JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2012, 45, 607–611 NUMBER 3 (FALL 2012)

SYSTEMATIC EVALUATION OF VARIABLES THAT CONTRIBUTE TO

The effects of time-out and escape extinction were examined with 2 preschoolers after we
identified variables that may have resulted in noncompliance. Results of a functional analysis
showed that noncompliance was highest in the escape condition for both participants. During
the treatment evaluation, escape extinction resulted in greater reductions in noncompliance
relative to time-out.
Key words: noncompliance, functional analysis, preschool children

Despite the relative abundance of indirect and vice versa. Results indicated that noncom-
and descriptive studies on noncompliance (e.g., pliance was maintained by attention for all
Crowther, Bond, & Rolf, 1981; Ndoro, participants.
Hanley, Tiger, & Heal, 2006), few studies In the current study, we sought to replicate
have evaluated appropriate methods for con- and extend Rodriguez et al. (2010) by (a)
ducting functional analyses of noncompliance assessing the function of preschoolers’ noncom-
(Kern, Delaney, Hilt, Bailin, & Elliot, 2002; pliance with common preschool curriculum
Wilder, Harris, Reagan, & Rasey, 2007). tasks and (b) comparing the effects of two
Rodriguez, Thompson, and Baynham (2010) potential treatments for noncompliance, escape
tested a procedure for comparing the effects of extinction (e.g., Cote, Thompson, & McKer-
attention and escape on noncompliance exhib- char, 2005) and time-out (e.g., American
ited by three preschoolers. During the attention Academy of Pediatrics, 1998).
condition, noncompliance was followed by
adult attention and physical guidance to METHOD
complete the demand, whereas compliance
was followed by escape from the demand and Participants, Setting, and Materials
removal of attention. In the escape condition, Jeane, a 4-year-old girl who had been
these contingencies were reversed such that diagnosed with tuberous sclerosis and infantile
escape was contingent on noncompliance and spasms, and Jayme, a typically developing 3-
attention was contingent on compliance. This year-old boy, participated. Sessions were con-
procedure permitted the escape condition to ducted at the participants’ preschool in a room
serve as the control for the attention condition (2.1 m by 1.5 m) equipped with a one-way
observation window, a table, two chairs, a baby
This research was completed in partial fulfillment of monitor, a timer, and other relevant materials
the requirements for the MS degree in Psychology at (see description below). Two to four sessions
Jacksonville State University by Layla Abby, who is now were conducted per day, no more than three
at Texas Tech University. We thank Todd McKerchar,
Eileen Roscoe, and the anonymous reviewers for their
times per week.
helpful comments on this paper. We also thank Suzette
Rosario, Amanda Nwosu, and Krista White for their Response Measurement and Interobserver
assistance with data collection. Agreement
Address correspondence to Paige M. McKerchar, Trained observers recorded the occurrence of
Department of Psychology, Jacksonville State Univer-
sity, 700 Pelham Road North, Jacksonville, Alabama compliance behind a one-way observation
36265 (e-mail: pmckerchar@jsu.edu). window using the ABC Data Pro application
doi: 10.1901/jaba.2012.45-607 for iPhone and iPod Touch. Compliance was

607
608 PAIGE M. MCKERCHAR and LAYLA ABBY

defined as initiating the task within 5 s of the attention and escape conditions, and control
experimenter’s initial (vocal) prompt and was demonstrated via a contingency reversal
continuously and correctly completing it with- that allowed noncompliance to occur in both
out interruption. Conversely, noncompliance1 conditions.
was defined as the participant not initiating the Prior to each session, the experimenter briefly
task within 5 s of the initial prompt or described the consequences for compliance and
initiating the task within 5 s of the prompt noncompliance to the participant to facilitate
with interruption (i.e., not continuously com- differential responding. During all sessions, the
pleting the task correctly). Initiating included participant was seated next to the experimenter
responses required for correctly completing the at a table. At the start of each demand
task (e.g., twisting a puzzle piece to fit in a
presentation, the experimenter placed one
spot). Participants typically completed tasks
sample stimulus (e.g., a dog picture), one
within 5 s of the experimenter’s prompts;
however, tasks always were completed within matching comparison stimulus (e.g., a dog
25 s. During the functional analysis, data were picture), and one nonmatching comparison
not collected on participants’ responses to stimulus (e.g., a fish picture) on the table in
instructions presented during the reinforcement front of the participant. The therapist pointed
interval. to the sample picture while stating, ‘‘match’’ (a
For each participant, a second observer mastered skill according to the Assessment of
independently recorded data during a mini- Basic Language and Learning Skills—Revised
mum of 40% of functional analysis sessions and for Jeane and based on teacher report for
35% of treatment sessions. Interobserver agree- Jayme). This demand was presented every 30 s
ment was calculated on a trial-by-trial basis by during each 5-min session (10 demands per
dividing the number of agreements by the total session), with the exception of Sessions 6, 7,
number of demands presented and converting and 8 for Jayme, in which fewer demand
this number to a percentage. An agreement was presentations occurred due to occasional diffi-
defined as both observers recording the same culty with returning him to the table after the
response following a demand. Mean agreement programmed escape.
across participants was 98% (range, 90% to Attention condition. Contingent on noncom-
100%) for functional analysis sessions and 99% pliance, the experimenter presented continuous
(range, 96% to 100%) for treatment sessions. encouragement (e.g., ‘‘Come on. I know you
can do it!’’) and prompts to match additional
Functional Analysis cards for the remainder of the 30-s interval (no
The functional analysis procedures were escape). That is, a new card and instruction
identical to those described by Rodriguez et were delivered as soon as the previous card was
al. (2010), except that a matching task was matched. If the participant stopped initiating
presented rather than a clean-up task. A the matching response at any time during the
multielement design was used to compare the reinforcement interval, the experimenter im-
1
To remain consistent with previous research on the
mediately provided hand-over-hand guidance
functional analysis of noncompliance, noncompliance is to ensure that the participant continued to
defined as the absence of compliance. Nevertheless, the match cards throughout the interval. Compli-
authors recognize that functional analyses typically
involve the delivery of potential reinforcers contingent
ance resulted in removal of experimenter
on the occurrence of a target behavior, and thus, there attention (no attention) and removal of task
may be conceptual and, possibly, methodological issues materials (escape) for the remainder of the 30-s
with providing consequences contingent on noncom- interval (i.e., until the next scheduled demand).
pliance when it is defined as the absence of behavior.
Discussion of this issue, however, is beyond the scope of Escape condition. Noncompliance with the
this report. initial instruction resulted in removal of
 EVALUATION OF NONCOMPLIANCE 609

experimenter attention (no attention) and could complete these tasks following a vocal
removal of task materials (escape) for the prompt only, the experimenter simultaneously
remainder of the 30-s interval. Contingent on presented vocal and model prompts during
compliance, the experimenter provided contin- initial treatment sessions (i.e., Sessions 17
uous praise (e.g., ‘‘You’re a great worker!’’) and through 20 for Jeane and Sessions 17 through
prompts to match additional cards for the 19 for Jayme) to ensure that the participants
remainder of the 30-s interval (no escape). That contacted the relevant contingency in effect.
is, a new card and instruction were delivered as During treatment, the experimenter sat next
soon as the previous card was matched. If the to the participant at a table and continuously
participant stopped initiating the matching presented demands during 5-min sessions.
response during the 30-s interval that followed Instructional tasks were presented in a quasi-
compliance with the initial instruction, the random order within and across sessions, such
experimenter immediately provided hand-over- that all tasks were presented in every session and
hand guidance to ensure that the participant two to three trials of one type of task (e.g., letter
continued to match cards throughout the identification) were presented before moving on
interval. to the next type (e.g., puzzle). For each task, the
Both participants exhibited higher levels of experimenter placed the necessary materials on
noncompliance in the escape condition than in the table and presented a vocal prompt (e.g.,
the attention condition, suggesting that non- ‘‘Point to the letter C.’’). In both conditions,
compliance was maintained, at least in part, by compliance resulted in experimenter praise and
negative reinforcement in the form of escape. immediate presentation of the next task.
Escape extinction. In this condition, the table
Treatment Evaluation
was moved to a corner of the room to enhance
Although the functional analysis included an
embedded treatment evaluation in that escape stimulus control and treatment integrity (i.e., this
extinction and time-out were in effect for placement allowed the experimenter to block
noncompliance in the attention and escape attempts to leave the table). Contingent on
conditions, respectively, we further evaluated noncompliance, the experimenter modeled the
the effects of these interventions under more correct behavior. If the participant initiated the
naturalistic conditions (e.g., with a variety of task within 5 s of the model prompt and
tasks and the removal of encouragement during completed the task correctly, the experimenter
escape extinction) using a multielement design. provided praise and presented the next task. If the
We presented common preschool curriculum participant did not initiate or correctly complete
tasks reported by teachers to result in noncom- the task after the model prompt, the experiment-
pliance during typical classroom routines. er used hand-over-hand guidance to have the
Specifically, these tasks included placing one participant complete the task. Noncompliance
piece into a four-piece puzzle with two pieces following model prompts was not included in
removed (two different puzzles were used), our data summary; only noncompliance after
matching visual stimuli by category (in a two- vocal prompts is depicted in our results.
stimulus array), and pointing to a visual Time-out. Contingent on noncompliance,
comparison stimulus (in a two-stimulus array) the experimenter removed the task materials
following an auditory stimulus. Stimuli includ- and turned away from the participant for 15 s.
ed categories (i.e., flowers, chairs, and apples),
actions (e.g., crawling, smiling, running),
RESULTS AND DISCUSSION
simple shapes, the letters A through E, and
the numbers 1 through 5. Although informal During Jeane’s functional analysis (Figure 1,
observations revealed that both participants top), noncompliance occurred at higher levels
610 PAIGE M. MCKERCHAR and LAYLA ABBY

Figure 1. Percentage of demands followed by noncompliance for Jeane (top) and Jayme (bottom) during the
functional analysis and treatment evaluation.

in the escape condition (M ¼ 85%) than in the escape from demands, whereas Rodriguez et al.
attention condition (M ¼ 56%). During found that noncompliance was maintained by
treatment, her noncompliance was consistently attention. This may be due to procedural
lower during escape extinction (M ¼ 42%) than differences between the studies (e.g., different
during time-out (M ¼ 66%). Jayme’s functional tasks, settings), but also underlines the impor-
analysis (Figure 1, bottom) shows that, with the tance of conducting functional analyses of
exception of Session 4, he consistently engaged noncompliance prior to treatment to avoid
in higher levels of noncompliance in the escape use of a contraindicated treatment.
condition (M ¼ 67%) than in the attention During treatment, we compared two inter-
condition (M ¼ 27%). During treatment, ventions used for treating noncompliance,
noncompliance decreased across the escape escape extinction and time-out. Escape extinc-
extinction condition (M ¼ 38%), whereas tion, the treatment informed by the results of
noncompliance was variable throughout the the functional analysis, resulted in lower levels
time-out condition (M ¼ 69%). of noncompliance for both participants. There-
In our replication of Rodriguez et al. (2010), fore, the results of the treatment analysis
noncompliance appeared to be maintained by validated the functional analysis outcome.
 EVALUATION OF NONCOMPLIANCE 611

Despite these findings, a few limitations ance, one could evaluate the relative effects of
should be noted. First, we did not conduct escape extinction alone and in combination
formal assessments to ensure that responses with differential reinforcement.
required for compliance were in the partici- In summary, we replicated the functional
pant’s repertoire. Therefore, it is possible that analysis procedures described by Rodriguez et
noncompliance may have been due to a skill al. (2010) and extended the results by conduct-
deficit rather than motivational variables. To ing a treatment informed by the functional
help rule out this potential confounding effect, analysis outcome. However, because we includ-
formal assessments could be conducted prior to ed only two participants with escape-main-
the functional analysis. tained noncompliance, the generality of our
Second, although the contingency reversal findings across populations, settings, and be-
provided an opportunity for noncompliance to
havioral functions is unclear. Therefore, addi-
occur in both functional analysis conditions, the
tional research on the assessment and treatment
absence of a separate control condition limited
of noncompliance is warranted.
detection of multiple control when noncompli-
ance occurred in attention and escape conditions
(as it did with Jeane). Future research should REFERENCES
evaluate alternative control conditions for func- American Academy of Pediatrics. (1998). Guidance for
tional analyses of noncompliance. effective discipline. Pediatrics, 101, 723–728.
Third, because the time-out condition in- Cote, C. A., Thompson, R. H., & McKerchar, P. M.
cluded a 15-s break from demands for (2005). The effects of antecedent interventions and
extinction on toddlers’ compliance during transitions.
noncompliance, it was associated with a lower Journal of Applied Behavior Analysis, 38, 235–238.
rate of demand presentations (M ¼ 3) than the Crowther, J. H., Bond, L. A., & Rolf, J. E. (1981). The
escape extinction condition (M ¼ 5.1). There- incidence, prevalence, and severity of behavior
fore, it is possible that time-out was less disorders among preschool-aged children in day care.
Journal of Abnormal Child Psychology, 9, 23–42.
effective because it was associated with fewer Kern, L., Delaney, B. A., Hilt, A., Bailin, D. E., & Elliot,
opportunities for noncompliance to contact the C. (2002). An analysis of physical guidance as
programmed contingency. In subsequent re- reinforcement for noncompliance. Behavior Modifi-
search that compares these interventions, one cation, 26, 516–536.
might increase the duration of time-out sessions Lerman, D. C., & Iwata, B. A. (1996). Developing a
technology for the use of operant extinction in
to correct for time spent in breaks. clinical settings: An examination of basic and applied
Finally, although lower levels of noncompli- research. Journal of Applied Behavior Analysis, 29,
ance occurred during escape extinction than in 345–382.
time-out, clinically significant reductions were Ndoro, V. W., Hanley, G. P., Tiger, J. H., & Heal, N. A.
(2006). A descriptive assessment of instruction-based
not observed. This may be due to limited interactions in the preschool classroom. Journal of
exposure to the treatment conditions (because Applied Behavior Analysis, 39, 79–90.
of time constraints) or the absence of a Rodriguez, N. M., Thompson, R. H., & Baynham, T. Y.
reinforcement contingency for compliance. (2010). A method for assessing the relative effects of
attention and escape on noncompliance. Journal of
For purposes of assessing the independent Applied Behavior Analysis, 43, 143–147.
effects of time-out and escape extinction, we Wilder, D. A., Harris, C., Reagan, R., & Rasey, A.
did not include a reinforcement component; (2007). Functional analysis and treatment of non-
however, research suggests that escape extinc- compliance by preschool children. Journal of Applied
tion is more effective when combined with Behavior Analysis, 40, 173–177.
other treatments (e.g., differential reinforce- Received July 20, 2011
ment of compliance; Lerman & Iwata, 1996). Final acceptance January 31, 2012
In further treatment evaluations of noncompli- Action Editor, Eileen Roscoe