JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1998, 31, 687–690 NUMBER 4 (WINTER 1998)

THE USE OF NONCONTINGENT ESCAPE TO

REDUCE DISRUPTIVE BEHAVIORS IN
CHILDREN WITH SPEECH DELAYS

CARRIE L. COLEMAN AND PETER A. HOLMES

EASTERN MICHIGAN UNIVERSITY

Noncontingent escape (NCE) was used to reduce disruptive behavior in 3 children during
regularly scheduled speech therapy sessions. Results showed rapid decreases in disruptive
behavior and accompanying increases in compliance across children. Findings suggest that
speech therapists with little expertise in behavior analysis can effectively implement NCE.
DESCRIPTORS: noncontingent escape, disruptive behavior, compliance, speech
therapy

Noncontingent escape (NCE), the remov- METHOD

al of aversive stimulation on a response-in- Participants and Setting
dependent schedule, has emerged as a prom-
ising treatment for escape-maintained prob- Three 4-year-old children (2 boys and 1
lem behavior (Vollmer, Marcus, & Ringdahl, girl) and their respective speech therapists
1995). The NCE treatment package, which participated. All children, who had received
typically involves fixed-time (FT) escape, ex- therapy for at least 7 months prior to this
tinction, and schedule thinning, has not study, engaged in high rates of disruptive be-
havior during training sessions. Neal had
been examined when used in a clinical con-
been diagnosed with pervasive developmen-
text by service providers who have little ex-
tal delay. Amy and Paul had been diagnosed
pertise in behavior analysis. Further, al-
with autism. All therapists possessed master’s
though the FT component of NCE should
degrees in speech and language pathology,
be relatively easy for care providers to im-
and their experience ranged from 2 to 25
plement, the schedule-thinning component years. Sessions took place at a speech clinic
used by Vollmer et al. is rather complicated. in rooms (3 m by 4 m) containing a table,
This study extends research on NCE by chairs, and instructional materials. Clinic
teaching speech therapists to implement sessions were usually scheduled in 30-min
NCE during regular sessions and by simpli- blocks. For various reasons (e.g., participant
fying the manner in which the schedule is late to session, therapist behind on sched-
thinned. ule), session length varied. Baseline consisted
of 10-min samples from regular therapy ses-
This report is based on a thesis submitted by the sions, and treatment sessions lasted 17 min
first author in partial fulfillment of the requirements on average. Sessions were usually conducted
for the MS degree at Eastern Michigan University. We 2 days per week.
thank the reviewers for their helpful comments on an
earlier version of this manuscript. The authors are Recording and Reliability
grateful to the therapists for their participation in this
study and to the parents who allowed their children Dependent variables included the per-
to participate. centage of 30-s intervals with (a) disruptive
Requests for reprints should be addressed to Peter
A. Holmes, Department of Psychology, Eastern Mich- behavior and (b) compliance. Disruptive be-
igan University, Ypsilanti, Michigan 48197. haviors included attempting to leave the

687
688 CARRIE L. COLEMAN and PETER A. HOLMES

chair, falling to the floor, kicking over the During baseline, the children sat at a ta-
chair, climbing on or under the table, run- ble with the therapists who delivered in-
ning to the door (Neal), screaming, crying structions in a one-on-one direct teaching
(Amy and Paul), hitting, and making inap- format. Therapists responded to compli-
propriate noises (Paul). Compliance was de- ance with praise, occasionally paired with
fined as correct responding within 5 s to a an edible item, and responded to noncom-
therapist’s instruction without physical guid- pliance and disruptive behavior with repet-
ance while not engaging in any disruptive itive verbal and physical prompts. In gen-
behavior. All sessions were videotaped and eral, each therapist delivered verbal com-
scored using continuous 30-s intervals in mands or requests to comply from 4 to 15
which disruptive behavior and compliance times per minute.
were recorded on an occurrence or nonoc- Prior to beginning NCE, therapists re-
currence basis. A second observer indepen- ceived vocal or written instructions, mod-
dently scored the videotapes during 46% of eling, and feedback. The therapist, cued by
the sessions. The number of agreements (in- a timer (Radio Shack Model 63-884 A),
terval by interval) was divided by the num- provided a 30-s break on an FT schedule.
ber of agreements plus disagreements and The instructional context, including con-
multiplied by 100%. Interobserver agree- sequences for compliance, was identical to
ment averaged 100% for disruptive behavior baseline. Sessions began with the therapist
and 80% for compliance. Procedural integ- prompting the child to ‘‘sit down.’’ The
rity was assessed on an ongoing basis to en-
prompting hierarchy consisted of a verbal
sure the therapists’ accurate implementation
request, model, and physical guidance as
of the treatment procedure. All therapists
needed, with 5 s between prompts. Rates of
provided 100% of scheduled breaks within
instructional requests ranged from 5 to 13
5 s of the FT schedule.
per minute (roughly similar to baseline).
Social Validity Assessment The initial FT schedule was arbitrarily set
The eight-item Client Satisfaction Ques- at 1 min and, as in Vollmer et al. (1995),
tionnaire (Attkisson, 1994), scored on a 4- was followed by 1.5 min, 2.0 min, 2.5 min,
point Likert-type scale, was administered at 3.0 min, and finally 4 min. The FT sched-
the conclusion of this study. Possible overall ule was advanced if the child engaged in no
scores range from 8 to 32, with higher scores disruptive behavior for three consecutive
indicating greater satisfaction. An example of FT intervals within a session. If the crite-
items on the scale is, ‘‘To what extent has rion was not met within two sessions, the
our program met your needs?’’ schedule was set back to the previous step.
The eventual goal for all children was to
Procedures and Experimental Design establish a 4-min FT interval, which, based
No pretreatment functional analysis was on therapists’ reports, was determined to be
conducted. However, based on interviews a socially valid interval.
with therapists and observation of baseline During follow-up, Neal’s and Paul’s ther-
videotape, the children’s behavior was hy- apists continued to implement the NCE
pothesized to be escape maintained because procedure using the terminal FT interval es-
the onset of disruptive behavior was highly tablished during treatment (FT 4 min). Fol-
correlated with instructional activity. The ef- low-up data were not collected for Amy be-
fects of treatment were evaluated using a cause her therapist left the clinic (for rea-
multiple baseline design. sons unrelated to this study).
 NONCONTINGENT ESCAPE 689

Figure 1. The percentage of 30-s intervals with disruptive behavior and compliance for each child during
baseline and NCE. Lines marked with time intervals indicate a change in the FT schedule during that session.

RESULTS AND DISCUSSION ment. Compliance increased from a baseline

Figure 1 shows that NCE suppressed dis- mean of 36.7% of intervals to a treatment
ruptive behavior and increased compliance mean of 70.9% of intervals. Because Amy’s
for all children. For Neal, the mean per- treatment phase was abbreviated, we exam-
centage of intervals with disruptive behavior ined her within-session response patterns
decreased from 100% in baseline to 16.7% across minutes during each of the three
during treatment. Compliance increased treatment sessions (box within center panel).
from a baseline mean of 35% of intervals to These data further support that her disrup-
a treatment mean of 96.4% of intervals. tive behavior rates decreased as treatment
Treatment effects were maintained at follow- progressed. For Paul, the mean percentage of
up. For Amy, the mean percentage of inter- intervals with disruptive behavior decreased
vals with disruptive behavior during baseline from 63% in baseline to 20% during treat-
was 65% compared to 24.3% during treat- ment. Compliance increased from a baseline
690 CARRIE L. COLEMAN and PETER A. HOLMES

mean of 75% of intervals to a treatment tion before treatment began, and treatment
mean of 92.6% of intervals. Treatment ef- goals were reached in relatively few sessions.
fects were maintained at follow-up. It should Paul’s unstable baseline and the brevity of
be noted that Paul’s disruptive behavior was Amy’s treatment phase preclude any conclu-
on a downward trend during baseline. Al- sive statements about the robustness of NCE
though it would have been appropriate for effects. However, overall, these results sup-
baseline to continue until stable patterns of port the practical utility of the NCE ap-
proach, considering issues of ease of imple-
responding were obtained, Paul’s therapist
mentation, rapidity of effect, and treatment
did not want to wait further before learning acceptability. Giving the children breaks led
how to decrease his disruptive behavior. to reductions in disruptive behavior, which
Social validity assessment results were 30 ultimately increased the therapists’ oppor-
for Neal’s therapist, 31 for Amy’s therapist, tunities to teach them skills.
and 32 for Paul’s therapist, suggesting that
all therapists were highly satisfied with the REFERENCES
intervention. Attkisson, C. C. (1994). Client satisfaction question-
Results of this study extend previous re- naire. In J. Fischer & K. Corcoran (Eds.), Mea-
search on NCE by demonstrating that the sures for clinical practice: A sourcebook (pp. 119–
121). New York: The Free Press.
technique can be implemented in a clinical Vollmer, T. R., Marcus, B. A., & Ringdahl, J. E.
context by service providers who have little (1995). Noncontingent escape as treatment for
self-injurious behavior maintained by negative re-
formal training in behavior analysis. The inforcement. Journal of Applied Behavior Analysis,
method for increasing the escape schedule 28, 15–26.
followed a simple rule to increase the FT Received August 20, 1997
Initial editorial decision October 3, 1997
interval following three problem-free inter- Final acceptance May 4, 1998
vals. Therapists required minimal instruc- Action Editor, Timothy R. Vollmer