JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2019, 52, 337–354 NUMBER 2 (SPRING)

An analysis of treatment integrity of response interruption and

redirection
CANDICE L. COLÓN
WESTERN NEW ENGLAND UNIVERSITY THE MAY INSTITUTE, MAY CENTER SCHOOL FOR AUTISM AND
DEVELOPMENTAL DISABILITIES

WILLIAM H. AHEARN
THE NEW ENGLAND CENTER FOR CHILDREN WESTERN NEW ENGLAND UNIVERSITY

Response interruption and redirection has been shown to effectively decrease stereotypy, but its
application outside an experimental setting has not been well studied. In Experiment 1, decreases
in automatically maintained vocal stereotypy were obtained following treatment in a controlled
setting for 3 participants diagnosed with autism spectrum disorder. Descriptive data on the con-
sistency and accuracy of response interruption and redirection were then collected in the class-
room setting. Results showed that the consistency of treatment implementation varied across
participants and staff members. Failure to implement the treatment was the most common
error. However, when response interruption and redirection was implemented, the components
were generally carried out as prescribed. In Experiment 2, we conducted a parametric analysis in
a controlled setting to test the impact of consistency errors on response interruption and redirec-
tion. The results indicated that response interruption and redirection was generally effective at
50% treatment implementation or higher. Furthermore, we observed treatment effects when
25% implementation sessions were interspersed with 100% treatment implementation sessions.
Application of response interruption and redirection in light of previous studies and clinical
implications are discussed.
Key words: automatic reinforcement, parametric analysis, response interruption and redirec-
tion, stereotypy, treatment consistency, treatment integrity

Response interruption and redirection behavior and redirecting to appropriate

(RIRD) is a variant of response blocking that
has been used to treat automatically maintained
stereotypic behavior. In general, RIRD is a
multiple component procedure that entails
interrupting each instance of the target
This paper was submitted to the Department of Psy-
chology in the School of Arts and Sciences at Western
New England University by the first author under the
supervision of the second author in partial fulfillment of
the requirements for a doctoral degree in behavior analy-
sis. We extend appreciation to Gregory Hanley, Amanda
Karsten, and Eileen Roscoe for their feedback on earlier
versions of this manuscript. We also thank Nicole Barlow,
Brittany Bolin, Nicole Lewis, Christina Mule, Veronique
Parent, and Ashley Wallace for their help with data collec-
tion and interobserver agreement. The first author is now
at Behavioral Concepts (BCI).
Address correspondence to: Candice Colón-Kwedor,
Behavioral Concepts, 345 Greenwood Street, Worcester,
MA 01607 (email: ccolonkwedor@bciaba.org)
doi: 10.1002/jaba.537
© 2019 Society for the Experimental Analysis of Behavior
337
responses (Ahearn, Clark, MacDonald, &
Chung, 2007). Martinez and Betz (2013)
reviewed eight RIRD studies published in the
Journal of Applied Behavior Analysis from 2007
through 2012 and concluded that RIRD con-
sistently produced a decrease in automatically
maintained stereotypic behavior.
Likewise, a meta-analysis conducted by Van-
derkerken, Heyvaert, Maes, and Onghena
(2013) indicated that single-case studies for
treating automatically maintained stereotypic
behavior (including several RIRD studies pub-
lished from 2007 through 2011) produced clin-
ically and statistically significant effects when
interventions included both skill training
(e.g., teaching appropriate replacement behav-
ior) and decelerative procedures. Two studies
not included in this meta-analysis (due to their
338 CANDICE L. COLÓN and WILLIAM H. AHEARN
publication dates) compared interventions with
and without an RIRD component. For
instance, Colón, Ahearn, Clark, and Masalsky
(2012) directly introduced verbal operant train-
ing to increase appropriate vocalizations and
decrease vocal stereotypy. Although appropriate
vocalizations increased, reductions in vocal ste-
reotypy were not clinically significant for two
of the three participants. Therefore, RIRD was
introduced and resulted in decreased vocal ste-
reotypy while appropriate vocalizations per-
sisted. Similarly, as a means of providing an
alternative behavior and decreasing the motiva-
tion for engagement in vocal stereotypy, Love,
Miguel, Fernand, and LaBrie (2012) arranged
noncontingent access to toys that provided
matched stimulation. Matched stimulation and
RIRD were then compared as a package and
alone. Suppression of vocal stereotypy was
obtained across all treatment conditions; how-
ever, appropriate vocalizations were highest
when RIRD was a component of the treat-
ment. These studies demonstrate that RIRD
optimizes treatment results.
Although RIRD has been shown to be effica-
cious, several researchers have suggested that
RIRD may require effortful responding to
implement (Duffy-Cassella, Sidener, Sidener, &
Progar, 2011; Kliebert, Tiger, & Toussaint,
2011; Miguel, Clark, Tereshko, & Ahearn,
2009). For example, if a student engages in
30 instances of stereotypy per hour, the teacher
must be available to implement RIRD 30 times
per hour. In most schools, staff members are
required to engage in various setting demands
(e.g., attending to other students, data collec-
tion, adherence to scheduled activities) and are
not likely to be trained in experimental
methods (McIntyre, Gresham, DiGennaro, &
Reed, 2007). These demands paired with
resource-intensive procedures (characterized by
long treatment durations or high rates of target
behavior) may increase the likelihood of treat-
ment integrity failures (Hinz, McGee, Hui-
tema, Dickinson, & Van Enk, 2014).
In addition, past RIRD studies have been con-
ducted in controlled settings; therefore, the
extent to which RIRD is effective and practical
in nonexperimental settings has not been thor-
oughly examined (Martinez & Betz, 2013).
Treatment integrity has been defined as the
extent to which an independent variable is
implemented as intended (Gresham, 1989;
Gresham, Gansle, & Noell, 1993). Clinicians
often choose evidence-based procedures that
have been evaluated in controlled settings with
near perfect integrity; however, the generality
of such treatment effects in the natural setting
under conditions of less than optimal treatment
integrity are often unstudied. Relatedly, Fry-
ling, Wallace, and Yassine (2012) emphasized
the need for descriptive treatment integrity data
collected in natural settings, to determine natu-
rally occurring errors (rather than arbitrarily
selected errors) in need of further empirical
evaluation. Hence, descriptive research may
help to ensure ecological validity, or the degree
to which an analysis is representative of errors
that occur in practice and predictive of phe-
nomena that may occur in the clinical or edu-
cational setting (Brooks & Baumeister, 1977;
Gaylord-Ross, 1979). For example, Carroll,
Kodak, and Fisher (2013) included a descrip-
tive assessment of treatment integrity errors
during discrete trial instruction in a classroom
setting. The experimenters demonstrated that
the three most commonly observed errors (fail-
ure to deliver a tangible following correct
responses, failure to deliver a controlling
prompt, failure to present the correct instruc-
tion once) resulted in less than optimal skill
acquisition outcomes for participants with
autism spectrum disorder (see also Donnelly &
Karsten, 2017).
Following a descriptive assessment, empirical
questions regarding treatment outcomes can be
answered by manipulating the consistency of
treatment implementation (McIntyre et al.,
2007; Peterson, Homer, & Wonderlich, 1982).
For instance, intermittent schedules of
 TREATMENT INTEGRITY
treatment delivery may reflect a pattern that is
representative of lapses in treatment integrity.
Parametric analyses investigate such phenom-
ena via the manipulation of a specific parameter
of an independent variable to determine the
extent to which it influences responding
(e.g., Ahrens, Lerman, Kodak, Worsdell, &
Keegan, 2011; Kliebert et al., 2011; St. Peter
Pipkin, Vollmer, & Sloman, 2010; Vollmer,
Roane, Ringdahl, & Marcus, 1999). Therefore,
parametric analyses can be used to determine
the effects of a range of treatment integrity
values on treatment outcome.
Lerman and Iwata (1996) and Smith, Russo,
and Le (1999) used parametric analyses to
investigate the behavioral mechanism responsi-
ble for the effects of response blocking, a proce-
dure similar to RIRD. Although it was not
their purpose, these studies illustrate the possi-
ble effects of degraded treatment integrity on
treatment outcome. A single participant was
exposed to various treatment implementation
values. The results indicated that response
blocking decreased hand mouthing across all
values. Smith et al. conducted a systematic rep-
lication of Lerman and Iwata in which they
attempted to implement continuous response
blocking following baseline. However, 100% of
responses could not be blocked due to the
rapid rate of eye poking emitted by the partici-
pant. Nonetheless, eye poking decreased under
this condition. Intermittent blocking showed
that eye poking decreased during implementa-
tion at 80%, but remained near baseline levels
during 50% and 67% integrity. Taken
together, these studies indicate that treatment
effects can be obtained at degraded levels of
treatment integrity, but they differ in the values
with which treatment effects were maintained.
In a manner similar to the aforementioned
studies, Ahrens et al. (2011) conducted a para-
metric analysis of RIRD using the values 50%,
25%, and 10% with one participant following
an RIRD treatment analysis. The results indi-
cated that RIRD was effective during the 50%
339
integrity condition, but treatment effects were
not maintained at lower integrity values.
The purpose of the current study was to
evaluate the extent to which treatment integrity
impacts RIRD treatment outcome. We evalu-
ated the effectiveness of RIRD on each partici-
pant’s automatically maintained behavior,
trained classroom staff to implement the RIRD
procedure, and conducted treatment integrity
observations in the classroom setting
(Experiment 1). We then conducted a paramet-
ric analysis in a controlled setting to determine
the impact of the errors observed during treat-
ment integrity observations (Experiment 2).
Specifically, we used the results of the treat-
ment integrity observations to inform the type
of error investigated (consistency rather than
accuracy) and the range of consistency errors
tested.
EXPERIMENT 1
Method
Participants, setting, and materials. All partici-
pants attended a private school program that
served approximately 160 day and residential
students. The program specialized in the educa-
tion and treatment of children diagnosed with
autism spectrum disorder. A request for partici-
pants was sent to the program’s clinical and
educational team leaders. Participants were
nominated by their clinical and educational
team due to engagement in vocal stereotypy
that interfered with daily activities
(e.g., interference with attention to task, social
interactions, and vocational jobs in the
community).
The participants selected were all diagnosed
with autism spectrum disorder by a clinician
unassociated with the clinical or research teams.
Individualized Education Plan goal summaries
and teacher reports were used to assess all par-
ticipants’ verbal repertoires. Kent was a
21-year-old male who engaged in prompted
echoics, but whose communication was
340 CANDICE L. COLÓN and WILLIAM H. AHEARN
primarily nonvocal. Therefore, his communica-
tion was supported with a voice output device.
Cora was a 16-year-old female who emitted
spontaneous one-word mands and occasional
tacts. Noah was a 15-year-old male who emit-
ted spontaneous mands in the form of full sen-
tences, participated in three conversational
volleys, and demonstrated prompted tacts. Each
participant was assigned to a different class-
room within the private school setting and was
taught in a 1:2 staff to student ratio.
We trained 10 staff members to implement
RIRD and subsequently observed the degree to
which the staff members implemented RIRD
following training. The staff members were
23 to 29 years old and had a bachelors or mas-
ter’s degree. They were accustomed to frequent
treatment integrity observations.
For Kent, we conducted experimental ses-
sions in a room (1.5 m by 3 m) equipped with
a video camera, a table, and chairs. For Cora
and Noah, we conducted experimental sessions
behind partitions in a classroom equipped with
a video camera, a table, and chairs. We con-
ducted treatment integrity observations in each
participant’s classroom as they participated in
regularly scheduled activities.
Response measurement and interobserver agree-
ment. During the treatment analysis, we
employed a second by second continuous dura-
tion data collection procedure to measure vocal
stereotypy, which is reported as the proportion
of the session with vocal stereotypy (i.e., total
duration with vocal stereotypy divided by total
session duration). Vocal stereotypy was defined
as any instance of noncontextual vocalizations,
including repetitive babbling, grunts, squeals,
and phrases unrelated to the present situation,
emitted at or above conversation level.
We recorded the frequency and duration of
each RIRD implementation to capture the
effort required to implement treatment. RIRD
implementation was defined as the therapist’s
immediate (i.e., within 2 s of the target behav-
ior) presentation of the RIRD procedure, and
duration was measured via a digital timer.
During treatment integrity observations, we
collected data on the frequency of opportunities
to implement RIRD, which was defined as
each episode of vocal stereotypy. Opportunities
to implement RIRD were separated by the
absence of vocal stereotypy for 2 s or more.
Vocal stereotypy that occurred during RIRD
implementation was excluded from data collec-
tion. To summarize treatment consistency
(Peterson et al., 1982; Vollmer, Sloman, &
St. Peter Pipkin, 2008), we divided the number
of occurrences of treatment implementation by
the total number of opportunities to implement
the procedure and multiplied by 100. Further-
more, each time the RIRD procedure was
implemented, accuracy data were collected. To
calculate treatment accuracy, we divided the
number of occurrences of each component by
the total number of opportunities to implement
the component and multiplied by 100. In addi-
tion, we calculated overall accuracy by dividing
the number of components implemented by
the total number of components in the RIRD
procedure and multiplied by 100.
An independent observer collected data
in vivo or from video recordings during a mini-
mum of 30% of sessions for interobserver
agreement purposes. Interobserver agreement
was calculated by dividing the number of 1-s
intervals with agreements by the total number
of intervals with agreements plus disagreements
and multiplying by 100%. Agreements were
defined as both observers indicating that vocal
stereotypy occurred during the same 1-s inter-
val. Disagreements were defined as only one
observer indicating that vocal stereotypy
occurred during a 1-s interval. During the
treatment analysis, the mean interobserver
agreement for vocal stereotypy was 99% for
Kent, 93% for Cora, and 89% for Noah (range
across participants, 78%-100%). The mean
agreement for the frequency of treatment
implementation was 100% for all participants.
The mean agreement for the duration of
 TREATMENT INTEGRITY
treatment implementation was 98% for Kent,
99% for Cora, and 97% for Noah (range across
participants, 94%-100%). The mean agree-
ment for accuracy was 98% for Kent, 97% for
Cora, and 95% for Noah (range across partici-
pants, 88%-100%) during treatment integrity
observations. The mean agreement for consis-
tency was 94% for Kent, 98% for Cora, and
91% for Noah (range across participants,
75%-100%).
Functional analysis. We conducted a func-
tional analysis of vocal stereotypy using a subset
of the conditions described in Iwata, Dorsey,
Slifer, Bauman, and Richman (1982/1994).
We omitted the play condition and alternated
5-min sessions across three conditions (atten-
tion, escape, and no interaction) to determine
whether the maintaining variable(s) of vocal
stereotypy were likely to be socially mediated
(similar to Roscoe, Carreau, MacDonald, &
Pence, 2008). Persistent responding in the no-
interaction condition relative to the attention
and escape conditions indicates presumed
maintenance by automatic reinforcement. Fur-
thermore, as noted in Ahearn et al. (2007),
RIRD would be contraindicated in the case of
attention-maintained behavior as the treatment
necessitates that attention in the form of vocal
redirection be provided contingent on the tar-
get behavior. For each participant, the results
of the functional analysis supported the
hypothesis that vocal stereotypy occurred inde-
pendent of social consequences (functional
analysis data available from first author).
Preference assessment. We conducted a paired-
stimulus assessment (Fisher et al., 1992) with
each participant to identify preferred leisure
items for use during the treatment analysis.
Classroom staff members who interacted regu-
larly with each participant identified the eight
leisure items that were presented. For each par-
ticipant, the most highly preferred item was
selected for use in the study.
Treatment analysis. A withdrawal design was
used to analyze the effects of RIRD on vocal
341
stereotypy. During each session, the participant
and the experimenter sat at a table. Sessions
lasted 5 min. Throughout baseline and treat-
ment sessions, each participant had access to
the preferred leisure item identified via the
preference assessment.
Baseline. There were no programmed conse-
quences for vocal stereotypy during baseline.
The experimenter sat at the table with the par-
ticipant and collected data. The experimenter
did not interact with the participant unless the
participant engaged in a mand. If the partici-
pant manded, the experimenter responded by
stating when the item would be available fol-
lowing the session.
RIRD. During treatment conditions, contin-
gent on each instance of vocal stereotypy, the
experimenter paused the session timer, removed
the preferred item, and provided vocal prompts.
Vocal-verbal behavior appropriate for each par-
ticipant’s verbal repertoire was prompted. For
example, a vocal prompt to engage in either
intraverbals (e.g., “Where do you live?”, “Where
do you go to school?”), to repeat a tact
(e.g., “The block is blue” in reference to the pre-
ferred item), or to engage in one-word echoic
responses (e.g., “go”, “dog”, “book”) were pro-
vided. For Kent, echoics of familiar words were
prompted. For Cora and Noah, tacts or intraver-
bals were prompted. If the participant did not
respond within 2 s of the prompt, the experi-
menter prompted a different vocal response. For
example, if the participant did not respond
within 2 s of the prompt, “Say dog,” the experi-
menter said, “Say book.” The procedure was
implemented until the participant responded to
three consecutive prompts in the absence of
vocal stereotypy. At the end of the procedure,
the experimenter delivered verbal praise for
using appropriate language (e.g., “Good job
talking to me!”), returned the preferred leisure
item, and resumed the session timer to ensure a
5 min sample for comparison to baseline.
Staff training. We provided staff members
with written RIRD guidelines (Flesch–Kincaid
342 CANDICE L. COLÓN and WILLIAM H. AHEARN
readability grade level of 9.5), which included
the definition of the target behavior, data col-
lection procedure, and treatment times
(e.g., work sessions, group activities, leisure
activities) versus nontreatment times
(e.g., toileting, meals, earned breaks). Staff
members could access these guidelines through-
out the day. Following a review of the written
guidelines, staff members were given the oppor-
tunity to ask questions. Next, staff members
viewed a 5-min video that depicted the lead
experimenter implementing a treatment analy-
sis session that included at least three instances
of the RIRD implementation with the relevant
participant. Following video modeling, we pro-
vided the staff members the opportunity to
practice RIRD implementation a minimum of
three times while the lead experimenter served
as the simulated consumer. While one staff
member role-played, the other staff members
observed the role play. During role plays, the
simulated consumer responded in ways that
exposed staff members to a range of scenarios
that may occur when implementing RIRD with
actual consumers. For example, after the staff
member initiated RIRD, the simulated con-
sumer engaged in two prompted responses,
then engaged in vocal stereotypy. This was
done to ensure staff members continued to
prompt responses until the simulated consumer
engaged in three consecutive prompted
responses before providing praise and ending
RIRD. At the end of each role play, we pro-
vided positive and corrective feedback specific
to steps implemented correctly and incorrectly,
respectively. Positive feedback specified the
steps implemented accurately, whereas correc-
tive feedback denoted steps implemented inac-
curately and included a description of how to
implement the step correctly in the future. To
complete the training, each staff member was
required to independently demonstrate 100%
accuracy for at least two consecutive role plays.
Treatment integrity observations. Following
staff training, we conducted observations to
document the accuracy and consistency with
which staff members implemented RIRD in
the classroom setting. During the treatment
integrity observations, staff members had access
to the written guidelines, but did not receive
feedback from the observer or their supervisor.
Each observation was approximately 15 min in
duration. Observations were conducted over a
period of approximately two months for each
participant. During that period, the 10 staff
members were observed an average of 7.7 times
each (range, 4-10) for an average total duration
of 115 min each (range, 60-150 min). Data
were collected at least twice per week across
various days, times of day, activities, and
trained staff members (Vollmer et al., 2008).
Results
Figure 1 shows the percentage of the session
duration with vocal stereotypy during the treat-
ment analysis. For Kent (top panel), Cora (mid-
dle panel), and Noah (bottom panel), an
ascending trend or moderate levels of vocal ste-
reotypy were observed during baseline
(M = 24.7%, 17.8%, and 43.5%, respectively),
followed by an immediate decrease in vocal ste-
reotypy to low levels when RIRD was intro-
duced (M = 0.7%, 2.5%, and 11.7%,
respectively). There was then an increase in
vocal stereotypy upon reintroduction of baseline
conditions for all participants (M = 34.2%,
22.8%, and 48.3%, respectively). We observed
a replication of the treatment effect in the sec-
ond RIRD condition as vocal stereotypy
decreased to low levels for all participants
(M = 2.7%, 1.8%, and 10.5%, respectively).
For Kent, the average frequency and duration
of RIRD implementation per session was 1.3
and 13 s, respectively. For Cora, the average fre-
quency and duration of RIRD implementation
per session was 5.5 and 18.6 s, respectively. In
comparison to Kent and Cora, the average fre-
quency of RIRD implementation per session for
Noah was more frequent (M = 27.2) and the
 TREATMENT INTEGRITY 343

Figure 1. Percentage of session duration with vocal stereotypy during baseline (BL) and the response interruption
and redirection (RIRD) treatment analysis (Experiment 1).

average duration per RIRD implementation
across RIRD conditions was 9 s.
Figure 2 depicts the percentage of treatment
implementation (consistency) and opportunities
to implement treatment (vocal stereotypy epi-
sodes) during classroom observations across staff
members. For Kent (top panel), the
opportunities to implement the procedure varied
(M = 14.3, range, 1-82), and treatment consis-
tency was variable (M = 60.0%, range, 0%-
100%). For Cora (middle panel), opportunities
to implement the procedure were similar to
Kent, but less variable (M = 4.2, range, 1-12).
Overall, treatment consistency was high
344 CANDICE L. COLÓN and WILLIAM H. AHEARN

Figure 2. Percentage of treatment consistency and opportunities to implement the RIRD procedure for all partici-
pants during treatment integrity observations (Experiment 1).

(M = 89.7%, range, 50%-100%). For Noah,
(bottom panel) there were frequent opportunities
to implement the procedure (M = 29.1, range,
2-49). Overall, treatment consistency was low
and variable (M = 41.1%, range, 0%-100%).
Table 1 shows the average accuracy with
which each RIRD component was
implemented across participants during obser-
vations. For Kent and Cora, all components
were implemented with high integrity
(M = 94%, range, 86%-100% and M = 99%,
range, 97%-100%, respectively). Although staff
members implemented component 5 (i.e.,
repeating the procedure until the participant
 TREATMENT INTEGRITY 345

Table 1 1. The participants were three teenagers diag-

Average Implementation Accuracy (%) - Experiment 1 nosed with autism spectrum disorder by an
independent clinician. Cora (from Experiment
Component Description Kent Cora Noah
1) participated in this experiment. Morris was a
1 Remove preferred item 92 97 ---
(if applicable)
14-year-old male who emitted spontaneous
2 Prompt/gain attention 100 100 86 one- and two-word mands and prompted tacts.
of participant Chad was a 15-year-old male who emitted
(if necessary)
3 Provide prompts for 99 100 96 echoics, mands, and prompted tacts consisting
appropriate of a combination of one-word approximations
vocalizations
4 If no response within 90 100 100 and picture exchanges. He did not participate
2 to 3 s provide the in conversational volleys.
next prompt
5 Repeat components 86 98 79 All participants attended the same private
3 and 4 until the school program, were assigned to different
participant complies
with three responses
classrooms, and were taught in a 1:2 staff to
in the absence of student ratio. We conducted sessions with Cora
the target response behind partitions in a classroom equipped with
6 Deliver praise for 97 100 85
appropriate a video camera, a table, and chairs. We con-
vocalizations ducted sessions with Morris and Chad in an
All components 94 99 89
experimental room (1.5 m by 3 m) equipped
with a video camera, a table, and chairs.
complies with three responses in the absence of Response measurement and interobserver agree-
the target response) below 80% integrity for ment. During the treatment analysis, we
Noah, they implemented all other components employed a second-by-second continuous dura-
with high integrity (M = 89%, range, 79%- tion data collection procedure to measure vocal
100%). Component 1 was not applicable for stereotypy, which is reported as the proportion
Noah, as his preferred activity was building of the session with vocal stereotypy (i.e., total
with blocks, and he independently paused duration with vocal stereotypy divided by total
engagement with the activity each time the pro- session duration). Similarly to Experiment
cedure was initiated. 1, we also collected data on the frequency and
To summarize, the results of Experiment duration of each RIRD implementation. In
1 suggest that, on average, staff members addition, we calculated the average latency to
implemented RIRD with low and variable con- the first instance of vocal stereotypy and the
sistency. However, when staff members imple- average intertreatment time (ITT) as a measure
mented RIRD, they did so with high accuracy of treatment potency. Latency was defined as
(except for component 5 for Noah). Therefore, the duration from the onset of the session to
the purpose of Experiment 2 was to conduct a the onset of the first instance of vocal stereo-
parametric analysis of RIRD implementation typy. ITT was defined as the duration from the
consistency wherein we evaluated vocal stereo- offset of each treatment implementation to the
typy across four levels of consistency. onset of subsequent occurrence of vocal
stereotypy.
An independent observer collected data dur-
EXPERIMENT 2
ing a minimum of 30% of sessions for interob-
Method server agreement purposes in vivo or from
Participants and setting. We recruited partici- video. Agreement for stereotypy duration was
pants in the same manner as for Experiment calculated as in Experiment 1. During the
346 CANDICE L. COLÓN and WILLIAM H. AHEARN
treatment analysis, the mean agreement for
vocal stereotypy was 97% for Chad and 98%
for Morris (range across participants, 92%-
99%). The mean agreement for the frequency
of treatment implementation was 100% for
Chad and Morris. The mean agreement for the
duration of treatment implementation was
100% for Chad and 97% (range, 96%-100%)
for Morris. During the parametric analysis, the
mean agreement for vocal stereotypy was 97%
for Cora, 95% for Chad, and 96% for Morris
(range across participants, 86%-100%). The
mean agreement for the frequency of treatment
implementation was 100% for both Cora and
Chad, and 99% (range, 99%-100%) for Mor-
ris. The mean agreement for the duration of
treatment implementation was 99% for Cora,
97% for Chad, and 97% for Morris (range
across participants, 94%-100%). The mean
agreement for the average latency and average
ITT was 98% for Cora, 97% for Chad, and
97% for Morris (range across participants,
95%-100%).
Functional analysis. As in Experiment 1, we
conducted a functional analysis of vocal stereo-
typy using a subset of the conditions described
in Iwata et al. (1982/1994). For each partici-
pant, the results supported the hypothesis that
vocal stereotypy occurred independent of social
consequences (functional analysis data available
from the first author).
Preference assessment. As in Experiment 1, we
conducted a paired-stimulus preference assess-
ment (Fisher et al., 1992) with each participant
to identify preferred leisure items for use during
the analysis, and the most highly preferred
items were selected for use in the study.
Experimental design. We employed a with-
drawal design to analyze the effects of RIRD
on vocal stereotypy and to evaluate the effects
of treatment integrity manipulations via para-
metric analysis conditions. During each session,
the participant and the experimenter sat at a
table together. Sessions lasted 5 min. Each par-
ticipant had access to the preferred leisure item
identified via the preference assessment during
sessions.
Baseline. Baseline was identical to that in
Experiment 1.
RIRD. Contingent on each instance of vocal
stereotypy, the experimenter paused the session
timer, removed the preferred item, and pro-
vided vocal prompts. Vocal-verbal behavior
appropriate for each participant’s verbal reper-
toire was prompted (echoics for Chad, tacts or
intraverbals for Morris). If the participant did
not respond within 2 s of the prompt, the
experimenter prompted a different vocal
response. The procedure was implemented
until the participant responded to three consec-
utive prompts in the absence of vocal stereo-
typy. At the end of the procedure, the
experimenter delivered verbal praise for using
appropriate language and returned the preferred
leisure item. The session timer was then
resumed to ensure a 5 min sample for compari-
son to baseline.
Parametric analysis. During the parametric
analysis, we implemented three treatment
integrity conditions (75%, 50%, and 25%)
sequentially, in alternation with baseline. We
implemented RIRD for three of every four
instances of vocal stereotypy during the 75%
condition, every other instance of vocal stereo-
typy during the 50% condition, and one of
every four instances of vocal stereotypy during
the 25% condition. Following the 25% condi-
tion, we reintroduced the RIRD treatment with
100% integrity to determine if treatment
effects could be recovered. In the last condi-
tion, we conducted a series of two 25% treat-
ment integrity sessions in alteration with one
100% integrity session (i.e., booster session) to
determine whether interspersing booster ses-
sions could also support recovery of previous
treatment effects. To assist with experimenter
adherence to condition-specific integrity levels,
the experimenter used a deck of cue cards that
indicated when to implement or withhold
RIRD (e.g., during the 50% treatment
 TREATMENT INTEGRITY 347

implementation condition, every other card
signaled that RIRD should be implemented).
After each instance of vocal stereotypy, the
experimenter moved the card on top of the
deck to the bottom. Because vocal stereotypy
was reported as a proportion of session time, if
the planned schedule was not met within a ses-
sion, the session was extended until the sched-
ule requirement was met. However, this rarely
occurred due to the frequency of vocal
stereotypy.
Results
Figure 3 shows the percentage of the session
duration with vocal stereotypy during the
treatment analysis. During baseline, Chad (top
panel) demonstrated an increasing trend for
vocal stereotypy (M = 14.4%) and Morris (bot-
tom panel) emitted vocal stereotypy at a high
level (M = 68.2%). Once RIRD was imple-
mented, there was a decrease in vocal stereo-
typy to low levels for Chad and Morris
(M = 3.5% and 6.6%, respectively). Upon the
removal of RIRD, we observed an immediate
increase in vocal stereotypy for Chad and Mor-
ris (M = 35.6% and 65.8%, respectively).
Once RIRD was reintroduced, vocal stereotypy
decreased to near zero levels (M = 1.0%) for
Chad and returned to the approximate level
observed in the initial RIRD condition

Figure 3. Percentage of session duration with vocal stereotypy during baseline (BL) and the response interruption
and redirection (RIRD) treatment analysis for Chad and Morris (Experiment 2).
348 CANDICE L. COLÓN and WILLIAM H. AHEARN

Table 2 average ITT per condition decreased from RIRD

Average Latency and ITT during Treatment Analysis for 1 (M = 25.5 s) to RIRD 2 (M = 13.8 s).
Experiment 2 Figure 4 shows the percentage of session
Participant Condition Average Latency Average ITT
duration with vocal stereotypy during the para-
metric analysis. For all three participants, the
Cora
Baseline 12.0 same pattern of responding was evident. How-
RIRD 1 14.7 48.4 ever, the results for Cora (top panel) and Chad
Baseline 36.7
RIRD 2 49.0 62.0
(middle panel) were more robust than those for
Chad Morris (bottom panel). Vocal stereotypy was
Baseline 33.0 emitted at a lower level during both the 75%
RIRD 1 14.3 29.9
Baseline 10.0 (M = 6.8%, M = 4.1%, and M = 13.8%,
RIRD 2 72.1 35.3 respectively) and the 50% (M = 2.6%,
Morris
Baseline 8.7 M = 5.3%, and M = 15.4 % respectively) con-
RIRD 1 22.8 25.5 ditions in comparison to the first two baseline
Baseline 0.3
RIRD 2 1.5 13.8 conditions (M = 29.2%, M = 42.3%;
M = 11.8%, M = 14.2%; and M = 40.2%,
Note. ITT = intertreatment time; RIRD = response inter- M = 38.4%, respectively). For all three partici-
ruption and redirection
pants, there was also an increase in vocal stereo-
typy when baseline was introduced for the
(M = 9.7%) for Morris. The average frequency third time (M = 22.3%, M = 30.3%, and
of RIRD implementation per session was 6.8 M = 39.2%, respectively). During the 25%
for Chad and 15.0 for Morris. The average condition, near-baseline levels of vocal stereo-
duration per RIRD implementation across con- typy were observed for all participants
ditions was 20 s for Chad and 13 s for Morris. (M = 21.3%, M = 18.3%, and M = 44.5%,
Table 2 shows the average latency to the first respectively). RIRD was then reintroduced at
instance of vocal stereotypy and the average ITT 100% integrity and low levels of vocal stereo-
for each condition of the treatment analysis. For typy were recovered (M = 1.7%, M = 3.4%,
Cora, the average latency increased across condi- and M = 5.8%, respectively). During the 25%
tions as the treatment analysis progressed. plus booster session condition, vocal stereotypy
Although the average latency increased in dura- increased in comparison to the 100% condition
tion overtime, a comparison of baseline and responding resembled the results obtained
(M = 12.0 s, M = 36.7 s) and treatment during the 50% condition (M = 4.2%,
(M = 14.7 s, M = 49.0 s) latency did not show M = 6.1%, and M = 15.5%, respectively).
a great difference. The average ITT per condi- For Cora, the average frequency of RIRD
tion increased from RIRD 1 (M = 48.4 s) to implementation per session was 3.6 and the
RIRD 2 (M = 62.0 s). For Chad and Morris, average duration per RIRD implementation
the difference between baseline and treatment was 9.7 s. There was a decreasing trend in the
latency was also not substantial. For Chad, the average duration per treatment implementation
average latency increased from RIRD as the parametric analysis progressed. For Chad,
1 (M = 14.3 s) to RIRD 2 (M = 72.1 s), and as the average frequency of RIRD implementation
for Cora, the average ITT per condition per session was 4.5 and the average duration
increased from RIRD 1 (M = 29.9 s) to RIRD per RIRD implementation was 8.8 s. Morris’
2 (M = 35.3 s). For Morris, however, the aver- average frequency of RIRD implementation per
age latency decreased from RIRD session was high (10.0) and the average dura-
1 (M = 22.8 s) to RIRD 2 (M = 1.5 s), and the tion per RIRD implementation (20.7 s) was
 TREATMENT INTEGRITY 349

Figure 4. Percentage of session duration with vocal stereotypy during baseline (BL) and the response interruption
and redirection (RIRD) parametric analysis for Experiment 2 participants. The percentages specify the percentage of
vocal stereotypy instances for which RIRD was implemented.

consistently longer in comparison to Cora and Furthermore, despite previous exposure to

Chad’s.
The parametric analysis indicated that RIRD
suppressed vocal stereotypy when implemented
during at least 50% of opportunities. When we
interspersed 100% treatment integrity booster
sessions with 25% treatment integrity sessions,
the level of vocal stereotypy was similar to that
observed during the 50% condition. However,
treatment effects for Morris were modest in
comparison to the results of Cora and Chad.
decreased levels of treatment integrity, optimal
treatment effects were observed once RIRD
with consistent 100% treatment integrity was
reintroduced for all participants.
GENERAL DISCUSSION
The current study evaluated the extent to
which treatment integrity impacts RIRD treat-
ment outcome. First, we determined the types
350 CANDICE L. COLÓN and WILLIAM H. AHEARN
of RIRD treatment integrity errors that occur
in the natural setting and then we analyzed
such errors in a controlled setting. Experiment
1 indicated that following training, staff mem-
bers implemented RIRD inconsistently contin-
gent on instances of stereotypy in the natural
setting. However, when staff members did
implement RIRD, they did so with high accu-
racy. In Experiment 2, we conducted a para-
metric analysis to determine the degree to
which systematic manipulations of RIRD
implementation consistency integrity would
influence RIRD effectiveness. The results of
Experiment 2 indicated that the impact of
treatment inconsistency was minimal if RIRD
was implemented following at least 50% of
instances of stereotypic behavior.
The results of the treatment analyses in
Experiments 1 and 2 replicated previous
research (e.g., Ahearn et al., 2007; Ahrens
et al., 2011) by demonstrating that RIRD was
effective in decreasing automatically maintained
vocal stereotypy. Previous studies (e.g., Ahrens
et al., 2011; Shawler & Miguel, 2015) have
indicated that one of the behavioral mecha-
nisms responsible for RIRD’s effectiveness may
be positive punishment. When using RIRD to
treat automatically maintained vocal stereotypy,
reinforcement remains available contingent on
each instance of vocal stereotypy because
extinction is not feasible. Given this concurrent
schedule arrangement, one might predict that
RIRD would be most effective if implemented
following each instance of the target response.
However, the results of the parametric analysis
conducted in Experiment 2 suggest that a mean
at or above 50% implementation consistency
may constitute a probability strong enough to
compete with continuous automatic reinforce-
ment. Although few studies have examined the
effects of intermittent punishment without the
use of extinction, Azrin, Holz, and Hake
(1963) demonstrated that intermittent punish-
ment without the use of extinction did not pro-
duce response suppression as rapid or
significant as continuous punishment. The
results of the parametric analysis support these
findings as continuous implementation of
RIRD produced greater response suppression in
comparison to intermittent implementation.
Given that RIRD was effective at suppres-
sing vocal stereotypy when implemented at
least 50% of the time, these findings replicate
those of Lerman and Iwata (1996) and Ahrens
et al. (2011), but not those of Smith et al.
(1999). The current study and those conducted
by Lerman and Iwata and by Ahrens et al. first
implemented treatment with 100% integrity
and continued to observe treatment effects
when integrity levels were decreased to as low
as 25%. However, as previously mentioned, the
participant in the Smith et al. study was not
exposed to a recent history of 100% treatment
integrity due to an inability to block every
response the participant emitted during the
continuous blocking condition. Subsequently,
Smith et al. only observed treatment effects
with integrity levels of 80% or 100%. There-
fore, one potential explanation for the differ-
ence in findings between Smith et al. and those
of Lerman and Iwata, Ahrens et al., and the
current study may be the difference in whether
experimenters were able to establish a history of
continuous punishment. Basic studies (Banks,
1966; Shemer & Feldon, 1984) have found
that exposure to intermittent punishment prior
to continuous punishment does not lead to suf-
ficient response suppression. Applied studies
(Rollings & Baumeister, 1981; Romancyzk,
1977) have shown intermittent punishment to
be effective only following recent exposure to
continuous punishment. Further research is
required to investigate whether treatment
effects would be stronger if treatment failure
conditions were directly preceded by a 100%
integrity condition rather than by baseline
(St. Peter Pipkin et al., 2010; Vollmer et al.,
1999). A second potential explanation for the
difference in findings may be related to the
behavioral mechanism responsible for treatment
 TREATMENT INTEGRITY
effects. That is, Ahrens et al. and Lerman and
Iwata attributed treatment effects for their par-
ticipants to positive punishment, whereas
Smith et al. indicated that extinction was likely
responsible for their results.
Although positive punishment is likely one
of the behavioral mechanisms responsible for
RIRD’s effectiveness, other variables may con-
tribute as well. First, access to preferred items,
in addition to RIRD, has been arranged in the
majority of previous RIRD studies
(Martinez & Betz, 2013). Access to preferred
items presents an opportunity for individuals to
allocate responding towards these stimuli that
in turn may suppress stereotypy (e.g., Ahearn,
Clark, DeBar, & Florentino, 2005; Ringdahl,
Vollmer, Marcus, & Roane, 1997). Although
we provided access to a preferred item during
baseline and RIRD, future researchers could
evaluate the effectiveness of RIRD with and
without the availability of preferred items or
conduct a competing items assessment to
decrease the likelihood of using items that may
compete with the target behavior. Second, akin
to several previous studies (e.g., Ahrens et al.,
2011, Duffy-Casella et al., 2011), we removed
the preferred item contingent on instances of
stereotypy prior to implementing RIRD. This
procedural element may function as negative
punishment and has been shown to suppress
stereotypy when implemented in isolation
(Toper-Korkmaz, Lerman, & Tsami, in press).
Future studies are needed to clarify the condi-
tions under which item removal should be used
in conjunction with RIRD. Third, praise for
compliance with demands, which potentially
functions as positive reinforcement for engage-
ment in the appropriate behavior, was included
in our study as well as in the majority of studies
reviewed by Martinez and Betz (2013). Future
researchers should systematically evaluate the
added value of providing access to a preferred
item and removing access to the item following
instances of stereotypy in conjunction with
RIRD; they might also determine the effects of
351
reinforcing compliance during RIRD. Such
studies will likely provide valuable information
that may influence the integrity with which
RIRD is implemented in clinical practice.
The results of the current studies provide
some preliminary support for the assertion that
the amount of time required to implement a
treatment may affect how susceptible it is to
lapses in treatment integrity (Gresham, Mac-
Millan, Beebe-Frankenberger, & Bocian,
2000). Specifically, Noah engaged in higher
rates of vocal stereotypy (M = 29.1), producing
an average of twice as many opportunities to
implement the RIRD procedure when com-
pared to the rates of vocal stereotypy emitted
by Kent and Cora (M = 14.3 and M = 14.0,
respectively). Noah’s treatment analysis data
also indicate that the frequency and the dura-
tion of RIRD were considerably higher for him
in comparison to Kent and Cora. These find-
ings in conjunction with the low rates of treat-
ment consistency observed during Noah’s
classroom observations highlight the concerns
set forth by past researchers regarding the
potential impact effortful treatment application
may have on treatment integrity. These results
suggest that additional research is needed to
generate solutions for such treatment chal-
lenges. One approach that may decrease the
duration of RIRD (thereby potentially decreas-
ing the response effort required by teachers)
without diminishing treatment effects is to
implement RIRD with one demand rather than
three demands. Saini, Gregory, Uran, and Fan-
tetti (2015) found RIRD with one demand to
be effective in reducing stereotypy and to be
less intrusive than RIRD requiring three
demands. Future researchers should determine
whether this procedural variation can bolster
treatment integrity in nonexperimental settings.
Implementing RIRD with matched stimulation
or directly training appropriate replacement
behavior may also decrease the duration of
RIRD. In regard to the former, Love et al.
(2012) demonstrated that arranging matched
352 CANDICE L. COLÓN and WILLIAM H. AHEARN
stimulation in conjunction with RIRD may
function as an abolishing operation, thereby
decreasing the rate of stereotypy and concomi-
tantly the frequency of RIRD implementation.
In regard to the latter, the direct training of
appropriate replacement behavior (e.g. Colón
et al., 2012) in conjunction with RIRD may
assist with suppression of vocal stereotypy by
making appropriate behavior more probable
and stereotypy less probable. Future studies
should also determine the effects of treatment
integrity failures on appropriate behavior
(e.g., appropriate vocalizations).
During the parametric analysis, we observed
a decrease in latency and ITT for Morris, but
not for Kent and Cora. It is possible that RIRD
was not potent enough to compete with a con-
tinuous schedule of automatic reinforcement
and habituation to RIRD may have occurred.
Relatedly, Cipani, Brendlinger, McDowell, and
Usher (1991) compared the effect of contin-
gent application of lemon juice to manual guid-
ance overcorrection without the use of
extinction on stereotypy and self-injurious
behavior. They found that lemon juice
decreased these behaviors to near-zero levels on
both a continuous schedule and a variable ratio
4 schedule. Although overcorrection decreased
the target responses when implemented contin-
uously, it did not produce sufficient response
suppression on a variable ratio 4 schedule. Fur-
ther investigation is necessary to clarify under
what conditions continuous and noncontinuous
applications of such procedures produce suffi-
cient treatment outcomes.
A potential limitation to the current study is
that the parametric analysis always occurred in
the following order: 75%-50%-25%. This
sequence may have led to outcomes that would
have differed had we conducted conditions in
different sequences (e.g., 25%-75%-50%). We
elected to conduct conditions in a fixed
sequence for the purpose of replication across
participants, and to mirror other parametric
analyses of treatment consistency that have
simulated systematic schedule thinning by
implementing variations of treatment integrity
on a continuum (Northup, Fisher, Kahng,
Harrell, & Kurtz, 1997; Stephenson & Hanley,
2010; Vollmer et al., 1999) and obtaining
behavioral stability before titrating up or down
the schedule. Future research could take steps
to identify potential sequence effects (see Hig-
gins Hains, & Baer, 1989). In addition, the
Experiment 1 results indicate that treatment
implementation in the natural environment is
variable. Future research investigating variable
or random levels of treatment consistency may
be important to understand how naturally
occurring patterns of treatment implementation
impact treatment outcome. Furthermore,
although we used cue cards to assist with
adherence during the parametric analysis, pro-
cedural integrity data were not collected.
The findings from the current study in con-
junction with existing research (e.g., Ahrens
et al., 2011; Lerman & Iwata, 1996; Roll-
ings & Baumeister, 1981; Romancyzk, 1977;
St. Peter Pipkin et al., 2010; Vollmer et al.,
1999) generally indicate that providing expo-
sure to 100% treatment integrity prior to the
onset of possible treatment integrity failures
may maximize the likelihood of clinically sig-
nificant treatment success. However, it
remains unknown how recent or how much
exposure to 100% integrity is necessary to
produce successful treatment outcomes.
Although further replication, particularly in
the natural setting, is necessary, the booster
condition in Experiment 2 offers a potential
strategy to ameliorate poor treatment integrity.
Moreover, although bolstering integrity to a
mean of 50% or greater sufficiently salvaged
previous treatment effects in the current study,
future research regarding the necessary param-
eters of effective booster sessions is warranted.
Until further research is conducted, it is
important for clinicians to monitor and ensure
that the level of ongoing treatment integrity
produces acceptable treatment effects.
 TREATMENT INTEGRITY
REFERENCES
Ahearn, W. H., Clark, K. M., DeBar, R., &
Florentino, C. (2005). On the role of preference in
response competition. Journal of Applied Behavior
Analysis, 38, 247-250. https://doi.org/10.1901/jaba.
2005.36-04
Ahearn, W. H., Clark, K. M., MacDonald, R. P. F., &
Chung, B. I. (2007). Assessing and treating vocal ste-
reotypy in children with autism. Journal of Applied
Behavior Analysis, 40, 263-275. https://doi.org/10.
1901/jaba.2007.30-06
Ahrens, E. N., Lerman, D. C., Kodak, T.,
Worsdell, A. S., & Keegan, C. (2011). Further evalu-
ation of response interruption and redirection as
treatment for stereotypy. Journal of Applied Behavior
Analysis, 44, 95-108. https://doi.org/10.1901/jaba.
2011.44-95
Azrin, N. H., Holz, W. C., & Hake, D. F. (1963).
Fixed-ratio punishment. Journal of the Experimental
Analysis of Behavior, 6, 141-148. https://doi.org/10.
1901/jeab.1963.6-141
Banks, R. K. (1966). Persistence to continuous punish-
ment following intermittent punishment training.
Journal of Experimental Psychology, 71, 373-377.
https://doi.org/10.1037/h0023016
Brooks, P. H., & Baumeister, A. A. (1977). A plea for
consideration of ecological validity in the experimen-
tal psychology of mental retardation: A guest edito-
rial. American Journal of Mental Deficiency, 81(5),
407-416.
Carroll, R. A., Kodak, T., & Fisher, W. W. (2013). An
evaluation of programmed treatment integrity errors
during discrete-trial instruction. Journal of Applied
Behavior Analysis, 46, 379-394. https://doi.org/10.
1002/jaba.49
Cipani, E., Brendlinger, J., McDowell, L., & Usher, S.
(1991). Continuous vs. intermittent punishment: A
case study. Journal of Developmental and Physical
Disabilities, 3, 147-156. https://doi.org/10.1007/
BF01045930
Colón, C. L., Ahearn, W. H., Clark, K. M., &
Masalsky, J. (2012). The effects of verbal operant
training and response interruption and redirection on
appropriate and inappropriate vocalizations. Journal of
Applied Behavior Analysis, 45, 107-120. https://doi.
org/10.1901/jaba.2012.45-107
Donnelly, M. G., & Karsten, A. M. (2017). Effects of
programmed teaching errors on acquisition and dura-
bility of self-care skills. Journal of Applied Behavior
Analysis, 50, 511-528. https://doi.org/10.1002/
jaba.390
Duffy-Cassella, M., Sidener, T. M., Sidener, D. W., &
Progar, P. R. (2011). Response interruption and redi-
rection for vocal stereotypy in children with autism:
A systematic replication. Journal of Applied Behavior
Analysis, 44, 169-173. https://doi.org/10.1901/jaba.
2011.44-169.
353
Fisher, W., Piazza, C. C., Bowman, L. G.,
Hagopian, L. P., Owens, J. C., & Slevin, I. (1992).
A comparison of two approaches for identifying
reinforcers for persons with severe and profound dis-
abilities. Journal of Applied Behavior Analysis, 25,
491-498. https://doi.org/10.1901/jaba.1992.25-491
Fryling, M. J., Wallace, M. D., & Yassine, J. N. (2012).
Impact of treatment integrity on intervention effec-
tiveness. Journal of Applied Behavior Analysis, 45,
449-453. https://doi.org/10.1901/jaba.2012.45-449
Gaylord-Ross, R. J. (1979). Mental retardation research,
ecological validity, and the delivery of longitudinal
education programs. The Journal of Special Education,
13, 69-80.
Gresham, F. M. (1989). Assessment of treatment integrity
in school consultation and pre-referral intervention.
School Psychology Review, 18, 37-50.
Gresham, F. M., Gansle, K. A., & Noell, G. H. (1993).
Treatment integrity in applied behavior analysis with
children. Journal of Applied Behavior Analysis, 26,
257-263. https://doi.org/10.1901/jaba.1993.26-257
Gresham, F. M., MacMillan, D. L., Beebe-
Frankenberger, M. E., & Bocian, K. M. (2000).
Treatment integrity in learning disabilities interven-
tion research: Do we really know how treatments are
implemented? Learning Disabilities Research &
Practice, 15, 198-205. https://doi.org/10.1207/
SLDRP1504_4
Higgins Hains, A., & Baer, D. M. (1989). Interaction
effects in multielement designs: Inevitable, desirable,
and ignorable. Journal of Applied Behavior Analysis,
22, 57-69. https://doi.org/10.1901/ jaba.1989.22-57
Hinz, K. L., McGee, H. M., Huitema, B. E.,
Dickinson, A. M., & Van Enk, R. A. (2014).
Observer accuracy and behavior analysis: Data collec-
tion procedures on hand hygiene compliance in a
neurovascular unit. American Journal of Infection
Control, 42, 1067-1073. https://doi-org.silk.library.
umass.edu/10.1016/j.ajic.2014.06.017
Iwata, B.A., Dorsey, M. F., Slifer, K. J.,
Bauman, K. E., & Richman, G. S. (1994). Toward a
functional analysis of self-injury. Journal of Applied
Behavior Analysis, 27, 197-209. (Reprinted from
Analysis and Intervention in Developmental Disabilities,
2, 3-20, 1982). https://doi.org/10.1002/bin.297
Kincaid, J. P., Fishburne, R. P., Rogers, R. L., &
Chissom, B. S. (1975). Derivation of new readability
formulas (automated readability index, fog count,
and Flesch reading ease formula) for Navy enlisted
personnel. Institute for Simulation and Training.
Paper 56.
Kliebert, M. L., Tiger, J. H., & Toussaint, K. A. (2011).
An approach to identifying the conditions under
which response interruption will reduce automatically
reinforced problem behavior. Behavior Analysis in
Practice, 4, 17-26. https://doi.org/10.1007/
BF03391771
354 CANDICE L. COLÓN and WILLIAM H. AHEARN
Lerman, D. C., & Iwata, B. A. (1996). A methodology
for distinguishing between extinction and punish-
ment effects associated with response blocking. Jour-
nal of Applied Behavior Analysis, 29, 231-233. https://
doi.org/10.1901/jaba.1996.29-231
Love, J. J., Miguel, C. F., Fernand, J. K., & LaBrie, J. K.
(2012). The effects of matched stimulation and
response interruption and redirection on vocal stereo-
typy. Journal of Applied Behavior Analysis, 45, 549-
564. https://doi.org/10.1901/jaba.2012.45-549.
Martinez, C. K., & Betz, A. M. (2013). Response inter-
ruption and redirection: Current research trends and
clinical application. Journal of Applied Behavior
Analysis, 46, 549-554. https://doi.org/10.1002/
jaba.38
McIntyre, L. L., Gresham, F. M., DiGennaro, F. D., &
Reed, D. D. (2007). Treatment integrity of school-
based interventions with children in the Journal of
Applied Behavior Analysis 1991-2005. Journal of
Applied Behavior Analysis, 40, 659-672. https://doi.
org/10.1901/jaba.2007.659-672
Miguel, C. F., Clark, K., Tereshko, L., & Ahearn, W. H.
(2009). The effects of response interruption and redi-
rection and sertraline on vocal stereotypy. Journal of
Applied Behavior Analysis, 42, 883-888. https://doi.
org/10.1901/jaba.2009.42-883
Northup, J., Fisher, W., Kahng, S. W., Harrell, R., &
Kurtz, P. (1997). An assessment of the necessary
strength of behavioral treatments for severe behavior
problems. Journal of Developmental and Physical
Disabilities, 9, 1-16. https://doi.org/10.1023/A:
1024984526008
Peterson, L., Homer, A. L., & Wonderlich, S. A. (1982).
The integrity of independent variables in behavior
analysis. Journal of Applied Behavior Analysis, 15,
477-492. https://doi.org/10.1901/jaba.1982.15-477
Ringdahl, J. E., Vollmer, T. R., Marcus, B. A., &
Roane, H. S. (1997). An analogue evaluation of envi-
ronmental enrichment: The role of stimulus prefer-
ence. Journal of Applied Behavior Analysis, 30,
203-216.
Rollings, J. P., & Baumeister, A. A. (1981). Stimulus con-
trol of stereotypic responding: Effects on target and
collateral behavior. American Journal of Mental
Deficiency, 86, 67-77.
Romanczyk, R. G. (1977). Intermittent punishment of
self-stimulation: Effectiveness during application and
extinction. Journal of Consulting and Clinical
Psychology, 45, 53-60. https://doi.org/10.
1037/0022-006X.45.1.53
Roscoe, E. M., Carreau, A., MacDonald, J., &
Pence, S. T. (2008). Further evaluation of leisure
items in the attention condition of functional ana-
lyses. Journal of Applied Behavior Analysis, 41,
351-364. https://doi.org/10.1901/jaba.2008.41-351
Saini, V., Gregory, M. K., Uran, K. J., & Fantetti, M. A.
(2015). Parametric analysis of response interruption
and redirection as treatment for stereotypy. Journal of
Applied Behavior Analysis, 48, 96-106. https://doi.
org/10.1002/jaba.186
Shawler, L. A., & Miguel, C. F. (2015). The effects of
motor and vocal response interruption and redirec-
tion on vocal stereotypy and appropriate vocaliza-
tions. Behavioral Interventions, 30, 112-134. https://
doi.org/10.1002/bin.1407
Shemer, A., & Feldon, J. (1984). Long-term partial rein-
forcement extinction effect and long-term partial
punishment effect in a one-trial-a-day paradigm.
Bulletin of the Psychonomic Society, 22, 221-224.
Smith, R. G., Russo, L., & Le, D. D. (1999). Distin-
guishing between extinction and punishment effects
of response blocking: A replication. Journal of Applied
Behavior Analysis, 32, 367-370. https://doi.org/10.
1901/jaba.1999.32-367
Stephenson, K. M., & Hanley, G. P. (2010). Pre-
schoolers’ compliance with simple instructions: A
descriptive and experimental evaluation. Journal of
Applied Behavior Analysis, 43, 229-247. https://doi.
org/10.1901/jaba.2010.43-229
St. Peter Pipkin, C., Vollmer, T. R., & Sloman, K. N.
(2010). Effects of treatment integrity failures during
differential reinforcement of alternative behavior: A
translational model. Journal of Applied Behavior
Analysis, 43, 47-70. https://doi.org/10.1901/jaba.
2010.43-47
Toper-Korkmaz, O., Lerman, D. C., & Tsami, L.
(in press). Effects of toy removal and number of
demands on vocal stereotypy during response inter-
ruption and redirection. Journal of Applied Behavior
Analysis.
Vanderkerken, L., Heyvaert, M., Maes, B., &
Onghena, P. (2013). Psychosocial interventions for
reducing vocal challenging behavior in persons with
autistic disorder: A multilevel meta-analysis of single-
case experiments. Research in Developmental
Disabilities, 34, 4515-4533. https://doi.org/10.1016/
j.ridd.2013.09.030
Vollmer, T. R., Roane, H. S., Ringdahl, J. E., &
Marcus, B. A. (1999). Evaluating treatment chal-
lenges with differential reinforcement of alternative
behavior. Journal of Applied Behavior Analysis, 32, 9-
23. https://doi.org/10.1901/jaba.1999.32-9
Vollmer, T. R., Sloman, K. N., & St. Peter Pipkin, C.
(2008). Practical implications of data reliability
and treatment integrity monitoring. Behavior Analysis
in Practice, 1, 4-11. https://doi.org/10.1007/
BF03391722
Received December 12, 2016
Final acceptance July 14, 2018
Action Editor, Jason Vladescu