JOURNAL OF APPLIED BEHAVIOR ANALYSIS 2007, 40, 685–689 NUMBER 4 (WINTER 2007)

TRAINING PARENT IMPLEMENTATION OF DISCRETE-TRIAL

TEACHING: EFFECTS ON GENERALIZATION OF PARENT TEACHING
AND CHILD CORRECT RESPONDING
MICHAEL LAFASAKIS
HOSPITAL CLINIC HOME CENTER INC.
QUEENS COLLEGE, CITY UNIVERSITY OF NEW YORK

AND

PETER STURMEY
QUEENS COLLEGE AND THE GRADUATE CENTER
CITY UNIVERSITY OF NEW YORK

Behavioral skills training was used to teach 3 parents to implement discrete-trial teaching with
their children with developmental disabilities. Parents learned to implement discrete-trial
training, their skills generalized to novel programs, and the children’s correct responding
increased, suggesting that behavioral skills training is an effective and efficient method of
teaching discrete-trial teaching to parents.
DESCRIPTORS: parent training, discrete-trial teaching, autism, behavioral skills training,
imitation, developmental disabilities
________________________________________

Many parents of children with developmen- to implement discrete-trial teaching correctly

tal disabilities now participate in their children’s and to correct idiosyncratic errors in teaching
early education, including teaching their chil- using similar brief behavioral skills training
dren using discrete-trial teaching (Sturmey & procedures. They showed that improvements in
Fitzer, 2007). This method has been used to staff behavior were accompanied by large
teach a wide range of social, language, and reductions in stereotypy in students with autism
academic behaviors to children with autism spectrum disorders. Sarokoff and Sturmey (in
spectrum disorders (Sturmey & Fitzer). Koegel, press) extended these studies by demonstrating
Russo, and Rincover (1977) showed that that teaching staff to conduct one discrete-trial
educational staff can acquire discrete-trial teaching program using instructions, modeling,
teaching skills; however, instruction took up rehearsal, and feedback produced generalization
to 25 hr. Sarokoff and Sturmey (2004) taught 3 of correct use of discrete-trial teaching to
teachers to implement discrete-trial teaching programs and students who they had not
using instructions, modeling, rehearsal, and previously been taught to work with. Further,
feedback. Intervention took only approximately these improvements in staff teaching were also
three 10-min sessions. Dib and Sturmey (2007) accompanied by increases in children’s correct
extended this study by teaching education staff responding in receptive language training
programs.
We thank the parents and their children for participa- Unfortunately, researchers have not applied
tion in this study. We also express appreciation to these methods extensively to teach parents of
Jonathan Karp, Frieda Spivack, John Ward-Horner,
Joseph Rosario Yanni, and Jason William Lantier. children with developmental disabilities to
Requests for reprints should be sent to Peter Sturmey, teach their own children. Hardy and Sturmey
Queens College, and Department of Psychology, 65-30 (1994) demonstrated that behavioral skills
Kissena Boulevard, Flushing, New York 11367 (e-mail:
peter_sturmey@QC.CUNY.edu). training was effective in teaching parents of 3
doi: 10.1901/jaba.2007.685–689 children with developmental disabilities to

685
686 MICHAEL LAFASAKIS and PETER STURMEY

teach their children. This study collected only a child-sized table and chairs present. During all
data on parents’ teaching skills. It did not assess sessions the parent sat across from the child. A
generalization of parents’ teaching and did not video camera on a tripod was used to record
collect data on children’s behavior. Thus, each session.
although parents are often involved in their
children’s early education, we do not know how Design, Data Collection, and Dependent Measure
to train them effectively and efficiently, whether A multiple baseline across parents design was
their teaching skills generalize to novel teaching used. The experimenter trained the parents to
programs, and what the effects are on their use discrete-trial teaching only during gross
children’s correct responding. Therefore, the motor imitation (GMI). To assess generaliza-
present study evaluated whether (a) behavioral tion, data were also collected during vocal
skills training was effective in teaching discrete- imitation (VIM). GMI was defined as the
trial teaching to parents of children with presentation of a motor movement by a parent,
developmental disabilities; (b) generalization in whose presence responding similarly to the
of parent teaching skills from trained to parent by the child produces reinforcement
untrained programs occurred; and (c) changes (Baer, Peterson, & Sherman, 1967). VIM was
in parent teaching were accompanied by defined as the presentation of a sound by
increases in children’s correct responding. a parent, in whose presence responding simi-
larly to the parent by the child produces
reinforcement. VIM took place during approx-
METHOD
imately 48% of all sessions, distributed across
Participants and Setting all instructions baseline and posttraining con-
Three parent–child dyads participated. Rosa ditions. The experimenter videotaped each
was a 35-year-old college graduate who was session and scored the videotape later.
born in Nigeria, had low income, and was There were four dependent measures. The
a registered nurse. Maria was a 48-year-old first two were the parents’ percentage of correct
American-born college graduate, had low in- use of 10 components during 10 consecutive
come, and was an administrator. Joanne was discrete trials of GMI and VIM. Observers
a 50-year-old high school graduate who was collected data using the same method as
born in Nigeria, had low income, and was Sarokoff and Sturmey (2004), in which parents
working as a direct-care worker with seniors. All could emit 10 correct teaching responses (e.g.,
these parents had no previous experience with obtaining eye contact), use a correction pro-
discrete-trial teaching. George was a 4-year-old cedure correctly, and collect data during each
boy with autism who exhibited no vocal trial. Percentage correct was calculated by
behavior and few imitative skills. Emmanuel dividing the total number of correct responses
was a 4-year-old boy with mental retardation, by the total number of correct and incorrect
severe self-injurious behavior, and no vocal responses and multiplying by 100%. Observers
behavior. Christian was a 4-year-old boy with also collected data on children’s correct re-
Down syndrome who had no vocal behavior sponses. A correct response was defined as any
and severe motor problems. The children had behavior that matched the modeled response or
been diagnosed by an in-house licensed psy- was an approximation of the modeled response.
chologist or an outside agency. All children For the GMI target behavior of clapping hands,
were not on any medications and had no prior a correct response was scored if the child
experience with discrete-trial teaching. touched his open hands together once or twice.
Parent training took place in a room (3 m by For the VIM target behavior ‘‘ah,’’ a correct
3 m) at a special education preschool with response was scored if the child’s mouth opened
 GENERALIZATION OF PARENT TEACHING 687

wide while he said ‘‘ah’’ or if his mouth opened nents performed correctly and informative
halfway while he said ‘‘ah.’’ The percentage of corrective feedback on components that the
correct responses was calculated by dividing the parent needed to practice. The experimenter
number of correct responses by the total then sat with the child and modeled three
number of responses and multiplying by 100%. additional discrete trials that included the
During GMI, parents presented three mod- specific components that were previously im-
els: clap hands, arms up, and touch nose. Clap plemented incorrectly. Rehearsal and modeling
hands was defined as moving both hands were repeated with the parent performing three
together and making an audible sound with trials and experimenter demonstrating three
the hands. Arms up was defined as raising both trials until 10 min elapsed. Following each
hands simultaneously above the head. Touch training session, the parent then performed 10
nose was defined as placing one finger on the tip uninterrupted discrete trials. This session was
of the nose. For VIM, parents presented three videotaped and scored at a later time. The
models: ‘‘ah,’’ ‘‘mm,’’ and ‘‘mmaa.’’ The criterion for completion of training was 90% or
instruction used in both GMI and VIM was more correct parent responses on two consec-
the parent’s verbal instruction ‘‘do this.’’ The utive training sessions.
parent then modeled the correct response. Posttraining. At the beginning of each
posttraining session, the experimenter stated,
Procedure ‘‘Do discrete-trial teaching to the best of your
Instructions baseline. During instructions ability.’’ During this phase, the experimenter
baseline, the experimenter gave the parent did not conduct any training. Each session
a typed list of definitions of the 10 components consisted of 10 trials and lasted approximately
of discrete-trial teaching. At the beginning of 5 min. Sessions were videotaped and scored
each session, the experimenter stated, ‘‘Do later.
discrete-trial teaching to the best of your
ability.’’ Each session consisted of 10 trials Interobserver Agreement
and lasted approximately 5 min. Sessions were Interobserver agreement was measured dur-
videotaped and scored later. ing 35% of sessions, randomly distributed
Training. The experimenter first provided across baseline and intervention phases and all
the same typed copy of 10 components of three parent–child dyads. Agreement was cal-
discrete-trial teaching and described each com- culated by dividing the total number of
ponent. Next, the experimenter gave the parent agreements by the total number of agreements
a copy of a graph of her instructions baseline plus disagreements and multiplying by 100%.
performance and a copy of the previous Agreement on correct use of discrete-trial
session’s data sheet. The experimenter stated teaching was 94% (range, 80% to 100%),
her average baseline score, described her 95% (range, 80% to 100%), and 93% (range,
performance during the last session, and 80% to 100%) for Rosa, Maria, and Joanne,
discussed her previous performance. The exper- respectively. Agreement was 100% for all
imenter asked the parent for any questions and correct children’s responses.
answered them. Next, the experimenter sat with
the child and modeled three discrete trials. The
RESULTS AND DISCUSSION
parent then sat with her child and performed
three discrete trials. The experimenter provided Figure 1 presents the data on parent and
the parent with descriptive spoken feedback child correct responding during both GMI and
immediately following the performance, in- VIM sessions. During the instructions baseline,
cluding positive comments on target compo- parents emitted few correct discrete-trial teach-
688 MICHAEL LAFASAKIS and PETER STURMEY

Figure 1. The percentage of correct implementation and generalization of discrete-trial teaching for parents and
correct gross-motor and vocal imitative responses for children during instructions baseline and posttraining.

ing responses. After training, parents’ use of portion of correct responses by 64%, 67%, and
discrete-trial teaching improved greatly. Simi- 57%, respectively. Similarly, during VIM,
larly, children emitted few and sometimes close George, Emmanuel, and Christian increased
to zero correct responses during instructions their proportion of correct responses by 45%,
baseline, but all 3 children emitted many more 58%, and 55%, respectively.
correct responses during both GMI and VIM. This study extended the results of Sarokoff
During GMI, Rosa, Maria, and Joanne and Sturmey (2004), Dib and Sturmey (2007),
increased their use of correct discrete-trial and Hardy and Sturmey (1994) by demonstrat-
teaching responses by 31%, 43%, and 40%, ing that behavioral skills training was highly
respectively. During VIM, Rosa, Maria, and effective and efficient in teaching discrete-trial
Joanne increased their use of correct discrete- teaching skills to parents. This study also
trial teaching responses by 31%, 43%, and demonstrated that improving correct imple-
33%, respectively. During GMI, George, mentations of discrete-trial teaching may result
Emmanuel, and Christian increased their pro- in generalization of correct parent teaching to
 GENERALIZATION OF PARENT TEACHING 689

untrained teaching programs and that their Hardy, N., & Sturmey, P. (1994). Portage guide to early
education: III. A rapid training and feedback system
children emitted more correct responses after to teach and maintain mother’s skills. Educational
their parents learned to teach effectively. These Psychology, 14, 345–358.
data show that this treatment package was Koegel, R. L., Russo, D. C., & Rincover, A. (1977).
Assessing and training teachers in the generalized use
effective, but this study did not identify which of behavior modification with autistic children.
component was responsible for change. Future Journal of Applied Behavior Analysis, 10, 197–205.
research should conduct a component analysis Sarokoff, R. A., & Sturmey, P. (2004). The effects of
behavioral skills training on staff implementation of
of the behavioral skills training package. discrete-trial teaching. Journal of Applied Behavior
Analysis, 37, 535–538.
Sarokoff, R., & Sturmey, P. (in press). The effects of
instructions, rehearsal, modeling and feedback on
REFERENCES acquisition and generalization of staff use of discrete-
trial teaching and child correct responses. Research in
Baer, D. M., Peterson, R. F., & Sherman, J. A. (1967). Autism Spectrum Disorders.
The development of imitation by reinforcing behav- Sturmey, P., & Fitzer, A. (Eds.). (2007). Autism spectrum
ioral similarity to a model. Journal of Experimental disorders: Applied behavior analysis, evidence and
Analysis of Behavior, 10, 405–416. practice. Austin, TX: Pro-Ed.
Dib, N., & Sturmey, P. (2007). Reducing student
stereotypy by improving teachers’ implementation of Received December 5, 2006
discrete-trial teaching. Journal of Applied Behavior Final acceptance May 14, 2007
Analysis, 40, 339–343. Action Editor, Mark Dixon