Teaching caregivers to implement video modeling imitation training

via iPad for their children with autism

Teresa A. Cardon*
Washington State University, United States
Autism is one of the fastest growing disabilities, with an estimated one in 88 children being diagnosed with an
Autism Spectrum Disorder (ASD; Center for Disease Control, 2012). While it is cautioned that these estimates apply to
the 14 sites analyzed in the aforementioned study, with this ever-increasing population, innovative new approaches are
needed to ensure optimal intervention strategies are being implemented for children with ASD (Lord & McGee, 2001). In
addition, it is important that new approaches be cost effective strategies that caregivers can implement in the home
environment to engage their child (Dawson et al., 2010; Koegel & Koegel, 2006). An innovative intervention strategy
with potential for caregiver implementation, specically Video Modeling Imitation Training (VMIT) via an iPad, was the
focus of this study.
1. Imitation and autism
For over 40 years, researchers have explored how to support skill development in children with ASD (Dawson & Adams,
1984; Lovaas, 1987; Lovaas, Freitas, Nelson, & Whalen, 1967; Rogers, Bennetto, McEvoy, & Pennington, 1996). One skill that
has garnered interest is imitation. Severity of autism is correlated with impaired imitation skills (Rogers, Hepburn,
Stackhouse, & Wehner, 2003). As such, children with autism fail to imitate from an early age and this lack of imitation is a
Research in Autism Spectrum Disorders 6 (2012) 13891400
A R T I C L E I N F O
Article history:
Received 5 April 2012
Received in revised form 11 June 2012
Accepted 12 June 2012
Keywords:
iPad
Caregiver training
Imitation
Video modeling
Autism
A B S T R A C T
Children with autismfail to imitate froman early age and this lack of imitation is a salient
diagnostic marker for the disorder. For children with Autism Spectrum Disorder (ASD),
increased imitation skills appear to be related to increased skill development in a variety of
areas. Video modeling was recently validated as a technique to support imitation
acquisition in young children with autism. The purpose of this research was to determine if
there is a functional relation between caregiver implemented Video Modeling Imitation
Training (VMIT) via iPad and increased imitation skills in young children with autism. In
addition, a secondary analysis of language development after exposure to VMIT was also
conducted. A multiple baseline design across four caregivers and their children with
autism was implemented. Results indicated that all four caregivers were able to
successfully create video models on an iPad when provided with minimal training and
implement VMIT with delity for their children. All four children made substantial gains in
their imitation skills during caregiver implemented treatment. Imitation skills maintained
post treatment and, to varying degrees, generalized to imitation of live models. Expressive
language skills increased to varying degrees for all participants.
2012 Elsevier Ltd. All rights reserved.
* Tel.: +1 509 358 7590; fax: +1 509 358 7600.
E-mail address: Teresa.cardon@wsu.edu.
Contents lists available at SciVerse ScienceDirect
Research in Autism Spectrum Disorders
Jo ur n al h o mepag e: ht t p : / / ees. el sevi er . co m/ RASD/ def aul t . asp
1750-9467/$ see front matter 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.rasd.2012.06.002
salient diagnostic marker for the disorder (Lord et al., 2000). For children with ASD, better imitation skills appear to be
related to improved language performance (Stone, Ousley, & Littleford, 1997), play skills (Libby, Powell, Messer, & Jordan,
1997), and social skills (Carpenter, Pennington, & Rogers, 2002; Ingersoll, 2011).
Until recently, imitation skills were taught in discrete, analog settings in adult-led exchanges (Cardon & Wilcox, 2011;
Ingersoll & Schreibman, 2006; Lovaas et al., 1967). Although some skill acquisition occurred in these settings, generalization
of skills was extremely limited (Dawson & Adams, 1984). Recently, researchers have found that for children with ASD,
imitation acquisition and generalization improves in naturalistic settings during child-motivated interactions (Cardon &
Wilcox, 2011; Charlop-Christy, Le, & Freeman, 2000; Ingersoll, Lewis, & Kroman, 2006; Ingersoll & Schreibman, 2006).
2. Video modeling and autism
Although video modeling (VM) has been described in the literature for over 50 years, it is only over the past decade that
VM has been utilized with children on the autism spectrum (Ayres & Langone, 2005; Bellini & Akullian, 2007; Buggey,
Toombs, Gardener, & Cervetti, 1999; Charlop-Christy et al., 2000). A seminal study designed to teach a variety of skills to
children with autism, compared the effectiveness of VM to live modeling (Charlop-Christy et al., 2000). Participants included
ve children with ASD with a chronological age range of 711 years old. Children with varying functioning levels (e.g.,
different mental ages, language ages, play skills) were purposefully selected to determine if VM would be effective in
supporting skill development. Results indicated that children in the VM condition acquired skills faster. Children also
demonstrated generalization of target behaviors after VM, but did not generalize target behaviors after live modeling
(Charlop-Christy et al., 2000). The researchers concluded that VM is an effective technique that can support the development
of a variety of behaviors such as play skills, expressive language, and self-help skills in children with ASD.
Ongoing research has shown video modeling to be an effective intervention tool for teaching preschool and school age
children with ASD a variety of behaviors, including play skills, social skills, and self-help skills (e.g., Ayres & Langone, 2005;
Bellini & Akullian, 2007; Carpenter, Charlop, Dennis, & Greenberg, 2010; DAteno, Mangiapanello, & Taylor, 2003; Nikopoulos
& Keenan, 2003). Both single and multi-step tasks have been successfully taught using video modeling (Tereshko,
MacDonald, & Ahearn, 2010). Previous research has also indicated that personalized video models are more effective than
commercially distributed video models (Palechka & MacDonald, 2010; Rosenberg, Schwartz, & Davis, 2010). The persons
used as models to present the actions in the videos has varied, with adults, children and siblings all being able to support
positive outcomes (DAteno et al., 2003; Reagon, Higbee, & Endicott, 2006). It has been proposed that VM is an effective
method because it capitalizes on characteristics associated with ASD (e.g., over-selectivity, social decits, preference for
visual stimuli; Corbett & Abdullah, 2005).
Several mediums have been successfully used to present the video model including television, computers, and portable
DVD players. Video modeling has been validated as a technique to facilitate the four key components (i.e., attention,
retention, production, and motivation; Bandura, 1977) required for observational learning to occur (Dowrick & Associates,
1991). The monitors (e.g., television, iPad screen, Portable DVD player) offer a restricted eld of vision and can therefore
direct childrens attention to relevant stimuli while decreasing their tendency to attend to irrelevant stimuli (e.g., Charlop-
Christy et al., 2000; Corbett, 2003). Retention is also supported via VM because of the consistent repetition of the modeled
behavior. In addition, children are given the opportunity to practice, or produce, the behavior they saw occurring in the
video. Finally, with regard to motivation, television has been found to be particularly motivating for children with autism
(e.g., Charlop-Christy et al., 2000; Corbett, 2003; Nally, Houlton, & Ralph, 2000; Nikopoulos & Keenan, 2003; Shane & Albert,
2008).
While video modeling has been used with children with autism for over a decade (Ayres & Langone, 2005; Bellini &
Akullian, 2007; Charlop-Christy et al., 2000), the use of personal computers (e.g., iPod touch, iPads, tablet computers, smart
phones) to deliver the video model has only recently been studied (Cardon, 2012; Cihak, Fahenkrog, Ayres, & Smith, 2010).
Studies supporting the benets of personal computers as a medium to deliver video modeling protocols are still emerging;
thus far, however, indicators suggest that iPod touches, smart phones, and iPads are viable alternatives to televisions, laptop
computers, and portable DVD players (Cihak et al., 2010). In addition, use of tablet computers, such as the iPad, to support
intervention has dramatically increased among children with ASD (Dunham, 2011; Sennet & Bowker, 2009). While there has
been reported success anecdotally with iPads, empirical evidence supporting the systematic use of iPads in treatment is
lacking.
3. Video Modeling Imitation Training
Video Modeling Imitation Training is a new imitation protocol designed to teach young children with autism to imitate
using iPads. Conceptually based on video modeling, VMIT supplements video modeling by including specic prompt and
praise procedures, similar to those used in a clinical setting. Developed from a video modeling protocol successfully used to
teach object imitation (Cardon & Wilcox, 2011), VMIT was recently analyzed as a tool to teach young children with autism to
imitate gestures. Results were promising with two out of three participants making substantial gains in imitation, as well as
receptive and expressive language skills (Cardon, 2012).
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1390
4. Purpose of the current research
While there is evidence that live imitation training can be taught to caregivers to support imitation development in
children with ASD (Ingersoll & Gergans, 2007), currently it is unknown if caregivers can be taught to use VMIT to teach
imitation. Caregiver implemented intervention is increasingly necessary as the number of autism diagnoses continues to rise
and intervention resources are limited. In addition, with the increasing popularity of tablet computers it is important that
research determine how and whether these devices can be used systematically during intervention. The purpose of this
research was to determine if there is a functional relation between caregiver implemented VMIT via iPad and increased
imitation skills in young children with autism. In addition, secondary analyses of language development after exposure to
VMIT, was also explored.
5. Method
5.1. Participants
Participants included two boys, ages 24 and 50 months, and two girls ages 26 and 42 months with an autism spectrum
diagnosis and their caregivers. Detailed participant characteristics for children and caregivers are shown in Tables 1 and 2
respectively. Participants were recruited from local autism agencies, school districts, support groups and doctors ofces. All
children met the following inclusionary criteria, a) diagnosis of autism from a developmental pediatrician, psychologist, or
psychiatrist and accompanying written report, b) conrmation of diagnosis by the principal investigator (PI) using the
Autism Diagnostic Observation Schedules (ADOS; Lord, Rutter, DiLavore, & Risi, 2001) and the Childhood Autism Rating Scale
(CARS; Schopler, Reichler, & Renner, 2002), c) reported television/movie watching for at least one hour per day, and d) no
participation in any outside intervention that specically taught imitation skills while enrolled in the study.
5.2. Setting and materials
All pre-assessments (i.e., ADOS, Vineland, Motor Imitation Scale, Preschool Language Scale-5) were conducted in a
university autism laboratory. The room contained two child-size tables, three metal cabinets, two le cabinets, a counter top
with wall cabinets, two adult size desks with computers, and a sink. Pre-assessments took place at one child size table that
had been partitioned off from the rest of the lab with cabinets, to minimize distraction and create a more child friendly
atmosphere. Baseline, treatment, post-assessment, and follow-up sessions took place in the participants homes. To protect
participant identity, pseudonyms are used in the following sections. Two participants, Mallory and Tessa, sat at a dining
room table during treatment, Nathan completed the tasks in his bedroom, and Joshua sat on the oor in his living room.
Three of the participants had siblings present for part or all of the in home sessions. The fourth participant was an only child
and only her caregivers were present during her in home sessions.
Materials used during baseline and VMIT sessions supported activities and routines that the caregivers selected for their
children (see Table 3 for a complete list of activities and routines). The others-as-models in the video clips included the
caregivers or siblings (see Table 3) and were created using second generation iPads and the standard video camera software
that is available on an iPad. The video self- model was created with the iPad video camera and manipulated using iMovie on
the iPad. Video clips averaged 13.13 s with a range of 630 s. A Sony HDR-CX550V HD digital camcorder was utilized by
research assistants to record the caregivers creation of the video models for later analysis. The video models created by the
caregivers were uploaded by the PI to a MacBook Pro and saved in case a technical failure of the video clips occurred on any of
the iPads.
5.3. Procedures
5.3.1. Caregiver training
Caregivers attended a 2-h training session. During the training, caregivers were instructed on how to create effective
video models. They received a training manual with instructions and picture descriptions of how to create their own video
Table 1
Child characteristics.
Child Chronological
age
Gender Adaptive Behavior
Composite (Vineland)
(SS)
Preschool Language
Scale Aud. Comp. (SS)
Preschool Language
Scale Expressive (SS)
Autism severity
(CARS)
Mallory 3.8 F 92 97 82 31
Nathan 4.2 M 92 92 86 32.5
Joshua 2.0 M 79 50 77 34.5
Tessa 2.3 F 70 50 66 42
Note: All names are pseudonyms. CARS: Childhood Autism Rating Scale; SS: Standard score. On the CARS, scores of 3036.5 indicate mild-moderate autism,
and scores above 36.5 indicate severe autism.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1391
models using an iPad. Caregivers reviewed examples of effective (e.g., limited distractions, clear target behavior, linguistic
mapping) and ineffective video models (e.g., poor lighting, distracting clothing, obstructed views) targeting a variety of
routines and activities. During the 2-h training, caregivers were also shown video examples and given instructions on how to
implement VMIT with their children. Written instructions for VMIT implementation were included in the manuals that were
provided to the caregivers (VMIT manuals are available upon request).
5.3.2. Caregiver measures
Caregiver creation of the video models was analyzed for delity using the Video Model Creation: Procedural Checklist
(Fig. 1). In addition, caregiver implementation of VMIT was analyzed for delity using the Video Modeling Imitation Training
and Caregivers: Procedural Reliability (Fig. 2).
5.3.3. Child measures
The ADOS and the CARS were administered by the PI, a certied SLP with over 19 years of experience working with
children with autism. The PI received training and met reliability requirements to administer formal autism assessments and
has diagnostic expertise with a variety of tools. Research assistants were trained to administer the outcome measures (i.e.,
Motor Imitation Scale and the Preschool Language Scale-5) and to conduct the interview portion of the Vineland Scales of
Adaptive Behavior (Sparrow, Balla, & Cicchetti, 2005).
5.3.3.1. Vineland Scales of Adaptive Behavior, Second Edition. The Vineland (Sparrow et al., 2005) is a standardized parent
interview that assesses a childs social and personal everyday living skills. The Vineland was used to assess four domains:
social, communication, daily living, and motor skills. The Vineland is recognized as an appropriate assessment tool for
individuals with autism (Perry & Factor, 1989).
5.3.3.2. Childhood Autism Rating Scale. The CARS (Schopler et al., 2002) is a diagnostic assessment tool used to identify
children who have autism and determine the severity of the diagnosis. The CARS was administered by the PI to conrm the
Table 2
Caregiver characteristics.
Caregiver Age range Gender # of caregivers
in the home
Highest level of
education completed
Annual range
of income
Mallorys mom 4049 F 2 Bachelors $80,001$100,000
Nathans mom 2939 F 2 High School Diploma $60,001$80,000
Joshuas mom 1828 F 2 Bachelors $20,001$40,000
Tessa s mom 1828 F 2 N/P N/P
Note: All names are pseudonyms. N/P: not provided.
Table 3
Caregiver selected actions for Video Modeling Imitation Training.
Participant Actions
Original actions New actions
Mallory 1. wave hi 1. state hi + name
2. pat baby doll 2. pat doll & feed doll bottle
3. wipe her face 3. More _____ please
4. hands up for where gesture 4. No thank you
5. clean up puzzle 5. clean up before nished
Nathan 1. pencil grip
2. scissor grasp
3. make the bed
4. respond to name
5. clean up
Joshua 1. touch his nose 1. blow kiss
2. hand cup to caregiver 2. hand bib to caregiver
3. allow mom to brush his teeth 3. hold moms hand while walking
4. keep a hat on his head 4. identify dog toy from eld of two
5. walk doll up stairs put in bed 5. clean up toys
Tessa 1. doll onto bed
2. sign for book
3. wave bye-bye
4. reach for juice
5. push toy car
Note: All actions were randomly presented to account for possible order effects. Joshuas video models included his ve-year-old sibling, and Nathans clean
up video included his twin sister. All other video models were modeled by the childs caregiver.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1392
reported autism diagnosis and determine the level of autism severity. Concurrent validity between the CARS and the DSM-IV
was shown to be high and the CARS has been deemed a useful instrument in diagnosing and designating autism severity
(Rellini, Tortolani, Trillo, Carbone, & Montecchi, 2004).
5.3.3.3. Autism Diagnostic Observation Schedule. A second measure used to conrm the reported autism diagnosis was the
ADOS (Lord et al., 2001). The ADOS is a semi-structured assessment of communication, socialization, play, and atypical or
repetitive behaviors. The PI met standard requirements for research reliability in administration and scoring of the ADOS.
The following measures were obtained pre- and post-treatment to assess changes associated with treatment.
5.3.3.4. Motor Imitation Scale. The MIS (Stone et al., 1997) is a structured imitation assessment consisting of 16 motor
imitation tasks: eight object imitation and eight body imitation tasks. During the assessment, tasks are presented in a playful
manner and may be modeled up to three times. The assessment instructions state that the examiner is to state, your turn,
after each action is modeled. For the purposes of this research, the your turn direction was omitted so that spontaneous
imitative acts could be assessed. Participants can earn a score of two if they imitate the action immediately, a score of one if
imitation of the action is emerging, and a score of zero if they fail to imitate the action, for a total possible score of 32. Internal
consistency for the MIS revealed sufcient scores (a= .88) with a two week, test-retest reliability score of .80 (Stone et al.,
1997). The MIS was administered to the participants as a generalization probe pre- and post-treatment to assess gains in
imitative abilities.
5.3.3.5. Preschool Language Scale-5th Edition. The PLS-5 (Zimmerman, Steiner, & Pond, 2011) is a diagnostic tool that
evaluates both receptive and expressive language in children up to six years of age. The PLS-5 allows for observation of skills
and/or parent report, and includes materials for dynamic assessment. Raw scores were used to characterize change over
time.
5.3.4. Baseline
Baseline sessions occurred three times per week, with Mallory (P1) attending ve sessions, Nathan (P2) attending six
sessions, Joshua (P3) attending seven sessions, and Tessa (P4) attending eight sessions. During baseline, all one-step
Client Code:_____________________ Clinician:_____________________
Date: _____/_____/_____ Session#:_________
Total Session Time: ________
Instrucons: Remind the parcipant to pick a single acvity that has a clear beginning, disnct movements, and a clear
ending point. They may use the VMIT manual and watch the video guide as many mes as they need to before making
their own video model using the iPad. Document the accuracy of the steps below for each recorded acon.
Imitaon Acon:
Start Time:__________ End Time:_______
Steps
Accuracy
(+ or )
Comments
1. Turn on the iPad
2. Slide bar to unlock
3. Press the camera icon
4. Slide to video mode (if camera is
set to faceme mode switch)
5. hold iPad steady and aim
a. Limited visual distracons
b. Limited audio distracons
c. Plenty of light
d. Neutral background
6. Press the record buon
a. Single key word/ phrase said
each me
b. Target acon is clear
c. Appropriate rate
7. Record the enre acon
8. Stop Recording
Fig. 1. Video Model Creation: Procedural Checklist.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1393
actions (Fig. 3) were presented live by the participants caregiver. After the action was modeled, the caregiver paused
for 10 s to give the child time to imitate the action. All ve of the selected actions were modeled during the
2030 min baselines sessions. Imitation was dened as the child copying the target action within 10 s with an action
that looked distinctly like the action being modeled. The child was expected to imitate the target action before they
performed a different action and before the adult modeled another action (Cardon & Wilcox, 2011; Ingersoll &
Schreibman, 2006).
5.3.5. Video Modeling Imitation Training
Treatment sessions occurred three times per week for a total of 12 sessions. Each session lasted no longer than 40 min.
During the rst treatment session, the caregiver received live coaching and feedback for implementation of VMIT. The
caregiver showed the child pre-recorded clips of the one-step actions on the iPad. One video clip included a video self-model,
as Nathans parents wanted one of his actions to be responding to his name. According to the VMIT protocol, the video clips
were presented and then paused while the caregiver stated a turn cue relevant to the action (e.g., Lets play, Mommy
brush, Lets clean up) and the child was given 10 s to imitate the action they had just viewed. If the child imitated the
action they saw on the clip, they were verbally praised and the next clip was shown. If the child did not imitate the action,
the clip was shown again and then paused. After the third demonstration of the clip in which the child did not imitate, the
caregiver physically prompted the child to perform the action and provided verbal praise before moving on to the next
action. The only action that was not physically prompted was the response-to-name task for Nathan. If Nathan did not
respond after three demonstrations of the video self-model, the caregiver moved on to the next clip. All ve actions were
presented twice during each session for a total of 10 actions to imitate each session. The presentation of clips was
randomized across sessions to account for possible order effects.
Criterion was met when imitation of the actions occurred 80% of the time across four consecutive sessions. Both
Mallory and Joshua met criterion and ve new actions were introduced. To increase the level of difculty, both one-step
and two-step actions were introduced. Mallorys caregivers requested that her second set of videos also include verbal
imitation targets, such as stating No thank you and responding to an introduction using the word Hi paired with a
persons name.
Client Code:_____________________ Clinician:_____________________
Date: _____/_____/_____Session#:_________ Start Time:__________ End Time:_______
Total Session Time: ________ Reliability/Observer____________
Denion of Imitaon: Imitaon has occurred when the child copies the target acon within 10 seconds with
an acon that looks disnctly like the acon being modeled. The child must imitate the target acon before
they perform a dierent acon and before the adult models another acon.
Direcons: Place the iPad in front of the child for all tasks. Please record a ( ) if the behavior/task occurs.
Record a (+) or () for the childs responses.
C = Caregiver, I = Incorrect
Trial
Smulus C plays
video
clip
C gives
turn cue
C waits
10 sec.
+ or
V
e
r
b
a
l
P
r
a
i
s
e
I = Physical
prompt
gesture
V
e
r
b
a
l
P
r
a
i
s
e
1
2
3
4
5
6
7
8
9
10
Reliability %
Praise = Good Listening, Nice Sing, Goodjob, Way to go, Nice playing, Good playing, Nice
work, You did it, Yah!
Fig. 2. Video Modeling Imitation Training and Caregivers: Procedural Reliability.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1394
5.3.6. Follow-up sessions
Follow-up sessions occurred in participants homes one and three weeks post-treatment. Follow-up sessions assessed
both maintenance and generalization of skills. To assess generalization of imitation skills, mastery probes conducted under
baseline conditions with live actions were implemented. The caregiver presented a live model of the action and gave the
child 10 s to imitate the action. If the child did not generalize and imitate live models of the actions with greater than 80%
accuracy, the iPad was reintroduced and the video model of the action was presented to assess maintenance of the skill.
5.3.7. Fidelity and reliability
To determine delity of implementation, the Video Model Creation: Procedural Checklist (Fig. 1) was analyzed by six
trained research assistants to determine if caregivers were able to accurately choose target actions and create the video
models to be used during treatment after attending one 2-h training session. Point by point comparisons across six trained
observers were made for 56% of the caregiver created video models (n = 50). Fidelity of implementation >80% was met by all
four caregivers in every category (Table 4).
To ensure the caregivers delity of VMIT, point by point comparisons were made across participants. Of the 26 baseline
sessions, 42% (n = 11) of them were scored by four trained research assistants in their entirety from video tape. Reliability
0%
20%
40%
60%
80%
100%
9 8 7 6 5 4 3 2 1 10 11 12 13 14 15 16 17 18 19
0%
20%
40%
60%
80%
100%
9 8 7 6 5 4 3 2 1 10 11 12 13 14 15 16 17 18
0%
20%
40%
60%
80%
100%
9 8 7 6 5 4 3 2 1 10 11 12 13 14 15 16 17 18 19 20 21
0%
20%
40%
60%
80%
100%
8 7 6 5 4 3 2 1 10 9 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Baseline
Treatment Follow Up
iPad acons
Mallory
Nathan
Joshua
Tessa
Criterion met,
new acons
on iPad
%

o
f

A
c

o
n
s

I
m
i
t
a
t
e
d
# of Sessions
Live acons
Fig. 3. Multiple baseline graph of child imitated actions after caregiver implemented VMIT.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1395
standards greater than 80% were easily met during baseline, with actions being modeled correctly by caregivers 100% of the
time, caregivers waiting 10 s 98%, and imitation rated as correct/incorrect 96% of the time.
During treatment, 58% of the sessions (n = 28) were scored in their entirety from video tape using the Video Modeling
Imitation Training and Caregivers: Procedural Reliability Checklist (Fig. 2). The following percentages for caregiver treatment
delity were obtained: presentation of the clip 98.8%, providing a turn cue 98.4%, waiting 10 s 97.5%, verbal praise for correct
imitation 95.7%, physical prompts 99.2%, and verbal praise after physical prompt 98.2%. To calculate scoring reliability for
imitation, the denition remained consistent across baseline, treatment and follow-up. Four trained research assistants
scored 58% of the sessions from video tape. Reliability for imitation was 95% during treatment.
5.4. Experimental design
The rst objective was to determine whether caregivers could be trained to create effective video models to teach their
children imitation skills using VMIT (Cardon & Wilcox, 2011). Caregivers attended one, 2 h training session, and received a
training manual and video guide with step-by-step instructions recorded on the iPad. The training manual and video guide
provided examples of actions/gestures that could be targeted during VMIT. Caregivers then identied ve specic actions/
gestures (e.g., teeth brushing) or play schemes (e.g., pushing a car) that they wanted to teach their child to imitate (Table 3).
Caregivers recorded the ve target actions/gestures that they identied as being important for their children on iPads that
were provided by the PI. Caregivers provided linguistic mapping during each video model (e.g., wipe my face; up, up, up;
hat on). Caregivers agreed that video models on the iPad would be shown to their child exclusively when the PI or a
graduate research assistant was present.
A multiple baseline design across the four participants was conducted. Participants attended baseline sessions until
stable trends were established. Mallory attended ve baseline sessions, Nathan six baseline sessions, Joshua seven baseline
sessions, and Tessa eight baseline sessions. At the completion of her baseline sessions, Mallory entered the treatment phase
for four weeks. When a positive trend was identied for Mallory, Nathan entered the treatment phase for four weeks and so
on for Joshua and Tessa. Treatment sessions occurred three times a week for up to 40 min/session depending on the number
of presentations required per session. Maintenance and generalization probes were conducted at one and three weeks post
treatment for all participants. The Motor Imitation Scale (Stone et al., 1997) and the Preschool Language Scale-5
(Zimmerman et al., 2011) were obtained pre- and post-treatment to assess changes over time associated with treatment.
6. Results
Caregivers used iPads to create video models and implement VMIT with their children on the autism spectrum. Analysis
of caregiver created video models indicated that all caregivers created personalized video models on the iPads with greater
than 90% delity across all categories (Table 4). In addition, caregivers implemented VMIT with high delity across sessions
(95.799.2%).
6.1. Visual analysis of multiple baseline design
A visual analysis was conducted to determine if there was a functional relation between caregiver implemented VMIT and
increased child imitation skills. When analyzing single case designs, such as the multiple baseline design used in this study,
the preferred method of analysis is a visual inspection of the data as this method does not require specic assumptions to be
met and participants act as their own control (Gast, 2010; Kratochwill et al., 2010; Kromrey & Foster-Johnson, 1996). One
aspect of visual analysis is identifying data levels and trends across and between phases of each participant.
All four participants exhibited no imitation (0%) to minimal levels (average 20%) of imitation during baseline, and stable
trends were established before staggered treatment sessions were implemented across all four participants (Fig. 3). Percent
Table 4
Caregiver delity of video model creation via iPad.
Category Percent correct
1. Hold iPad steady and aim at target 98%
2. Limit visual distractions 98%
3. Limit auditory distractions 96%
4. Plenty of light 100%
5. Neutral background 98%
6. Press record button 98%
7. Single work/key phrase stated 92%
8. Target action clear 96%
9. Appropriate rate 100%
10. Record entire action 100%
11. Stop recording 100%
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1396
imitated was calculated as the total number of actions imitated divided by the total number of actions presented (# of
actions imitated/total # of actions). All four participants demonstrated an increased level of imitation once treatment was
implemented (Fig. 3) and maintained higher than baseline levels of imitation during treatment. Three of the four participants
demonstrated immediate increases by the second treatment session. After immediate increases early on, data for three
participants remained stable, with two participants meeting a priori criteria (>80% imitation over four consecutive sessions)
for the introduction of new target actions (Table 3). New target actions for Mallory and Joshua consisted of expanded targets
(i.e., one step to two step actions, action with verbal imitation) or generalization of target actions (e.g., give me cup to give
me bib).
Nathan did not meet a priori criteria for the introduction of new actions because he exhibited problem behaviors during
the sessions with his caregiver (e.g., refusals to participate, demands for specic video clips, tantrums). A token economy
system, a sticker received for each video watched, was implemented during the last two treatment sessions to address
Nathans problem behaviors. With the external sticker reward system, Nathans problem behaviors decreased and his video
watching and imitation increased.
Tessa gradually increased her imitation skills after the onset of VMIT; however, during ongoing visual analysis of the data
it was noted that her progress was limited. A decision was made to increase the number of sessions she received during the
last two weeks of treatment from three sessions per week to four sessions per week. With the increased number of sessions,
Tessa demonstrated an increased positive response to treatment.
6.2. Pre- and post-treatment assessments
All four participants made positive gains in expressive communication as measured by increases in raw scores on the PLS-
5 post treatment (see Fig. 4). Three out of four participants also demonstrated gains in auditory comprehension. All four
children made gains on the Motor Imitation Scale (Stone et al., 1997) post-treatment (see Fig. 5). Consistent with past
research, participants with lower scores on the CARS indicating less autism severity (i.e., Mallory and Nathan) demonstrated
larger gains on the Motor Imitation Scale, while participants with higher scores on the CARS (i.e., Joshua and Tessa)
demonstrated fewer imitation gains. While some improvement in language and imitation skills could be attributed to
maturation, there was a limited amount of time between pre- and post-assessment sessions (maximum of 8 weeks).
6.3. Maintenance and generalization
Mastery probes, a return to baseline conditions, were conducted at one and three weeks post-treatment in participants
homes to determine if imitation skills had maintained and/or generalized. All four participants demonstrated mastery of the
imitation skills to varying degrees at the initial one week follow-up session. Mallory imitated 80% of her target actions when
presented with a mastery probe. Nathan imitated 100%, Joshua 90%, and Tessa 40% of her target actions. Because Tessas
imitation of the mastery probes was below 80%, she was given an opportunity to view the actions she did not imitate on the
iPad to determine if the ability to imitate the video models had maintained. With the iPad, Tessa was able to imitate 40% of
the target actions.
At the second follow-up session, three weeks post-treatment, Mallory refused to imitate the mastery probe actions and, in
fact, stated, no and requested the iPad. With the iPad, her imitation skills of previously viewed target actions maintained at
100%. Nathan continued to generalize his skills and imitated 100% of his target actions in response to a mastery probe. Given
a mastery probe during his second follow-up session, Joshua imitated 70% of his target actions. Because his imitation during
the second follow-up was below 80%, the iPad was reintroduced and he demonstrated maintenance of imitation skills at 90%.
0
5
10
15
20
25
30
35
40
45
50
Treatment Pre
Post Treatment
Auditory Comprehension Expressive Communicaon
Fig. 4. Preschool Language Scale-5 raw scores pre- and post-treatment.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1397
Tessa did not imitate any mastery probes during her second follow-up; however, she did maintain the ability to imitate the
video models presented on the iPad at 40%.
7. Discussion
This study investigated caregivers abilities to create effective video models with iPads and examined the functional
relation between caregiver implemented VMIT and imitation gains in their young children with autism. Caregivers created
effective video models on an iPad with the training provided and use those video models to teach their children with ASD to
imitate a variety of actions/gestures. Treatment delity was high and caregivers proved to be effective implementers of the
intervention approach. In addition, all four children generalized imitation skills to mastery probes to varying degrees after
discontinuation of treatment. Expressive and receptive language skills also increased as reported on the PLS. While language
skills were not explicitly targeted, these ndings add support to existing research indicating a possible relationship between
imitation skills and language development (Cardon, 2012; Ingersoll & Schreibman, 2006; Stone et al., 1997).
Previous research has shown that caregivers can be successful implementers of autism intervention (Dawson et al., 2010;
Ingersoll & Gergans, 2007; Koegel & Koegel, 2006; Yoder & Warren, 1998) and results from this study support the use of
caregiver implemented VMIT for autism intervention. Further, while iPad use among children with autism has been
anecdotally reported, this study among the rst to provide empirical evidence to support the use of iPads as intervention
agents in the home environment. In addition, this research furthered support for VMIT as an effective tool to teach both
object and gestural imitation to children with ASD, and suggests that verbal imitation is a viable target behavior for VMIT.
Three of four participants in the current study showed immediate gains once their caregivers introduced the video model,
conrming that video modeling is an effective method for teaching a variety of skills (Cardon & Wilcox, 2011; Charlop-
Christy et al., 2000). This immediacy effect is a relative strength of video modeling in general and was conrmed with these
results. Participants were highly motivated to attend to the iPads, thereby increasing their attention to the learning
opportunity. The use of electronic mediums, such as the iPad, to focus childrens attention on relevant stimuli supports the
use of VMIT as an intervention tool (Cardon & Azuma, 2012; Corbett & Abdullah, 2005). Participants were able to learn ve
individual target behaviors concurrently. Participants abilities to attend to, retain, and produce actions viewed on an iPad
with caregiver support is an important component of this research.
The results of this work have important clinical implications with regard to the use of iPads in intervention settings. The
demonstrated effectiveness of iPads to deliver VMIT, and the relative ease with which the caregivers were taught to create
their own video models and implement VMIT, provides tentative evidence for clinicians considering iPad use with children
on the autism spectrum. Caregiver implemented programs (e.g., Pivotal Response Treatment, Early Start Denver Model,
Reciprocal Imitation Training) typically require extensive training by highly skilled professionals and are very expensive. In
contrast, VMIT required one, 2-h training session and a one hour coaching session for an estimated cost of $300 per child. The
need for a highly skilled professional was limited once the initial training was complete. Future research should examine the
extent to which VMIT could be implemented in rural settings via telemedicine with pre-recorded, training sessions and
online coaching.
While VMIT is not meant to replace intervention with trained clinicians, the present results suggest that it may enhance
existing interventions. For example, Nathans parents chose two ne motor tasks, pencil grip and scissor grasp, which he had
been struggling with in traditional occupational therapy. While Nathans occupational therapist was concerned, she agreed
to defer continued training of these tasks while Nathan completed the study. After VMIT, Nathan was able to produce both
actions when presented with a live model 100% of the time at follow-up and his occupational therapist was reportedly
thrilled with the results. Similarly, Joshua had been working on identifying body parts in speech therapy for over 8 months
with limited success. His mother wanted him to imitate touching his nose as one of his target behaviors. After three sessions,
he was able to imitate the action after watching the video of his sister touching her nose. In addition, teaching verbal
imitation through VMIT had not previously been addressed until Mallorys mother requested that it be included in her
0
5
10
15
20
25
30
35
Mallory Nathan Joshua Tessa
Pre Treatment
Post Treatment
Fig. 5. Motor Imitation Scale pre- and post-treatment.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1398
second set of target behaviors. While previous research with video modeling has demonstrated that adults and adolescents
can learn verbal targets through video modeling (Charlop & Milstein, 1989; Taylor, Levin, & Jasper, 1999), given Mallorys
success, teaching verbal imitation via VMIT to young children should be explored further in future studies. These results
strongly suggest that focused attention on relevant stimuli through VMIT can help some children with autism learn target
behaviors that had previously been resistant to training using traditional methods.
While the overall results of this study are promising, there are some limitations. The four participants in the current study
are appropriate for the single-case design methodology employed in this research; replication with a larger group in a
randomized control trial is warranted. Physical prompts were used as part of the VMIT protocol; however, this is a limitation
of the study as the increase in imitative behaviors may have been inuenced by the physical prompts. In addition, the
introduction of a token economy system to address Nathans problem behaviors during his last two sessions could be
considered a limitation because of its impact on Nathans increased response to VMIT. Tessas need for additional training is a
limitation of this study and it would be benecial for future research to determine the extent to which autism severity
inuences a childs response to video modeling in general and VMIT specically (Cardon & Wilcox, 2011; Rogers et al., 2003).
Future research might also ask if children with certain prerequisite skills (e.g., increased interest in play, increased verbal
language, etc.) respond more consistently to VMIT.
This study offers important contributions to the growing eld of autism intervention research. Results indicate that
caregivers of children with autism can be taught to utilize iPads to record effective video models. Caregivers require only
minimal training to implement VMIT effectively with their children in the home environment. Finally, iPads can be effective
tools to support imitation development in very young children with autism.
References
Ayres, K. M., & Langone, J. (2005). Intervention and instruction with video for students with autism: A review of the literature. Education and Training in
Developmental Disabilities, 40(2), 183196.
Bandura, A. (1977). Self-efcacy: Toward a unifying theory of behavioral change. Psychological Review, 84, 191215.
Bellini, S., & Akullian, J. (2007). A meta-analysis of video modeling and video self-modeling interventions for children and adolescents with autism spectrum
disorders. Exceptional Children, 73, 261284.
Buggey, T., Toombs, K., Gardener, P., & Cervetti, M. (1999). Training responding behaviors in students with autism: Using videotaped self-modeling. Journal of
Positive Behavior Interventions, 1, 205214.
Cardon, T. (2012). Video modeling imitation training to support gestural imitation acquisition in young children with ASD. Manuscript submitted for publication.
Cardon, T., & Azuma, T. (2012). Visual attending preferences in children with autism spectrum disorders: A comparison between live and video presentation
modes. Research in Autism Spectrum Disorders, 6, 10611067.
Cardon, T., &Wilcox, M. J. (2011). Promoting imitation in young children with autism: A comparison of reciprocal imitation training and video modeling. Journal of
Autism and Developmental Disorders, 41(5), 654677.
Carpenter, M. H., Charlop, M. H., Dennis, B., & Greenberg, A. (2010). Teaching socially expressive behaviors to children with autism through video modeling.
Education and Treatment of Children, 33(3), 371393.
Carpenter, M., Pennington, B. E., &Rogers, S. J. (2002). Interrelations among social-cognitive skills in young children with autism. Journal of Autism&Developmental
Disorders, 32, 91106.
Center for Disease Control. (2012, March 31). Autism Spectrum Disorders. Retrieved from http://www.cdc.gov/ncbddd/autism/data.html.
Charlop-Christy, M. H., Le, L., & Freeman, K. (2000). A comparison of video modeling with in vivo modeling for teaching children with autism. Journal of Autismand
Developmental Disorders, 30, 537552.
Charlop, M. H., & Milstein, J. P. (1989). Teaching autistic children conversational speech using video modeling. Journal of Applied Behavior Analysis, 22, 275285.
Cihak, D., Fahrenkrog, C., Ayres, K. M., & Smith, C. (2010). The use of video modeling via a video ipod and a system of least prompts to improve transitional
behaviors for students with autism spectrum disorders in the general education classroom. Journal of Positive Behavior Interventions, 12(2), 103115.
Corbett, B. A. (2003). Video modeling: A window into the world of autism. The Behavior Analyst Today, 4, 8896.
Corbett, B. A., & Abdullah, M. (2005). Video modeling: Why does it work for children with autism? Journal of Early and Intensive Behavior Intervention, 2(1), 28.
DAteno, P., Mangiapanello, K., & Taylor, B. A. (2003). Using video modeling to teach complex play sequences to a preschooler with autism. Journal of Positive
Behavior Interventions, 5, 511.
Dawson, G., & Adams, A. (1984). Imitation and social responsiveness in autistic children. Journal of Abnormal Child Psychology, 12, 209226.
Dawson, G., Rogers, S., Munson, J., Smith, M., Winter, J., Greenson, J., et al. (2010). Randomized controlled trial of an intervention for toddlers with autism: The
Early Start Denver Model. Pediatrics, 125, 1623.
Dowrick, P. W., & Associates, (1991). Practical guide to using video in behavioral sciences. New York: John Wiley & Sons Inc.
Dunham, G. (2011, April 05). The future at hand: Mobile devices and apps in clinical practice. The ASHA Leader.
Gast, D. L. (2010). Single subject research methodology in behavioral sciences. New York: Routledge.
Ingersoll, B. (2011). Brief report: Effect of a focused imitation intervention on social functioning in children with autism. Journal of Autism and Developmental
Disabilities http://dx.doi.org/10.1007/s10803-011-1423-6 (Published online).
Ingersoll, B., & Gergans, S. (2007). The effect of a parent-implemented imitation intervention on spontaneous imitation skills in young children with autism.
Research in Developmental Disabilities, 28, 163175.
Ingersoll, B., Lewis, E., & Kroman, E. (2006). Teaching the imitation and spontaneous use of descriptive gestures in young children with autismusing a naturalistic
behavioral intervention. Journal of Autism and Developmental Disorders, 37, 14461456.
Ingersoll, B., & Schreibman, L. (2006). Teaching reciprocal imitation skills to young children with autism using a naturalistic behavioral approach: Effects on
language, pretend play and joint attention. Journal of Autism and Developmental Disorders, 36, 487505.
Koegel, R., & Koegel, L. (2006). Pivotal response treatments for autism. (pp. 141159) Baltimore, MD: Paul H Brookes Publishing Co.
Kratochwill, T. R., Hitchcock, J., Horner, R. H., Levin, J. R., Odom, S. L., Rindskopf, D. M., et al. (2010). Single-case designs technical documentation. Retrieved from
What Works Clearinghouse website: http://ies.ed.gov/ncee/wwc/pdf/wwc_scd.pdf.
Kromrey, J. D., &Foster-Johnson, L. (1996). Determining the efcacy of intervention: The use of effect sizes for data analysis insingle subject research. The Journal of
Experimental Education, 65, 7393.
Libby, S., Powell, S., Messer, D., & Jordan, R. (1997). Imitation of pretend play acts by children with autism and Down syndrome. Journal of Autism and
Developmental Disorder, 27, 365383.
Lord, C., & McGee, J. P. (2001). Educating children with autism. Washington, DC: National Academy Press.
Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Jr., Leventhal, B., DiLavore, P. C., et al. (2000). The autismdiagnostic observation schedule generic: A standard measure
of social and communication decits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30, 205223.
Lord, C., Rutter, M., Di Lavore, P. C., & Risi, S. (2001). Autism Diagnostic Observation Schedule. Los Angeles, CA: Western Psychological Services.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1399
Lovaas, O. I. (1987). Behavioral treatment and normal educational and intellectual functioning in young autistic children. Journal of Consulting and Clinical
Psychology, 55, 39.
Lovaas, O. I., Freitas, K., Nelson, K., & Whalen, C. (1967). The establishment of imitation and its use for the development of complex behavior in schizophrenic
children. Behaviour Research and Therapy, 5, 171182.
Nally, B., Houlton, B., &Ralph, S. (2000). Researches in brief: The management of television and video by parents of children with autism. Autism, 4(3), 12361254.
Nikopoulos, C. K., & Keenan, M. (2003). Promoting social initiation in children with autism using video modeling. Behavioral Interventions, 18, 87108.
Palechka, G., & MacDonald, R. (2010). A comparison of the acquisition of play skills using instructor-created video models and commercially available videos.
Education and Treatment of Children, 33(3), 457474.
Perry, A., & Factor, D. C. (1989). Psychometric validity and clinical usefulness of the Vineland Adaptive Behavior Scales and the AAMD Adaptive Behavior Scale for
an autistic sample. Journal of Autism and Developmental Disorders, 19, 4155.
Reagon, K. A., Higbee, T. S., & Endicott, K. (2006). Teaching pretend play skills to a student with autism using video modeling with a sibling as model and play
partner. Education and Treatment of Children, 29(3), 517526.
Rellini, E., Tortolani, D., Trillo, S., Carbone, S., & Montecchi, F. (2004). Childhood AutismRating Scale (CARS) and AutismBehavior Checklist (ABC) correspondence
and conicts with DSMIV criteria in diagnosis of autism. Journal of Autism and Developmental Disorders, 34, 703708.
Rogers, S., Bennetto, L., McEvoy, R., & Pennington, B. (1996). Imitation and pantomime in high-functioning adolescents with autism spectrum disorders. Child
Development, 67, 20602073.
Rogers, S. J., Hepburn, S. L., Stackhouse, T., & Wehner, E. (2003). Imitation performance in toddlers with autism and those with other developmental disorders.
Journal of Child Psychology and Psychiatry, 44, 763781.
Rosenberg, N. E., Schwartz, I. S., & Davis, C. (2010). Evaluating the utility of commercial videotapes for teaching hand washing to children with autism. Education
and Treatment of Children, 33(3), 443455.
Schopler, E., Reichler, R. J., & Renner, B. R. (2002). The Childhood Autism Rating Scale. Los Angeles, CA: Western Psychological Services.
Sennet, S., & Bowker, A. (2009). Autism, AAC and Proloquo2Go. Perspectives on Augmentative and Alternative Communication, 18, 137145.
Shane, H. C., & Albert, P. D. (2008). Electronic screen media for persons with autismspectrumdisorders: Results of a survey. Journal of Autismand Developmental
Disorders, 38(8), 14991508.
Sparrow, S. S., Balla, D. A., &Cicchetti, D. V. (2005). Vineland Adaptive Behavior Scales InterviewEdition survey form. Circle Pines, MN: American Gui-Johnce Service.
Stone, W., Ousley, O., & Littleford, C. (1997). Motor imitation in young children with autism: Whats the object? Journal of Abnormal Child Psychology, 25, 475485.
Taylor, B. A., Levin, L., & Jasper, S. (1999). Increasing play-related statements in children with autismtoward their siblings: Effects of video modeling. Journal of
Developmental & Physical Disabilities, 11(3), 253264.
Tereshko, L., MacDonald, R., & Ahearn, W. (2010). Strategies for teaching children with autismto imitate response chains using video modeling. Research in Autism
Spectrum Disorders, 4, 479489.
Yoder, P. J., & Warren, S. F. (1998). Maternal responsivity predicts the prelinguistic communication intervention that facilitates generalized intentional
communication. Journal of Speech, Language and Hearing Research, 41, 12071219.
Zimmerman, I. L., Steiner, V. G., & Pond, R. E. (2011). Preschool Language Scale (5th ed.). San Antonio, TX: The Psychological Corporation.
T.A. Cardon / Research in Autism Spectrum Disorders 6 (2012) 13891400 1400