The Arts in Psychotherapy 39 (2012) 333341

Contents lists available at SciVerse ScienceDirect

The Arts in Psychotherapy

Music therapy techniques as predictors of change in mental health care

Karin Mssler, Dr. sc. mus. a, , Jrg Assmus, PhD b , Tor Olav Heldal, MA c ,
Katharina Fuchs, Mag. art. d , Christian Gold, PhD b
a
GAMUT, University of Bergen, Lars Hilles Gate 3, 5015 Bergen, Norway
b
GAMUT, Uni Health, Uni Research, Postboks 7810, 5020 Bergen, Norway
c
Nordfjord Psychiatric Centre, Frde Health Trust, Sjukehusvei 14, 6770 Nordfjordeid, Norway
d
State Psychiatric Clinic Wagner-Jauregg, Wagner Jauregg Weg 15, 4020 Linz, Austria

a r t i c l e i n f o a b s t r a c t

Keywords: The application of music in therapy is realised through different working modalities which can be cat-
Music therapy techniques egorised into three types of techniques: production, reception, and reproduction. These techniques are
Reproducing music commonly used in mental health settings in music therapy practice and previous research suggests that
Outcome predictors
specic working modalities might be important predictors of change in music therapy. However, little
Therapy motivation
is known about which ingredients specically contribute to the outcomes of music therapy. This study
Process-orientation
aimed to investigate the application of music therapy techniques and whether they predict changes
in clinical outcomes in mental health settings with individuals displaying a low therapy motivation.
Participants (N = 31) were assessed before, during, and after participating in individual music therapy.
Music therapy techniques were assessed for three selected therapy sessions per participant. Associations
between music therapy techniques and outcomes were calculated using linear models with repeated
measures. Results showed that reproduction techniques were used most intensely. In addition, rela-
tional competencies (interpersonal and social skills) amongst the participants improved when focusing
on reproducing music (e.g. singing or playing familiar songs, learning musical skills). Results indicated
that reproduction music therapy techniques may foster the development of relational competencies in
individuals with low motivation.
2012 Elsevier Inc. All rights reserved.

Individuals aficted with a mental illness may often display
difculty in the areas of emotional responsiveness and social inter-
action, verbally and non-verbally. Challenges in interpersonal and
intrapersonal skills can result in individuals experiencing low moti-
vation for verbal therapies such as psychotherapy or psychological
counselling. To that end, mental health care clients with low
motivation are frequently referred to music therapy, as the pri-
mary medium of interaction is music rather than verbal language
(Hannibal, 2005). Although, music therapy has been indicated as
an effective intervention with regard to fostering motivation, emo-
tional expression, and relatedness (Gold, Mssler, et al., submitted
for publication) additional process-outcome research is needed to
identify the most effective music therapy techniques with this par-
ticular client group.
[ClinicalTrials.gov Identier: NCT00137189].
Corresponding author. Present address: GAMUT, Uni Health, Uni Research, Post-
boks 7810, 5020 Bergen, Norway. Tel.: +47 97479289.
E-mail addresses: karin.moessler@uni.no (K. Mssler), joerg.assmus@uni.no
(J. Assmus), toroheld@gmail.com (T.O. Heldal), katharina.fuchs@gespag.at
(K. Fuchs), christian.gold@uni.no (C. Gold).
In music therapy, musical experiences and the therapeutic rela-
tionship developing through them are used as dynamic forces
producing therapeutic change (Bruscia, 1998). Relational musical
experiences that deepen the therapeutic relationship are fostered
by the use of music therapy techniques which are applied within a
systematic process between the client and therapist. Music therapy
techniques can be understood as working modes offering different
musical and relational experiences. These techniques work in tan-
dem with psychotherapeutic techniques (e.g. mirroring, holding,
confronting) within specic use of musical parameters (e.g. rhythm,
sound, tonality) (Storz, 2000a; Wigram, 2004). Music therapy tech-
niques can be assigned to the following categories: Production
techniques focus on emotional expression and the creation of the
relationship through musical improvisation (e.g. structured, the-
matic, communicative, trying out, free improvisation) in which the
client and therapist create something musically new. Reproduction
techniques involve the client and therapist playing or singing pre-
composed pieces of music as well as learning or practicing musical
skills (e.g. guitar chords, melodies on the piano). They may pro-
vide a holding structure and framework in which the actualisation
of memories can be supported and explored within the context
of relationship. Reception techniques involve the client listening to
live (e.g. music played by the therapist) or recorded music. These

0197-4556/$ see front matter 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.aip.2012.05.002
334 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341
musical experiences may be used to focus on conscious awareness
of the clients current mental state, emerging associations, as well
as to facilitate relaxation or pain management (Storz, 2000a, 2011).
In each of the three categories of music therapy techniques, both
the therapist and client contribute mutually and are engaged in a
therapeutic process within the context of relating and communi-
cating. Thus, the choice of techniques will also be the result of this
mutual process.
Previous research in mental health care has provided supporting
evidence that specic factors in music therapy are of importance.
According to these ndings, techniques focusing on musical com-
munication, expression, and transformation may be of relevance
(Danner & Oberegelsbacher, 2001) when working with clients with
psychosomatic problems in mental health care. Another study
found that music therapy-specic techniques (e.g. free improvisa-
tion, songs) were associated with bigger improvement compared to
less specic techniques not unique to music therapy (e.g. free play,
puppet play) in children and adolescents (Gold, Wigram, & Voracek,
2007). Although psychotherapy research has demonstrated that
specic factors are of little importance to produce psychotherapeu-
tic change (Cooper, 2008; Lambert & Ogles, 2004; Wampold, 2001),
music therapy research is sparse in this area and more research
is needed in order to support or challenge this notion. Outcome
studies in adult mental health care have mainly investigated music
therapy techniques in one of the two ways: (a) either several tech-
niques were applied within individual or group music therapy but
were not statistically tested in terms of their particular impact on
outcomes (De lEtoile, 2002; Tang, Yao, & Zheng, 1994; Yang, Li,
Weng, Zhang, & Ma, 1998), or (b) one main technique applied within
a session was investigated for various outcomes (Grocke, Bloch, &
Castle, 2009; Silverman & Marcionetti, 2004; Talwar et al., 2006).
In terms of external validity, focusing on only one technique when
investigating music therapy may not appropriately reect effec-
tive common clinical practice in mental health care. In this setting
it is more common to implement production, reproduction as well
as reception techniques as they are related to differing therapeutic
topics and goals which may all be vital within the same therapeutic
process (Storz, 2004).
In the present study we aimed to examine whether different
types of music therapy techniques can explain some of the vari-
ance in outcomes of clients in mental health care with low therapy
motivation. In this context, it was important to rst understand
which music therapy techniques are being used within this client
population. Examining music therapy techniques in this eld is of
clinical interest as it may contribute to a better understanding of
how clients can improve their engagement into therapeutic pro-
cesses.
Specically, the main objectives of this study were to explore:
which music therapy techniques are applied within music ther-
apy in clients with mental illness presenting with a low therapy
motivation, and
whether music therapy techniques predict changes in clinical
outcomes related to the development of ego-strength, relational
competencies, and quality of life.
Method
This was an exploratory study using a naturalistic, observa-
tional design with pre, post and intermediate tests of clients who
began music therapy. It used data from an international multicen-
tre randomised controlled trial (Gold, Mssler, et al., submitted for
publication; Gold et al., 2005) that had shown positive effects as
well as additional material concerning the use of music therapy
techniques in the same clients. It should be emphasised that the
present study only used data from the experimental group of that
trial. Furthermore, it only used data from those sites where the
additional data were available.
Music therapy in this study was guided by a manual of resource-
oriented principles (Rolvsjord, Gold, & Stige, 2005). That manual
describes desirable attitudes of the therapist but does not impose
any restrictions on particular techniques. The use of techniques was
allowed to vary from client to client, similar to clinical practice
outside the study, and therefore it was possible to examine their
application and potential impact.
Data collection was carried out at one decentralised psychiatric
centre in Nordfjordeid (Norway) and one psychiatric clinic in Linz
(Austria). Data were collected from a range of different outcome
measures used within the overall multicentre study as well as ther-
apy journals. Therapy journals were used as source for assessing
music therapy techniques quantitatively.
Participants
We included clients with any non-organic mental disorder
according to ICD-10 criteria (F1F6). Within this population, we
included only those who met one or more types of low motivation
(Gold, Mssler, et al., submitted for publication; Gold et al., 2005)
because having difculties to engage in verbal therapy is typical
reason for referral to music therapy (Hannibal, 2005). Specically,
participants had to meet at least one of the following criteria:
The client is lacking or has insufcient insight into illness.
The client has difculties talking about feelings or problems.
The client wants a medication cure, s/he does not believe in
talking.
The client has not achieved sufcient improvement in previous
psychotherapy.
This was based on the judgement of the interdisciplinary team,
i.e. of all professionals directly involved in the clients treatment.
The nal decision to refer a client to the study was made and signed
by the ward clinician, i.e. the person with primary responsibility
of the clients treatment plan. Clients suffering from severe men-
tal retardation or a life-threatening illness were excluded from the
study (Gold et al., 2005).
We initially included 40 participants for the present study
(Austria: 23; Norway: 17). However, due to drop out and result-
ing missing values in outcome assessments, the usable sample size
was reduced to N = 31 (Austria: n = 22; Norway: n = 9).
The most frequent diagnoses were schizophrenia, schizotypal or
delusional disorder (ICD-10: F2, n = 14), affective disorder (F3, n = 8),
and personality disorder (F6, n = 6). The remaining three partici-
pants presented either a neurotic/somatoform disorder (F4, n = 2)
or a mental and behavioural disorder due to psychoactive sub-
stance abuse (F1, n = 1). The sample thus included similar numbers
of participants with psychotic (n = 14) and non-psychotic (n = 17)
disorders. The most frequent types of low motivation were not
having achieved sufcient improvement in previous psychother-
apy (n = 21) and difculties in talking about feelings or problems
(n = 20; overlaps possible). There were 19 male and 12 female
participants. The mean age was 37 with a range from 18 to 59.
Most of the participants (n = 29) received music therapy as inpa-
tients; two were out-patients and one attended a day-clinical
setting.
Participants were offered up to 26 sessions of individual music
therapy over a course of three months in addition to standard
care. The sessions were offered twice a week, lasting 45 min each
(Gold et al., 2005). The average number of music therapy ses-
sions received was 19, ranging from 12 to 25. Music therapy was
conducted by three therapists trained at masters level in music
 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341
therapy. One had graduated from the Sogn og Fjordane University
College, Sandane, Norway, which is a training focused on inter-
action theory and resource-orientation and based on humanistic
perspectives. The two others had graduated from the University of
Music and Performing Arts Vienna, Austria, a training with a focus
on humanistic and psychodynamic theory construction. In both
institutions music therapists were trained in the conscious use of
music as a therapeutic medium in relation to various music therapy
techniques. Ethical approval was granted by the Regional Commit-
tee for Medical Research Ethics in Western Norway (REK VEST) as
well as the Ethics Committee of Upper Austria. All participants gave
written informed consent.
Outcome measures
When selecting outcome measures it is of importance to con-
sider the client population, the therapeutic approach and the goals
that should be achieved within therapy (Gold, in press; Kazdin,
1999, 2001). Therefore, we aimed at selecting a maximum of ve
outcome measures from a larger pool of measures available (Gold
et al., 2005) that we felt were most relevant to our particular
population and the goals of music therapy. According to our clini-
cal experience these clients rarely have concrete ideas about what
should improve or change in their lives when starting therapy. They
typically do not come with an urgent need to talk about problems
or conicts, but are interested in doing something with music.
Their needs and wishes may initially be connected to: (a) a joint
activity such as playing something together; (b) experiencing joy;
(c) experiencing mastery; or (d) experiencing a safe place through
playing or singing familiar songs. Interpreting these wishes, thera-
peutic goals may be formulated as follows: (a) nurturing social and
relational abilities; (b) improving quality of life; (c) supporting self-
esteem and self-efcacy; and (d) gaining self-condence (Mssler,
2011; Mssler et al., 2011). Furthermore, working with music may
be a goal in itself for clients attending music therapy. Considering
these goals, we were most interested in outcomes related to the
development of ego-strengths, relational competences and quality
of life, rather than symptoms or functioning. Therefore, we chose
the following outcomes:
Self-esteem was measured using the Rosenberg Self-Esteem
Scale, a self-report measure containing 10 items with demon-
strated reliability (Rosenberg, 1989).
Self-efcacy was assessed using the General Perceived Self-
Efcacy Scale. This self-report scale includes 10 items. The scale
has demonstrated reliability in clinical as well as in non-clinical
samples (Schwarzer & Jerusalem, 1995).
Interpersonal problems, as a proxy measure for (lack of) relational
competence, were measured with the Inventory of Interpersonal
Problems (short version), a self-report measure with 32 items.
The scale has shown reliability and validity (Barkham, Hardy, &
Startup, 1996).
Actual social relationships were assessed using the 11-item Qual-
ity of Life Enjoyment and Satisfaction Questionnaire. Reliability
and validity have been demonstrated for this scale (Endicott, Nee,
Harrison, & Blumenthal, 1993).
Interest in music measured by the Interest in Music Scale, sub-
scale Musical Activity and Emotional Engagement with Music.
This 10-item scale was especially established for assessing the
clients access to music and how s/he is making use of it for
either being in contact with oneself and others through music,
or express oneself through music. Reliability has been conrmed
for this scale (Gold, Rolvsjord, Mssler, & Stige, 2012).
In all scales except the Inventory of Interpersonal Problems, a
high score represents a favourable outcome. Measures were taken
335
before inclusion into the study as well as after one month (interme-
diate test, early during music therapy) and three months (post-test,
end of music therapy).
Music therapy techniques
After each music therapy session, the music therapist wrote a
detailed report of the session using a semi-structured format. Guid-
ance was given for what type of information the journals should
contain, but without restricting them in any way. Because of this
format, the journals contained all the relevant information from
which data about the use of music therapy techniques could be
extracted. Under the journal headings therapy process and tech-
niques information were provided about which techniques had
been used and how much the client had been engaged in the partic-
ular working mode. From the available pool of data for all sessions,
we selected those sessions that were closest to the time points
of outcome measurement. For each participant, the rst session,
the session prior to intermediate test (Mdn = 8th session, range:
215) and prior to post-test (Mdn = 18th session, range: 1125)
were analysed. To categorise and rate therapy techniques we used
the Questionnaire for the Assessment of Music Therapeutic Work-
ing Modes (Mssler, 2008a; Mssler & Oberegelsbacher, 2004). This
scale was originally developed for evaluating music therapeutic
working modes in psychiatry, psychosomatic, and children with
special needs based on the music therapy practice in Austria. It rep-
resents the use of each technique on a ve-point Likert scale and
was chosen because it reects the broad spectrum of production,
reproduction and reception techniques.
After discussing the usefulness of the scale with all music ther-
apists collaborating on this study, the scale was slightly revised for
improved applicability in the context of this study. One technique
that was more relevant to the work with children (situation songs)
was removed from the scale. We added one technique (learning
or practicing musical skills), which was important reecting our
clinical practice with this population. Furthermore, the original
scale suggested a separation of vocal improvisation techniques and
improvisation techniques in general. This separation was viewed as
redundant and we therefore merged the vocal improvisation tech-
niques into the improvisation techniques. The nal scale included
11 items with a ve point intensity each (0 = not at all, 4 = very
much). Intensity was dened primarily by the amount of time
spent on each technique, but also taking into account the degree
of engagement with it. We also included an additional question
about other techniques. This category was never used and there-
fore not relevant for the analysis. The following techniques were
represented in the nal scale.
The production subscale included ve techniques: structured,
thematic, communicative, trying-out, and free improvisation.
Singing or playing pre-composed songs, and learning or practic-
ing musical skills, were the two items of the reproduction subscale.
The four items of the reception subscale were listening to live
music, listening to recorded music, listening in combination with
body perception, or with movement and dance. Although the tech-
niques included in this scale do not claim completeness, they
were considered as comprehensive and valuable according to the
study population as well as the cross-cultural implementation of
the study. More detailed information on the scale is provided in
Appendix 1.
The music therapists who had conducted the sessions rated
the items of the scale on the basis of their reexive therapy jour-
nals. For gaining consistency, the music therapists were trained in
developing a common understanding of the items and their de-
nitions (Appendix 1). A shared understanding was established by
working on concrete examples. Raters were asked to assess the
intensity with which each technique was applied within a session.
336 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341

We relied on therapist self-ratings because we expected the music Table 1

Frequency and effective intensity of music therapy techniques.
therapists who had conducted the sessions to be able to provide
a more accurate account of the techniques used than independent Technique Frequency Effective
observers who would read and interpret the journal entries with- N (%) intensity Rate
out any personal recollection of the session. Therefore, we refrained Production techniques 49 (52.7) 0.76
from adding independent observer ratings or assessing inter-rater Structured improvisation 29 (31.2) 1.97
reliability for any such ratings. Thematic improvisation 11 (11.8) 2.00
Communicative improvisation 24 (25.8) 1.88
Trying out improvisation 1 (1.1) 1.00
Data analysis music therapy techniques Free improvisation 29 (31.2) 2.07
Reproduction techniques 58 (62.4) 1.92
The rst aim of the statistical analysis was to describe the Singing or playing pre-composed songs 49 (52.7) 2.86
Learning or practicing musical skills 35 (37.6) 2.37
application of the different music therapy techniques and the
Reception techniques 26 (28.0) 0.67
three subscales (production, reproduction, reception), respectively. Listening to live music (therapist plays) 12 (12.9) 1.67
Techniques and subscales were analysed in terms of their frequency Listening to recorded music 19 (20.4) 2.26
of appearance and their effective intensity which were dened as Listening to music and body perception 3 (3.2) 2.33
follows. Listening to music and dance/movement 0 (0.0) 0.00

We dened the frequency of a music therapy technique as the

number of sessions (of the three selected) where the technique was
used. Correspondingly, the frequency of a subscale was dened as
the number of sessions where at least one of the techniques in that
category was used. That is, any session was only counted once, so
that it did not matter how many of the techniques belonging to the
subscale were used in the session.
In contrast, we dened the effective intensity of a music therapy
technique as the mean of the intensities of those sessions where the
technique was used. (Recall from above that intensity was the score
from 0 to 4 in a given session.) This is, for a single technique, the sum
of all intensities divided by the frequency of the technique. Corre-
spondingly, for a subscale it is the sum of the intensities divided by
the frequency and the number of techniques belonging to the sub-
scale. In other words, effective intensity always ranged from 0 to 4,
just like the original items, but represented how much a technique
was used if and when it was used.
We used nonparametric statistical tests, Chi-squared for fre-
quencies and Wilcoxon for effective intensities, to test whether
there was a different use of the techniques between clients with
a psychotic versus a non-psychotic disorder. Because the three
subscales were of equal importance to us, we used Bonferroni
adjustment to correct for multiple testing (i.e. we regarded effects
as signicant when p .05/3 = .0167).
Data analysis outcome predictors
We used a linear model with repeated measures (Generalised
Equations Estimation, GEE) (Verbeke & Molenberghs, 2009) to
account for the time dependence of the measures (i.e. measures
for the same participant at more than one time points) and to
control for important confounding variables. The change scores
in outcomes were the dependent variables. The intensities of the
three subscales (i.e. the mean intensities of the techniques belong-
ing to the subscales production, reception, reproduction) were used
as predictors. Age, sex, diagnosis (psychotic/non-psychotic), ther-
apist, and time point were entered as potential confounders. We
calculated two types of models: a fully adjusted model, containing
all predictors and confounders, and an unadjusted model, contain-
ing only one of the predictors (i.e. we calculated one unadjusted
model for each predictor). While the association between predic-
tors and outcomes is adjusted for confounding in the fully adjusted
models, the unadjusted models show the crude association. Formu-
las of the models are shown in Appendix 2. Here, multiple testing
was accounted for as follows: each combination of approaches
(Appendix 2) was implemented for each outcome both in the fully
adjusted and the unadjusted model, thus there were four com-
binations of approaches and a total of eight coefcient estimates
for each variable and each outcome. Bonferroni adjustment there-
fore led to a highly conservative signicance level of p .0063
(=.05/8). In all linear models, positive coefcients indicated that an
increasing intensity of a music therapeutic technique was associ-
ated with an increasing value for the outcome variable. Therefore,
for all outcomes except interpersonal problems a positive coef-
cient implied that greater use of the technique was associated with
greater improvement, whereas for interpersonal problems this was
reversed. All computations were done in R (version 2.12.1, www.r-
project.org).
Results
Music therapy techniques
Both frequency and effective intensity were highest for the
reproduction techniques and subscale, respectively (Table 1). In
other words, the techniques singing or playing pre-composed songs
and learning musical skills were used most frequently per client,
and if they were used in a session, they were also used most
intensely. The frequency of the production subscale was slightly
lower, and reception techniques were applied least frequently.
The effective intensities were fairly similar between the reception
and production subscale. Within the production subscale empha-
sis was given to structured, communicative and free improvisation,
while the therapists play and recorded music were prominent on
the reception subscale. Some techniques were not or hardly ever
used. One reception technique (listening to music in combination
with dance/movement) was not used at all and another reception
technique (listening to music in combination with body perception)
was only used three times. This is similar for one production tech-
nique (trying-out improvisation) which was only used once. All
frequencies and effective intensities are listed in Table 1. No signif-
icant differences were found between participants with psychotic
and non-psychotic disorders.
Outcome predictors
Linear models with repeated measures were calculated for
the Baseline model (changes from baseline) and the Inter-
cept model (changes between time points; Appendix 2). No
signicant effects of music therapy techniques were found
for the Intercept model. The results of the GEE calculations
for the Baseline model are presented in Table 2, excluding
the outcome variable for self-efcacy as no effects could be
identied for this variable. Music therapy techniques as pre-
dictor variables are highlighted with grey background. From
this table it can be seen that many signicant effects were
found when applying the strict Bonferroni-corrected signif-
icance level (p < .0063; highlighted in bold font), but most
 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341 337

Table 2
Music therapy techniques as outcome predictors: coefcients and p-values of the Baseline model.

Point approach Interval approach

Adjusted model Unadjusted model Adjusted model Unadjusted model

b (SE) p b (SE) p b (SE) p b (SE) p

Self-esteem
Time
Production techniques
Reproduction techniques
Reception techniques 5.91 (2.47) .0165
Age 0.11 (0.04) .0117 0.13 (0.06) .0360 0.12 (0.05) .0217 0.13 (0.06) .0360
Sex: male
Diagnosis 3.61 (1.24) .0037 3.85 (1.70) .0232 3.70 (1.28) .0037 3.85 (1.70) .0232
Therapist 2.96 (0.94) .0015 2.75 (0.77) .0004 2.62 (0.95) .0057 2.75 (0.77) .0004

Interest in music
Time
Production techniques
Reproduction techniques
Reception techniques
Age
Sex: male
Diagnosis
Therapist 3.24 (0.78) .0000 2.43 (0.73) .0009 2.91 (0.81) .0003 2.43 (0.73) .0009

Interpersonal problems
Time
Production techniques
Reproduction techniques 2.58 (1.23) .0360
Reception techniques 7.85 (3.57) .0280 8.46 (3.16) .0075 12.4 (4.3) .0041
Age
Sex: male
Diagnosis
Therapist

Social relationship
Time
Production techniques 2.10 (0.98) .0280 2.16 (0.98) .0286
Reproduction techniques 1.13 (0.45) .0120 1.25 (0.48) .0097
Reception techniques 3.34 (1.55) .0310
Age 0.092 (0.045) .0400 0.092 (0.045) .0400
Sex: male
Diagnosis
Therapist

Note: The music therapy techniques (predictors) are marked by grey background. All non-signicant predictors (p > .05) are removed. Coefcients and p-values are signicant
after the Bonferroni-adjustment (p < .0063). The models are dened in Eqs. (1) and (2).

of them concerned confounding variables (age, diagnosis, and
therapist) that were not of substantial interest for this study.
There was one exception: Reception techniques showed a signi-
cant negative effect on interpersonal problems (b = 2.75, p = .0004),
suggesting that greater intensity in the use of reception tech-
niques was associated with less improvement in interpersonal
problems. Additionally, we found a number of predictor vari-
ables with p-values between the Bonferroni-corrected level and
the conventional signicance level (p < .05), marked by plain num-
bers in Table 2. These cannot strictly be interpreted as signicant
effects but can be seen as tendencies that might indicate use-
ful hypotheses for future research. While changes in self-efcacy,
self-esteem, and interest in music were largely unrelated to the
use of any particular techniques, some tendencies were found
for interpersonal problems and particularly social relationships.
Tendencies of positive effects of reproduction techniques on inter-
personal problems and social relationships could be seen. The
unadjusted models of the Interval approach indicated that social
relationships tended to increase when reproduction techniques
were used more intensely (b = 1.25, p = .0097). Concurrently, inter-
personal problems tended to decrease more when focusing on
reproduction techniques more intensely (b = 2.58, p = .0360). In
contrast, both reception techniques and production techniques
might be associated negatively with improvements in social
relationships (Table 2).
Discussion
The main nding of this study is that reproduction techniques
such as singing or playing pre-composed songs and learning musi-
cal skills (e.g. practicing a melody on the piano) might play an
important role in music therapy when interacting with clients in
mental health care with low therapy motivation. Reproduction
techniques were used more frequently than production and recep-
tion techniques, and if they were used, the amount of time, interest
and motivation spent on this working mode was higher in com-
parison to other techniques. Reproduction techniques might also
play an important role in predicting outcomes related to relational
aspects. In contrast to production and reception techniques, repro-
duction techniques tend to decrease interpersonal problems and
increase social relationships. Getting in contact with others, to feel
comfortable in the presence of other people, to open up and show
feelings to another person, but also to disagree with others when a
situation calls for it are examples of abilities that might increase
more when working actively with familiar and known musical
pieces. Reproducing music might help people in mental health care
338 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341
to build up a stronger relationship with themselves as well as with
others.
Clinical implications
Previous qualitative and quantitative research on music therapy
techniques in mental health care has concentrated on the inves-
tigation of production techniques (improvising, also composing)
(Albornoz, 2011; De Backer, 2008; Erkkil et al., 2011; Grocke,
Bloch, & Castle, 2009; Solli, 2008), which might underline the
importance of these types of techniques in this specic eld. Our
ndings concerning the application of music therapy techniques
have shown that production techniques are of certain relevance.
However, the use of reproduction techniques appeared even more
relevant. Our sample was representative for typical music ther-
apy practice in that it consisted of clients who were hard to reach
verbally, who had difculties articulating their problems or feel-
ings, and who showed a lack of motivation to participate in verbal
psychotherapy. When working with this client population, our
experiences were that participants especially struggled to con-
nect to themselves (i.e. to their inner world) and hence showed
difculties in perceiving and articulating needs, feelings or prob-
lems. Their main interest was at rst to do something with music
without having more psychologically minded ideas about ther-
apy goals or what to change related to their current life situation
(see case examples in Mssler et al., 2011). Our ndings suggest
that the involvement and engagement with therapy for this client
population could be supported by the experience of reproducing
music embedded in a therapeutic relationship. This activity might
provide a holding structure, a secure musical space in which the
client dares to get involved with her/his musical and personal top-
ics (Rolvsjord, 2001). It can be assumed that the structure, provided
by a pre-composed piece, offers those holding and containing qual-
ities clients are seeking when the creation of own (musical) ideas,
wishes or needs are not accessible or perceivable, and therefore
cannot yet be expressed.
As an illustrative example, a participant in this study sang and
practiced the Harry Belafonte song Island in the Sun on instru-
ments many times and eventually was able to describe that this
island represented her husband who had died recently. On the
safe basis of the familiar and repeating song structure, she recog-
nised all her feelings of sadness, longings and despair which were
represented through the song almost long before she was able to
talk about them (the full case history was presented in Mssler,
2008b). Another participant redened her self-image by practicing
pieces on the recorder, which had been a hobby in her youth. She
found access to hidden resources and started to build up new rep-
resentations of herself, rather than focusing on decits and failures
determining her self-image at that time. For a third participant, it
was important to master the rst line of Beethovens Fr Elise on
the piano, and she discovered herself as being self-efcient by being
able to reproduce this piece. Her self-condence improved, and this
enabled her to solve difculties also outside the therapeutic room
(the full case history was presented in Mssler, 2011).
These short glimpses into cases may help to elucidate poten-
tial meanings of reproduction techniques. These techniques may
be understood as music experiences promoting the building up of
a clients self-concept on the basis of musical resources. Through
reproducing music, clients may be able to better connect to their
own ideas and feelings as well as to better understand and adapt
to the ideas and feelings of others while still retaining their own
identity (Bruscia, 1998). It therefore has a resource-activating
(Decker-Voigt, Oberegelsbacher, & Timmermann, 2008) as well as
a reassuring function. The pre-composed music might act as a rep-
resentation into which own feelings can be transferred. Due to the
known or even famous musical piece that has already found social
recognition, clients may feel conrmed in their emotions which
are being reinforced or doubled (Bolterauer, 2006) when repro-
ducing the musical piece. Own emotions may be validated by the
pre-composed substitute which might also act as a transitional
object (Wiesmller, 2005). Furthermore, the achievement and mas-
tering of musical goals can serve as a basis for building up images
about other goals that could be achieved in life.
One last but important aspect concerns the meaning of repro-
duction techniques in terms of a protection function. Holding on to
the reproduction of a musical piece might also indicate a resistance
to get in contact with oneself and potential threatening emotions
as well as in contact with the therapist, but still guarantees to keep
the therapeutic framework. In this sense, reproducing something
familiar musically might help the client with low motivation to
continue attending the therapy. We believe that the process of
reproducing does not only support the actualisation of memories,
but may also help to build up inner representations which were not
accessible before. These may support the clients ability to relate
to oneself as well as to others with respect to their own feelings
and personal boundaries. In this sense, reproducing music might
also strengthen the ability to mentalise (Fonagy, Gergely, Jurist, &
Target, 2002). In the music therapy literature, mainly improvisation
techniques have so far been investigated in relation to mentalisa-
tion and object theories (De Backer, 2008; Metzner, 2010; Strehlow,
2009). Even though improvisation techniques take an important
place in music therapy in general and mental health care in partic-
ular, the meaning of reproducing music in this eld embedded in
this particular theoretical framework needs to be explored quali-
tatively. This has long been underrepresented in the literature but
recently there seems to be an upcoming interest in this working
mode (Smetana, 2012).
Limitations and directions for future research
This study was based on previous research suggesting that spe-
cic ingredients can be of importance in music therapy (Danner &
Oberegelsbacher, 2001) and that music therapy may be more suc-
cessful when these are emphasised (Gold et al., 2007). The present
study showed tendencies in the same direction to support this body
of literature, even though most of these were not statistically sig-
nicant in this study due to limited sample size and power (Gold,
2004). Therefore, the meaning of reproduction techniques and their
potential importance as outcome predictors in individual music
therapy as outlined in the rst part of the discussion need to be
investigated more in depth.
One might ask if reproduction techniques were preferred by
the therapists conducting the sessions in this study. As all thera-
pists were trained in the application of all techniques, and two of
them were especially educated in a music therapy tradition empha-
sising improvisation techniques (Mssler, 2010), our ndings do
not seem to reect a one-sided use of techniques steered by the
therapists. Music therapy techniques were rated retrospectively on
the basis of therapy journals. Future research would benet from
more objective assessments of techniques, for example using video
recordings and independent raters. However, there is no reason to
believe that this could have led to any systematic bias in the study
results. Importantly, the outcomes were rated by clients and were
thus independent of the ratings of techniques.
The lack of signicant effects in this study might be caused by
the absence of an impact of music therapy techniques on outcome
variables, but it could also be due to a lack of power in the tests.
The sample of this study was relatively small (about half as many
participants as in Gold et al., 2007); furthermore, we used Bon-
ferroni adjustment which is extremely conservative and allows
only very strong predictors to become signicant. It is also reason-
able to assume that the clients presentation, represented by the
 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341 339

outcomes, is a result of the entire therapy up to the time point Technique Explanation
when it is measured. As our data collection only included three
Thematic improvisation Musical performance and expression of
sessions per participant (of an average total of 19 sessions), one a certain topic (e.g. images, memories,
may question whether the whole therapy was adequately repre- emotions)
sented. Taking into account the occurrence of differences in the Communicative improvisation Musical dialogue and interaction (e.g.
use of techniques and assuming that outcomes are inuenced by partner play, call-response) between
client and therapist resulting out of a
several sessions that have taken place before measurement, a larger psychodynamic context
sample of sessions may be worthwhile to include. To summarise, Trying out improvisation Important situations of the future or
increased sample size, more objective data concerning techniques, from the past are visualised by the
as well as extracting data from more sessions may all lead to a client and tried to be performed,
expressed, and transformed musically
more conclusive answer of the potential impact of techniques on
Free improvisation Musical play without rules in the sense
outcomes. The individual therapist, as well as the clients diagno- of the free association; can include
sis, also seemed to play an important role. This may be related to aspects of composing
common factors theory (Wampold, 2001) especially emphasising Reproduction techniques
factors in therapy that are connected to the therapists being rather Singing or playing Singing or playing known and familiar
pre-composed songs songs to build up trust, to support the
than doing as well as the variables inherent to the client. Future clients personal musical resources and
research should also take these factors into account. preferences, and to promote the
The examination of mechanisms of change in music therapy is actualisation of memories
still in its infancy, currently raising more questions than providing Learning or practicing musical Learning or practicing musical skills is
skills (e.g. guitar chords, used in the sense of practice without
satisfying answers. Further process-outcome research is needed to
playing melodies on the piano) practicing (Schmlz, 1974). It does not
explore in greater depth what it is that makes music therapy work. aim at perfectionism through
It is especially important to nd working modes for those peo- repetition by using a behavioural
ple in mental health care who are hard to reach verbally and hard oriented training method. Moreover,
to engage in psychotherapeutic treatment. Motivation for therapy the opportunity to practice on a
self-motivated basis is given where
may be a crucial point for the process and success of therapy as
failed performances are recognised as
it may not work if clients are not motivated (Schneider, Klauer, valuable as the successful ones. The
Janssen, & Tetzlaff, 1999; Wampold, 2001). focus is on the clients conscious
Findings of this study suggest that reproduction techniques awareness of her/his behaviour while
practicing
might be of particular importance when working with clients with
Reception techniques
low motivation. Relational competences tend to increase more Listening to live music The therapists play intents to support
when focusing on reproducing music. Consequently, this working (therapist plays for the client) the clients sensual awareness for
mode might help to get in contact with these clients and engage creating a willingness to listen and
them in therapy. Furthermore, it seems to be relevant that clients perceive. Furthermore, it is used in
moments of containment, where live
are actively involved in a musical activity rather than focusing
music can create an immediate
on working modalities operating with music listening. Working respond to the clients needs
actively with a familiar song, concurrently representing something Listening to recorded music Listening to recorded music can be
of the clients inner world, may help people in mental health care used similarly to the use of live music
but lacking the possibility to react
more sufciently to express themselves in a time of speechlessness
musical spontaneously on the clients
and nd ways to reconnect to this inner world as well as to their upcoming needs. Relaxation as well as
social environment. free association can be supported
while listening to music
Listening to music in Relaxation as well as free association
combination with body can be supported while listening to
perception music
Acknowledgements Listening to music in Exercises to support the clients body-
combination with and self-perception while listening to
We would like to thank our colleagues at the Grieg Academy dance/movement music (e.g. sound journey through the
body, muscle relaxation, yoga
Music Therapy Research Centre and Monika Smetana for their exercises)
feedback on a previous version of the article. This study was con- Others (specify)
ducted with institutional support from the State Psychiatric Clinic Note: The amount of time spent on a particular technique; additionally considering
Wagner-Jauregg Linz, Austria and the Nordfjord Psychiatric Centre, how much the client was interested to engage with the particular technique should
Nordfjordeid, Norway. This study was supported by a grant from the inuence the rating decision. Categories are partly based on Oberegelsbacher (1997)
and Storz (2000b). Response categories: very much (4), much (3), somewhat (2),
Research Council of Norway (project no. 186025, program Mental
little (1), not at all (0).
Health).

Appendix 2.

Appendix 1.
Statistical methods

Rating scale music therapy techniques

The adjusted model for the change scores of each outcome Y
Technique Explanation and the session i at the time point t was specied as
Production techniques
Structured improvisation Structured musical games based on  
musical parameters such as rhythm or Yit = b0 + b1 t + bi Xijt + bk Cik + eit , (1)
melody (e.g. rhythmic games like techniques confounding
imitating a rhythm or a melody,      
playing contrasts) predictors control
340 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341
Dependent

Intercept approach
Outcomes Y Y
variables

1 2

Baseline approach Y
1 Y
2

Outcome: Y0 Outcome: Y1 Outcome: Y2

Months 0 1 3
Time axis
Session: S0 Session: S1 Session: S2

Point approach
Therapeutic techniques

Intercept

X =S X =S
1 1 2 2

Interval approach
Predictors

X =(S + S )/2 X =S
1 0 1 2 2

Point approach
X =S
Baseline

X =S 2 2
1 1
Interval approach
X =(S + S + S )/3
X =(S + S )/2 2 0 1 2
1 0 1

Fig. 1. Time line and approaches used in the linear model with repeated measures. Above time line: approaches for outcomes. Below time line: approaches for predictors.
Notation: Yi , outcome variables at time point i; Yi , change of the outcome variables at time point i; Xi , predictor variable in the model at time point i; Si , intensity of the use
of therapy techniques in session i. Arrows show the covered time interval of the outcome change score in the different approaches.

where X denotes the music therapy techniques (these vary over
time), C, the confounding variables (these are time independent)
and e, the error term. The coefcients b describe the association
between each variable and the outcome. The indices j and k denote
the techniques or confounders that are included in the model. The
unadjusted model was
Yit = b0 + b1 t + eit (time)
or Yit = b0 + bj Xijt + eit (predictors)
or Yit = b0 + bk Cik + eit (confounders)
specied as using the same notation as in the fully adjusted model.
Fig. 1 shows further how Y and X were dened in the different
models. The change scores Y were conceptualised in two ways:
change from baseline to each time point (Baseline approach),
and change between two neighbouring time points (Intercept
approach). For the technique variables X, we also conceptualised
the predictor in two different (though related) ways: rst, we sim-
ply used the intensity in the session at the corresponding time point
(Point approach). Second, because the rst two selected sessions
both occurred before the one-month assessment and may therefore
both inuence the change of the outcomes within this time inter-
val, we also used the mean intensity of both sessions during the
time interval the change in outcome covered (Interval approach).
Having dened the models, we estimated the coefcients b with the
corresponding condence interval. Additionally, we tested for each
variable if it was associated with the outcome (i.e., if the coefcient
for the predictor was signicantly different from zero).
References
Albornoz, Y. (2011). The effects of group improvisational music therapy on depres-
sion in adolescents and adults with substance abuse: A randomized controlled
trial. Nordic Journal of Music Therapy, 20(3), 208224.
Barkham, M., Hardy, G. E., & Startup, M. (1996). The IIP-32: A short version of the
Inventory of Interpersonal Problems. British Journal of Clinical Psychology, 35,
2135.
Bolterauer, J. (2006). Die Macht der Musik. Psychoanalytische berlegungen zur
Wirkungsweise von Musik und ihren Wurzeln in der frhkindlichen Entwick-
lung [Power of music. Psychoanalytical reections on the effects of music and
its roots in early childhood development]. Psyche, 60, 11731204.
Bruscia, K. E. (1998). Dening music therapy (2nd ed.). Gilsum: Barcelona Publishers.
Cooper, M. (2008). Essential research ndings in counselling and psychotherapy. The
facts are friendly. Los Angeles/London/New Delhi/Singapore/Washington, DC:
Sage Publications.
Danner, B., & Oberegelsbacher, D. (2001). Spezische und unspezische Wirk-
faktoren der Musiktherapie Katamnestische Erhebung an psychosomatischen
Patienten einer Psychiatrischen Klinik [Specic and nonspecic factors of music
therapy. A catamnestic survey with psychosomatic patients at a psychiatric
clinic]. Nervenheilkunde, 8, 434441.
De Backer, J. (2008). Music and psychosis. Nordic Journal of Music Therapy, 17(2),
89104.
De lEtoile, S. (2002). The effetiveness of music therapy in group psychotherapy for
(2) adults with mental illness. The Arts in Psychotherapy, 29, 6978.
Decker-Voigt, H.-H., Oberegelsbacher, D., & Timmermann, T. (2008). Lehrbuch Musik-
therapie. Mnchen: Ernst Reinhardt Verlag.
Endicott, J., Nee, J., Harrison, W., & Blumenthal, R. (1993). Quality of Life Enjoyment
and Satisfaction Questionnaire: A new measure. Psychopharmcology Bulletin, 29,
321326.
Erkkil, J., Punkanen, M., Fachner, J., Ala-Ruona, E., Pnti, I., Tervaniemi, M., et al.
(2011). Individual music therapy for depression: Randomised controlled trial.
British Journal of Psychiatry, 199(2), 132139.
Fonagy, P., Gergely, G., Jurist, E., & Target, M. (2002). Affect regulation, mentalization
and the development of the self. New York: Other Press.
Gold, C. (2004). The use of effect sizes in music therapy research. Music Therapy
Perspectives, 22(2), 9195.
Gold, C., (in press). Methodology of psychotherapy outcome research: Some general
issues. In A. Pritz, O. Gelo & B. Rieken (Eds.), Psychotherapy research: Historical
and scientic foundations, outcome and process. Vienna, AT: Springer.
Gold, C., Mssler, K., Grocke, D., Heldal, T. O., Tjemsland, L., Aare, T., et al. (2011)
Individual music therapy for mental health care clients with low therapy motivation:
Multi-centre randomised controlled trial, submitted for publication.
Gold, C., Rolvsjord, R., Aaro, L. E., Aarre, T., Tjemsland, L., & Stige, B. (2005). Resource-
oriented music therapy for psychiatric patients with low therapy motivation:
Protocol for a randomised controlled trial [NCT00137189]. BMC Psychiatry, 5(39)
Gold, C., Rolvsjord, R., Mssler, K., & Stige, B. Reliability and validity of a scale to
measure interest in music among clients in mental health care. Psychology of
Music, submitted for publication.
Gold, C., Wigram, T., & Voracek, M. (2007). Predictors of change in music therapy
with children and adolescents: The role of therapeutic techniques. Psychology
and Psychotherapy: Theory, Research and Practice, 80, 577589.
Grocke, D., Bloch, S., & Castle, D. (2009). The effect of group music therapy on quality
of life for participants living with a severe and enduring mental illness. Journal
of Music Therapy, 46(2), 90104.
Hannibal, N. (2005). Beskrivelse af patientpopulationen i klinisk musikterapi. In H.
M. Ridder Ochsner (Ed.), Musikterapi i psykiatrien. rsskrift (pp. 6475). Aalborg,
DK: Aalborg Psychiatric Hospital/Aalborg University.
Kazdin, A. E. (1999). The meanings and measurement of clinical signicance. Journal
of Consulting and Clinical Psychology, 67(3), 332339.
Kazdin, A. E. (2001). Almost clinically signicant (p > .10): Current measures may
only approach clinical signicance. Clinical Psychology: Science and Practice, 8(4),
455462.
Lambert, M. J., & Ogles, B. M. (2004). The efcacy and effectiveness of psychother-
apy. In M. J. Lambert (Ed.), Bergin and Garelds handbook of psychotherapy and
behavior change (5th ed., Vol. 4, pp. 139193). New York: John Wiley & Sons, Inc.
 K. Mssler et al. / The Arts in Psychotherapy 39 (2012) 333341
Metzner, S. (2010). About being meant. Music therapy with an in-patient suffering
from psychosis. Nordic Journal of Music Therapy, 19(2), 133150.
Mssler, K. (2008a). Wiener Schule der Musiktherapie. Auf den Spuren einer Iden-
titt (Viennese School of Music Therapy. On the traces of an identity) (PhD, Music
Therapy). University of Music and Theatre, Hamburg.
Mssler, K. (2008b, September). Music therapy practice and research in acute psy-
chiatric settings. Paper presented at the Seminary Musikkterapi i psykiatri et
seminar om hva klinisk praksis og effektforskning har si til hverandre? (Music
therapy in psychiatry What can clinical practice and outcome research tell
each other). Grieg Academy, University of Bergen, Norway.
Mssler, K. (2010). I am a psychotherapeutically oriented music therapist. The-
ory construction and its inuence on professional identity formation under the
example of the Viennese School of Music Therapy. Nordic Journal of Music Ther-
apy, http://dx.doi.org/10.1080/08098131003768115. First published on: 18 May
2010 (iFirst)
Mssler, K. (2011, January). I would like to do something with music. Effects of resource-
oriented music therapy on negative symptoms in psychiatric patients with a low
therapy motivation. Paper presented at the Conference Forskning nytter. Oslo,
Norway.
Mssler, K., Fuchs, K., Heldal, T. O., Inger-Marie, K., Kenner, J., Nsheim, S., et al.
(2011). The clinical application and relevance of resource-oriented principles in
music therapy within an international multicentre study in psychiatry. British
Journal of Music Therapy, 25(1), 7291.
Mssler, K., & Oberegelsbacher, D. (2004). Fragebogen zur Erhebung musiktherapeutis-
cher Arbeitsweisen (Questionnaire for the assessment of music therapeutic working
modes) (unpublished questionnaire).
Oberegelsbacher, D. (1997). Musiktherapeutisches Improvisieren als Mittel der
Verdeutlichung in der Psychotherapie. In E. Fitzthum, D. Oberegelsbacher, &
D. Storz (Eds.), Wiener Beitrge zur Musiktherapie. Weltkongresse Wien Hamburg
1996 (pp. 4266). Vienna: Edition Praesens.
Rolvsjord, R. (2001). Sophie learns to play her songs of tears. A case study exploring
the dialectics between didactic and psychotherapeutic music therapy practices.
Nordic Journal of Music Therapy, 10(1), 7785.
Rolvsjord, R., Gold, C., & Stige, B. (2005). Research rigour and therapeutic exibility:
Rationale for a therapy manual developed for a randomised controlled trial.
Nordic Journal of Music Therapy, 14(1), 1532.
Rosenberg, M. (1989). Society and the adolescent self-image (Revised edition). Mid-
dletown, CT: Wesleyan University Press.
Schmlz, A. (1974). Kreativitt in der Musiktherapie. In W. J. Revers, G. Harrer, &
W. C. M. Simon (Eds.), Neue Wege der Musiktherapie. Grundzge einer alten und
neuen Heilmethode (pp. 173177). Dsseldorf: Econ.
Schneider, W., Klauer, T., Janssen, P. L., & Tetzlaff, M. (1999). Zum Einu
der Psychotherapiemotivation auf den Psychotherapieverlauf [Inuence of
psychotherapy motivation on the course of psychotherapy]. Nervenarzt, 70,
240249.
341
Schwarzer, R., & Jerusalem, M. (1995). Generalized Self-Efcacy Scale. In J. Wein-
mann, S. Wright, & M. Johnston (Eds.), Measures in health psychology: A users
portfolio. Causal and control beliefs (pp. 3537). Windsor, UK: NFER-NELSON.
Silverman, M. J., & Marcionetti, M. J. (2004). Immediate effects of a single music
therapy intervention with persons who are severely mentally ill. The Arts in
Psychotherapy, 31, 291301.
Smetana, M. (2012). Die Wiederkehr des hnlichen. Zur Bedeutung musikalischer
Objekte in der Musiktherapie bei Jugendlichen mit strukturellen Strungen (The
recurrence of the similar. On the meaning of musical objects in music therapy with
structural disordered adolescents. Unpublished doctoral dissertation, University
of Music and Performing Arts, Vienna.
Solli, H. P. (2008). Shut up and play! Improvisational use of popular music for a man
with schizophrenia. Nordic Journal of Music Therapy, 17(1), 6777.
Storz, D. (2000a). Musiktherapeutische Techniken. In G. Stumm, & A. Pritz (Eds.),
Wrterbuch Psychotherapye (pp. 444445). Wien/New York: Springer.
Storz, D. (2000b). Improvisation. In G. Stumm, & A. Pritz (Eds.), Wrterbuch der
Psychotherapie. Vienna: Springer, p. 303.
Storz, D. (2004). Fokale Musiktherapie. Entwicklung eines Modells psychodynamisch
musiktherapeutischer Kurztherapie Vienna: Edition Praesens.
Storz, D. (2011). Musiktherapeutische Techniken. In J. Illner, & M. Smetana (Eds.),
Wiener Schule der differenziellen klinischen Musiktherapie ein Update (pp.
4764). Vienna: Praesens.
Strehlow, G. (2009). The use of music therapy in treating sexually abused children.
Nordic Journal of Music Therapy, 18(2), 167183.
Talwar, N., Crawford, M. J., Maratos, A., Nur, U., McDermott, O., & Proc-
ter, S. (2006). Music therapy for in-patients with schizophrenia:
Exploratory randomised controlled trial. British Journal of Psychiatry, 189,
405409.
Tang, W., Yao, X., & Zheng, Z. (1994). Rehabilitative effect of music therapy for
residaul schizophrenia: A one-month randomised controlled trial in Shanghai.
British Journal of Psychiatry, 165(Suppl. 24), 3844.
Verbeke, G., & Molenberghs, G. (2009). Linear mixed models for longitudinal data. New
York: Springer.
Wampold, B. (2001). The great psychotherapy debate: Models, methods, and ndings.
Mahwah, NJ: Lawrence Erlbaum Associates.
Wiesmller, E. (2005). Zum Umgang mit den Begriffen bergangsobjekte und Objek-
tbesetzungen in der Musiktherapie [On the use of the terms transitional objects
and object cathexis in music therapy]. Musiktherapeutische Umschau, 26(1),
3949.
Wigram, T. (2004). Improvisation. Methods and techniques for music therapy clinicians,
educators and students. London/New York: Jessica Kingsley Publishers.
Yang, W.-Y., Li, Z., Weng, Y.-Z., Zhang, H.-Y., & Ma, B. (1998). Psychosocial rehabil-
itation effects of music therapy in chronic schizophrenia. Hong Kong Journal of
Psychiatry, 8(1), 3840.