The Effect of Improvisational

Music Therapy on the

Communicative Behaviors of
Autistic Children1

Cindy Lu Edgerton

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Michigan State University

The purpose of this study was to determine the effectiveness

of improvisational music therapy, based on Nordoff and Rob­
bins’ (1977) Creative Music Therapy approach, on autistic
children’s communicative behaviors. Eleven autistic children,
ranging in age from 6 to 9 years, participated in individual
improvisational music therapy sessions for a period of 10
weeks. A reversal design was applied. The Checklist of Com­
municative Responses/Acts Score Sheet (CRASS), designed
specifically for this study, was used to measure the subjects’
musical and nonmusical communicative behaviors. Results
strongly suggest the efficacy of improvisational music therapy
in increasing autistic children’s communicative behaviors.
Significant differences were found between the subjects’ first
session CRASS scores and those of their last sessions (p
< .01). Also, abrupt and substantial decreases inscores were
noted for all 11 subjects when reversal was applied.

In the first description of children diagnosed with early in­

fantile autism, Kanner (1943) devoted a large amount of atten­
tion to their communication deficits. Since then, research has
continued to support Kanner’s observations of the numerous
problems related to autistic children’s communication devel­
opment. According to the National Society for Autistic Children
(1978) and the American Psychiatric Association (1987), lan­
guage development deficit is one of the symptoms necessary
for a diagnosis of autism.
The acquisition of language is crucial to the prognosis for
autistic children. The presence of useful speech by 5 years of

1The author gratefully acknowledges Roger Smeltekop, M.M., RMT-BC, and

Dale Bartlett, Ph.D., for their support and guidance throughout this study.
Correspondence regarding this article should he addressed to Cindy Lu Edger­
ton, 443 Forest, Charlotte, Ml 48813.
32 Journal of Music Therapy

age was one of the most significant distinguishing characteristics

between autistic children rated as making poor adjustment and
those who made good adjustment (Eisenberg & Kanner, 1956).
Further research confirmed the finding that the degree of lan­
guage development by ages ranging from 30 months to 6 years
is predictive of later development (Bagley & McGeein, 1989;

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Brown, 1963; DeMyer, Barton, DeMyer, Norton, Allen, & Steel,
1973; Kurita, 1985). Rutter (1978) and Gillberg and Steffenburg
(1987) found that functional language skills demonstrated by
early school age were as powerful as intelligence in predicting
autistic children’s later skills.
In Kanner’s (1946) description of autistic children, he noted
the following specific characteristics of their language deficits:
muteness, immediate and delayed echolalia, metaphorical sub­
stitution, literalness, simple verbal negation, repetitions, and
pronoun reversals. He concluded that both the mute and the
verbal children were the same as far as meaningful commu­
nication was concerned. Current views of these characteristics
focus on difficulties in pragmatics, or how language is used for
various purposes, which are now seen as a universal feature of
autism (Frith, 1989). Autistic children have been found to have
deficits in means-end behaviors (Abrahamsen & Mitchell, 1990),
verbal, gestural, and motor imitative skills (Dawson & Adams,
1984; Stone & Lemanek, 1990), spontaneous speech (Shapiro,
Chiarandini, & Fish, 1974), initiation of contact with others
(Watson, 1985), intentional communicative behaviors/vocali­
zations (Ball, 1978), communicative functions (Landry & Love­
land, 1989; Wetherby & Prutting, 1984), social communication
skills (Attwood, 1984; Kubicek, 1980), and prosodic develop­
ment (Baltaxe, Simmons, & Zee, 1984). Frith (1989) points out
that, even though more has been written on the language im­
pairments of autistic people than on any other of their deficit
areas, numerous questions remain unanswered.
At least three-quarters of all speaking autistic children dem­
onstrate echolalia, but it is not yet known why (Frith, 1989).
The belief that echolalic behaviors serve no significant com­
municative purposes has been challenged. Echolalia and stereo­
typical language of autistic children can be analyzed according
to communicative intent, comprehension, and structural changes.
 Vol. XXXI, No. 1, Spring, 1994 33

The results of two major studies indicated that four of the seven
types of immediate echolalia (Prizant & Duchan, 1981) and 9
of the 20 categories of delayed echolalia (Prizant & Rydell,
1984) are interactive. Tager-Flusberg (1985) stated, “Echolalia
and stereotyped language are now seen as primitive strategies
for communicating, especially in the context of poor compre­

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

hension” (p. 72).
Throughout the research and literature pertaining to autistic
children, structured intervention approaches are frequently rec­
ommended. A high degree of structure is seen as an essential
element in autistic children’s treatment plans (Thaut, 1980). In
reviewing the research of education approaches, Clarizio &
McCoy (1983) found general agreement that autistic children
learn best in structured environments, where both the stimuli
and the child’s responses are determined by the adult.
Current trends in language intervention programs with au­
tistic children are numerous. There has been a change in em­
phasis from teaching language skills to teaching communication
skills due to the importance placed on functional communi­
cation. This shift focuses on accepting each child’s language
impairment and working toward his/her optimum potential in
communication development (Schopler & Mesibov, 1985). The
most recent intervention technique for people with autism is
facilitated communication, a method developed by Rosemary
Crossley (Biklen, 1990). This method is based upon a praxis
theory, which presumes that autistic people have a neurologi­
cally based deficit, not in comprehension, but in expression.
Facilitated communication uses an electronic typing device and
allows for education through dialogue and personal expression.
Only one research study on facilitated communication, con­
ducted by the Intellectual Disability Review Panel (1989) in
Melbourne, Australia, has been published to date. This study
produced support both for people who claimed facilitated com­
munication was valid and for those who doubted its validity.
Intervention programs have demonstrated success in autistic
children’s socialization and communication achievements.
However, treatment appears to have only a modest effect on
long-term language adjustment.
Many reports from clinical observations and experiments with
34 Journal of Music Therapy

autistic children emphasize their special responsiveness toward

and unusual interest in musical stimuli (Applebaum, Egel, Koe­
gel, & Imhoff, 1979; DeMyer, 1979; Kolko, Anderson, & Camp­
bell, 1980; Rimland, 1964; Sherwin, 1953; Thaut, 1980). In
studies investigating the use of structured music therapy tech­
niques with autistic children, positive effects were found in the

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

children’s prosocial behaviors (Stevens & Clark, 1969); attention
span (Farmer, 1963; Goldstein, 1964; Mahlberg, 1973; Saper­
ston, 1982); self-expression (Cecchi, 1990; Goldstein, 1964;
Mahlberg, 1973); mental age (Goldstein, 1964); spontaneous
speech (Miller & Toca, 1979; Watson, 1979); vocal imitation
skills (Miller & Toca, 1979; Saperston, 1982); interpersonal re­
lationships (Goldstein, 1964; Saperston, 1982); task accuracy
(Burleson, Center, & Reeves, 1989); and shopping skills (Staum
& Flowers, 1984).
The literature most relevant to this investigation has focused
on improvisational music therapy techniques. Saperston (1973)
used improvised music to establish communication with an au­
tistic child who had not previously appeared to experience any
type of communication. Alvin and Warwick (1992) reported
on the use of improvisation to facilitate interactions, to provide
for self-expression, and to help develop a relationship of trust
and enjoyment for autistic children and their mothers. Hollan­
der and Juhrs (1974) used Orff Schulwerk activities to help
severely autistic children invest in a meaningful group expe­
rience.
Nordoff and Robbins (1964, 1968a. 1971, 1977) have done
extensive work using improvisation with autistic children. Their
technique, Creative Music Therapy, emphasizes the creation of
musical improvisations which serve as a nonverbal means of
communication between the therapist and the child. Numerous
case studies by Nordoff and Robbins have illustrated and cor­
roborated the many values of Creative Music Therapy. Working
with a 3½-year-old autistic boy, vocal and instrumental impro­
visation aided in establishing communication, providing a means
of self-expression, improving interpersonal relationships, and
decreasing pathological behaviors (Nordoff & Robbins, 1964).
In another study, Nordoff and Robbins (1968a) used improvised
music with a 6-year-old autistic girl who demonstrated progress
Vol. XXXI, No. 1, Spring, 1994 35

in increased vocabulary, self-expression, and the spontaneous

use of personal pronouns. In establishing a therapist-client re­
lationship with a 5-year-old boy who exhibited autistic features,
Nordoff and Robbins (1977) used improvisational techniques
through both drum/cymbal-piano and vocal interaction activ­
ities. Progress was noted in several areas, including an increase

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

in vocabulary, development of spontaneous and communicative
speech, development of conversational jargon, and acceptance
of change and novel situations. Although numerous case studies
have demonstrated many therapeutic values of Creative Music
Therapy, no controlled experimental studies of improvisational
techniques based on Creative Music Therapy were found.
Therefore, the purpose of this study was to examine the effects
of improvisational music therapy upon the communicative be­
haviors of autistic children. Specifically, the following research
questions were proposed:
1. Is there a significant difference between the number of
total communicative behaviors as measured by the Checklist of
Communicative Responses/Acts Score Sheet (CRASS) demon­
strated by autistic children in their first improvisational music
therapy sessions and the number demonstrated in their last
sessions?
2. Is there a significant difference between the number of
Communicative Responses/Acts demonstrated by autistic chil­
dren in their first sessions and the number demonstrated in their
last sessions in each of the following subcategories of the CRASS:
tempo, rhythm, structure/form, pitch, speech production, com­
municative-interactive, and communicative intent?
3. IS there a significant relationship between the subjects’
musical vocal behavior score changes and their nonmusical
speech production score changes as recorded on the CRASS?
4. Will any changes in the autistic children’s communicative,
social/emotional, and musical behaviors be observed by the
parents, teachers, or speech therapists outside of the music ther­
apy setting at the conclusion of the IO-week period?
5. Are there significant relationships between the autistic
children’s overall CRASS score changes and the parents’, teach­
ers’, or speech therapists’ total ratings on the Behavior Change
Survey?
36 Journal of Music Therapy

Method
Subjects
Eleven autistic impaired subjects (ten males and one female),
ranging in age from 6 to 9 years, participated in this study.
Subjects’ diagnosis of autistic impaired ranged from severely to
mildly impaired. Deficits in communication skills were common

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

to all of the subjects. Language ages, measured by standardized
tests and/or observation and reported by speech therapists and/
or teachers, ranged from “no formal means of intentional com­
munication” to 5 years. Five subjects were nonverbal, and four
subjects demonstrated limited functional language skills.

Materials/Settings
The following musical materials were used in the treatment
procedure: a piano, a snare drum with the snare removed, and
a 16-inch cymbal. The snare drum and cymbal were mounted
on adjustable stands. The height and tilt of the snare drum and
cymbal were adapted for each child to allow for successful
attempts at beating. A chair was available for the subjects, and
the experimenter was seated on a piano bench. A variety of
beaters were also available, including regular medium-weight
drumsticks, both heavy and light tympani mallets, and one pair
of brushes. A videocamera was also in the room.
The study was conducted in three different settings: two
elementary schools and a music therapy clinic. Two of the rooms
were similar in size and content; the other was a music education
room that was larger and contained a variety of musical ma­
terials/instruments. An area within this room which resembled
the other two rooms was set up for the experiment. The ex­
perimenter and each child were alone in the room with the
exception of a few unanticipated interruptions.

Measurement
The dependent variable in this study was communicative
behaviors. Consultations with speech therapists and a search of
relevant literature revealed no standardized test that evaluates
musical and nonmusical communicative behaviors for autistic
children. Nonmusical communicative responses of autistic chil­
dren are evaluated through a variety of standardized tests. Nor-
Vol. XXXI,NO.1. spring, 1994 37

doff and Robbins (1977) developed a musical communicative­

ness scale with established reliability; however, this reliability
was established using music therapists who were specifically
trained in the Creative Music Therapy technique (C. Robbins,
personal communication, May 26, 1992). Therefore, an original
checklist, Checklist of Communicative Responses/Acts Score
Sheet (CRASS), was constructed by the experimenter. Each of

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

the behaviors listed on the score sheet was operationally defined.
The CRASS was based on items from numerous rating scales
and assessments for musical communicativeness, autism, and
communication skills (Brigance, 1978; Bzoch & League, 1970;
Krug, Arick, & Almond, 1979; Nordoff & Robbins, 1977; Rut­
tenberg, Dratman, Fraknoi, & Wenar, 1966; Stillman, 1978;
Uzgiris & Hunt, 1975; Wetherby & Prutting, 1984).
The CRASS was divided into two categories: musical and
nonmusical. Communicative Responses were defined as verbal,
vocal, gestural, or instrumental behaviors demonstrated by the
child which are influenced by the experimenter’s improvisation,
e.g., matches a fast basic beat, simultaneously imitates the rhythm
of a melodic motif, participates in a rhythmic give-and-take,
etc. Verbal, vocal, or instrumental behaviors initiated by the
child in an attempt to influence the experimenter’s improvi­
sation/behaviors or for the purpose of independent expression
were categorized as Communicative Acts (e.g., creates a rhyth­
mic pattern, develops a melodic give-and-take, spontaneously
creates a new melodic phrase, etc.). Behaviors which served as
prerequisite skills necessary for musical communication were
also categorized as Communicative Acts (e.g., beats within a
tempo range, vocalizes, etc.).
Within the musical category, operationally defined Com­
municative Responses and Acts were listed under four sub­
categories: tempo, rhythm, structure/form, and pitch. In the
nonmusical section, operationally defined behaviors were cat­
egorized according to speech production skills, communicative­
interactive skills, and communicative intent skills.
The CRASS contained a total of 107 items, with 91 items in
the musical category and 16 items under the nonmusical cat­
egory. Sixty-nine items were categorized as Communicative
Responses, and thirty-eight items were categorized as Com­
municative Acts.
38 Journal of Music Therapy

Time interval sampling was used, with one lo-minute interval

randomly selected prior to each 30-minute session. The sessions
were videotaped for data collection purposes. During the one
lo-minute interval, two observers independently recorded the
communicative behaviors of each child using the CRASS. The
observers were senior undergraduate music therapy students.

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

A check was given for each behavior observed, with a maximum
of one check recorded for each behavior, even if that particular
behavior was repeated. The checks were then tallied, resulting
in a total Communicative Responses/Acts score per subject per
session. The second observer served as a reliability check
throughout the study.
Interobserver agreement was calculated for both occurrences
and nonoccurrences using the following formula: agreements
divided by the sum of agreements and disagreements. Inter­
observer reliability for occurrences ranged from 75% to 100%,
with a mean of 86.2%. For nonoccurrences, interobserver agree­
ment ranged from 77% to 100%, with a mean of 94%.
The second measurement device used was the Behavior
Change Survey, which was given to parents, teachers, and speech
therapists for each subject immediately following the conclusion
of the study. There were five questions related to communi­
cative behaviors, six questions related to social/emotional be­
haviors, and two questions related to musical behaviors. A seven­
point rating scale was used to indicate the number of changes
seen in the subject’s communicative, social/emotional, and mu­
sical behaviors. The numbers, in sequence from “1” to “7,”
represented the following descriptions: much less, somewhat
less, slightly less, same, slightly more, somewhat more, and much
more.

Procedure
A reversal design was used, consisting of the following phases:
(a) intervention, (b) one-session withdrawal of intervention after
a level of consistency in responses was achieved, and (c) rein­
troduction of the intervention.
Each subject was scheduled for one 30-minute session per
week for 10 weeks. Due to illnesses and unforeseen circum­
stances, two subjects were not able to attend all 10 sessions. One
Vol.XXXI, No. 1, Spring. 1994 39

subject attended eight sessions, and the other subject attended

nine sessions.
Intervention consisted of improvisational music therapy, based
on Nordoff and Robbins’ (1977) Creative Music Therapy ap­
proach. The experimenter created music to establish contact
with the child, to enable the child to respond, and to facilitate

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

development of the child’s musical communicativeness. The
experimenter played the piano and/or sang, and each child had
opportunities to play instruments and to sing. The following
two basic principles were followed with all of the subjects: (a)
Each child was treated as competent, and it was assumed that
he/she understood all that was said and was capable of musically
expressing him/herself; and (b) total emotional support was
provided for each subject, with the experimenter remaining as
responsive as possible to each child and conveying acceptance
of him/her.
A hierarchy of musical experiences/activities was provided
as a guiding reference for ongoing decisions made by the ex­
perimenter throughout the intervention sessions. Specific tech­
niques used were decided in the course of the music therapy
sessions, dependent upon the child’s responses, capacities, and
needs. Numerous techniques listed in the hierarchy (114 specific
techniques) were available to the experimenter to allow for
flexibility within each session in creating an atmosphere for the
child in which optimal growth and development could occur.
The experimenter worked freely within the hierarchy of mu­
sical experiences/activities. Many of these techniques were taken
from the book, Creative Music Therapy, by Nordoff and Rob­
bins (1977). A complete copy of the hierarchy is available upon
request.
The first intervention phase continued until consistency in
responses was noted. Due to the fact that all subjects’ measured
responses showed an ascending baseline by the sixth session, all
reversal sessions occurred in Session 6. Reversal consisted of the
experimenter playing and singing structured precomposed mu­
sic as opposed to improvised music. During this phase, the
experimenter continued to evoke, maintain, and/or develop the
child’s responses. Gestural invitations, verbal invitations, and
reinforcements remained the same. The use of written music
40 Journal of Music Therapy

added additional materials and decreased the amount of eye

contact during the reversal phase.
During the first 10 minutes of the reversal session, preselected
music therapy activity songs were played and sung. The songs,
chosen prior to implementation of the study, provided oppor­
tunities for each child to respond in all of the areas listed in

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

the CRASS. Songs used during reversal included “I Have a Song
to Sing” (Cross, 1989), “Charlie Knows How to Beat the Drum”
(Nordoff & Robbins, 1962), “Drum Talk” (Nordoff & Robbins,
1968b), “3/4 and Strong” (Dubesky, 1982), and “It’s Music”
(Dubesky, 1989). These songs were then repeated during the
lo-minute data collection interval, which was randomly chosen
prior to the session.
Following the reversal, intervention was continued as ex­
plained above for the remaining sessions.

Results
Figure 1 shows group mean Communicative Responses/Acts
for each session. An overall increase in total scores was noted
for the group as a whole, along with an abrupt decrease in the
total group mean score during the reversal (Session 6). Figures
2-12 show individual graphs for each subject.
These figures reveal individual differences in the total num­
ber of Communicative Responses/Acts and in the degree of
improvement in the CRASS scores over the 10 sessions; however,
an overall trend was demonstrated, showing an increase in the
CRASS scores during both intervention phases and a decrease
in these scores during reversal for each individual.
In the initial intervention phase, the level of change for all
subjects was in an improving direction ranging from 8 to 40
points, with a mean of 18.3. Withdrawal of the intervention
resulted in an abrupt and substantial decrease in the quantity
of Communicative Responses/Acts. Decreasing level changes
ranged from 9 to 37 points, with a mean of 19.3. This level was
reversed immediately upon reintroduction of treatment pro­
cedures. Increasing level changes from the reversal to the re­
instatement of intervention ranged from 10 to 43 points, with
a mean of 22.6 During the second intervention phase, the level
of change was in an improving direction and ranged from 6 to
17 points, with a mean of 11.3.
vol. xxxI, NO.1, spring, 1994 41

Group Mean Scores

Responses

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Sessions
Figure 1
Group Mean Communicative Responses/Acts across 10 Sessions

A positive acceleration trend was noted in both intervention

phases for all 11 subjects. Trend stability within conditions was
determined for both intervention phases for each subject. Cri­
teria for trend stability was set at 80% of the data points falling
within 15% along the trend line (Tawney & Gast, 1981). Ail 22
conditions showed a stable trend.
The Wilcoxon Matched-Pairs Signed-Ranks Test was used to
 Subject A

70 i

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Figure 2
Communicative Respones/Acts of Subject A across 10 Sessions.

determine if a significant difference existed between subjects’

scores of their first and last sessions. Figures 2-12 show that all
of the subjects’ last session scores were greater than their first
session scores. Consequently, the differences between the scores
were significant at the .01 level (T = 0).
Taking into account that one of the characteristics of autistic
children is resistance to change, questions emerged concerning
the validity of the first session scores. Therefore, a statistical
 vol. xxxi, NO. 1, spring, 1994 43

Subject B
Responses

70

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

I
4 5 6 7 6 9 10
Sessions
Figure 3.
Communicative Responses/Actsof SubjectB across 10Sessions,

analysis was computed to determine whether a significant dif­

ference existed between the subjects' third session scores and
their last session scores. Using the third session scores instead of
the first session scores in the analysis was based on the assump­
tion that, by the third session, the subjects were not viewing
music therapy as a change in their routine. Therefore, the third
session scores may have been more accurate in portraying the
communicative abilities of the children at the beginning of the
44

Subject C

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

30 I

1 2 3 4 5 6 7 6 9 10
Sessions
Figure 4
Communicative Responses/Acts ofSubjects C across 10 sessions

Study. Significance was achieved at the .01 level (T =0), thus

supporting the original analysis completed
Figure 13 shows group mean scores in each of the four musical
subcategories of the CRASS across 10 sessions; Figure 14 shows
the group mean scores in each of the three nonmusical subcate­
gories of the CRASS across 10 sessions. Statistical analyses were
applied to the subjects’ first and last session scores in all of the
subcategories of the CRASS. The Wilcoxon Matched-Pairs
vol. XXXI, NO.1, spring. 1994 45

Subject D

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

10

Sessions
Figure 5.
Communicative Responses/Acts of Subject D across 10 Sessions.

Signed-Ranks Test indicated significant differences at the .01

level between first session scores and last session scores for tempo
(T = 0), rhythm (T = 0), structure/form (T = 0), pitch (T =
0), speech production (T = 0), and communicative-interactive
(T = 0). Significant differences at the .05 level were found
between first session scores and last session scores for commu­
nicative intent (T = 2.5).
A Spearman Rank Correlation Coefficient was calculated be-
 Subject E

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Figure 6
Communicative Responses/Acts of Subjects Eacross 10 Sessions

tween the musical vocal behavior gains and the nonmusical

speech production gains as recorded on the CRASS. Table 1
shows each subject’s total point gain in both of these categories.
The coefficient corrected for ties was .645, which was significant
at the .05 level (t = 2.532). These results indicate that, as musical
vocal behaviors increased, nonmusical speech production be
haviors also increased.
The Behavior Change Survey was completed by 11 parents,
Vol. XXXI, No. 1, Spring, 1994 47

Subject F
Responses

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

1 2 3 4 5 6 7 8
Sessions
FIGURE 7.
Communicative Responses/Acts of Subject F across 8 Sessions.

4 teachers, and 2 speech therapists. Thirty-three surveys were

distributed and 29 were returned (58% return rate). Table 2
shows the means for each of the three categories as answered
by the parents, teachers, and speech therapists. Most of the
means fell between 4, which indicated no change, and 5, which
indicated a slight change. Overall, the parents gave the highest
ratings (M = 4.8), followed by the teachers (M = 4.7), and
finally the speech therapists (M = 4.2). Changes were seen in
all three categories by parents, teachers, and speech therapists.
48

Subject G
Responses
70 1

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

60

Sessions
FIGURE 8.
Communicative Responses/Acts of Subject G across 10 Sessions.

The highest rankings were given in the musical category (M =

4.9). Both the communication and the social/emotional cate­
gories received a mean of 4.5.
The Spearman Rank Correlation Coefficient was used to de­
termine whether there were any correlations between (a) gains
in CRASS scores and parent Behavior Change Survey ratings,
(b) gains in CRASS scores and teacher Behavior Change Survey
Vol. XXX/, No. 1, Spring, 1994 49

Subject H
I*+ Responses
70
7
I

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

FIGURE9.
Communicative Responses/Actsof Subject H Across 10 Sessions.

ratings, and (c) gains in CRASS scores and speech therapist

Behavior Change Survey ratings.
Table 3 shows each subject’s gain in CRASS scores and his/
her total ratings obtained from the Behavior Change Survey.
A significant correlation was found between the gains in CRASS
scores and the parent ratings. The rho was ,773, which obtained
significance at the .01 level (t = 3.658). The rho corrected for
ties for gains in CRASS scores and teacher ratings was .217 and
did not reach significance. The Spearman Rank Correlation
50

Subject I

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

10 hi

1
Sessions
Figure 10
Communicative Responses/Acts ofSubject Iacross 10 Sessions.

Coefficient, corrected for ties for gains in CRASS scores and

speech therapist ratings, was .387 and did not obtain signifi­
cance.

Discussion
Results of this study suggest that improvisational music ther­
apy is effective in eliciting and increasing communicative be­
haviors in autistic children within a musical setting. These re-
 Vol. XXXI, No. 1, Spring. 1994

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

123458789

Sessions

Figure 11
Communicative Responses/Acts of Subject, Jacross
9 Sessions

sults support numerous case studies and clinical experiences

which suggest the effectiveness of improvisational music ther­
apy (Alvin & Warwick, 1992; Hollander & Juhrs, 1974; Nordoff
& Robbins, 1964, 1968a, 1971, 1977; Saperston, 1973). This
study differs from current research available in the area of
improvisational music therapy and communicativeness in au­
tistic children in that objective methods of control, observation,
and data reporting were applied.
 Subject K
1.3’Responses

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Figure 12
Communications Responses/Acts of Subject K across 10 Sessions.

Results of this study show that autistic children can make

gains in communication when participating in a low-structured
intervention. These findings are contradictory to current lit­
erature and research, in which structured approaches are fre­
quently recommended. Do autistic children need more oppor­
tunities to experience spontaneity and creativity successfully?
The improvisational approach not only allows for spontaneity
and flexibility, but also allows for successful experiences. Within
Vol. XXXI, No. 1, Spring, 1994

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

FIGURE 13.
Group Mean Scoresin Tempo, Rhythm, Structure/Form, and Pitch across 10
Sessions.

this spontaneity, music provides for sufficient predictability to

give the child the amount of support he/she needs.
Significant differences were noted between the number of
each of the four musical communicative modalities used by the
autistic children in their first sessions and the number used in
their last sessions. In comparing these four modalities (tempo,
rhythm, structure/form, and pitch), the group as a whole used
54 Journal of Music Therapy

Nonmusical Subcategories of CRASS

10

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

6

12 3 4 5 6 7 8 9 10
Sessions
Figure 14.
Group Mean Scores in Speech Production, Cummunicative-interactive, and
Communicative Intent across 10 Sessions.

tempo most frequently in both the first and the last sessions.
Also, the largest point gain from the first to the last session was
noted in tempo for the group as a whole (M = 9.9). The other
three modalities, in order of group mean decreasing point gains,
were pitch (M = 7.7). rhythm (M = 5), and form (M = 4.6).
One interpretation of these data concerns the rhythmic re­
petitive behaviors characteristic of autistic children. Colman
 Vol. XXXI, No. 1, Spring, 1994 55

TAble 1
CRASS Musical Vocal Behavior Gains and Nonmusical Speech Production
Gains

Subjects
A B c D E F G H I J K Means
Vocal 23 26 9 32 21 8 19 12 36 11 3 18.2

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Speech 2 3 0 2 2 1 3 1 7 5 1 2.5
Note. CRASS = Checklist of Communicative Responses/Acts Score Sheet.

and his colleagues (1976) assert that there is a stability in the

frequency at which repetitive behaviors occur. The modality
of tempo consisted of beating/vocalizing in a steady tempo,
matching the experimenter’s tempo, and beating/vocalizing
and matching tempo variations. Because of the fundamentally
rhythmic behaviors of autistic children, tempo may initially be
one communicative modality in which autistic children can
immediately experience success.
Thaut (1980) suggested the possibility of rhythm being ab­
sorbed on a physiological level and bypassing the cognitive
deficits of autistic children. His definition of rhythm encom­
passed both the tempo and the rhythmic modalities measured
in this study. This could provide one possible explanation for
the high levels of communicativeness found in the tempo mo-

Table 2
Mean Scores for Behavior Change Survey Categories as Rated by Parents
Teachers, and Speech Therapists

Communicative Behaviors
4.5 4.2
Social/Emotional Behaviors
4.7 4.2
Musical
5.1 4.4
Overall
4.7 4.2
56 Journal of Music Therapy

Table 3
CRASS Gain Scores and Total Behavior Change Survey Ratings Work-up for
the Spearman Rank Correlation Coefficient

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Means 32.8 62.9 60.6 55.1
Note. CRASS = Checklist of Communicative Responses/Acts Score Sheet. A
dash indicates no score was available.

dality. However, as noted subsequently, subjects demonstrated

minimal increases in the rhythm modality.
Another interpretation of the increased amount of commu­
nicativeness found in the tempo modality concerns interactional
synchrony. Condon (1976) reported that listeners move in exact
synchronous relationships with speakers. In researching autistic
children, however, Condon found that this synchrony is dis­
torted (1975). The subjects in the present study were able to
synchronize their drum beating with the ongoing music to vary­
ing degrees. One could posit that this synchrony facilitated
communicative interaction through the music. Also, the music
was synchronized with the children’s repetitive movements and
vocalizations, using their levels of intensity, their rhythms, and
their tempi. This could have created a sense of awareness, sense
of control over their environment, and a new means of corn­
munication.
In the modality of pitch, gains ranging from 3 to 15 points
were noted. All of the items in this modality were vocal re­
sponses. It is significant that all 11 subjects made gains in this
modality.
Vol.XXX/, No. 1, Spring, 1994 57

Overall gains were also noted in rhythm and structure/form.

However, the gains were smaller than with the previous two
modalities discussed. Both modalities demand more cognitive
involvement as compared to tempo. Perhaps there is a connec­
tion between the cognitive deficiencies found in autistic chil­
dren and the modalities of rhythm and structure/form.

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

The significant correlation coefficient obtained between the
musical vocal behavior gains and the nonmusical speech pro­
duction gains (rs = .645, p < .05) indicates that, as musical
vocal behaviors increased, nonmusical speech production be­
haviors also increased on the average. It has been stated that
communication through music bypasses the speech and lan­
guage barriers of autistic people; this could be one possible
explanation for the observed increases in musical vocal behav­
iors. However, the significant relationship found between in­
creases in musical vocal skills and increases in speech production
skills leads to the question as to whether there is a cause-and­
effect relationship. Further research is needed to examine this
question.
Results of the Behavior Change Survey indicated change in
subjects’ behaviors; however, the change was minimal. Parents
and teachers reported more changes than did speech therapists
in all three categories. This could be due to the limited amount
of time the speech therapists see the subjects as compared to
the teachers and parents. Also, the placebo effect must be taken
into consideration when interpreting these data. It is possible
that the changes observed in the subjects’ behaviors could have
been attributed to changes in the parents’ and teachers’ attitudes
and expectations of the subjects since they were aware of the
purpose of this study.
The musical category was the highest rated category overall.
This could have been influenced by knowledge of the subjects’
participation in the music therapy research study. It is possible
that, due to this knowledge, parents, teachers, and speech ther­
apists became increasingly aware of the subjects’ attraction to
musical stimuli and demonstration of musical behaviors.
A significant correlation was found between the subjects’
CRASS gains and the parents’ total Behavioral Change Survey
ratings, which indicates that, on the average, parents of subjects
who demonstrated the most CRASS gains rated their children
58 Journal of Music Therapy

higher on the Behavior Change Survey than did the other par­
ents. The question of possible generalization or transfer of learn­
ing from one setting to another emerges from this finding.
The teachers’ and speech therapists’ total ratings were not
significantly correlated with subjects’ CRASS gains. One possible
reason for these differences might be that teachers and speech
therapists use a more structured setting than do parents. Within

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

the structured environment, opportunities for subjects to dem­
onstrate spontaneity and use new skills may be limited.
If this study is replicated, some changes should be considered.
First, a larger number of subjects would increase the validity
of the study. Also, a male:female ratio which is proportional
to the actual ratio found in autism, which ranges from 1.4 to
4.8:1 (Gillberg, 1989), would help increase the generalizability
of the results. Uncontrolled factors, including the various set­
tings of the study and the decrease in eye contact and increase
in materials during reversal, should be controlled in future
studies. Another suggestion would be to increase the total num­
ber of sessions for each subject, which would allow for more
sessions during the reversal phase.
One change in the CRASS is also recommended for future
research. An increase in the number of items in the nonmusical
category would provide a more comprehensive look at the com­
municative behaviors of autistic children in music settings and
would balance the nonmusical and musical sections of this mea­
surement device.
Future research studies in improvisational music therapy could
be designed to study both the effects of specific techniques
within improvisational music therapy and autistic children’s
specific responses in the various musical communicative mo­
dalities. Other studies could focus on pragmatic aspects of au­
tistic children’s communication in an improvisational music
therapy setting, identifying and comparing the specific com­
municative functions in both singing and speaking contexts.
Further experimentation might increase understanding of the
communication deficits of autistic children. This increased
knowledge could facilitate the development of intervention pro­
grams in which autistic children could express themselves and
experience the joys of communication.
Vol. XXXI, No. 1, Spring. 1994 59

References
Abrahamsen, E. P., & Mitchell, J. R. (1990). Communication and sensorimotor
functioning in children with autism. Journal of Autism and Developmental
Disorders, 20, 75-85.
Alvin, J., & Warwick, A. (1992). Music therapy for the autistic child (2nd
ed.). London: Oxford Press.
American Psychiatric Association. (1987). Diagnostic and statistical manual
of mental disorders (3rd. ed., rev.). Washington, D.C.: Author.

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Applebaum, E., Egel, A. L., Koegel, R. L., & Imhoff, B. (1979). Measuring
musical abilities of autistic children. Journal of Autism and Developmental
Disorders, 9, 279-285
Attwood, A. (1984). The gestures of autistic children. Unpublished doctoral
thesis, University of London.
Bagley, C., & McGeein, V. (1989). The taxonomy and course of childhood
autism. Perceptual and Motor Skills, 69, 1264-1266.
Ball, J. (1978). A pragmatic analysis of autistic children’s language with
respect to aphasic and normal language development. Unpublished doc­
toral dissertation, Melbourne University, Melbourne, Australia.
Baltaxe, C. A. M.. Simmons, J, A., & Zee, E. (1984). Intonation patterns in
normal, autistic and aphasic children. In A. Cohen & M. van de Broecke
(Eds.), Proceedings of the Tenth International Congress of Phonetic SCi­
ences (pp. 713-718). Dordrecht, The Netherlands: Foris Publications.
Biklen, D. (1990). Communication unbound: Autism and praxis. Harvard Ed­
ucational Review, 60, 291-314.
Brigance, A. H. (1978). Brigance diagnostic inventory of early development.
Boston, MA: Curriculum Associates.
Brown, J. L. (1963). Fallow-up study of preschool children of atypical devel­
opment (infantile psychosis): Later personality patterns in adaptation to
maturational stress American Journal of Orthopsychiatry, 33, 336-338.
Burleson, S. J., Center, D. B., & Reeves, H. (1989). The effect of background
music on task performance in psychotic children. Journal of MusicTher­
apy, 26, 198-205.
Bzoch, K., & League, R. (1970). Receptive-expressive emergent language scale.
Gainesville, FL: Tree of Life Press.
Cecchi, V. (1990). Analysis of a little girl with an autistic syndrome. Inter­
national Journal of Psycho Analysis, 71(3), 403-412.
Clarizio, H., & McCoy, G. F. (1983). Behavior disorders In children. New
York: Harper Bow.
Colman, R. S., Frankel F., Ritvo, E., & Freeman, B. J. (1976). The effects of
Fluorescent and incandescent illumination upon repetitive behaviors in au­
tistic children. Journal of Autism and Childhood Schizophrenia, 6, 157­
162.
Condon, W. S. (1975). Multiple responses to sound in dysfunctional children.
journal of Autism and Childhood Schizophrenia, 5, 37-56.
Condon, W. S. (1976). An analysis of behavioral organization. Sign Language
Studies, 13, 285-318.
60 Journal of Music Therapy

Cross,R.(1989).“I havea songto sing.”In Movin’ along: A treasuryofmusic

activitiesfor growth (p. 10),Compiledby Michigan Music Therapists.
Dawson, G., & Adams, A. (1964). imitation and social responsiveness in autistic
children. Journal of Abnormal Child Psychology, 12, 209-226.
DeMyer, M. K. (1979). Parents and children in autism. Washington, D.C.: V.
H. Winston
DeMyer, M. K., Barton, s., DeMyer, W. E., Norton, J. A., Allen, J., & Steel, R.
(1973). Prognosis in autism: A follow-up study. journal of Autism and

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Childhood Schizophrenia, 3, 199-246.
Duhesky, C. L. (1982). "3/4 and strong.” Unpublished sung.
Dubesky, C, L. (1989). “It’s music.” In Movin’ along: A treasury of music
activitiesfor growth (p. 14). Compiled by Michigan Music Therapists.
Eisenberg, L., & Kanner, L. (1956). Early infantile autism 1943-55. American
Journal of Orthopsychiatry, 26, 556-566.
Farmer, R. (1963). A musical activities program with young psychotic girls.
American Journal of OccupationalTherapy, 17, 116-119.
Frith, U. (1989). Autism. Cambridge, MA: Basil Blackwell.
Gillberg, C. (Ed.). (1989). Diagnosis and treatment of autism.New York:
Plenum.
Gillberg, C., & Steffenburg, S. (1987). Outcome and prognostic factors in
infantile autism and similar conditions: A population-based study of 46
cases followed through puberty. Journal of Autism and Developmental
Disorders, 17, 273-287.
Goldstein, C. (1964). Music and creative arts therapy for an autistic child.
Journal of Music Therapy, 1, 135-138.
Hullander, F. N., & Juhrs, P. D. (1974), Orff-Schulwerk, an effective treatment
tool with autistic children. Journal of Music Therapy, 11, 1-12.
Intellectual Disability Review Panel. (1989). Investigation into the reliability
and validity of the assisted communication technique. Melbourne, Aus­
tralia: Department of Community Services, Victoria.
Kanner, L. (1943). Autistic disturbances of affective contact. Nervous Child,
2, 217-250.
Kanner, L. (1946). Irrelevant and metaphorical language in early infantile
autism American Journalof Psychiatry, 103, 242-246.
Kolko, D. J., Anderson. L., & Campbell, M (1980). Sensory preference and
overselective responding in autistic children. Journal of Autism and De­
velopmental Disorders, 10, 259-271.
Krug, D., Arick, J., & Almond, P. (1979). Autism screening instrument for
educational planning. Portland, OR: ASIEP Educational.
Kubicek, L. F. (1980). Organization in two mother-infant interactions involving
a normal infant and his fraternal twin brother who was later diagnosed as
autistic. In T. M. Field, S. Goldberg, D. Stern, & A. M. Sostek (Eds.), High­
risk infants and children: Adult and peer interactions. New York: Aca­
demic Press.
Kurita, H. (1985). Infantile autism with speech loss before the age of thirty
months. Journal of The American Academy of Child Psychiatry, 24, 191­
196.
Vol. XXX/, No. 1. Spring, 1994 61

Landry, S. H., & Loveland, K. A. (1989). The effect of social content on the
functional communication skills of autistic children. Journal of Autism and
Developmental Disorders, 19, 283-299.
Mahlberg, M. (1973). Music therapy in the treatment of an autistic child.
Journal of Music Therapy, 10, 184-188.
Miller, S. B., & Toca, J. M. (1979). Adapted melodic intonation therapy: A
case study of an experimental language program for an autistic child.
Journal of Clinical Psychiatry, 40, 201-203.

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

National Society for Autistic Children (1978). National Society for Autistic
Children definition of the syndrome of autism. Journal of Autism and
Childhood Schizophrenia, 8, 162-167.
Nordoff, P., & Robbins, C. (1962). The first book of children’s play-songs.
Bryn Mawr, PA: Theodore Presser.
Nordoff, P., & Robbins, C. (1964). Music therapy and personality change in
autistic children. Journal of the American Institute of Homeopathy, 57,
305-310.
Nordoff, P., & Robbins, C. (1968a). Improvised music as therapy for autistic
children. In E. T. Gaston (Ed.), Music in therapy (pp. 191-193). New York:
MacMillan Publishers.
Nordoff, P., & Robbins, C. (1968b). The second book of children’s play-songs.
Bryn Mawr, PA: Theodore Presser.
Nordoff, P., & Robbins, C. (1971). Therapy in music for handicapped children.
New York: St. Martin’s Press.
Nordoff, P., & Robbins, C. (1977). Creative music therapy. New York: John
Day.
Prizant, B. M., & Duchan, J. F. (1981). The functions of immediate echolalia
in autistic children. Journal of Speech and Hearing Disorders, 48, 241­
249.
Prizant, B. M., & Rydell, P. (1984). An analysis of the functions of delayed
echolalia in autistic children. Journal of Speech Hearing Research, 27,
183-192.
Rimtand, B. (1964). Infantile autism. New York: Appleton-Century.
Ruttenberg, B. A., Dratman, M. L, Fraknoi, J., & Wenar, C. (1966). An in­
strument for evaluating autistic children. Journal of the American Acad­
emyof Child Psychiatry, 5, 453-478.
Rutter, M. (1978). Developmental issues and prognosis. In M. Rutter & E.
Schopler (Eds.), Autism: A reappraisal of concepts and treatment (pp.
497-505). New York: Plenum Press.
Saperston, B. (1973). The use of music in establishing communication with an
autistic mentally retarded child. Journal of Music Therapy, 10, 184-188.
Saperston, B. (1982). Case study: Timmy. In D. W. Paul (Ed.), Music therapy
for handicapped children: Emotionally disturbed (pp. 42-57). Washington
D.C.: Office of Special Education and the National Association for Music
Therapy.
Schopler, E., & Mesibov, G. (Eds.) (1985). Communication problems in autism.
New York: Plenum Press.
Shapiro, T., Chiarandini, I., & Fish, B. (1974). Thirty severely disturbed chil­
dren. Archives of General Psychiatry, 30, 819-826.
62 Journal of Music Therapy

Sherwin, A. C. (1953). Reactions to music of autistic (schizophrenic) children.

AmericanJournalof Psychiatry, 109, 823-831.
Staum, M. J., & Flowers. P. J. (1984). The use of simulated training and music
lessonsin teaching appropriate shopping skills to an autistic child. Music
Therapy Perspectives, 1(3), 14-17.
Stevens, E.. & Clark, F. (1969). Music therapy in the treatment of autistic
children. Journal of Music Therapy, 6, 98-104.

Downloaded from http://jmt.oxfordjournals.org/ at Florida Atlantic University on November 18, 2014

Stillman, R. (Ed.) (1978). The Callier-Azusa scale. Unpublished manuscript,
The University of Texas at Dallas.
Stone, W. L., & Lemanek, K. L. (1990). Parental report of social behaviors in
autistic preschoolers. Journal of Autism and Developmental Disorders, 20,
513-522.
Tager-Flusberg, H. (1985). Psycholinguistic approaches to language and com­
munication in autism. In E. Schopler & G. B. Mesibov (Eds.). Communi­
cation problems in autism (pp. 69-88). New York: Plenum Press.
Tawney, J. W., & Gast, D. L. (1984). Single subject research In special edu­
cation. Columbus, OH: Charles E. Merrill.
Thaut, M. (1980). Music therapy as a treatment tool for autistic children.
Unpublished master’s thesis, Michigan State University, East Laming.
Uzgiris, I., & Hunt. J. (1975). Assessmentin infancy: Ordinal scales of psy­
chological development. Urbana University of Illinois Press.
watson, D. (1979). Music as reinforcement in increasing spontaneous speech
among autistic children. Missouri Journal of Research in Music Education,
4(3), 8-20.
Watson, L. B. (1985). The TEACCH communication curriculum. In E. Schop­
ler & G. G. Mesibov (Eds.), Communication problems in autism (pp. 187­
206). New York: Plenum Press.
Wetherby, A. M., & Prutting, C. A. (1984). Profiles of communicative and
cognitive-social abilities in autistic children. Journal of Speech and Hearing
Research, 27, 364-377.