Content uploaded by Sabina Alispahic
Author content
All content in this area was uploaded by Sabina Alispahic on Feb 05, 2016
Content may be subject to copyright.
Available via license: CC BY-SA 4.0
Content may be subject to copyright.
VOJNOSANITETSKI PREGLED Page 1
Correspondence to: Nina Hadžiahmetović, Faculty of Philosophy, University of Sarajevo, Franje Račkog 1, Sarajevo, Bosnia and Herze-
govina. E-mail: n.hadziahmetovic@gmail.com
ORIGINAL ARTICLE UDC: 615.851
DOI: 10.2298/VSP140911141H
Therapist’s interpersonal style and therapy benefit as the
determinants of personality self-reports in clients
Način ophođenja terapeuta prema pacijentima i korist od psihoterapije kao
odrednice samoprocene ličnosti klijenata
Nina Hadžiahmetović, Sabina Alispahić, Djenita Tuce,
Enedina Hasanbegović-Anić
Department of Psychology, Faculty of Philosophy, University of Sarajevo, Sarajevo,
Bosnia and Herzegovina
Abstract
Background/Aim. In (counter)transference relationship thera-
pist’s interpersonal style, implying the perceived relation of
therapist to a client (patient) in terms of control, autonomy, care
and positive feedback, has been shown to be important. The aim
of our study was to assess the relationship between therapist’s in-
terpersonal style and clients’ personality self-reports. Within
therapist’s interpersonal style, preliminary validation of the
Therapist’s Interpersonal Style Scale has been conducted, which
included double translation method, exploratory factor analysis,
confirmatory factor analysis, as well as the reliability tests of the
derived components. Methods. This research was conducted on
a group of 206 clients, attending one of the four psychotherapy
modalities: psychoanalysis, gestalt therapy, cognitive-behavioral
and systemic family therapy. Beside Therapist’s Interpersonal
Style Scale, Big Five Questionnaire and Therapy Benefit Scale
were administered, showing good internal consistency. Results.
Principal component analysis of therapist’s interpersonal style
singled out two components Supportive Autonomy and Ignoring
Control, explaining 42% of variance. Two-factor model of the
therapist’s styles was better fitted in confirmatory factor analysis
than the original 4-factor model. Structural model showing indi-
rect and direct effects of therapist’s interpersonal styles on self-
reports in clients indicates good fitness (χ2(12) = 8.932, p = 0.709;
goodness-of-fit index = 0.989), with Ignoring Control having di-
rect effect on Stability, Supportive Autonomy on Therapy Bene-
fit, and Therapy Benefit on Plasticity. Conclusion. The results
of this study indicate the importance of further research on
therapist’s interpersonal style, as well as further validation of the
instrument that measures this construct. Besides, a client’s per-
ception that the therapy is being helpful could instigate more ex-
plorative and approach-oriented behavior, what indirectly might
contribute to a client’s stability.
Key words:
psychotherapy; physician-patient relations; personality;
personality assessment; questionnaires.
Apstrakt
Uvod/Cilj. U (kontra)transfernom odnosu značajan je stil terape-
uta, koji podrazumeva poimanje relacije klijenta sa terapeutom u
smislu kontrole, autonomije, brige i pozitivne povratne informaci-
je. Cilj našeg istraživanja bio je ispitivanje povezanosti interperso-
nalnog stila terapeuta i samoprocene ličnosti klijenta. U okviru na-
čina ophođenja terapeuta prema klijentima sprovedeno je i pretho-
dno vrednovanje skale načina ophođenja terapeuta prema klijenti-
ma koje je obuhvatilo metod dvostrukog prevoda, eksploratornu
faktorsku analizu, konfirmatornu faktorsku analizu i ispitivanje po-
uzdanosti izdvojenih faktora. Metode. Istraživanje je rađeno na
grupi od 206 klijenata, koji su bili na psihoterapiji primenom jed-
nog od četiri psihoterapijska modaliteta: psihoanalize, geštalt terapi-
je, kognitivno-bihverioralne i porodične sistemske terapije. Pored
skale interpersonalnog stila terapeuta, primenjene su i skala proce-
ne za Velikih pet i skala percepcije koristi psihoterapije, sa zadovo-
ljavajućom unutrašnjom stabilnošću. Rezultati. Analizom glavnih
komponenti načina ophođenja terapeuta izdvojena su dva faktora,
suportivna autonomija i ignorišuća kontrola, koji objašnjavaju 42%
varijanse. Model načina ophođenja terapeuta sa dva faktora poka-
zao je bolje uklapanje u konfirmatornu faktorsku analizu od origi-
nalnog modela četiri faktora. Strukturalni model, koji prikazuje di-
rektne i indirektne efekte načina ophođenja terapeuta na samopro-
cenu ličnosti klijenta pokazuje dobru podešenost (χ2(12) = 8,932, p
= 0,709; goodness-of-fit index = 0,989), pri čemu ignorišuća kontrola
direktno doprinosi stabilnosti, suportivna autonomija percepciji
koristi terapije, a percepcija koristi terapije plastičnosti. Zaključak.
Rezultati ove studije upućuju na značaj daljeg istraživanja načina
ophođenja terapeuta prema klijentima i vrednovanja instrumenta
kojim se meri ovaj odnos. Pored toga, osećaj klijenta da je terapi-
ja korisna mogao bi potaknuti više istraživačkog ponašanja i
ponašanja orijentisanog na cilj, što bi indirektno moglo dopri-
neti stabilnosti klijenta.
Ključne reči:
psihoterapija; lekar-bolesnik odnosi; ličnost; ličnost,
procena; upitnici.
Page 2 VOJNOSANITETSKI PREGLED
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Introduction
Self-determination theory (SDT) 1, the theory of basic
psychological needs promotes autonomy as a sense of voli-
tion and psychological freedom 2, what is of quintessential
significance for client satisfaction in the process of
psychotherapy. Clients have differential motivations for
therapy that is susceptible to change depending on external
factors such as therapist’s interpersonal style (TIS) 3. The-
rapist’s style can be self-determination oriented when pro-
moting support, involvement and information, or control-
ling, when manifested as the opposite. Recent studies have
come to interesting discovery that it is not the visible and
superficial trademarks of a therapist, such as sociodemo-
graphic variables, professional experience or sex, the level
of training and type of orientation that contribute to thera-
peutic outcome 4, 5 as expected, what directs researchers to
pay more attention to therapist-client matching 4. The the-
rapist-client communication varies from autonomy to pater-
nalism 6. The more perceived coercion increases, the more
positive evaluation of therapeutic relationship decreases 7.
Clients reported being more intrinsically motivated when
therapist provided the opportunity for them to make a decisi-
on, expressed sincere care, provided constructive feedback or
did not exert pressure for specific activities, and more amoti-
vation was evident, when therapists were controlling 3. In a
meta-analysis of the pooled data on interaction styles, inclu-
ding control and negotiation as option, caring interaction
style (e.g. sensitive, friendly, relaxed and open) had a mode-
rate and positive correlation with satisfaction with consulta-
tion 6. Personal therapeutic attributes that turned out to
positively impact therapeutic alliance include an array of
characteristics such as conveying a sense of being
trustworthy, affirming, interested, alerted, affiliative type be-
havior as helping and protecting, coherent communication
style, and attunement to patient 8. As addition, active, enga-
ging and extraverted therapists produced faster symptom re-
duction in short-term therapy, but also non-intrusive thera-
pists generated better outcome in long-term therapy within
the range of 3 years of follow-up 9. In a research of the cli-
ent-oriented existential therapy failure, the main factors of
negative outcome were the lack of therapeutic attunement
and inflexibility 10. In another research in the domain of dif-
ferent psychoanalytic orientation, technical adherence and
directivity was shown in the therapists with hostile and con-
trolling introjects. These kinds of therapists were most likely
to monitor their own behavior as control for potential
external disapproval of their skills 11. But not all researches
argue against control in therapy. Taking into account cultural
framework, Chinese clients perceive directive therapist’s
style to be the most effective, finding concrete homework to
be more useful than only talking to therapist. Leading con-
versation guided by the therapist was also considered to be
appropriate and the rest focused on therapist-client match,
where therapists were regarded as someone who needed to
know how to click with others 12. The question whether cli-
ents benefit from directive counseling is yet to be addressed,
since controlling does not necessarily subsume coercion.
It most often relates to a structure, especially if promoted in a
rather autonomy-supportive manner 2. When expressed as a
support, recommendation is likely to be experienced as in-
formational, leaving the client to make a decision for
him/herself 13. Whether it is autonomy or control in therapy
that matters, therapist-client matching has been stressed out
on numerous occasions. A study shows that clients who were
matched with their preferred treatments had a 58% chance of
outcome improvement, so it is recommendable to include
client preferences into treatment 14.
Even though every therapy has its own effectiveness
criteria, the measure of client satisfaction was introduced as
the part of the broader scope approach to assess the quality
of service and some of previous client satisfaction measures
encompassed subscale ranging from relevance (fitting the
service with the problem), impact (effect of services on the
problem) and gratification (effect of service on client’s self-
efficacy) 15.
Aside from attending therapy and therapist’s characteri-
stics, the great deal of research indicates personality disposi-
tions to best predict personal wellbeing. The findings of posi-
tive contribution of personality are very consistent
16, 17. As
stated by Steel et al. 18, personality and wellbeing have much
stronger correlation than previously recognized. Studies
consistently show extraversion to have a positive and neuro-
ticism to have a negative influence on wellbeing with spillo-
ver effect on the overall wellbeing. Findings reveal genetic
dispositions in personality and long-lasting influence of
personality has been shown in longitudinal studies 19–22. Ba-
sic personality traits are described through the Big Five di-
mensions, replicable independently of culture: neuroticism,
extraversion, openness to experience, agreeableness and con-
scientiousness. Albeit there is a certain amount of published pa-
pers contending these traits are structurally organized into two
higher-order factors, usually labeled stability (neuroticism, that
is emotional stability, agreeableness and conscientiousness) and
plasticity (extraversion and openness to experience)
23 with
underlying biological substrates, Ashton et al. 24 argue this could
just be a methodological artifact, representing two or more
blends of the Big Five factors.
Scholarly attention has been paid so far to the therapeutic
matching and alliance itself, but it appears that the quality of
the alliance is more the result of therapist’s actions or charac-
teristics playing the most important role in achieving benefici-
al outcome, since only the variability within therapist and not
the client was significantly predictive of outcome, as the recent
study shows 25. Furthermore, it is also still unknown what cli-
ents think is important for psychotherapy 12. In view of previo-
us finding advising not to interpret heritability of personality
as the impossibility to change
19, this research included
personality as the outcome variable; especially since the traits
organized as higher-order factors named stability and plasticity
may represent socially desirable self-presentation behavior 24.
Since the personality as described could be the protective or
impairing factor to the wellbeing by itself, it was interesting to
examine whether these features are at least to a certain degree
susceptible to be determined by the exerted therapist’s style.
VOJNOSANITETSKI PREGLED Page 3
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Therapist’s style has previously been operationalized as
the four-scale construct 3 (including support of autonomy,
control, care, and support of competence) pertaining to inter-
personal behaviors with the function of motivational antece-
dents. This scale was originally adapted from Pelletier et al. 26
where it was first administered to capture motivational ante-
cedents in sports. It was shown that autonomy supportive be-
haviors providing opportunities for choice foster intrinsic
motivation, while the coach’s behaviors manifesting the lack
of care for the athlete undermine self-determined motivati-
on 26. Applied to the psychotherapy context, the same pattern
of results occurred showing the perception of the therapists
as providing opportunity to make decisions, carrying for cli-
ents, giving constructive feedback, or not putting pressure on
clients to be related to self-determined motivation toward
therapy 3. No previous studies to our knowledge considered
contribution of TIS to personality self-reports. Therefore, the
general aim of this study was to assess the contribution of
therapist-relevant variables, TIS specifically, to the Big Fi-
ve personality self-report in clients. In light of what is
known of interpersonal style in general, we were interested
to examine the direct and indirect effect of extracted thera-
pist’s styles on personality self-report. Indirect effect was
assessed through therapy benefit. The latter was invoked as
the mediator into the model, as the measure of satisfaction
with therapy, since interpersonal style is usually regarded
as motivational antecedent and therapy satisfaction as mo-
tivational consequence 3.
However, since TIS was previously measured by ad hoc
constructed scale (TIS) for the purpose of motivation for
therapy scale validation, without previous history of a thoro-
ugh psychometric validation per se, the first and foremost
purpose of this study was to preliminary validate this scale
into Bosnia and Herzegovina (BH) languages. We were first
interested in the translation of scale from the original English
into BH languages, then to conduct the double translation
procedure, which was followed by exploratory and
confirmatory analyses of TIS structure, as well as the
reliability testing of the extracted factors.
In previous research 26, it was recommended that the
degree to which parents and coaches adhere to supporting the
children and spending time with them should facilitate self-
determination in children. We are not certain whether the
patterns of behaviors pertaining to coach or parents
(specifically the ones including spending time or providing
permanent feedback) are totally applicable to the patterns of
the therapist’s behavior, especially since many therapy scho-
ols have different rationale as to how to approach a client.
For this reason, and being aware that the 4-factor structure
has not been confirmed yet, we did not make any definite as-
sumptions regarding the preset number of factors to be
extracted. But we did expect that if autonomy/support prone
styles were extracted, these should have positive, and con-
trolling prone styles, also in case of exploratory extraction,
should have negative contribution to therapy benefit and
personality self-report, the latter defined by two presumably 24
secondary factor loaded personality variables Stability and
Plasticity.
Methods
The study sample included 206 clients (154 females, 47
males, 3 participants did not specify their gender; mean age
33.99 ± 10.17), attending 4 psychotherapy schools:
psychoanalysis (n = 28), gestalt therapy (n = 76), cognitive-
behavioral therapy (n = 75) and systemic family therapy (n =
27). Most of clients had university degree (n = 109), following
high school diploma (n = 84), year degree (n = 12) and only one
participant completed primary school. As for employment sta-
tus, 82 clients had full-time job, 57 were unemployed, 23 had
part-time employment, 21 fixed term employment, 7 were reti-
red and 16 did not provide information about their employment.
With respect to marital status, 89 clients lived in formal marri-
age or with a partner, 84 were never married, 25 were divorced
and 1 widowed. The rest 7 did not provide information in refe-
rence to their marital status. Beside therapy in which 140 cli-
ents were enrolled without taking medication, 65 clients also
had joint medication treatment and one left out the information
about medication intake.
In order to assess TIS and personality traits, two self-
report measures were administered: The Adapted Therapist’s
Interpersonal Style Scale and the Big Five Questionnaire 27.
The first instrument was adapted for the purpose of this re-
search. The fundament upon which the adaptation was made
was the original TIS scale 3. The scale was originally construc-
ted in English and is made of 4 subscales, consisting of 3 items
each, aimed to assess 4 different types of interpersonal styles.
The 12 items forming 4 subscales, originally adapted from Pel-
letier et al. 26, include Support of Autonomy (e.g. “My therapist
provides me with opportunity to take personal decisions”), Con-
trol (e.g. “My therapist pressures me to do what he/she wants.”),
Care and Support of Competence. The answers are given on a
Likert type scale ranging from 1 to 7. TIS Scale was originally
adapted, although not psychometrically validated, from the simi-
lar scale administered in sports domain, the Coach’s Interperso-
nal Style (CIS)
26. The latter consisted also of four scales:
Autonomy Supportive Climate (e.g. “My coach accepts that mi-
stakes I make are part of a learning process.”), Caring (e.g. My
coach cares about me.”), Providing Structure (e.g. When my co-
ach asks me to do something, he or she gives me a rationale for
doing it.”), and Competence Feedback (e.g. “The feedback I re-
ceive from my coach is constructive in helping me make impro-
vements.”). Sample items in both scales are very similar, except
that the “coach” was switched by the “therapist” to accommo-
date more to the therapist’s style.
The adapted TIS scale was first translated from English in-
to BH languages by two psychologists in clinical domain. Then
back translation process was conducted in which bilingual En-
glish/BH language(s) speaking psychologist independently tran-
slated the BH version of the scale back to English. This transla-
tion was again thoroughly checked by English professor. The
back translation process was followed by double translation pro-
cedure. It included sending the back translation version of the
instrument to the author of the original scale 3 for further con-
firmation about the equivalence of the original and the translated
items. After the check of double translation it turns out that 7
items were identical in meaning as in the original TIS, while
Page 4 VOJNOSANITETSKI PREGLED
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
the rest 5 items were slightly or considerably changed, and
different in meaning comparing to the original TIS.
Analyzing the rest 5 items by its content, it was noticeable
that changes were made in a more control and problem-
solving direction, since our study was more focused on
autonomy or control manifesting therapist’s behavior in ge-
neral. So we did not further consider our instrument we cal-
led the Adapted TIS Style Scale to be the equivalent to the
original TIS, but we conducted a psychometric validation on
our version of the instrument to establish whether it is in its
own right applicable in other analyses. This process we dis-
cuss more in the Discussion section and the adapted version
of the instrument is given in the Appendix. In this research,
two factors pertaining to therapist’s styles into which all
items were aggregated, were singled out, what is explained
in more detail in the Results section. The factors showed,
though not perfect, acceptable reliability of the subscales:
Control (Cronbach’s alpha 0.63) and Autonomy (Cronbach’s
alpha 0.78).
The Big Five Questionnaire consists of 50 adjective
items on a 5-point Likert scale forming 5 subscales intended
to capture the Big Five personality traits: emotional stability
(inverted Neuroticism), extraversion, intellect (openness to
experience), pleasantness (agreeableness) and conscientious-
ness. The subscale reliabilities in this research were very go-
od ranging from 0.82 to 0.87.
Therapy Benefit Scale consists of 3 very simple questions
measuring the satisfaction with therapy: “To what extent is the
therapeutic treatment you are currently involved in important for
you?”, “How much do you consider therapy helps you?”, and
“To what extent are you satisfied with the therapy you are in-
volved in?” All questions were responded on a 4-point Likert
scale adding up to a total score of therapy benefit. Principal
component analysis revealed all items to load on one factor,
enabling to add individual responses to a common score. Besi-
des, reliability of this measure, in spite of containing only 3
items, was very good (Cronbach’s alpha 0.79).
All data were collected on a voluntary basis, respecting
the anonymity of clients. Clients gave consent to participate
in the study and could withdraw from research at any point.
They were already enrolled to therapy for a substantial time
to be able to evaluate their perception of the relationship with
the therapist. Prior to the questionnaire distribution, the the-
rapists of the four above mentioned therapy schools were
contacted to recruit interested clients into the research. The
questionnaires sealed in envelopes were mailed to different
locations in BH. Clients were given the questionnaires they
opened on the site, filled it in the waiting-room with no sug-
gestions from the therapist, sealed the filled forms again, and
handed it to the therapist whereupon they were returned to
the researcher.
Data were analyzed using Principal Component Analysis
(PCA) to derive components of the therapist’s interpersonal
style. The components derived in exploratory factor analysis
(EFA) were also fitted for confirmation in subsequent
confirmatory factor analysis (CFA). Correlational analysis was
engaged to examine whether therapist’s styles have any relati-
ons to personality self-report measures. Finally, structural
equation modeling was conducted to shed further light on the
direct and indirect effects of therapist’s styles on personality
self-report as mediated through therapy benefit, as well as to
examine secondary factor loadings personality model. Desc-
riptive analyses are also provided in the following section. To
obtain the analyses, two statistical software packages were
used: IBM SPSS Statistics for Windows, Version 19.00 and
IBM SPSS Amos, Version 19.00.
Results
To assess how many components would be suitable to be
extracted in order to best explain the variance of the interper-
sonal therapist’s styles, PCA was conducted (Table 1).
Table 1
Principal Component Analysis (PCA) Pattern Matrix of the items
Items Supportive
Autonomy
Ignoring
Control
h2
3. My therapist gives me the feedback about the way I make a progress.* 0.87 0.18 0.69
1. The feedback I receive from my therapist is constructive in helping me make
improvements.
0.86 0.72
8. My therapist consults me before (s)he decides how to address
my problems.*
0.80 0.10 0.59
11. My therapist provides me with lots of opportunities to take personal
decisions in what I do.
0.55 -0.12 0.37
2. When I ask my therapist to help me solve a problem, he or she asks me what I
think before giving me his or her opinion.
0.48 -0.19 0.33
10. I feel that my therapist doesn’t care how much I improve through therapy. -0.20 0.05
7. My therapist pressures me to do what he or she wants. 0.11 0.76 0.53
4. My therapist is trying to impose her/his ideas on me.* 0.66 0.44
6. I feel that my therapist is indifferent towards me. -0.12 0.61 0.44
12. My therapist leaves me with little choice about the ways in which my
problems could be resolved.*
-0.10 0.55 0.35
5. The feedback I get from my therapist is basically useless criticism. -0.31 0.44 0.38
9. My therapist is being harsh to me.* 0.44 0.18
Note: The items written in boldface are changed and adapted in translation from English to BH languages, so due to a
change, they are left in the translated form. Other items are written as in the original Therapist’s Interpersonal Style.
VOJNOSANITETSKI PREGLED Page 5
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Firstly, we tried to conduct PCA with 4 preset factors, as
in the original TIS Scale to examine whether 4 subscales can
be singled out. By applying first the default orthogonal rotati-
on, since there were no similar previously published results
upon which to build our analysis, it was evident from the valu-
es in component transformation matrix that in fact 4 factors
were intercorrelated. The 4-factor solution is by the predeter-
mined extraction rationale the solution that best corresponds to
the Kaiser-Guttmann’s criterion
28 for extraction. However,
this solution was not substantiated in structure matrix where
only two items saturated component in some cases. The same
pattern of results occurred independently from the type of rota-
tion. But, the general criteria in analysis for extraction were
fulfilled 28 including Kaiser-Meyer-Olkin (KMO) = 0.731 and
Bartlett’s Test of Sphericity χ2
(66) = 570.75, p < 0.001, indica-
ting that manifest items are correlated to a certain degree to be
able to capture a latent component, but not as singular as not to
be discernable as distinct entities.
Another criterion for extraction is shown in Figure 1.
Cattell’s Scree test shows that 2 distinct components
probably best explain the latent structure of the therapist’s
interpersonal style with eigenvalue far exceeding 1. Even
though on the basis of visual inspection it is visible that 4
components could be extracted, what corresponds to the Kai-
ser-Guttmann’s extraction criterion as well, it is also visible
that the last 2 factors exceeding 1 on the ordinate explain the
variance far less than the first 2. Besides, the point of
infexion happens on the third component, what makes this
component a surplus 28, which is usually not retained in
analysis. This is the reason why we actually chose to explore
the latent structure of the therapist’s style if we suppose 2 in-
tercorrelated components for extraction applying direct ob-
limin rotation on the components.
Fig. 1 – Cattell’s Scree Test for component extraction.
After preliminary check for the appropriate number, 2
components have been extracted, accounting for 42.17% of
variance. The first component labeled Supportive Autonomy
explained 28.98% and the second labeled Ignoring Control
explained 13.19% of variance. Both components had accep-
table reliabilities as outlined in instrument description, even
though ignoring control had a bit lower Cronbach’s alpha
reliability value of 0.63. It is stated in the literature that
reliability as low as 0.61 could be acceptable when conduc-
ting exploratory studies 29, so this component was further re-
tained in the analysis. Reliability analysis was also conduc-
ted on the original four subscales, but internal consistency of
the 2 derived components outperformed original scales
which had the reliabilities ranging from 0.71 for the support
of competence to 0.33 for care, which was the least value.
Unlike the original TIS having four subscales, in this rese-
arch all items were aggregated into two plausible compo-
nents relating to the conglomerate of the four therapist’s
styles from the original scale. These were labeled Supportive
Autonomy, since the items originally belonging to support of
competence and support of autonomy all saturated one com-
ponent. Ignoring Control was labeled upon the conglomera-
tion of the items originally pertaining to the lack of care and
control, which in this exploratory analysis all saturated the
second extracted component. Supportive autonomy could be
described as giving useful feedback, support for independent
decision-making, and consultation prior to expressing own
opinion. On the contrary, ignoring control is not captured as
giving directive instructions and maintaining structure, but
more like negative controlling of client and being ignoring
and non-empathetic.
To ascertain these results more thoroughly, we decided
to further conduct confirmatory factor analysis. This analysis
required post hoc updates to the original model outlined in
Figure 2.
Fig. 2 – Confirmatory model of the Adapted Therapist’s In-
terpersonal Style Scale.
This model was tested for fitting with maximum likeli-
hood method comparing to the 4-factor model as would be in
the original scale. The parameters showed, even though the
Page 6 VOJNOSANITETSKI PREGLED
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Table 2
Correlations among the extracted therapist’s interpersonal styles, therapy benefit and personality self-evaluations
Scale 1 2 3 4 5 6 7 8
1. Supportive Autonomy 1
2. Ignoring Control -0.44** 1
3. Therapy Benefit 0.50** -0.39** 1
4. Neuroticism -0.10 0.26** -0.13 1
5. Extraversion 0.13 -0.14* 0.21** -0.33** 1
6. Intellect/Openness to Experience 0.14* -0.14* 0.17* -0.35** 0.59** 1 .
7. Pleasentness/Agreeableness 0.12 -0.13† 0.12 -0.26** 0.19** 0.29** 1
8. Conscientiousness 0.06 -0.19** 0.10 -0.26** 0.25** 0.39** 0.52** 1
Mean ± standard deviation 29.69
± 4.77
8.38
± 3.53
6.87
± 1.51
32.79
± 6.94
34.56
± 8.19
35.46
± 7.59
42.81
± 5.56
39.95
± 6.59
**statistically significant (p ≤ 0.01); *statistically significant (p ≤ 0.05); † marginally significant (p = 0.059; p ≤ 0.06).
2-factor model is not fitted on chi-square level χ2
(48) = 89.85,
p < 0.001 (the significant p means a difference between the
observed and default theoretical model) according to which
the null hypothesis was rejected, it well outperformed the
original model on other parameters. It is also worth noting
that χ2/df was 1.87. The Goodness-of-fit index 30, measuring
the fit between the observed and hypothesized covariance
matrix (acceptable cut-off over 0.90) for the 2-factor model
had a value of 0.935, while other parameters including com-
parative fit index (CFI), what is recommendable to check 31
and root mean square error of approximation (RMSEA) 32
had the values of 0.923, and 0.065 (PCLOSE = 0.112),
respectively, approaching the recommended cut-offs. RM-
SEA less than 0.05 is usually taken as the indicator of the
good model fit 32, which is here almost the case, very closely
approaching the desired value. Some of other parameters
such as the normed fit index (NFI) 33 = 0.853 argue for the
necessary improvement of this model, but considering this
analysis to be preliminary and sensitivity of the NFI to the
sample size 31, these parameters are not negligible. Unlike the
2-factor model, the 4-factor model showed less favorable pa-
rameters as follows: χ2
(47) = 121.46, p < 0.001 with much lar-
ger χ2/df ratio 2.58. In the 4-factor model the null hypothesis
was also rejected, and the rest of parameters had values farther
from acceptable comparing to the posed 2-factor model inclu-
ding: GFI = 0.916, CFI = 0.864, RMSEA = 0.088 the close-
ness of-fit statistic [(the closeness of-fit statistics PCLOSE) =
0.001), and NFI was even smaller 0.802. What is more interes-
ting, post hoc analysis showed that the 4-factor model could be
better fitted if we supposed the regression weight on the item 5
(“The feedback I get from my therapist is basically useless cri-
ticism”.) from Control to be unconstrained, what implies that
the item originally belonging to Care has a probable secondary
loading on Control, as well. This correlation was pretty high
(0.46) for the secondary loading, and was significant (p <
0.01). With this modification, the 4-factor model also reached
its fitting maximum.
Turning back to the 2-factor model, all estimates were
significant, and it also turned out for the extracted compo-
nents to be inter-correlated (r = -0.50, p < 0.001), so these
components were further retained in other analyses that con-
sidered the contribution of interpersonal therapist’s styles to
personality self-reports.
In the forthcoming analyses, first the correlation of all
variables included in structural equation model is shown in
Table 2.
In this analysis, the relationship between relevant therapy
variables (such as therapist’s styles and therapy benefit) and
personality self-evaluations was of the greatest interest. As can
be observed, the correlations between therapist’s styles and
therapy benefit are patterned in a predictive way. Supportive
autonomy is positively and ignoring control negatively corre-
lated with therapy benefit (Spearman’s rho(205) = 0.50; Spear-
man’s rho(205) = -0.39, respectively, both significant at p <
0.01). The correlation between therapist’s styles also shows
inverse pattern, the more perceived supportive autonomy is
expressed, the less perceived ignoring control and vice versa
(Spearman’s rho(206) = -.44, p < 0.01). Although correlations
between personality self-evaluations and therapist’s styles are
fair, but significant, it is visible that supportive autonomy is
almost not at all related to personality, but ignoring control
shows fair, and significant correlations to all personality self-
reports, Spearman’s rho ranging from -0.13 to -0.39. Neuroti-
cism correlates positively with Ignoring Control, Spearman’s
rho = 0.26 (p < 0.01). To obtain the correlations, Spearman’s
rho as the robust method was engaged for precaution, since
few distributions showed a certain asymmetry, especially
therapy variables supportive autonomy and therapy benefit
being negatively asymmetrical.
To examine the predictability of the relationship
between therapist’s styles and personality, further regression
analyses were conducted, not reported in this research, but
they shed further light on potential directionality of the con-
tribution, revealing that it were the therapist’s styles, and
not the reverse, that contributed to personality self-report.
This finding led to setting the structural model to assess the
direct and indirect effects of therapist’s styles on personality
self-report. As outlined earlier, the model was set which in-
cluded therapist’s styles as exogenous variables, therapy be-
nefit as the mediator and personality factors as endogenous
variables. The model was also set to be fitted for secondary
factor loadings for stability and plasticity, instead of presu-
ming higher-order factor structure. The hypothesized model
with the standardized coefficients of the effects is presented
in Figure 3.
VOJNOSANITETSKI PREGLED Page 7
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Fig. 3 – The hypothesized structural model of the Therapist’s Interpersonal Style effects on personality self-report.
The maximum likelihood model was very well fitted
with the following parameters: χ2
(12) = 8.93, p = 0.709; GFI =
0.989, RMSEA = 0.00 (the insignificant p here means no dif-
ference between the observed and default theoretical model,
indicating good fitness of the model). The model provides
some direct and indirect effects of TISs to be observed. Sup-
portive autonomy is moderately positively related to therapy
benefit (r = 0.40, p < 0.01) and ignoring control is fairly
negatively related to therapy benefit (r = -0.25, p < 0.01).
Supportive autonomy and ignoring control show moderate
inverse correlation as two opposite therapist’s styles (r = -
0.38, p < 0.01). The interesting finding is that supportive
autonomy has no any significant direct effect on either
stability (agreeableness, conscientiousness and neuroticism)
or plasticity (extraversion and openness), all values approac-
hing 0. On the contrary, ignoring control, though not having
direct effect on plasticity, has a direct marginally significant
effect on stability (r = -0.21, p = 0.073). Also indicating of
the secondary factor loading in the measurement model is
that much of the variance pertaining to stability (agreeable-
ness, conscientiousness and neuroticism) can be explained by
Plasticity as the latent variable. As hypothesized, and
contrary to the higher-order model, agreeableness and con-
scientiousness are also explained by plasticity (r = 0.40, r =
0.29 respectively), independent from being set to regress on
stability. And neuroticism is more explained by plasticity (r
= -0.43), than by stability (-0.11) that was preset in the mo-
del. When we analyze standardized indirect effects in more
detail, for both supportive autonomy and ignoring control on
personality self-reports, mediated by therapy benefit, these
effects are very low tending to be zero. Even though therapy
benefit has a significant direct contribution to plasticity
(r = 0.23, p < 0.01), the indirect effect both therapist’s styles
have, mediated by the perception of therapy benefit, is very
low. As for the total effects of therapist’s styles, the total
contribution of ignoring control to plasticity, comparing to
supportive autonomy approaching 0, is larger and has the va-
lue of -0.13. The total effect of ignoring control on stability
(almost completely attributable to direct effect) is also twice
as high as for supportive autonomy. Additional important
finding, which is in line with the previous assertion that the-
rapist’s styles could have contribution to self-presentation in
personality, is also a lower total effect on the sole personality
traits than on their blends (stability and plasticity), but again,
that effect is a bit larger for ignoring control than for suppor-
tive autonomy and in range of -0.11, -0.16 and -0.21 for
openness, agreeableness and conscientiousness, respectively.
All autonomy effects approach 0.
Discussion
Therapeutic relationship and coercion are both impor-
tant in clinical practice and have driven a lot of attention in
the clinical literature 7. According to some findings 8, allian-
ce is a pan-theoretical construct impacting psychotherapy
independently from therapeutic approach. Even though not
dealing with alliance by itself, we were interested to examine
whether there were some important characteristics of the
very therapist that could contribute to the way clients descri-
bed themselves. Considering the fact that researches such as
this are not that common, which is the probable reason for
lacking the adequate instruments to measure some concepts,
the administration of the Adapted TIS Scale was a pretty
challenge. In exploratory factor analysis (PCA) two compo-
Page 8 VOJNOSANITETSKI PREGLED
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
nents we called supportive autonomy and ignoring control
were derived in which all items, previously belonging to 4
components, were aggregated. Two components extracted in
this research had also far better reliabilities than four original
subscales, which is also the reason why they were retained in
further analysis. Even more so, CFA showed the 2-factor
model outperformed the 4-factor model. However, there are
a few limitations of the procedure used that should be emp-
hasized. Firstly, some items of the original TIS scale were
changed and administered with the changes in translation
that were more inclining to what authors considered to be
therapeutic language. For instance, item 4 was completely
changed corresponding more to the controlling style. Since
the original TIS has not to our knowledge been validated be-
fore, we are not certain whether 4 factors are confirmed in
the first place in other studies, or whether the results would
have been different, had we not made changes in the transla-
tion process. Furthermore, the fitting indexes for both models
show both could be more improved. This is also one of the
reasons why the content of the items should be regarded with
more care in future studies. We consider our exploratory
analysis to be preliminary and certainly, further replications
that could contribute to the validation process are advised.
Secondly, since there is no explicit theoretical background as
for the number of TISs, confirmatory factor analysis is
definitely not the method of choice for this confirmation. Be-
ing aware of this limitation and considering that our 2-factor
model underwent subsequent post-hoc modifications intro-
ducing covariances between error terms (Figure 2), this
confirmatory analysis is in its essence another case of
exploration, and not confirmation to be more precise. To be
able to fully conduct the process of the validation of TIS sca-
le, some other validity check such as convergent or predicti-
ve, which were not specifically considered in this study, sho-
uld also be introduced in future research. Since interpersonal
styles are not only applicable to therapists, as we are aware
they were also mentioned in a scale relating to coaches, and
probably could be adapted to various domains, more research
is needed on how many interpersonal styles are mentioned in
other scientific resources and whether all styles could be ap-
plied generally, or to only certain domains. Under this ratio-
nale the transferability issue of styles studied in coaches to
other domains should also be addressed in the future.
Considering the rest of our findings, the unambiguous
result was that supportive autonomy had a direct contribution
to the perception of therapy benefit, but what happens next
and how it affects other outcome variables is yet to be asses-
sed. Contrary to our hypothesis, supportive autonomy did not
have any effect on personality variables. Such a result
actually coincides with findings that it was the therapist’s
characteristics that contributed more to the alliance itself
than to the outcome, even though some studies report the po-
sitive impact of autonomy continued to persist in abstinence
behavior long after the treatment was over 13. We know that
behaving in autonomy-supportive manner will not provide
any side-effects, but do we know what autonomy essentially
is? SDT makes a distinction between autonomy and inde-
pendence, for autonomy supports a volitional treatment-
adherence, while independence implies independent decisi-
on-making. In the case of the latter, the opposite would be a
total dependence on counselor’s direct advices. According to
the authors 2, the opposite of autonomy is heteronomy, en-
compassing free will to adhere whether behavioral changes
are induced by internal or external influences. But in this re-
search, we came up with the correlational analysis of -0.38 (p
< 0.01) between supportive autonomy and ignoring control,
implying this is probably not unipolar, but bipolar construct.
Other studies also identify autonomy with coercion absence 7.
Since therapeutic alliance is very often given supremacy in
the research, the unique contribution of the therapist or the
client has often been concealed. But if autonomy as bipolar
construct holds truth, then it could be presumed that support
is self-understanding in any therapy benefit, but it is
probably the active absence of coercion or control what has
the impact on the outcome variables. In this research, the re-
duction of ignoring control, independently from therapy be-
nefit, had positive contribution to stability. We presume that
many clients enrolled in therapy with predominant anxiety
and depression problems come instable by default, since both
of these states are characterized by perceived lack of control
to influence external circumstances or one’s life. Directive
orders without any consultation with a client or treating the
client with neglect and lack of care might not be different
from the outer therapeutic conditions. Stability also referred
to as social propriety or socialization seems to reflect rever-
ted neuroticism, but is also a broader construct for encom-
passing other two traits 23. In this model the score of neuroti-
cism, instead of emotional stability, was left purposely, for
most of the problems clients come to therapy with, include
some sort of neurotic symptoms, so it was important to keep
neuroticism as the outcome variable. In other research 34,
stability negatively predicted externalizing behavior (correla-
tion -0.71), including aggression, vandalism, drug abuse, op-
position and hyperactivity. Considering the results obtained
in this research it could be surmised that the lack of control
and restraint coming from the therapist could contribute
positively to the reduction of behaviors such as aggression
and substance abuse, what should be investigated in further
research. But one has to keep in mind precaution by not prec-
luding the possibility that it is just the self-presentation res-
trained ignoring control contributes to and certainly not the
change in the trait neuroticism. One has to be aware that this
study does not argue for changing the personality as the out-
come variable, especially for it being the input variable in
many other cases, but rather that might have contribution to
the way clients present themselves in the self-report. Other
factor that can account for this assumption is the instrument
used to assess personality. The Big Five Questionnaire was
administered, consisting of adjectives as personality descrip-
tors derived from lexical studies. When self-evaluating on
adjectives that are, according to lexical hypothesis, the words
that are most important for capturing the individual differen-
ces, and as such, the most frequent words in the vocabulary
of many cultures, it is well possible that these words are go-
od representatives of the well-behaved expressions, such as
being “stable” or “flexible”. As the other authors 24 also ar-
VOJNOSANITETSKI PREGLED Page 9
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
gue, it is possible that so-called higher-order factors (here
secondary loaded factors) in fact represent moralistic bias
rather than substantive dimensions of personality. Anyway,
this finding is important since it shows that the lack of coer-
cion and control by the therapist could stabilize a client in
the well-behaved and socialized manner. Therapy Benefit
contributing to plasticity, the latter, also referred to as
dynamism or personal growth 23 was positively correlated
with externalizing behavior (correlation r = 0.75) in other re-
search 34, which denotes not only instability and lack of res-
traint, but also exploratory and approach-oriented behavior.
Therefore, therapy benefit could have a direct effect on ta-
king responsibility for own actions and incite client on more
exploration in his/her life. Other important finding in this
study was the better confirmation of the personality model
with blended variables, instead of the higher-order factors.
Another model including higher-order factors was also tested
(but not reported in this study), and was outperformed by the
blended-variables personality model. The result is in accor-
dance with another study testing only the confirmation of the
latent factors as the higher-order vs blended variables vs ort-
hogonal factors 24, where the model presuming orthogonal
factors was exceeded by the higher-order model, but the lat-
ter was exceeded by the blended variables model in three
samples from Ontario, Oregon and Alberta. Other studies ar-
guing for stability and plasticity as the higher-order factors
do not provide unambiguous results for the theoretical re-
gression of the factors onto higher-order factors (e.g. in a
study 23 using also adjective personality markers, the two
higher-order factors model fitted the data well, but open-
ness/intellect did not load significantly on it). Such a result
authors ascribe to the variation in markers descriptions. In
our study of the blended variable model, where stability and
plasticity are the blends of the Big Five factors, it is shown,
that agreeableness, conscientiousness and neuroticism, besi-
des loading on stability, substantially loaded on plasticity, as
well. Neuroticism was better regressed onto plasticity (r = -
0.43, p < 0.01), than stability, but this in fact may be due to
plasticity feature encompassing some behaviors indicating
instability. More studies in this domain considering personality
models independently from exogenous variables, what this
study actually has not specifically dealt with, could be useful.
This study has a few limitations that should be overco-
me in future research. Since this is the first study to bring
therapist’s styles into relation to personality, as well as to
confirm the blended variable personality model, the study
would be recommendable to replicate on a bit larger sample.
Although the confirmation of personality model can be repli-
cated in general population, therapist’s styles assessment
requires participants to be therapy-involved and that is why
this study is unique in terms of the sample engaged. But furt-
her caution is advised when engaging clients involved in
therapy process. What was not controlled for in this study
and could have contributed to the ways questionnaires were
fulfilled or the ways client perceived his/her therapist is the
level of therapist’s education. Some therapists in this study
were able to do their own practice without supervision, while
the others were in the process of doing the practice under su-
pervision. In this research the level of therapist’s education
was not considered as the variable, but should be taken in re-
gard in researches where therapists are in fact the subject of
evaluation by client. This could even more objectify the eva-
luation independently given by client. Considering that
psychotherapy in BH is still under intensive development,
more researches that should follow the practice in counseling
and psychotherapy should be welcomed. Another shortco-
ming is that we did not have pretest data for personality self-
evaluation before the therapy process started, so it cannot be
with certainty argued that therapy actually changed somet-
hing in clients’ personality traits. But this model gives gene-
ral pattern as for the potential paths that should be given at-
tention when demonstrating certain interpersonal styles. Cor-
relational analyses showed significant results that did not
hold in the direct paths of the model. This implies taking in
account some other variables that were not included into this
model. This research did not measure variables such as client
wellbeing directly, but it would be interesting to examine
whether personality or therapist’s styles have better unique
or common contribution to personal wellbeing. Instead of
measuring personality as the disposition to behave, future
studies should consider more concrete behaviors for the out-
come variables. Personality inventories containing state-
ments instead of one-word trait markers should be used in fu-
ture research, since the latter can be more susceptible to self-
presentation.
Conclusion
This study shows the underlying structure of therapist’s
styles could be best accounted for by the two preliminary
extracted opposite styles labeled supportive autonomy and
ignoring control, also confirmed in preliminary confirmatory
analysis. Further analysis shows ignoring control and no
supportive autonomy was correlated to personality self-
evaluations on 5 personality traits. The more elaborate model
gives further insight into the relationship, showing a few im-
portant relations: therapy benefit is predicted positively by
supportive autonomy and negatively by ignoring control.
Therapy benefit directly contributes to the account of
plasticity, and ignoring control has a direct marginal effect
on stability. Personality traits explained by stability also had
secondary factor loadings on plasticity. Supportive autonomy
has no direct or indirect effect either on stability or plasticity.
This shows that Supportive autonomy is necessary but not
enough condition for improvement in terms of manifesting
more stable or exploring behavior. It is important to be care-
ful when demonstrating directivity, especially if it leaves
possibility to be blurred by control without support. A cli-
ent’s perception that the therapy is being helpful could insti-
gate more explorative and approach-oriented behavior, what
indirectly might contribute to client stability.
Acknowledgements
The authors would like to thank Luc Pelletier, Full Profes-
sor at School of Psychology, University of Ottawa for providing
Page 10 VOJNOSANITETSKI PREGLED
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
us the original Therapist’s Interpersonal Style Scale, as well as
for his help in the double translation process.
The authors would also like to thank Saša Drače As-
sociate Professor at the Department of Psychology, Faculty
of Philosophy at the University of Sarajevo for providing
important additional resources of scientific databases and
Prof. Nazifa Savčić from English Department at the Faculty
of Philosophy, University of Sarajevo for the help in back
translation of the instrument.
The authors would like to thank the therapists who hel-
ped by recruiting their clients, as well as to the clients who
participated in the study.
REFERENCES
1. Ryan RM, Deci EL. Self-determination theory and the facilitation
of intrinsic motivation, social development, and well-being. Am
Psychol 2000; 55(1): 68−78.
2. Vansteenkiste M, Wiliams GC, Resnicow K. Toward systematic inte-
gration between Self-Determination Theory and Motivational In-
terviewing as examples of top-down and bottom-up intervention
development: Autonomy or volition as a fundamental theoretical
principle. Int J Behav Nutr Phy 2012; 23(9): 1−11.
3. Pelletier LG, Tuson KM, Haddad NK. Client Motivation for Therapy
Scale: a measure of intrinsic motivation, extrinsic motivation, and
amotivation for therapy. J Pers Assess 1997; 68(2): 414−35.
4. Dinger U, Strack M, Leichsenring F, Wilmers F, Schauenburg H. Thera-
pist effects on outcome and alliance in inpatient psychotherapy. J
Clin Psychol 2008; 64(3): 344−54.
5. Okiishi JC, Lambert MJ, Eggett D, Nielsen L, Dayton DD, Vermeersch
DA. An Analysis of Therapist Treatment Effects: Toward Provid-
ing Feedback to Individual Therapists on Their Cli-
ents‘Psychotherapy Outcome. J Clin Psychol 2006; 62(9):
1157−72.
6. Oliveira VC, Refsdhauge KM, Ferreira ML, Pinto RZ, Beckenkamp PR,
Filho RF, et al. Communication that values patient autonomy is as-
sociated with satisfaction with care: a systematic review. J Physio-
ther 2012; 58(4): 215−29.
7. Theodoridou A, Schlatter F, Ajdačić V, Rössler W, Jäger M. Therapeutic
relationship in the context of perceived coercion in a psychiatric
population. Psychiatry Res 2012; 200(2−3): 939−44.
8. Ackerman SJ, Hilsenroth MJ. A review of therapist characteristics
and techniques positively impacting the therapeutic alliance. Clin
Psychol Rev 2003; 23(1): 1−33.
9. Heinonen E, Lindfors O, Laaksonen MA, Knekt P. Therapists' profes-
sional and personal characteristics as predictors of outcome in
short- and long-term psychotherapy. J Affect Disord 2012; 138(3):
301−12.
10. Watson JC. Treatment failure in humanistic and experiential psy-
chotherapy. J Clin Psychol 2011; 67(11): 1117−28.
11. Henry WP, Schacht TE, Strupp HH, Butler SF, Binder JL. Effects of
training in time-limited dynamic psychotherapy: mediators of
therapists' responses to training. J Consult Clin Psychol 1993;
61(3): 441−7.
12. Ng CT, James S. "Directive approach" for Chinese clients receiving
psychotherapy: is that really a priority. Front Psychol 2013; 4(49):
1−13.
13. Patrick H, Williams GC. Self-determination theory: its application
to health behavior and complementarity with motivational inter-
viewing. Int J Behav Nutr Phy 2012; 9(18): 1−12.
14. Swift JK, Callahan JL. The impact of client treatment preferences
on outcome: a meta-analysis. J Clin Psychol 2009; 65(4): 368−81.
15. Mcmurtry SL, Hudson WW. The Client Satisfaction Inventory. Re-
sults of an Initial Validation Study. Res Social Work Prac 2000;
10(5): 644−63.
16. Landa AJ, Martos MP, Lopez-Zafra E. Emotional intelligence and
personality traits as predictors of psychological well-being in Span-
ish undergraduates. Soc Behav Personal 2010; 38(6): 783−94.
17. Romero E, Gomez-Fraguela JA, Villar P. Life aspirations, personality
traits and subjective well-being in a Spanish sample. Eur J Personal-
ity 2012; 26(1): 45−55.
18. Steel P, Schmidt J, Shultz J. Refining the relationship between per-
sonality and subjective well-being. Psychol Bull 2008; 134(1):
138−61.
19. Lucas RE, Diener E. Personality and subjective well-being. In: John
OP, Robins R, Pervin LA. , editors. Handbook of personality : the-
ory and research. 3rd ed. New York: Guilford; 2008. p. 795−814.
20. Weiss A, Bates TC, Luciano M. Happiness is a personal(ity) thing:
the genetics of personality and well-being in a representative sam-
ple. Psychol Sci 2008; 19(3): 205−10.
21. Gutiérrez JL, Jiménez BM, Hernández EG, Puente CP. Personality and
subjective well-being: big five correlates and demographic vari-
ables. Pers Indiv Differ 2005; 38: 1561−9.
22. Gale CR, Booth T, Mõttus R, Kuh D, Deary IJ. Neuroticism and Ex-
traversion in Youth Predict Mental Wellbeing and Life Satisfaction
40 Years Later. J Res Pers 2013; 47(6): 687−97.
23. DeYoung CG. Higher-order factors of the Big Five in a multi-
informant sample. J Pers Soc Psychol 2006; 91(6): 1138−51.
24. Ashton MC, Lee K, Goldberg LR, de Vries RE. Higher order factors
of personality: do they exist. Pers Soc Psychol Rev 2009; 13(2):
79−91.
25. del Re AC, Flückiger C, Horvath AO, Symonds D, Wampold BE.
Therapist effects in the therapeutic alliance-outcome relationship:
a restricted-maximum likelihood meta-analysis. Clin Psychol Rev
2012; 32(7): 642−9.
26. Pelletier LG, Fortier MS, Vallerand RJ, Tuson KM, Brìere NM, Blais
MR. Toward a New Measure of Intrinsic Motivation, Extrinsic
Motivation, and Amotivation in Sports: The Sport Motivation
Scale (SMS). J Sport Exerc Psychol 1995; 17: 35−53.
27. Kardum I, Smovjer I. Five Factor Model of Personality Structure:
Item Selection descriptors in Croatian language. Godišnjak
Zavoda za psihologiju 1993; 2(1): 91−100. (Croatian)
28. Field A. Discovering Statistics Using SPSS. 3rd. London: Sage
Publications; 2009.
29. Hair FH, Black WC, Babin BJ, Anderson RE. Multivariate Data
Analysis. 7th ed. New Jersey (US): Pearson Prentice Hall; 2010.
30. Jöreskog KG, Sörbom D. LISREL-VI user‘s guide. 3rd ed. Moores-
ville (IN): Scientific Software; 1984.
31. Byrne BM. Structural Equation Modeling with AMOS. 2nd ed.
New York (US): Routledge, Taylor and Francis Group; 2010.
32. Browne MW, Cudeck R. Alternative ways of assessing model fit. In:
Bollen KA, Long JS, editors. Testing Structural Equation Models.
Newbury Park: Sage 1993. p. 136−62.
33. Bentler PM, Bonett DG. Significance tests and goodness of fit in the
analysis of covariance structures. Psychol Bull 1980; 88(3):
588−606.
34. de Young CG, Peterson JB, Séguin JR, Tremblay RE. Externalizing be-
havior and the higher order factors of the Big Five. J Abnorm
Psychol 2008; 117(4): 947−53.
Received on September 11, 2014.
Revised on November 10, 2014.
Accepted on December 4, 2014.
Online First December, 2015.
VOJNOSANITETSKI PREGLED Page 11
Hadžiahmetović N, et al. Vojnosanit Pregl 2015; Online First December (00): 141–141.
Appendix
The Adapted Therapist's Interpersonal Style Scale (ATIS)
1-------------2-------------3-------------4-------------5-------------6-------------7
Never Often Always
1. The feedback I receive from my therapist is constructive in helping me make improvements.
2. When I ask my therapist to help me solve a problem, he or she asks me what I think before giving me his or her
opinion.
3. My therapist gives me the feedback about the way I make a progress.
4. My therapist is trying to impose her/his ideas on me.
5. The feedback I get from my therapist is basically useless criticism.
6. I feel that my therapist is indifferent towards me.
7. My therapist pressures me to do what he or she wants.
8. My therapist consults me before (s)he decides how to address my problems.
9. My therapist is being harsh to me.
10. I feel that my therapist doesn’t care how much I improve through therapy.
11. My therapist provides me with lots of opportunities to take personal decisions in what I do.
12. My therapist leaves me with little choice about the ways in which my problems could be resolved.
