Content uploaded by Laura A. Guli
Author content
All content in this area was uploaded by Laura A. Guli on Jan 13, 2021
Content may be subject to copyright.
The
Arts
in
Psychotherapy
40 (2013) 37–
44
Contents
lists
available
at
SciVerse
ScienceDirect
The
Arts
in
Psychotherapy
Social
Competence
Intervention
Program
(SCIP):
A
pilot
study
of
a
creative
drama
program
for
youth
with
social
difficulties
Laura
A.
Guli,
PhDa,∗,
Margaret
Semrud-Clikeman,
PhDb,
Matthew
D.
Lerner,
MAc,
Noah
Britton,
MAd
aUniversity
of
Texas
at
Austin,
SZB
504,
1
University
Station,
D5800,
Austin,
TX
78712,
United
States
bUniversity
of
Minnesota
Medical
School,
420
Delaware
St.
SE,
MMC
486,
Minneapolis,
MN
55455,
United
States
cUniversity
of
Virginia,
Department
of
Psychology,
102
Gilmer
Hall,
PO
Box
400400,
Charlottesville,
VA
22904-4400,
United
States
dBunker
Hill
Community
College,
250
New
Rutherford
Avenue,
Boston,
MA
02129,
United
States
a
r
t
i
c
l
e
i
n
f
o
Keywords:
Social
skills
Autism
Asperger’s
Group
Treatment
Drama
a
b
s
t
r
a
c
t
This
study
explored
the
effects
of
participation
in
the
Social
Competence
Intervention
Program
(SCIP),
an
innovative
creative
drama-based
group
intervention,
of
children
diagnosed
with
autism
spectrum
disor-
der
(ASD),
nonverbal
learning
disability
(NLD)
and/or
attention
deficit
hyperactivity
disorder
(ADHD).
Eighteen
participants
in
SCIP
were
compared
to
a
clinical
control
group
of
16
on
changes
in
meas-
ures
of
social
perception,
social
competence,
and
naturalistic
observed
social
behavior.
Hierarchical
multiple
regression
model
was
used
for
all
primary
quantitative
analyses.
Interviews
were
conducted
post-treatment
to
provide
qualitative
data.
The
treatment
group
showed
significant
improvement
in
key
domains
of
observed
social
behavior
in
a
natural
setting
compared
to
the
clinical
control
group.
Parents
and
children
in
the
SCIP
condition
reported
multiple
positive
changes
in
social
functioning.
These
findings
provide
preliminary
support
for
the
use
of
a
creative
drama
program
for
children
with
social
competence
deficits
related
to
social
perception
problems.
© 2012 Elsevier Inc. All rights reserved.
Introduction
Deficits
in
social
competence,
or
the
ability
to
function
effec-
tively
in
interpersonal
situations
and
perform
competently
on
social
tasks,
are
a
defining
characteristic
of
youth
with
autism
spec-
trum
disorder
(ASD;
Koenig,
De
Los
Reyes,
Cicchetti,
Scahill,
&
Klin,
2009).
Social
competence
difficulties
have
been
documented
not
only
in
youth
with
ASD,
but
those
with
a
nonverbal
learning
dis-
ability
(NLD),
and
attention
deficit
hyperactivity
disorder
(ADHD),
as
well
(Little
&
Clark,
2006;
Woodbury-Smith
&
Volkmar,
2009).
A
key
element
necessary
for
social
competence
is
social
perception,
defined
as
the
ability
to
identify,
recognize,
and
interpret
the
mean-
ing
and
significance
of
the
behavior
of
others
(Lipton
&
Nowicki,
2009).
The
process
of
social
perception
can
be
broken
down
into
the
input
of
sensory
cues,
integration
of
these
cues
and
output
of
an
appropriate
behavioral
response
(Johnson
&
Myklebust,
1967).
Children
and
adolescents
with
ASD
and
NLD
have
difficulty
with
each
of
these
steps
(Rourke,
1995;
Semrud-Clikeman,
Walkowiak,
Wilkinson,
&
Minne,
2010;
Woodbury-Smith
&
Volkmar,
2009).
∗Corresponding
author
at:
3625
Manchaca
Rd.,
Suite
202,
Austin,
TX
78704,
United
States.
Tel.:
+1
512
522
4093;
fax:
+1
512
685
1514.
E-mail
addresses:
laura@drlauraguli.com,
mdl6e@virginia.edu
(L.A.
Guli),
semrudcl@msu.edu
(M.
Semrud-Clikeman),
mlerner@virginia.edu
(M.D.
Lerner),
noahbritton@gmail.com (N.
Britton).
Specific
deficits
have
been
found
in
these
populations’
ability
to
accurately
decode
facial
cues,
voice
tone,
and/or
prosody
(Deruelle,
Rondan,
Gepner,
&
Tardif,
2004).
Research
has
also
begun
to
identify
attention
issues
in
children
with
ASD
and
NLD
(Fine,
Semrud-
Clikeman,
Butcher,
&
Walkowiak,
2008)
as
well
as
social
perceptual
difficulties
in
children
with
ADHD
(Corbett
&
Constantine,
2006).
Emerging
evidence
is
present
in
the
literature
that
ADHD
and
ASD
may
share
not
only
common
behaviors
but
also
a
common
deficit
in
the
frontostriatal
pathways
as
a
basis
of
their
disorders.
Thus,
it
has
been
strongly
suggested
that
studies
include
children
with
ADHD
as
well
as
those
with
ASD
within
the
same
groups
for
intervention
(Corbett,
Constantine,
Hendren,
Rocke,
&
Ozonoff,
2009).
Empirical
studies
indicate
that
while
inattention
relates
to
social
functioning
difficulties
(Fine
et
al.,
2008),
a
more
fundamental
variable
under-
lying
these
social
skills
difficulties
is
deficits
in
social
perception
(Semrud-Clikeman
et
al.,
2010).
Social
skills
interventions
Though
many
social
skills
interventions
exist,
many
programs
have
demonstrated
inconsistent
efficacy
in
addressing
the
social
competence
needs
of
children
with
ASD,
NLD,
and/or
ADHD
(Matson,
Matson,
&
Rivit,
2007).
These
populations
may
bene-
fit
from
social
skills
programs
that
are
experiential
rather
than
didactic,
and
developed
for
their
specific
needs
(Davis
&
Broitman,
2011;
Koenig
et
al.,
2009;
Lerner
&
Levine,
2007).
It
has
been
0197-4556/$
–
see
front
matter ©
2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.aip.2012.09.002
38 L.A.
Guli
et
al.
/
The
Arts
in
Psychotherapy
40 (2013) 37–
44
recommended
that
interventions
for
ASD
and
NLD
focus
on
shar-
ing
relationships,
and
break
down
complex
social
behaviors
into
concrete
steps
(Kransny,
Williams,
Provencal,
&
Ozonoff,
2003).
In
recent
years,
various
interventions
targeting
ASDs
and
related
disorders
have
emerged
that
emphasize
social-cognitive,
relation-
ship
building
and
social
perception
skills.
While
promising
findings
using
these
skills
have
been
found,
they
often
used
very
small
sam-
ples,
failed
to
employ
controlled
group
designs,
and
have
yielded
often
inconsistent
results
(see
Lerner,
Hileman,
&
Britton,
in
press;
Matson
et
al.,
2007).
Drama
as
effective
intervention
An
increasing
number
of
scholars
are
realizing
that
drama
ther-
apies
may
be
well
suited
to
ASD
and
related
populations.
Several
programs
using
drama
therapies
are
currently
being
used
with
chil-
dren
with
ASD
and
have
gained
notoriety
as
treatments
for
social
difficulties,
including
programs
developed
by
Loretta
Gallo-Lopez
(www.playandcreativetherapy.com/services/actproject)
and
Lee
Chasen
(Chasen,
2011).
More
specifically,
some
have
considered
whether
drama
activities
(rather
than
entire
drama
therapy
pro-
grams),
may
have
unique
efficacy
for
remediating
social
challenges.
Drama
activities
are
interactive,
emphasize
relationships,
emo-
tions,
communication,
cooperation
and
imagination,
in-context
learning,
and
emphasize
the
give
and
take
of
interpersonal
non-
verbal
cues
(Spolin,
1986).
A
number
of
studies
have
already
used
drama
successfully
to
address
various
aspects
of
social
competence
(de
la
Cruz,
Lian,
&
Morreau,
1998;
Goldstein
&
Winner,
2012;
Lerner
&
Levine,
2007).
Drama
activities
have
been
promoted
specifically
as
an
intervention
for
ASD
and
related
disor-
ders
because
they
effectively
address
social-cognitive
processes,
emphasize
relationships
and
tap
into
social
perceptual
abilities
(Attwood,
2007;
Sherratt
&
Peter,
2002;
Warger,
1984).
In
fact,
many
drama
activities
and
games
in
the
public
domain
originally
developed
as
means
for
actors
to
become
skilled
in
reading
each
other’s
nonverbal
cues
for
the
stage,
and
thus
directly
address
the
social
perception
difficulties
experienced
by
children
with
ASD,
NLD,
and
ADHD
(Schneider,
2007).
Few
controlled
research
studies
have
examined
the
use
of
drama
activities
with
these
populations.
Recently,
Socio-Dramatic
Affective
Relational
Intervention
(SDARI;
Lerner,
Mikami,
&
Levine,
2011),
a
program
using
some
creative
drama
techniques,
was
developed
on
the
principle
that
children
with
ASD
will
benefit
from
programs
that
are
highly
moti-
vating
and
focused
on
relationships.
Pilot
study
results
indicated
generalized,
maintained
gains
for
participants
in
social
assertion,
the
ability
to
accurately
perceive
nonverbal
social
cues
(Lerner
et
al.,
2007),
as
well
as
faster
increases
in
within-group
peer
liking
and
interaction
relative
to
a
more
traditional
model
(Lerner
&
Mikami,
2012).
However,
studies
of
SDARI
have
used
relatively
small
sam-
ples
(<10
participants/condition),
have
not
examined
changes
in
peer
interaction
in
naturalistic
settings,
have
only
used
ASD
partic-
ipants,
and
do
not
use
a
pure
creative
drama
approach.
Similarly,
a
modified
and
abbreviated
version
of
the
Social
Competence
Inter-
vention
Program
for
children
aged
6–8
(SCIP;
Guli,
Wilkinson,
&
Semrud-Clikeman,
2008)
resulted
in
positive
outcomes
in
a
qualitative
study
(Minne
&
Semrud-Clikeman,
in
press).
All
child
participants
demonstrated
improvements
in
social
interactions
as
measured
by
parent
report
post-intervention.
The
program
used
play
therapy
and
sociodramatic
play
as
the
primary
therapy
modal-
ity.
Current
study
The
purpose
of
this
study
was
to
evaluate
the
efficacy
of
the
full
and
manualized
version
of
the
Social
Competence
Intervention
Program
(SCIP;
Guli
et
al.,
2008).
SCIP
is
a
manualized
creative
drama
intervention
program
designed
for
use
with
children
with
ASD
and
NLD.
Our
first
hypothesis
was
that
the
participants,
relative
to
a
population-matched
comparison
group
of
youth
who
did
not
receive
the
intervention,
would
display
improved
parent-
reported
social
functioning
at
the
end
of
the
intervention
period
as
measured
by
a
standardized
rating
scale.
Second,
we
hypothesized
that
SCIP
participants,
relative
to
the
comparison
group,
would
display
decreased
errors
in
receptive
nonverbal
cue
reading
on
an
objective
computer-based
task
at
the
end
of
the
intervention.
Third,
we
hypothesized
that
a
subsample
of
SCIP
participants
would
improve
in
observed
naturalistic
social
interaction
relative
to
the
comparison
group,
indicating
generalization
of
improved
social
skills
to
non-clinical
settings.
Finally,
we
hypothesized
that
improvements
in
participants’
social
competence
would
be
reflected
in
parents’
and
participants’
perceptions
as
indicated
in
post-treatment
interviews.
Methods
Participants
Thirty-nine
youth
(31
male),
8–14-years-old
(M
=
10.97),
par-
ticipated.
Nineteen
children
had
a
diagnosis
of
ASD
that
had
been
provided
by
a
licensed
community
psychologist
following
a
comprehensive
neuropsychological
evaluation.
Nine
children
were
previously
diagnosed
with
NLD
following
a
comprehensive
neuropsychological
evaluation
conducted
by
community
neu-
ropsychologists.
Eleven
children
had
a
primary
diagnosis
of
ADHD
based
on
DSM
IV-TR
criteria
(APA,
2000).
Thirty
participants
(76.9%)
with
ASD
or
NLD
were
reported
by
parents
to
have
a
comorbid
diag-
nosis
of
ADHD
and
20
(51.3%)
were
reported
to
take
prescription
medication.
Thirty-six
participants
(92%)
were
Caucasian,
two
were
Hispanic,
and
one
was
African-American.
The
participants
were
of
middle
to
upper
class
socioeconomic
status
and
attended
numer-
ous
schools
in
and
around
a
major
southwestern
American
city.
While
all
participants
provided
informed
consent,
children
were
unaware
of
the
hypotheses
being
tested.
Inclusion
criteria
were
(a)
overall
intelligence
above
80
(range
80–122)
as
measured
by
the
Kaufman
Brief
Intelligence
Test
(KBIT,
Kaufman
&
Kaufman,
1990)
or
the
Wechsler
Intelligence
Test
for
Children,
Third
Edition
(WISC-III,
Wechsler,
1991),
and
(b)
evidence
of
social
competence
difficulties
documented
by
either
previous
diagnosis
of
ASD
or
NLD
by
a
licensed
psychologist
or
neuropsychol-
ogist,
or
a
primary
diagnosis
of
ADHD
along
with
significant
deficits
as
indicated
on
the
Social
Skills
Ratings
System
(SSRS;
Gresham
and
Elliott,
1990;
see
below).
Exclusion
criteria
were
the
presence
of
a
history
of
head
injury,
psychosis,
oppositional
defiant
disorder
or
conduct
disorder,
and
a
primary
spoken
native
language
other
than
English.
The
first
23
children
meeting
all
inclusion
and
exclusion
crite-
ria
were
assigned
to
the
treatment
group
on
a
consecutive,
rolling
entry
basis,
in
order
to
best
serve
the
needs
of
the
community,
yield-
ing
a
pseudo-random
assignment
procedure.
Participants
were
referred
by
parents,
school
district
personnel,
and
a
neurological
clinic
in
the
community
in
a
large
southwestern
city
in
the
U.S.
Five
participants
dropped
the
program
early,
one
of
whom
was
placed
in
the
clinical
comparison
group.
All
of
the
participants
who
dropped
out
of
the
intervention
program
had
a
primary
diagnosis
of
ADHD.
After
attrition,
the
intervention
group
(n
=
18)
contained
11
children
with
a
primary
diagnosis
of
ASD
(7
of
whom
had
a
secondary
diagnosis
of
ADHD),
2
with
a
primary
diagnosis
of
NLD
(both
of
whom
had
a
secondary
diagnosis
of
ADHD),
and
5
with
a
sole
diagnosis
of
ADHD
(see
Table
1).
Children
were
assigned
to
the
comparison
group
consecutively
with
attempts
to
match
for
age,
gender
and
cognitive
ability.
Par-
ticipants
in
this
group
included
children
placed
on
a
waitlist
for
L.A.
Guli
et
al.
/
The
Arts
in
Psychotherapy
40 (2013) 37–
44 39
Table
1
Baseline
measures
between
groups
(after
attrition).
Clinical
control
group
Treatment
group
Baseline
group
difference
Gendera13
male
15
male
.03
(.874)
Primary
diagnosisa7
ASD,
6
NLD,
3
ADHD 11
ASD,
2
NLD,
5
ADHD 3.28
(.194)
Medication
statusa4
yes
12
yes
5.90
(.015)*
Age
(months)
133.00
(24.63),
16
125.28
(23.05),
18
−.94
(.352)
IQ 104.17
(15.47),
12
107.50
(14.04),
18
.61
(.546)
Social
skills
rating
system
71.75
(11.31),
12
78.05
(15.98),
17
1.17
(.251)
DANVA2
–
faces
3.15
(1.86),
13
5.22
(2.71),
18
2.37
(.025)*
DANVA2
–
voices
6.62
(2.63),
13
7.28
(3.44),
18
.58
(.566)
BASC
– withdrawal
59.93
(14.01),
15 64.17
(17.14),
18 .77
(.449)
BASC
–
social
skills
33.33
(11.07),
12
35.29
(9.95),
17
.50
(.62)
Observed
positive
interactions 12.56
(10.91),
9
12.25
(10.21),
8
−.06
(.953)
Observed
solitary
behaviors
19.89
(14.42),
9
17.63
(14.08),
8
.33
(.748)
aSimple
counts.
Numbers
in
table
are
raw
group
means
with
standard
deviations
in
parentheses
followed
by
n’s.
In
the
baseline
group
difference
column,
t-test
or
chi-square
value
and
associated
p-value
(in
parentheses)
for
the
significance
of
the
difference
between
the
baseline
means
is
listed.
*p
<
.05.
treatment,
those
who
could
not
participate
due
to
scheduling
difficulties,
and
children
who
dropped
after
two
or
fewer
ses-
sions,
a
procedure
that
has
been
used
in
other
treatment
studies
(McCloskey,
Noblett,
Deffenbacher,
Gollan,
&
Caccaro,
2008).
The
comparison
group
(n
=
16)
included
7
children
with
a
primary
diag-
nosis
of
ASD
(3
of
whom
had
a
secondary
diagnosis
of
ADHD),
6
with
a
primary
diagnosis
of
NLD
(5
of
whom
had
a
secondary
diagnosis
of
ADHD),
and
3
with
a
sole
diagnosis
of
ADHD
(see
Table
1).
Measures
SSRS
(Gresham
and
Elliott,
1990)
is
a
widely
used
parent-report
measure
of
socially
appropriate
behaviors
in
social
skills
interven-
tions
for
youth
with
social
competence
deficits
(White
et
al.,
2007).
In
the
present
study,
it
was
administered
only
at
baseline
to
ensure
participants
evidenced
clinically
significant
deficits
in
social
skills
relative
to
typically
developing
peers.
Behavioral
Assessment
System
for
Children
(BASC;
Reynolds
&
Kamphaus,
1992)
is
a
parent-report
measure
of
social
and
behav-
ioral
adjustment
in
children.
Only
the
withdrawal
and
social
skills
subscales
were
used,
as
these
scales
have
been
found
to
be
particu-
larly
sensitive
to
social
competence
difficulties
found
in
youth
with
ASD
(Lindner
&
Rosen,
2006).
The
BASC
subscales
were
used
as
a
pre-
and
post-test
measure.
Diagnostic
analysis
of
nonverbal
accuracy
2
(DANVA2;
Nowicki,
2004)
is
a
computer-based
objective
measure
of
receptive
non-
verbal
cue
reading
through
facial
expression
and
prosodic
cues.
This
measure
was
used
as
a
pre-
and
post-test
measure
and
it
has
demonstrated
treatment
sensitivity
with
children
with
ASDs
(Lerner
et
al.,
2011;
Solomon,
Goodlin-Jones,
&
Anders,
2004).
Observed
social
interaction
For
17
of
the
participants
(43.6%)
at
pre-
and
post-treatment,
observations
were
carried
out
for
one
20-min
interval
by
one
of
three
trained
observers
(38%
of
treatment
group
and
50%
of
control
group).
Of
the
original
sample
8
parents
in
the
intervention
group
and
9
parents
in
the
control
group
consented
to
an
observation
of
their
child
in
the
school
setting.
Among
these,
for
the
intervention
group
there
were
5
children
with
a
primary
diagnosis
of
ASD,
2
with
a
sole
diagnosis
of
ADHD,
and
1
with
NLD.
For
the
comparison
group,
there
were
3
children
with
a
primary
diagnosis
of
ASD,
1
with
a
sole
diagnosis
of
ADHD,
and
5
children
with
a
diagnosis
of
NLD.
Observations
were
conducted
using
the
partial
interval
recor-
ding
method
with
an
audiotape
consisting
of
40
intervals
of
20
s
for
each
observation,
separated
by
10-s
intervals
for
recording
data.
The
partial
interval
recording
method,
which
notes
whether
or
not
a
behavior
occurs
during
specific
time
periods,
was
chosen
because
the
type
of
the
social
behaviors
being
observed
occurred
at
vary-
ing
rates,
and
thus
it
was
believed
to
provide
a
useful
estimate
of
the
frequency
of
behaviors
in
the
natural
setting.
The
percentage
of
intervals
in
which
the
following
variables
were
observed
was
recorded:
positive
social
interaction,
solitary
behavior
or
neutral
behavior.
These
variables
were
created
based
on
previous
studies
investigating
social
interactions
in
school
aged
children
with
ASD
(Owen-Deshryver,
Carr,
Cale,
&
Blakeley-Smith,
2008).
Variables
were
operationally
defined
as
follows:
Positive
social
interaction
was
defined
as
a
communicative
exchange,
verbal
or
nonverbal,
between
the
observed
child
and
peers
during
which
they
demonstrated
one
or
more
of
the
fol-
lowing:
playing
cooperatively,
sharing,
conversing
pleasantly
(as
evidenced
by
smiling,
speaking
with
respect,
taking
turns
when
speaking),
or
socially
appropriate
exchanges
such
as
greetings,
say-
ing
please
and
thank
you.
Any
social
interaction
that
reflected
that
the
child
is
accepted
by
and
accepts
peers
was
coded
positive.
Solitary
behavior
referred
to
a
behavior
performed
in
isolation,
whether
intentional
or
unintentional.
For
example,
a
child
may
iso-
late
him
or
herself
on
the
playground
and
wish
to
play
alone,
or
play
alone
because
no
other
peer
wants
to
be
with
him
or
her.
Any
of
these
behaviors
or
interactions
was
coded
as
solitary.
Neutral
behavior
referred
to
behaviors
that
were
not
coded
in
the
above
categories.
Examples
included:
behaviors
instructed
by
an
adult
or
demanded
by
the
setting
(bouncing
a
basketball
to
someone
in
gym
class,
or
saying
‘thank
you’
when
instructed
by
a
teacher),
or
interactions
resulting
from
clumsiness
(bumping
into
someone
by
accident).
To
complete
training
on
this
observational
coding
system,
the
three
observers
coded
a
30-min
test
tape
of
child
interactions
cre-
ated
by
the
principal
investigator
to
establish
inter-rater
reliability.
A
two-way
mixed
effects
intra-class
correlation
(ICC)
of
individ-
ual
ratings,
ICC(3,1),
was
used
to
calculate
reliabilities,
since
the
3
raters
coded
all
intervals,
and
the
reliability
of
each
individual
rater
was
the
relevant
unit
of
analysis
(Shrout
&
Fleiss,
1979).
These
ICCs
are
asymptotically
equivalent
to
multi-rater
kappa
for
categorical
data
over
large
datasets,
and
may
in
fact
produce
more
conserva-
tive
and
appropriate
standard
errors
relative
to
kappa
(see
Fleiss
&
Cohen,
1973;
Sawa
&
Morikawa,
2007).
Reliabilities
on
the
test
tape
were
excellent
(ICC
>
.75;
Cicchetti,
1994)
for
positive
(ICC
=
.779),
solitary
(ICC
=
.904),
and
neutral
(ICC
=
.898)
behaviors.
Final
behavior
observations
(used
herein
for
primary
analyses)
took
place
during
recess,
lunch
or
a
cooperative
activity
period
at
the
child’s
school.
Pre-
and
post-observations
for
a
single
child
were
completed
by
the
same
observer
and
observations
were
conducted
40 L.A.
Guli
et
al.
/
The
Arts
in
Psychotherapy
40 (2013) 37–
44
during
the
same
activity
pre-
and
post-treatment.
No
observer
overlap
took
place
during
the
final
coding.
Parent
and
child
interviews
Treatment
participants
(n
=
18)
and
most
parents
of
treat-
ment
participants
(n
=
15)
were
interviewed
individually,
for
approximately
30
min
each,
with
a
semi-structured
interview
post-intervention
to
measure
treatment
satisfaction,
assess
the
intervention’s
efficacy
and
establish
social
validity.
Interview
ques-
tions
were
developed
by
the
author
of
this
study.
Interviews
were
audio
taped
and
coded
using
grounded
theory
methodology
qual-
itative
procedure
(Strauss
&
Corbin,
1998),
which
groups
data
according
to
similar
concepts.
Procedure
Data
collection
Participants
and
parents
completed
the
measures
as
adminis-
tered
by
an
advanced
doctoral
student
immediately
prior
to,
and
two
months
after,
the
intervention,
Behavioral
observations
were
conducted
pre-
and
post-intervention
for
the
treatment
group
and
before
and
after
an
8–12
week
gap
for
the
comparison
group.
Intervention
content
SCIP
is
a
16-session
manualized
intervention
program
devel-
oped
from
creative
drama
activities
(Guli
et
al.,
2008).
The
program’s
goal
is
to
improve
participants’
social
perception
of
non-
verbal
cues
and
to
thus
improve
social
competence
in
the
natural
setting.
An
earlier
and
shorter
version
of
SCIP
was
piloted
twice
with
positive
responses
from
participants
and
parents
(Glass,
Guli,
&
Semrud-Clikeman,
2000;
Guli
&
Semrud-Clikeman,
2002).
The
intervention
utilized
in
this
study
is
the
same
as
was
published
later.
Content
of
the
manual
was
modified
from
or
inspired
by
tra-
ditional
creative
drama
activities
(Bowell
&
Heap,
2001;
O’Neill,
1995;
Spolin,
1986),
including
cooperative
games,
story
dramati-
zation
and
process
drama
improvisations
(Spolin,
1986).
Activities
emphasize
process
the
interaction
between
participants,
rather
than,
production
or
performance.
SCIP
sessions
initially
focus
on
children’s
own
experience
of
emotion
and
expand
to
understanding
of
others’.
Sessions
1
through
7
target
the
input
stage
of
social
perception,
and
cover
the
following
topics:
establishing
group
cohesion,
emotional
knowl-
edge,
focusing
attention,
facial
expression
and
body
language,
vocal
cues
and
putting
several
cues
together.
For
example,
in
the
game
“Say
it
With
Feeling”,
participants
say
a
short
sentence
several
dif-
ferent
times
with
different
emotions,
and
take
turns
guessing
what
emotion
is
being
conveyed.
This
may
train
them
to
hear
the
sub-
tle
differences
in
tone
and
prosody
and
learn
how
these
contribute
to
meaning.
In
the
game
“Jell-O
Room,”
participants
move
across
a
room
pretending
that
the
room
is
filled
with
different
substances
or
different
emotions.
Thus
they
practice
using
body
language
to
communicate
feeling.
Sessions
8
through
12
focus
on
the
integra-
tion
and
interpretation
of
nonverbal
cues.
These
sessions
include
activities
that
focus
on
taking
others’
points
of
view
and
interpre-
ting
conflicting
cues.
Participants
engage
in
several
improvisations,
or
process
dramas,
through
which
they
practice
perspective
tak-
ing
and
cognitive
flexibility.
Leaders
help
participants
break
down
complex
social
interactions
into
simpler
steps,
discuss
the
emotions
present,
and
act
out
a
variety
of
possible
responses.
For
example,
in
the
improvisation,
“Detective
Agency,”
children
are
detectives
while
leaders
take
the
roles
of
crime
witnesses.
Participants
inter-
view
witnesses
and
examine
nonverbal
and
contextual
cues
to
solve
a
mystery.
Sessions
13
through
16
address
the
output
stage
of
social
perception,
and
focus
on
effective
ways
to
respond
to
others.
Session
topics
include
handling
teasing,
beginning
conversations,
resolving
conflict
and
reviewing
overall
session
content.
Session
structure
Each
session
began
with
a
warm-up
activity
(usually
a
cooper-
ative
game
to
motivate
and
focus
attention,
such
as
slow-motion
freeze
tag),
followed
by
a
review
of
a
home
challenge
assignment
(i.e.,
observing
how
peers
begin
conversations
and
writing
in
a
journal
about
it),
2–3
main
drama
activities
that
focused
on
the
session’s
topic,
and
discussions
to
process
the
experience.
Par-
ticipants
were
encouraged,
but
not
required,
to
complete
home
challenges
and
this
was
rewarded
with
small
tangible
reinforcers
(i.e.,
stickers,
etc.)
in
the
following
session.
Parents
were
given
a
written
overview
of
intervention
objectives,
target
goals
and
detailed
schedule
on
the
first
day
of
the
intervention.
Parents
were
also
asked
to
encourage
their
children
to
complete
home
challenges
after
each
session.
Behavior
management
A
response-cost
system
of
behavior
management
was
used
to
reinforce
positive
behavior.
Specifically,
each
child
was
entitled
to
choose
3
small
rewards
or
toys
at
the
end
of
each
session
from
an
assortment.
After
several
warnings
for
misbehavior,
a
child
could
lose
two
of
these
rewards.
Appropriate
behavior
was
rein-
forced
with
positive
verbal
praise
from
adults,
as
well
as
with
extra
intermittent
rewards
offered
to
the
group
for
participation
and
cooperation.
Also,
group
leaders
were
trained
in
cooperative
discipline
techniques,
addressing
one
of
four
outlined
common
underlying
causes
of
misbehavior
with
specific
strategies.
During
discussion,
when
a
child
spoke,
he
or
she
held
a
“talking
stick”
to
remind
children
to
speak
one
at
a
time.
Finally,
each
child
with
the
help
of
group
leaders
identified
a
specific
goal
to
work
on
(i.e.,
not
interrupting).
When
children
worked
toward
the
goal,
they
were
praised.
Similar
individual
goals
have
been
identified
in
other
intervention
programs
(i.e.,
Lerner
&
Levine,
2007).
Treatment
fidelity
Treatment
fidelity
was
monitored
by
reviewing
session
objec-
tives
in
weekly
supervision
meetings.
Group
leaders,
who
were
trained
in
six
1-h
sessions
prior
to
intervention,
recorded
after
each
session
whether
objectives
were
met.
When
an
objective
was
not
sufficiently
met,
as
determined
by
discussion
and
review
in
weekly
meetings,
it
was
re-addressed
in
a
following
session.
This
level
of
review
may
be
related
to
improved
outcomes
across
psychosocial
treatments
(Herschell,
Kolko,
Baumann,
&
Davis,
2010).
Treatment
conditions
The
intervention
was
run
twice.
In
fall
2002,
eight
children
were
placed
in
a
younger
subgroup
(8–10)
with
four
leaders
(3
female,
1
male)
and
7
in
an
older
subgroup
(11–14)
with
three
female
group
leaders.
Four
children
ended
participation
after
two
weeks.
In
spring
2003,
seven
boys
(none
of
whom
participated
in
fall
2002)
aged
9–14
participated.
One
13-year-old
boy
dropped
out
after
4
weeks.
Some
changes
to
treatment
conditions
were
made
in
the
spring.
In
fall,
sessions
were
held
in
large
university
classrooms;
in
spring,
they
were
in
an
elementary
school
gymnasium.
To
accom-
modate
schedules,
spring
groups
were
longer,
and
held
once
a
week
for
12
weeks
(12
longer
sessions,
each
held
for
2
h),
as
opposed
to
shorter
sessions
twice
a
week
for
8
weeks
in
fall
(16
sessions,
each
held
for
an
hour
and
a
half).
Thus,
both
fall
and
spring
ses-
sions
consisted
of
24
h
of
intervention.
In
the
spring,
sessions
were
modified
and
condensed
to
ensure
that
all
objectives
were
covered.
Other
variations
from
the
manual
naturally
occurred
as
group
lead-
ers
adjusted
to
the
needs
of
individual
group
members
(i.e.,
optional
activity
skipped
to
resolve
a
conflict
between
participants).
L.A.
Guli
et
al.
/
The
Arts
in
Psychotherapy
40 (2013) 37–
44 41
Data
analytic
plan
Baseline
descriptive
characteristics
of
completers
were
com-
pared
between
groups
using
chi-square
tests
for
non-continuous
data
(gender,
number
who
had
ADHD
or
ASD
diagnoses,
medication
status)
and
t-tests
for
continuous
data
(age,
IQ)
and
baseline
scores
on
all
recorded
measures.
Next,
participants
who
dropped
out
were
compared
to
completers
using
t-tests
to
assess
potential
differences
in
age
and
IQ.
The
Levene
Test
for
Equality
of
Variance,
as
well
as
observation
of
relative
size
of
standard
deviations
between
groups,
were
used
to
determine
whether
t-tests
assuming
equal
or
unequal
variances
would
be
used
and
reported.
Third,
intent-to-treat
anal-
yses
comparing
change
between
the
comparison
and
intervention
groups
were
conducted
on
all
measures
of
social
ability.
As
the
intervention
and
comparison
groups
were
matched
on
observed
measures
at
baseline,
ANCOVA
of
change
was
used
for
all
primary
analyses.
That
is,
we
used
two
step
hierarchical
multiple
regression
models
predicting
endpoint
values
of
each
social
ability
measures.
We
placed
corresponding
baseline
values
of
each
social
ability
mea-
sure
on
Step
1.
Then,
we
placed
intervention
status
(dummy
coded:
0
=
comparison,
1
=
intervention)
on
Step
2,
such
that
higher
values
of
this
coefficient
correspond
to
relative
increases
in
the
variable
of
interest.
This
procedure
produces
unbiased
results
for
inferring
treatment
effects
with
the
present
methodological
design
(Van
Breukelen,
2006).
Power
to
detect
large
effects
(see
Lerner
et
al.,
2011)
was
sufficient
(>.84)
for
baseline
and
attrition
analyses,
and
somewhat
low
(>.61)
for
primary
analyses.
For
all
regression
analy-
ses,
Cohen’s
(1992)
conventions
for
qualification
of
effect
size
were
used.
Results
Descriptive
analyses
There
were
no
differences
between
groups
in
terms
of
gen-
der,
primary
diagnosis,
age
or
IQ
between
groups
(see
Table
1).
However,
the
groups
did
differ
on
the
number
of
individuals
on
medications,
with
more
medicated
children
in
the
intervention
group.
Likewise,
treatment
and
comparison
groups
did
not
differ
on
any
baseline
measures
except
the
DANVA-2
Faces,
on
which
the
intervention
group
showed
greater
baseline
impairment.
On
the
SSRS
at
baseline,
mean
of
both
groups
was
more
than
one
standard
deviation
below
population
mean,
t(28)
=
−9.208,
p
<
.001,
supporting
the
presence
of
clinically
significant
social
skills
deficits
in
participants.
Analysis
of
attrition
data
The
Levene
Test
for
equality
of
variances
indicated
equal
vari-
ances
between
dropout
and
complete
groups
for
IQ
(F
=
1,89,
p
=
.185)
but
not
age
(F
=
4.49,
p
=
.048;
see
below
for
SD);
thus,
a
t-test
assuming
equal
variances
was
used
to
compare
dropout
and
completer
groups
for
IQ,
but
assuming
unequal
variances
for
age.
These
analyses
did
not
reveal
differences
in
IQ
(t
=
.48,
p
=
.635),
but
demonstrates
significant
group
difference
in
age
(t
=
−2.937,
p
=
.02).
While
mean
age
of
treatment
completers
in
the
treatment
group
after
attrition
was
125.28
months
(SD
=
23.05),
mean
age
of
those
children
who
dropped
out
was
148.00
months
(SD
=
9.54).
Primary
outcome
analyses
Table
2
displays
group
means
at
post-test,
while
Table
3
summa-
rizes
the
results
of
the
primary
analyses.
Significant
effects
were
not
found
for
the
BASC
or
DANVA2
measures.
However,
medium
effects
(Cohen,
1992)
were
obtained
for
observed
increases
in
positive
Table
2
Posttest
measures
between
groups.
Clinical
control
group
Treatment
group
DANVA2
–
faces
2.46
(1.81),
13
3.37
(2.45),
18
DANVA2
– voices
5.69
(2.21),
13 6.16
(2.52),
18
BASC
– withdrawal
58.79
(13.61),
14
61.89
(13.55),
18
BASC
–
social
skills
35.07
(7.87),
14
36.22
(9.37),
18
Observed
positive
interactions
6.89
(7.57),
9
14.75
(9.04),
8
Observed
solitary
behaviors
21.11
(9.03),
9
10.13
(13.21),
8
Numbers
in
table
are
raw
group
means
with
standard
deviations
in
parentheses
followed
by
n’s.
No
posttest
values
are
reported
for
the
social
skills
rating
system,
as
it
was
not
administered
at
posttest.
No
comparison
statistics
are
reported
in
this
table,
as
between-group
comparisons
in
change
over
time
are
addressed
in
Table
3.
interactions
(R2=
.21)
and
decreases
in
solitary
play
(R2=
.17),
indi-
cating
improved
social
interaction
in
a
naturalistic
setting
among
intervention,
but
not
clinical
control,
participants.
Qualitative
findings
The
interview
data
is
consistent
with
the
observational
data.
When
parents
were
asked
if
they
observed
any
improvements
in
their
child’s
social
functioning,
75%
reported
one
or
more
specific
positive
changes,
including
improvement
in
interpersonal
rela-
tions,
better
understanding
of
nonverbal
dues,
increased
empathy
and
improved
self-control.
Several
excerpts
from
parent
interviews
regarding
their
child’s
experience
are
as
follows:
“I’ve
noticed
that
Table
3
Hierarchical
multiple
regression
analyses
for
main
treatment
effects.
Measure
B
SE
B
Beta
p
R2
BASC-withdrawal
Step
1
.48
Pre-treatment .62
.12
.69*** .000
Step
2
.001
Pre-treatment
.61
.12
.69*** .000
Clinical
vs.
controla.67
3.70
.02
.859
BASC-social
skills
Step
1 .52
Pre-treatment
.61
.11
.72*** .000
Step
2 .01
Pre-treatment
.61
.11
.73*** .000
Clinical
vs.
control
−1.84
2.25
−.11
.420
DANVA-2
child
faces
Step
1
.36
Pre-treatment
.51
.12
.60*** .000
Step
2
.001
Pre-treatment
.53
.14
.62** .001
Clinical
vs.
control
−.17
.71
−.04
.812
DANVA
2
child
voices
Step
1
.43
Pre-treatment
.50
.10
.66*** .000
Step
2
.01
Pre-treatment
.49
.11
.64**** .000
Clinical
vs.
control
.56
.64
.12
.390
Observed
positive
interactionsb
Step
1
.30
Pre-treatment
.48
.19
.55*.022
Step
2
.21
Pre-treatment
.49
.16
.56*.010
Clinical
vs.
control 8.13
3.30
.46*.028
Observed
solitary
behaviorsb
Step
1
.46
Pre-treatment
.60
.17
.68** .003
Step
2
.17
Pre-treatment .57
.15
.64*** .002
Clinical
vs.
control
−9.85
3.97
−.41*.026
aClinical
group
dummy
coded
to
1,
control
group
dummy
coded
to
0.
bSubset
n
=
17.
*p
<
.05.
** p
<
.01.
*** p
<
.001.
42 L.A.
Guli
et
al.
/
The
Arts
in
Psychotherapy
40 (2013) 37–
44
he
seems
to
be
reaching
out
more
to
kids
in
the
neighborhood
to
try
to
befriend
them.
.
.I
think
the
thing
I
see
is
that
he
is
more
inter-
ested
in
people,
in
relating
to
them.
.
.”
(mother
of
11-year-old
boy
with
ASD)
“.
.
.(I)t
was
like
night
and
day.
.
.his
face
seemed
a
lot
more
animated
and
he
seemed
to
make
a
big
effort
to
communi-
cate
and
actually
do
a
give
and
take
in
communication.
.
.”
(mother
of
12-year-old
boy
with
NLD).
“He
was
looking
at
her,
making
faces
to
her,
making
her
laugh,
he
was
laughing.
I
noticed
his
face
shows
more
expression”
(mother
of
8-year-old
boy
with
ASD).
“I
have
seen
her
showing
more
empathy
toward
(her
sister).
.
.if
she
falls
down
or
hurts
herself
and
is
crying
I
have
noticed
that
she
is
showing
more
attention
to
her”
(mother
of
8-year-old
girl
with
ASD).
“He
seems
to
have
improved
in
being
able
to
perhaps
think
about
what
the
other
person
may
be
feeling,
something
that
never
entered
his
mind
before.
He
even
showed
compassion
for
a
friend
yesterday
with
severe
learning
disabilities”
(mother
of
an
11-year-old
boy
with
ASD).
Similarly,
82%
of
children
reported
one
or
more
positive
effects,
including
the
belief
that
they
could
read
nonverbal
cues
better,
make
friends,
learn
about
feelings,
and
get
along
better
with
oth-
ers.
Child
participants
had
much
to
say,
including
the
following:
“I
can
focus
on
other
people’s
body
language
a
little
bit
clearer;
I
can
understand
what
they’re
saying
with
their
body
language
a
little
clearer”
(11-year-old
boy
with
ASD).
“I’m
a
bit
calmer.
I
can
under-
stand
people
now.
I
have
a
social
life
now”
(12-year-old
boy
with
NLD).
“Now
I
know
a
lot
about
feelings
and