previous week and asked the
participant to read the word as a measure of maintenance. Finally,
the interventionist conducted the condition that was randomly
selected to occur second.
Third session. During the third session, the interventionist
administered the AWM subtest. Then, the interventionist assessed
generalization and maintenance of words taught the previous week
as described above. Finally, the interventionist conducted the
condition that was randomly selected to occur last.
Fourth session. During the fourth session, the interventionist
assessed generalization and maintenance of words taught the pre-
vious week as described above.
Implementation Fidelity and Interobserver
Agreement (IOA)
Information on implementation fidelity was collected during 11
of 41 sessions (26.8%) for each IR variant based on observer
availability. A 17-item checklist was used that detailed both gen-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
THEORY-BASED MODIFICATIONS TO IR
eral IR procedures (e.g., last known word is removed when a new
unknown word is added) and procedures specific to the current
study (e.g., asking the participant to read the word in a sentence
during IR-C, starting a timer once the interventionist started teach-
ing the first word). Mean implementation fidelity was 99.2%
(range 94.1% to 100%). The only step interventionists missed
during observed sessions was starting a timer when the interven-
tionist began teaching the first word. For one participant in one IR
variant, efficiency data was missing. The only procedural steps
observed to be not appropriately followed were related to the
collection of efficiency data; therefore, it is possible that these data
are not entirely accurate.
IOA was collected on 30.1% of maintenance and generalization
sessions, 24.3% of AWM administrations, and 26.8% of WA
administrations by dividing agreements by agreements plus dis-
agreements. Average IOA was 97.5% (range 71.4% to 100%)
across all maintenance and generalization sessions; IOA for main-
tenance (98.7%) was slightly higher than IOA for generalization
(96.3%). Disagreements were based on whether the participant
said the word within 2 s and slight variations in pronunciation.
Average agreement was 96.6% (range 79.2% to 100%) for
AWM administrations, and 95.0% (range 87.5% to 100%) for
WA administrations.
Data Analyses
Descriptive statistics were calculated for all measures. To in-
vestigate the first and second hypotheses regarding the effect of IR
variants on maintenance, generalization and efficiency, we used
one-way repeated measures analyses of variance (ANOVAs), with
three repeated measures (one for each condition) and one group.
The third and fourth hypotheses were addressed with a series of
linear regressions as recommended by Baron and Kenny (1986)
and Aiken and West (1991). Maintenance and generalization for
IR-V and IR-C served as the dependent variables, and maintenance
or generalization for the IR condition, decoding skill, or auditory
working memory, and an interaction effect served as the predic-
tors. All predictor variables were mean centered as recommended
by Keith (2006). The interaction effects were computed by mul-
tiplying the centered means of the other predictor variables
(Fairchild & McQuillin, 2010). All analyses were conducted with
IBM SPSS Statistics 24.
Results
Descriptive statistics for the primary variables are included in
Table 1. Overall, maintenance and generalization across all con-
ditions were high. Additionally, maintenance and generalization
were negatively skewed and leptokurtic, but only one measure
(IR-C maintenance) had an absolute skew value (3.13) that
exceeded 2.00. Thus, most of the data were likely distributed
normally enough to conduct parametric analyses, but the IR-C data
were transformed with logarithmic 10 (Howell, 2007) before con-
ducting the regressions. The resulting transformed IR-C data were
no longer negatively skewed (skewness 0.89, SE 0.37).
In terms of session length, IR was the shortest on average,
followed by IR-V and IR-C, which were nearly equal. Session
length was positively skewed and leptokurtic across all three
conditions, indicating that session length did not vary consider-
ably, but a few sessions took longer than average. Descriptive
statistics for AWM and WA are shown in Table 2. The average
AWM standard score was approximately 109; the group was at the
high end of the average range overall. The average WA standard
score was approximately 106. AWM and WA standard scores were
nearly normally distributed as indicated by skewness and kurtosis
values near 0.
Maintenance and Generalization
The first hypothesis regarding the effects of adding theory-based
components to IR was investigated using a repeated measures
ANOVA, with condition as the within-subjects factor. Mauchly’s
test of sphericity was not significant,
2
(2) 1.53, p.47. There
was a significant effect among the three IR variants on mainte-
nance, F(2, 76) 3.54, p.05. Partial eta-squared was moderate,
2
.09. Guidelines for interpretation suggest that a small effect
is
2
.01, a medium effect is
2
.06, and a large effect is
2
.14 (Cohen, 1988). Post hoc pairwise comparisons using a Bon-
ferroni correction (p.017) suggested that there were no signif-
icant differences in maintenance among IR variants when utilizing
adjusted pvalues. Cohen’s deffect sizes comparing IR variants
were small but favored IR-V and IR-C, as shown in Table 1.
Mauchly’s test of sphericity for generalization was not signifi-
cant,
2
(2) 0.75, p.69. The effect among the three IR variants
was not significant, F(2, 76) 2.55, p.09. Effect sizes com-
paring generalization between IR variants were small, as shown in
Table 1.
Table 1
Descriptive Statistics for Number of Words Maintained and Generalized, Session Length in
Minutes, and Efficiency in Words Retained per Instructional Minute for the Three Conditions
Dependent
measures
IR IR with vocabulary
IR with contextual
reading practice
NM SDNMSD d NM SD d Statistic
Maintenance 41 5.78 1.49 40 6.25 1.06 .37 40 6.15 1.27 .27 F3.54
Generalization 41 5.29 1.42 40 5.78 1.39 .35 40 5.60 1.52 .21 F2.55
Session length 40 11.08 2.77 41 16.14 3.77 1.55 41 15.91 3.88 1.45 NA
Efficiency 40 .55 .17 41 .41 .11 1.02 41 .41 .13 .94
2
31.16
Note.dis compared with incremental rehearsal (IR).
p.05.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6PETERSEN-BROWN AND BURNS
Efficiency
The second hypothesis addressed efficiency of IR variants.
Efficiency in words maintained per minute of instructional time
was greater in IR (M0.55, SD 0.17) than in IR-V (M0.41,
SD 0.11, d1.02) and IR-C (M0.41, SD 0.13, d0.94).
Mauchly’s test of sphericity was significant,
2
(2) 12.69, p
.002. F(1.54, 57.05) 31.08, p⫽⬍.001, so a nonparametric
analysis was used. The Friedman’s test of ranks was significant,
2
(2) 31.16, p.001, with IR being more efficient than the
other two variants.
Role of Working Memory and Decoding Skills
The third and fourth hypotheses regarding the role of auditory
working memory and decoding skills on maintenance and gener-
alization across IR variants were investigated using regression
analyses. The results presented in Tables 3 and 4 showed that the
number of words maintained in the IR condition significantly
predicted words maintained in the two IR variant conditions and
accounted for 24% and 34% of the variance. The number of words
generalized in the IR condition significantly predicted the number
of words generalized in the IR-C and IR-V conditions and ac-
counted for 17% and 31% of the variance.
The data in Table 3 suggest a potential moderating effect for
decoding skills on generalization. Adding in WA scores accounted
for 3% and 10% more variance for maintenance or words across
the IR-V and IR-C conditions and 9% to 17% more variance for
generalization of the words. WA also significantly predicted main-
tenance and generalization of words for IR-C, but did not predict
either for IR-V. The interaction between maintenance with IR and
WA scores was significant in predicting words maintained and gen-
eralized for the IR-V condition. Thus, decoding skill is a potential
moderating variable for maintenance and generalization of words with
IR-C and IR-V.
Table 2
Descriptive Statistics for Potential Moderators
Potential moderator NMin Max MSDSkew Kurtosis
Auditory working memory 41 91 134 109.49 10.50 .49 .35
Word Attack 41 83 132 105.98 11.19 .25 .13
Note. Data are grade-based standard scores.
Table 3
Regression Analyses for Decoding Skills (Word Attack) as a Moderator of the Effect of Incremental Rehearsal With Vocabulary (IRV)
or Contextual Reading Practice (IRC) on Maintenance (M) and Generalization (G) of Words
Condition and dependent measure
Model 1 Model 2 Model 3
BSEtBSEtBSEt
IRV-M
Constant 6.26 .15 42.57
6.26 .15 42.82
6.44 .15 42.73
IR-M centered .35 .10 .49 3.51
.27 .12 .38 2.28 .01 .15 .01 .05
WA centered .02 .02 .20 1.20 .02 .02 .25 1.62
IR-M centered WA centered .02 .01 .50 2.75
R
2
.24, Fchange 12.291
R
2
.03, Fchange 1.03 R
2
.13, Fchange 7.55
IRV-G
Constant 5.79 .20 28.58
5.80 .19 29.82
6.10 .19 32.30
IR-G centered .40 .15 .41 2.77
.22 .17 .22 1.32 .04 .16 .04 .25
WA centered .04 .02 .35 2.08 .05 .02 .39 2.40
IR-G centered WA centered .04 .01 .52 3.67
R
2
.17, Fchange 7.67
R
2
.09, Fchange 4.33 R
2
.21, Fchange 6.77
IRC-M
Constant .20 .03 6.88
.20 .03 7.34
.18 .03 6.03
IR-M centered .09 .02 .58 4.41
.06 .02 .37 2.51
.02 .03 .16 .84
WA centered .01 .01 .37 2.51
.01 .01 .41 2.77
IR-M centered WA centered .01 .01 .28 1.68
R
2
.34, Fchange 19.44
R
2
.10, Fchange 6.37
R
2
.04, Fchange 2.80
IRC-G
Constant 5.61 .20 27.88
5.61 .18 31.67
5.79 .19 30.31
IR-G .59 .14 .56 4.17
.31 .15 .29 2.08 .18 .16 .17 1.13
WA .07 .02 .49 3.46
.07 .02 .51 3.75
IR-G WA centered .02 .01 .26 2.01
R
2
.31, Fchange 17.36
R
2
.17, Fchange 11.95
R
2
.05, Fchange 4.03
Note. Maintenance and generalization were measured after 1 week. IR incremental rehearsal; WA Word Attack.
p.01.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
THEORY-BASED MODIFICATIONS TO IR
The results of the analyses in Table 4 did not indicate a potential
moderating effect for auditory working memory on maintenance or
generalization of words with either IR variant. As shown in Table
4, adding in AWM scores did not account for significantly more
variance, accounting for 2% to 3% of additional variance for
maintenance of words, and 1% to 5% additional variance for
generalization of words. The interaction between maintenance
for the IR condition and auditory working memory was not sig-
nificant for predicting generalization, but did significantly predict
maintenance of words with IR-V.
Discussion
Both maintenance and generalization were strong in the current
study, with maintenance and generalization averaging 82.6% to
89.3% and 75.6% to 82.6%, respectively, across IR variants. We
found small effect sizes for maintenance in favor of IR-V and IR-C
(d0.37 and 0.27, respectively), but these differences were not
significant when considering Bonferroni-corrected pvalues. The
current study extended previous research, which also found greater
maintenance for IR-V than traditional IR (Petersen-Brown &
Burns, 2011). The prior study found large and significant effect
sizes in favor of IR-V, while effect sizes in the current study were
small and not statistically significant per adjusted pvalues. There
was no significant difference in generalization, which was incon-
sistent with the previous study. IR was found to be more efficient
than IR-V (d1.02) and IR-C (d0.94). Finally, some evidence
of a potentially moderating effect of decoding skills was noted.
Working memory did not moderate maintenance or generalization.
Effect of Additional Components to IR
Both maintenance and generalization were relatively high across
conditions, with median values at the maximum in many cases.
This suggests that IR was an effective intervention for enhancing
maintenance and generalization of taught words, regardless of
modification. Differences in maintenance and generalization
across IR variants were generally nonsignificant, although small
effect sizes favored IR-V and IR-C.
These results provided limited support for the depth of process-
ing framework (Craik & Lockhart, 1972) and Stokes and Baer’s
(1977) generalization framework. Stokes and Baer’s (1977) sug-
gestion to train multiple exemplars was not supported, perhaps due
to possible ceiling effects in which differences were difficult to
detect, given high performance across conditions. Additionally,
results may also be due to subtle differences across variants and
insufficient statistical power given the relatively small sample size
and smaller than anticipated effect sizes.
IR-V was meant to enhance maintenance and IR-C meant to
enhance generalization. However, this differentiation was not ob-
served. Both IR-V and IR-C showed small and generally nonsig-
nificant impacts on maintenance and generalization. It is possible
Table 4
Regression Analyses for Auditory Working Memory (WM) as a Moderator of the Effect of Incremental Rehearsal (IR) With
Vocabulary (IRV) or Contextual Reading Practice (IRC) on Maintenance (M) and Generalization (G) of Words
Condition and dependent measure
Model 1 Model 2 Model 3
BSEtBSEtBSEt
IRV-M
Constant 6.26 .15 42.57
6.26 .15 42.59
6.31 .13 47.60
IR-M centered .35 .10 .49 3.51
.34 .10 .48 3.41
.28 .09 .39 3.04
WM centered .02 .02 .14 1.01 .02 .01 .15 1.12
IR-M WM centered .04 .01 .41 3.17
R
2
.24, Fchange 12.29
R
2
.02, Fchange 1.03 R
2
.16, Fchange 10.08
IRV-G
Constant 5.79 .20 28.58
5.79 .21 28.30
5.91 .20 29.29
IR-G centered .40 .15 .41 2.77
.39 .15 .39 2.57 .33 .15 .34 2.29
WM centered .01 .02 .08 .51 .02 .02 .11 .75
IR-G centered WM centered .04 .02 .32 2.24
R
2
.17, Fchange 7.68
R
2
.01 Fchange .26 R
2
.10, Fchange 5.00
IRC-M
Constant .20 .03 6.88
.20 .03 6.98
.20 .03 6.80
IR-M centered .09 .02 .58 4.41
.08 .02 .56 4.29
.08 .02 .56 4.10
WM centered .01 .01 .18 1.38 .01 .01 .18 1.36
IR-M centered WM centered .01 .01 .04 .30
R
2
.34, Fchange 19.44
R
2
.03, Fchange 1.89 R
2
.01, Fchange .09
IRC-G
Constant 5.61 .20 27.88
5.60 .20 28.43 5.65 .21 27.22
IR-G centered .59 .14 .56 4.17
.54 .14 .51 3.74
.52 .25 .49 3.58
WM centered .03 .02 .22 1.61 .03 .02 .23 1.66
IR-G centered WM centered .01 .02 .10 .75
R
2
.31, Fchange 17.36
R
2
.05, Fchange 2.59 R
2
.01, Fchange .56
Note. Maintenance and generalization were measured after 1 week.
p.01.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8PETERSEN-BROWN AND BURNS
that IR-V resulted in enhanced maintenance through deeper pro-
cessing, and a corresponding increase in generalization resulted.
IR-C may have also oriented participants toward a deeper level of
processing (semantic or contextual are the deepest levels; Craik &
Tulving, 1975) and thus resulted in similar effects.
Previous research on IR’s efficiency was inconclusive (Burns &
Sterling-Turner, 2010; Nist & Joseph, 2008). The current study
showed that IR variants were less efficient than IR, which sug-
gested that IR variants were likely to be less efficient than other
flashcard methods as well. However, it is possible that efficiency
may vary over longer retention intervals. Nist and Joseph (2008)
suggested that IR was the best procedure when students did not
respond to other flashcard methods. Likewise, IR-V and IR-C may
be useful when students have not responded to other flashcard
interventions including IR, but this is a recommendation in need of
future research.
Decoding and Working Memory as Moderators for
Effects of IR
Decoding skills showed a significant relationship to mainte-
nance and generalization of words for both IR variants. WA or the
interaction between WA and the maintenance/generalization for IR
was a significant predictor for all four conditions listed in Table 3.
The significant interaction effects between decoding skills and
maintenance highlights the importance of decoding skills for sight
word learning. Perhaps decoding skills facilitate processing of
target words during teaching, or perhaps decoding skills enable
students to efficiently decode a word whether or not it was main-
tained. Overall, the current finding is also consistent with research
suggesting that decoding skills facilitate word recognition (Pikul-
ski & Chard, 2005).
Verbal working memory was not a significant predictor of
words maintained or generalized for either IR variant. Perhaps the
nonsignificant relationship between maintenance or generalization
and verbal working memory could be attributed to all participants
having average to above average verbal working memory skills on
the selected assessment. Students with poor verbal working mem-
ory may have had greater difficulties processing and later recalling
target words (e.g., Swanson & Howell, 2001). Although verbal
working memory predicted reading achievement in previous re-
search (Alloway & Alloway, 2010; Dixon et al., 1988), these
researchers did not study maintenance or generalization of words.
Moreover, the current research is consistent with a previous study
that found that the number of words learned was significantly
correlated with decoding skills (␤⫽.56, R
2
.32), but not with
working memory (␤⫽.21, R
2
.04; Burns, Davidson, Zaslofsky,
Parker, & Maki, 2018). The magnitudes of the estimates of the
relationships closely matched those found in the current study.
Practical Implications
Theory-based modifications may be a promising avenue for
practitioners hoping to enhance the effectiveness of or modify
interventions to support nonresponders. Theory should be used to
inform practice (Burns, 2011), and the depth of processing and
Stokes and Baer’s (1977) generalization frameworks are theories
that provide a systematic process for modifying interventions to
enhance their effectiveness. However, theoretical frameworks of-
ten have considerable basic research supporting them but lack
applied research support. Therefore, interventions modified ac-
cording to theory should be carefully monitored to ensure their
effectiveness in a specific situation for a specific student. In an
applied setting, this intervention is likely to be implemented with
students struggling with word identification. Students struggling
with word identification, and possibly decoding skills, may show
lower maintenance and generalization results overall. However,
they may show a greater benefit of theory-based modifications,
which may provide a way of accommodating for a lack of basic
reading skills. This suggests an area for future research, as sug-
gested below.
Theory-based modifications may also impact important inter-
vention variables, including efficiency. Efficiency is important to
consider when selecting interventions (Cates et al., 2003), and
modifying interventions may significantly affect efficiency.
Within this investigation, modifying interventions using theoreti-
cal frameworks greatly reduced the efficiency of the intervention.
Practitioners considering theory-based modifications IR should
monitor the impacts of these modifications on both outcomes and
efficiency. Practitioners may wish to employ theory-based modi-
fications, despite their impact on efficiency, when a student strug-
gling with word identification has been nonresponsive to less
intensive intervention efforts. In these cases, it may be deemed
appropriate to allocate more resources and individualized interven-
tion.
Limitations and Future Research
This study builds upon the evidence base for IR and is an
investigation of the influence of theory-based modifications. While
this investigation suggests possible implications for practice, re-
search, and theory, its limitations should be considered. First, the
lack of statistical significance when comparing maintenance and
generalization across IR variants suggests a greater sample size
may be needed to see significant effects. The lack of significance
and small effect sizes regarding maintenance and generalization
were also potentially due to ceiling effects; at the very least, the
lack of variability in outcomes may have obscured any effects of
IR variants. Moreover, the data represented some violations of
distributional assumptions, which were corrected for the regression
analyses or analyzed nonparametrically. However, the first hy-
pothesis was analyzed parametrically with uncorrected data, the
results of which should be interpreted cautiously. Finally, small
effect sizes may also result from subtle differences between IR
variants. Future investigations may pursue various paths to avoid
these issues: increasing the number of unknown words, increasing
maintenance intervals, establishing greater differences between IR
variants, and/or working with struggling students who are less
likely to remember a substantial number of words.
Related to study participants, a convenience sample was used in
which participants were second and third graders from a specific
school in a suburban area who were among the first to return
consent forms. Therefore, this sample may not be representative of
broader populations. Future investigations may focus on students
with lower achievement levels because these are the students who
are most likely to participate in IR.
With regard to study materials and procedures, unknown words
used in the current study may not reflect words students are
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
THEORY-BASED MODIFICATIONS TO IR
expected to recognize. Many words students must recognize can-
not be succinctly defined and therefore may not be presented
through IR-V in particular; this limitation may limit applications of
IR-V. Unknown words were located prior to the first session and
taught across the first three sessions; it is possible that students
may have learned the words previously identified as unknown
prior to teaching. However, randomly selecting words for each
week and randomly sequencing IR variants should control for any
effects this may have had. In addition, the WA and AWM assess-
ments were administered prior to IR sessions which could have
resulted in fatigue during the IR session that day. In addition,
generalization was always measured prior to maintenance, which
may account for why it was consistently lower than maintenance.
The small intervention dosage of one session per IR variant is also
a limitation of this study. This brief exposure is likely insufficient
for making determinations about the effectiveness of an interven-
tion and is certainly insufficient to yield appreciable results rele-
vant in an educational setting.
Future research could investigate the role of the depth of pro-
cessing framework or generalization frameworks with other types
of information previously taught through IR (such as math facts or
letter sounds), with other populations (such as students struggling
with decoding), or in other interventions. In addition, this study
could be implemented so that the potential for ceiling effects is
reduced and effects of modifications maximized (possibly through
longer retention intervals, by presenting more words, by recruiting
struggling readers, or by using a less effective intervention).
Future research may also investigate whether IR variants can be
reconfigured to be more efficient (e.g., by using fewer known
words and therefore fewer OTRs) while still yielding favorable
results. Although outcomes would be expected to decrease, given
that OTRs have been previously identified as a potential causal
mechanism in IR (Szadokierski & Burns, 2008), this effect may be
eliminated or reduced by using deeper processing and high quality
OTR, but that is a hypothesis in need of future research.
Conclusions
IR and the two variants led to high maintenance and general-
ization. There was some evidence of a moderating effect of de-
coding skills, but not for working memory. The current study adds
to the growing literature supporting IR as an intervention that can
support a wide variety of outcomes in the schools. It also marks an
attempt to apply educational theory to a practical application and
contributes to the body of research related to the depth of process-
ing framework, which has primarily been investigated in basic
research. Finally, the current investigation suggests that both cog-
nitive and behavioral theory can be integrated in interventions
utilized in educational settings.
References
Aaron, P. G., Joshi, R. M., Ayotollah, M., Ellsberry, A., Henderson, J., &
Lindsey, K. (1999). Decoding and sight-word naming: Are they two
independent components of word recognition skill? Reading and Writ-
ing, 11, 89 –127. http://dx.doi.org/10.1023/A:1008088618970
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Alberto, P. A., & Troutman, A. C. (2003). Applied behavioral analysis for
teachers (6th ed.). Upper Saddle River, NJ: Merrill Prentice Hall.
Alloway, T. P., & Alloway, R. G. (2010). Investigating the predictive roles
of working memory and IQ in academic attainment. Journal of Exper-
imental Child Psychology, 106, 20 –29. http://dx.doi.org/10.1016/j.jecp
.2009.11.003
Archibald, L. M. D., & Gathercole, S. E. (2006). Short-term and working
memory in specific language impairment. International Journal of Lan-
guage & Communication Disorders, 41, 675– 693. http://dx.doi.org/10
.1080/13682820500442602
Ardoin, S. P. (2006). The response in response to intervention: Evaluating
the utility of assessing maintenance of intervention effects. Psychology
in the Schools, 43, 713–725. http://dx.doi.org/10.1002/pits.20181
Ardoin, S. P., Eckert, T. L., & Cole, C. A. S. (2008). Promoting general-
ization 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. http://dx.doi.org/10.1007/s10864-
008-9066-1
Baddeley, A. (1992). Working memory. Science, 255, 556 –559. http://dx
.doi.org/10.1126/science.1736359
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Personality and Social Psychology,
51, 1173–1182. http://dx.doi.org/10.1037/0022-3514.51.6.1173
Burns, M. K. (2004a). Age as a predictor of acquisition rates as measured
by curriculum-based assessment: Evidence of consistency with cognitive
research. Assessment for Effective Intervention, 29, 31–38. http://dx.doi
.org/10.1177/073724770402900203
Burns, M. K. (2004b). Empirical analysis of drill ratio research: Refining
the instructional level for drill tasks. Remedial and Special Education,
25, 167–173. http://dx.doi.org/10.1177/07419325040250030401
Burns, M. K. (2007). Reading at the instructional level with children
identified as learning disabled: Potential implications for response-to-
intervention. School Psychology Quarterly, 22, 297–313. http://dx.doi
.org/10.1037/1045-3830.22.3.297
Burns, M. K. (2011). School psychology research: Combining ecological
theory and prevention science. School Psychology Review, 40, 132–139.
Burns, M. K., & Boice, C. H. (2009). Comparison of the relationship
between words retained and intelligence for three instructional strategies
among students with below-average IQ. School Psychology Review, 38,
284 –292.
Burns, M. K., Davidson, K., Zaslofsky, A. F., Parker, D. C., & Maki, K. E.
(2018). The relationship between acquisition rate for words and working
memory, short-term memory, and reading skills: Aptitude-by-treatment
or skill-by-treatment interaction? Assessment for Effective Intervention,
43, 182–192. http://dx.doi.org/10.1177/1534508417730822
Burns, M. K., Dean, V. J., & Foley, S. (2004). Preteaching unknown key
words with incremental rehearsal to improve reading fluency and com-
prehension with children identified as reading disabled. Journal of
School Psychology, 42, 303–314. http://dx.doi.org/10.1016/j.jsp.2004.04
.003
Burns, M. K., & Sterling-Turner, H. (2010). Comparison of efficiency
measures for academic interventions based on acquisition and mainte-
nance. Psychology in the Schools, 47, 126 –134.
Burns, M. K., Zaslofsky, A. F., Kanive, R., & Parker, D. C. (2012).
Meta-analysis of incremental rehearsal using phi coefficients to compare
single-case and group designs. Journal of Behavioral Education, 21,
185–202. http://dx.doi.org/10.1007/s10864-012-9160-2
Cates, G. L., Skinner, C. H., Watson, T. S., Meadows, T. J., Weaver, A.,
& Jackson, B. (2003). Instructional effectiveness and instructional effi-
ciency as considerations for data-based decision making: An evaluation
of interspersing procedures. School Psychology Review, 32, 601– 616.
Codding, R. S., Archer, J., & Connell, J. (2010). A systematic replication
and extension of using incremental rehearsal to improve multiplication
skills: An investigation of generalization. Journal of Behavioral Educa-
tion, 19, 93–105. http://dx.doi.org/10.1007/s10864-010-9102-9
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 PETERSEN-BROWN AND BURNS
Cohen, J. (1988). Statistical power analysis for the behavioral sciences
(2nd ed.). New York, NY: Academic Press.
Coleman, M., & Liau, T. L. (1975). A computer readability formula
designed for machine scoring. Journal of Applied Psychology, 60, 283–
284. http://dx.doi.org/10.1037/h0076540
Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A
framework for memory research. Journal of Verbal Learning & Verbal
Behavior, 11, 671– 684. http://dx.doi.org/10.1016/S0022-5371(72)
80001-X
Craik, F. I. M., & Tulving, E. (1975). Depth of processing and the retention
of words in episodic memory. Journal of Experimental Psychology:
General, 104, 268 –294. http://dx.doi.org/10.1037/0096-3445.104.3.268
Dixon, P., LeFevre, J., & Twilley, L. C. (1988). Word knowledge and
working memory as predictors of reading skill. Journal of Educational
Psychology, 80, 465– 472. http://dx.doi.org/10.1037/0022-0663.80.4
.465
Fairchild, A. J., & McQuillin, S. D. (2010). Evaluating mediation and
moderation effects in school psychology: A presentation of methods and
review of current practice. Journal of School Psychology, 48, 53– 84.
http://dx.doi.org/10.1016/j.jsp.2009.09.001
Frye, E. B., & Kress, J. E. (2006). The reading teacher’s book of lists (5th
ed.). San Francisco, CA: Jossey-Bass.
Gough, P. B., & Tunmer, W. E. (1986). Decoding, reading, and reading
disability. RASE: Remedial & Special Education, 7, 6 –10. http://dx.doi
.org/10.1177/074193258600700104
Haegele, K., & Burns, M. K. (2015). Effect of modifying intervention set
size with acquisition rate data among students identified with a learning
disability. Journal of Behavioral Education, 24, 33–50. http://dx.doi.org/
10.1007/s10864-014-9201-0
Howell, D. C. (2007). Statistical methods for psychology (6th ed.). Bel-
mont, CA: Thomson Wadsworth.
Hughes, J. N. (2015). Integrating theory and empirical science in school
psychology: Progress and remaining challenges. School Psychology
Review, 44, 262–270. http://dx.doi.org/10.17105/spr-15-0060.1
Hupp, S. C. (1986). Use of multiple exemplars in object concept training:
How many are sufficient? Analysis & Intervention in Developmental
Disabilities, 6, 305–317. http://dx.doi.org/10.1016/S0270-4684(86)
80011-8
Keith, T. Z. (2006). Multiple regression and beyond. Boston, MA: Pearson.
MacQuarrie, L. L., Tucker, J. A., Burns, M. K., & Hartman, B. (2002).
Comparison of retention rates using traditional drill sandwich, and
incremental rehearsal flash card methods. School Psychology Review,
31, 584 –595.
Manis, F. R. (1985). Acquisition of word identification skills in non-
disabled readers. Journal of Educational Psychology, 77, 78 –90. http://
dx.doi.org/10.1037/0022-0663.77.1.78
McGrew, K. S., & Woodcock, R. W. (2001). Woodcock-Johnson III
technical manual. Itasca, IL: Riverside.
Nist, L., & Joseph, L. M. (2008). Effectiveness and efficiency of flashcard
drill instructional methods on urban first-graders’ word recognition,
acquisition, maintenance, and generalization. School Psychology Re-
view, 37, 294 –308.
Northwest Evaluation Association. (2011). Measures of Academic Prog-
ress. Retrieved from http://www.nwea.org/
Petersen-Brown, S., & Burns, M. K. (2011). Adding a vocabulary compo-
nent to incremental rehearsal to enhance retention and generalization.
School Psychology Quarterly, 26, 245–255. http://dx.doi.org/10.1037/
a0024914
Pikulski, J. J., & Chard, D. J. (2005). Fluency: Bridge between decoding
and reading comprehension. The Reading Teacher, 58, 510 –519. http://
dx.doi.org/10.1598/RT.58.6.2
Samuels, S. J. (1988). Decoding and automaticity: Helping poor readers
become automatic at word recognition. Reading Teacher, 41, 756 –760.
Skinner, C. H. (2008). Theoretical and applied implications of precisely
measuring learning rates. School Psychology Review, 37, 309 –314.
Skinner, C. H., Belfiore, P. J., & Watson, T. S. (2002). Assessing the
relative effects of interventions in students with mild disabilities: As-
sessing instructional time. Journal of Psychoeducational Assessment,
20, 207–220.
Stokes, T. F., & Baer, D. M. (1977). An implicit technology of general-
ization. Journal of Applied Behavior Analysis, 10, 349 –367. http://dx
.doi.org/10.1901/jaba.1977.10-349
Swanson, H. L., & Howell, M. (2001). Working memory, short-term
memory, and speech rate as predictors of children’s reading performance
at different ages. Journal of Educational Psychology, 93, 720 –734.
http://dx.doi.org/10.1037/0022-0663.93.4.720
Szadokierski, I., & Burns, M. K. (2008). Analogue evaluation of the effects
of opportunities to respond and ratios of known items within drill
rehearsal of Esperanto words. Journal of School Psychology, 46, 593–
609. http://dx.doi.org/10.1016/j.jsp.2008.06.004
Tucker, J. A. (1989). Basic flashcard technique when vocabulary is the
goal. Unpublished teaching materials, University of Tennessee at Chat-
tanooga. Chattanooga, TN: Author.
Volpe, R. J., Burns, M. K., DuBois, M., & Zaslofsky, A. F. (2011).
Computer-assisted tutoring: Teaching letter sounds to kindergarten stu-
dents using incremental rehearsal. Psychology in the Schools, 48, 332–
342. http://dx.doi.org/10.1002/pits.20557
Woodcock, R. W., McGrew, K. S., & Mather, N. (2001). Woodcock-
Johnson III. Itasca, IL: Riverside.
Received December 1, 2017
Revision received August 13, 2018
Accepted August 26, 2018
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
THEORY-BASED MODIFICATIONS TO IR
... Over the last 50 years or so, the cognitive psychology literature has identified a broad range of techniques that can support cognitive functioning. Many of these strategies are focused on memory processes such as rehearsal [60], chunking [61] and use of interactive imagery [62]. The development of the current intervention was based on the premise that once difficulties are profiled and understood, they can be paired with the use of these internal strategies to support them. ...
Article
Full-text available
Background Interventions focused on cognitive function in neurodivergent children typically focus on single functions, e.g. working memory training. They are often focused on ‘deficit’ models and lack an emphasis on understanding areas of individual strengths and difficulties as a prerequisite to appropriate support. The multidimensional nature and phenotypic variability of cognitive profiles in these children indicate a need for a multicomponent-tailored intervention programme focused on understanding and supporting an individual child’s cognitive functioning. Aims The ‘EPIC’ intervention (Edinburgh Psychoeducation Intervention for Children and Young People) is focused on improving cognition, learning and behaviour in neurodivergent children such as those with attention deficit hyperactivity disorder (ADHD) or who are autistic. Building on our previous co-production work, this study aimed to use a participatory methods approach to develop EPIC practices and materials in relation to our key principles which include psychoeducation, multicomponent, individualised approach, strengths and difficulties profiling and pairing of a child’s individual strengths and difficulties with internal and external strategies. We also set out to assess the feasibility and acceptability of EPIC, and pilot this novel tool-kit intervention with neurodivergent children and their parents and teachers. Methods The intervention practices, materials and strategies of EPIC were co-produced with neurodivergent children, their parents, teachers and clinicians taking a strengths and difficulties approach. Identification of psychoeducation activities and strategy practices (e.g. mind-maps, chunking), testing of feasibility and collection of pilot data were conducted over a bi-weekly 8-week programme. Eleven neurodivergent children aged 7 to 12 completed the 16-session individualised programme. Acceptability and feasibility were ascertained via qualitative reports elicited within child and teacher interviews and child ratings of enjoyment. Pilot evaluation data was collected pre- and post-intervention participation, and across cognitive assessments (CANTAB, BRIEF), educational attainment (WIAT) and parent and teacher questionnaires measuring clinical symptoms and behaviour (Conners, AQ, SDQ, self-perception). Data was compared with a matched neurodivergent treatment-as-usual control group (N = 9). Results The co-produced EPIC intervention was both feasible to deliver and acceptable to children, parents and their teachers. Pilot data identified that the 8-week intervention improved cognition (short-term and working memory) and literacy (receptive vocabulary, oral word fluency, listening comprehension). Improvements in the intervention group were also found for parent-reported child behavioural difficulties and aggression, and teacher-reported scholastic competence. Effect sizes generated (Cohen’s d) ranged from 0.65 to 2.83. Parents reported continuing to use EPIC strategies when interviewed over a year after participating in the programme. Conclusion The current study met our objectives fully. ‘EPIC’ (Edinburgh Psychoeducation Intervention for Children and Young People) is feasible in home and school contexts and improves a range of aspects of cognition, learning and behaviour in neurodivergent children. Our findings show EPIC is suitable to be assessed within a full-scale trial.
... Researchers have successfully implemented IR with various populations, including English language learner students, students with learning disabilities, and preschool-to middle-school-age students (Erbey et al., 2011;Klingbeil et al., 2017;MacQuarrie et al., 2002). Findings for IR are promising and have led to increases in word recognition, word reading fluency, and oral reading fluency (Klingbeil et al., 2017;Matchett & Burns, 2009;Petersen-Brown & Burns, 2019). ...
Article
Full-text available
Flashcard interventions are commonly used to increase sight-word acquisition and reading ability among students. However, most research on flashcard interventions is implemented in person and not using virtual instruction. The current study compared two flash card interventions, incremental rehearsal (IR) and strategic incremental rehearsal (SIR), on sight-word acquisition when administered via Zoom® video sharing technology. Participants included four struggling readers in first grade. An adapted alternating treatments single-case design exposed participants to the IR, SIR, and a control condition for 10 sessions with the interventionist providing all sessions via Zoom®. Findings indicated all participants learned more words with IR and SIR than in the control condition. However, SIR resulted in more gains in sight-word acquisition compared to IR for all participants. Implications for implementing individualized interventions via virtual modalities are discussed.
... During this inner monitoring phase, a form of inner speech related to verbal working memory might be used [56]. When we hear or read a word the meaning of which we are not certain, we might rehearse possible other words to trace its meaning [57]. The presence of this process of word search is compatible with the activation of the left inferior frontal gyrus, often associated with phonological working memory and other high-level motor control functions, during abstract concept processing [49,50]. ...
Article
Full-text available
We explore the role of inner speech (covert self-directed talk) during the acquisition and use of concepts differing in abstractness. Following Vygotsky, inner speech results from the internalization of linguistically mediated interactions that regulate cognition and behaviour. When we acquire and process abstract concepts, uncertainties about word meaning might lead us to search actively for their meaning. Inner speech might play a role in this searching process and be differentially involved in concept learning compared with use of known concepts. Importantly, inner speech comes in different varieties—e.g. it can be expanded or condensed (with the latter involving syntactic and semantic forms of abbreviation). Do we use inner speech differently with concepts varying in abstractness? Which kinds of inner speech do we preferentially use with different kinds of abstract concepts (e.g. emotions versus numbers)? What other features of inner speech, such as dialogicality, might facilitate our use of concepts varying in abstractness (by allowing us to monitor the limits of our knowledge in simulated social exchanges, through a process we term inner social metacognition )? In tackling these questions, we address the possibility that different varieties of inner speech are flexibly used during the acquisition of concepts and their everyday use. This article is part of the theme issue ‘Concepts in interaction: social engagement and inner experiences’.
... During this inner monitoring phase, a form of inner speech related to verbal working memory might be used [56]. When we hear or read a word the meaning of which we are not certain, we might rehearse possible other words to trace its meaning [57]. The presence of this process of word search is compatible with the activation of the left inferior frontal gyrus, often associated with phonological working memory and other high-level motor control functions, during abstract concept processing [49,50]. ...
Preprint
Full-text available
We explore the role of inner speech (covert self-directed talk) during the acquisition and use of concepts differing in abstractness. Following Vygotsky, inner speech results from the internationalization of linguistically mediated interactions that regulate cognition and behaviour. When we acquire and processabstract concepts, uncertainties about word meaning might lead us to search actively for their meaning.Inner speech might play a role in this searching process and be differentially involved in concept learning compared to use of known concepts. Importantly, inner speech comes in different varieties – e.g., it can be expanded or condensed (with the latter involving syntactic and semantic forms of abbreviation). Do we useinner speech differently with concepts varying in abstractness? Which kinds of inner speech do we preferentially use with different kinds of abstract concepts (e.g., emotions vs. numbers)? What other features of inner speech, such as dialogicality, might facilitate our use of concepts varying in abstractness(byallowing us to monitor the limits of our knowledge in simulated social exchanges, through a process we term inner social metacognition)? In tackling these questions, we address the possibility that different varieties of inner speech are flexibly used during the acquisition of concepts and their everyday use.
Article
The importance of understanding the mechanics of the processes of transfer and perception of knowledge among students of technical universities is still relevant today. This article discusses the possibility of activating various interests among students-programmers on the basis of creating a special artificial environment for the interaction of the teacher with the students. Some limiting properties of the forms of transmission and assimilation of knowledge are studied: minimization of the terminological base, maximization of the definiteness, concreteness and repetition of the term. The research methodology included a three-stage semester survey, as well as a final survey of students. Conclusions were made about the unconditional possibility of transferring knowledge to those students-programmers who either have already formed some skill and have a corresponding cognitive interest in the subject, or when through the creation of special artificial environments, interest can be ignited. This is achieved by reducing the number of meaningful terms, as well as through their well-understood and practical use. A direct relationship between the quality of teaching and the degree of personal interaction between teacher and student was also found.
Article
Full-text available
The Strategic Incremental Rehearsal (SIR) intervention, a modified version of Incremental Rehearsal, is an efficient flashcard procedure that has demonstrated effectiveness on sight word acquisition for children who exhibit reading difficulties. However, to date, the procedure has not been evaluated with children identified with a reading disability who have a history of receiving special education services. This study uses a multiple baseline design across participants to examine the effects of SIR with a modified criterion for removal on sight word reading with three third-grade participants receiving special education services for a specific learning disability in reading. Results indicated sight word reading increased for all 3 participants at the onset of intervention compared to baseline. The total intervention time for each participant ranged from 16 to 48 min. All 3 participants correctly read a minimum of 21 out of 25 targeted words at a 5-week maintenance check. The results indicated that SIR with a modified criterion of removal is a potentially effective and efficient intervention for sight word reading for participants with specific learning disabilities in reading.
Article
Full-text available
The amount of information that students successfully learn and later recall from each intervention session is limited and is called the acquisition rate (AR). Research has consistently supported the effects of modifying intervention set sizes with AR data, but research with AR is in its infancy. The current study compared the relationship between AR while learning words with working memory, short-term memory, and reading skills. Participants were 52 fourth- and fifth-grade students with and without learning disabilities (LDs). Working memory (r = .34), short-term memory (r = .41), and word reading skills (r = .57) all moderately correlated with AR, but word reading skills accounted for 32% of the variance and the other two scores added little unique variance. The corrected correlation coefficients were higher for the word reading with AR than with any other variable and were essentially equal for both groups (r = .73 for average readers and r = .75 for students with an LD in reading). Thus, the data not only support the validity of making decisions with AR data but also suggest that AR is more consistent with a skill-by-treatment interaction framework than an aptitude-by-treatment interaction approach. Potential applications, directions for future research, and limitations are discussed.
Article
Full-text available
The current study replicated MacQuarrie, Tucker, Burns, and Hartman (2002) with a sample of 20 students who had been identified with a disability and had an IQ score that was between 1 and 3 standard deviations below the normative mean. Each student was taught 27 words from the Esperanto International Language with the following conditions: (a) traditional drill in which unknown words were rehearsed until correctly stated three times, (b) three unknown words interspersed among six known words and repeated three times (interspersal), and (c) incremental rehearsal involving the rehearsal of unknown words among nine known words so that each new word was rehearsed nine times. Consistent with the previous study the condition with the most opportunities to respond led to the best retention. The correlation between IQ and the number of words retained 1 to 2 weeks later for the most effective condition (incremental rehearsal) was.03 and.15 after correcting for range restriction. Moderate correlation coefficients between IQ and number of words retained were found for the other two conditions. Implications for research and practice are discussed.
Article
Full-text available
The current article comments on the importance of theoretical implications within school psychological research, and proposes that ecological theory and prevention science could provide the conceptual framework for school psychology research and practice. Articles published in School Psychology Review should at least discuss potential implications for theory, should be written from an ecological and preventative perspective, and should discuss implications for policy when appropriate. Moreover, intervention studies published as general articles in School Psychology Review should address moderating or mediating variables so that researchers can better understand the ecology and causal mechanisms. Manuscripts that address strength-based competence enhancement, multitiered systems of support for a variety of issues, school-based coordination of services, or multicultural competence will be consistent with an ecological approach to prevention.
Article
Full-text available
Assessment data are used to make treatment recommendations for students with mild disabilities. However, no assessment procedures exist that allow one to predict with certainty that one academic intervention will be more effective than another academic intervention for a particular student with mild disabilities. Therefore, hypotheses regarding intervention effectiveness should be tested by assessing students' learning rates under different instructional procedures. In this demonstration, alternating treatment designs are used to show how more precise measurement of instructional time can impact the assessment of relative learning rates when students are exposed to more than one intervention. Discussion centers around the importance of time as a contextual variable when assessing the effects of academic interventions.
Article
The current study replicated MacQuarrie, Tucker, Burns, and Hartman (2002) with a sample of 20 students who had been identified with a disability and had an IQ score that was between 1 and 3 standard deviations below the normative mean. Each student was taught 27 words from the Esperanto International Language with the following conditions: (a) traditional drill in which unknown words were rehearsed until correctly stated three times, (b) three unknown words interspersed among six known words and repeated three times (interspersal), and (c) incremental rehearsal involving the rehearsal of unknown words among nine known words so that each new word was rehearsed nine times. Consistent with the previous study, the condition with the most opportunities to respond led to the best retention. The correlation between IQ and the number of words retained 1 to 2 weeks later for the most effective condition (incremental rehearsal) was .03 and .15 after correcting for range restriction. Moderate correlation coefficients between IQ and number of words retained were found for the other two conditions. Implications for research and practice are discussed.
Book
Análisis aplicado de conducta para maestros, el mítico texto que durante 40 años ha sido el manual de referencia sobre el uso de aplicaciones conductuales en el aula, se edita por primera vez en español. El libro te muestra cómo utilizar los principios del análisis de conducta para crear tu propia receta hacia el éxito educativo. Lleno de prácticas educativas basadas en la evidencia, estrategias concretas y ejemplos divertidos que podrás identificar con tu práctica docente. El texto te ofrece además poderosos métodos para el manejo eficaz y ético de conductas desafiantes en el aula.
Article
This article reviews progress made since Hughes (2000a) criticized intervention research in school psychology for insufficient consideration of theory. A review of contemporary intervention research published in journals specifically identified with the specialty of school psychology supports the conclusion that current school psychology intervention research is much more theory driven, less insular, and more developmental than it was in 2000. The author highlights intervention studies that exemplify these changes and suggests areas for continued improvement in four areas: (a) articulating and testing the presumed processes responsible for treatment effects, (b) testing boundaries of treatment effectiveness, (c) broadening evaluation of treatment effects to include potential negative effects of intervention, and (d) testing maintenance of intervention effects at least 1 year after treatment.